Mon 22 Nov
16:00 - 17:00
On-Demand Podium Sessions
View anytime on this date, after 16:00, until January 2. Impact of the COVID-19 Pandemic: Healthcare Utilisation and Outcomes
This session covers the latest research on impacts of COVID-19 on healthcare utilisation and outcomes using different data sources.
Roisin Adams, MPharm, MSc, PhD
National Centre for Pharmacoeconomics, Dublin, Ireland
Dr. Roisin Adams is Head of HTA Strategy and External engagement for the NCPE. Dr. Adams led the HTA team for a number
of years before being seconded to the HSE to lead a new unit tasked with overseeing and managing high cost drugs in acute hospitals. Dr. Adams co-chairs the HTA domain of the Beneluxa initiative and oversees the EUNetHTA work of the NCPE. She also is a Director on the Board of the International Horizon Scanning Initiative. She has been awarded a number of grants from the Health Research Board and held advisory positions for Department of Health, the Health Information and Quality Authority and policy direction at EU level. In recent years she has led the COVID-19 specific assessments for the NCPE.
P22: COVID-19 Pandemic Impacts Volume of Evaluation & Management (E&M) Telehealth Visits within Community Oncology Practices
4:15PM - 4:30PM
Karhade M 1, Mohammad N 2, Robert N 3, Wu N 4, Heller B 5, Alwardt S 6, Neubauer M 6, Smith H 6, Moore L 7 1Ontada, Houston, TX, USA, 2Ontada, Cypress, TX, USA, 3Ontada, Irving, TX, USA, 4US Oncology Network, The Woodlands, TX, USA, 5Southern Cancer Center, Mobile, AL, USA, 6Ontada, The Woodlands, TX, USA, 7Ontada, Shaker Heights, OH, USA
OBJECTIVES: The USA declared the COVID-19 pandemic a national emergency on 03/13/20. On 03/17/20, CMS expanded telehealth rules, allowing Medicare to cover telehealth visits like regular visits. This study aims to analyze the utilization of Evaluation & Management (E&M) telehealth options in community oncology pre and post pandemic. METHODS: Deidentified patient visits data were obtained from iKnowMed electronic health records between 01/01/18 to 05/24/2021 from 20 US Oncology practices. A combination of patient MRN and date was used as an identifier to report number of visits for all measures. Patient visits with modifiers –GT, –95, and –GQ were classified as telehealth visits. Visit dates without modifiers were defined as non-telehealth (in-office) visits. E&M visits were defined based on standard CPT codes. RESULTS: A total of 5,914,125 unique E&M patient visits were analyzed during the study period. Between Jan-2018 and Mar-2020 (pre-COVID-19), E&M visits rose from 30,000/week to 36,000/week (20%). Fewer than 0.01% of these visits were telehealth. By April 12, 2020, overall E&M visits had dropped 35%, but the telehealth visits had risen to 16%. Since then, the overall E&M visit count remained approximately 5% lower as compared to the pre-COVID-19 trend, and telehealth visits averaged approximately 6% thereafter. Corresponding to the 2 nd wave, in Dec-2020 the telehealth proportion rose again to 10%. As of 05/23/2021, telehealth E&M visits represented approximately 5% of the total E&M visits within US Oncology practices. CONCLUSIONS: This study provides a timeline of how COVID-19 has impacted E&M visits and telehealth utilization among community oncology practices. The pandemic has led to an increase in E&M telehealth visits that may remain post pandemic. Continued research is necessary to monitor telehealth utilization and its impact on the quality of care, provider finances, and future of community oncology considering rising vaccination rates, CDC guidance, and public sentiment. P21: Social Distancing and Trends in Influenza Hospitalization during the COVID-19 Outbreak: A Difference-in-Difference Analysis of German Claims Data
4:00PM - 4:15PM
Pacis S 1, Maywald U 2, Wilke T 3, Ghiani M 4 1Cytel Inc, Berlin, BE, Germany, 2AOK PLUS, Dresden, Germany, 3IPAM e.V., Wismar, Germany, 4IPAM, University of Wismar, Berlin, BE, Germany
OBJECTIVES: As COVID-19 spread worldwide, indicators of influenza activity in the Northern Hemisphere began to decline by mid-to-late February. In Germany, federal lockdown measures were introduced to contain the outbreak on 22/03/2020 (week 12). We used claims data from AOK PLUS, a regional sickness fund covering around half the population in Saxony and Thuringia (6.2 million inhabitants), to examine the trend of influenza hospitalizations in 2020 compared to 2019. METHODS: Using data from 01/01/2019 to 31/05/2020 (weeks 1-22), influenza hospitalizations were identified using ICD-10-GM codes J10-J11. We estimated changes in the number of influenza hospitalizations using a “difference-in-differences” model including variables for age group (<18, 18-44, 45-64, 65-79, 80+), gender, week, year, and outbreak status (interaction variable between year 2020 and week 12 or later). Adjusted incidence rate ratios (aIRRs) were estimated using Poisson regression with heteroskedasticity-robust standard errors. RESULTS: During weeks 1-22, we observed 5,174 influenza hospitalizations in 2019 and 2020. Influenza hospitalizations in 2020 showed similar trends until week 12 and then showed a relative decline compared to 2019. The average number of influenza hospitalizations per week during weeks 12-22 significantly decreased in 2020 compared to 2019 (1.6 vs. 5.2; aIRR: 0.45; 95% CI: 0.34-0.59; p<0.001). When stratified by age group, all groups except age 18-44 had a similar decrease in average influenza hospitalizations per week in 2020 compared to 2019, with large relative declines in patients age 80+ (2.2 vs. 5.8; aIRR: 0.36; 95% CI: 0.28-0.46; p<0.001) and children <18 (1.8 vs. 8.0; aIRR: 0.38; 95% CI 0.32-0.46; p<0.001). CONCLUSIONS: The number of influenza hospitalizations saw a relative faster decline in 2020 compared to 2019 after the introduction of federal lockdown measures in Germany, possibly due to the effectiveness of non-pharmaceutical interventions like social distancing and the use of facemasks. P24: Telehealth Access and Use by the U.S. Medicare Population during the Pandemic
4:45PM - 5:00PM
Swenson T Des Moines University, Des Moines, IA, USA
OBJECTIVES: Telehealth access and reimbursement varied by payer and regionally prior to COVID-19. and its limited availability expanded in response to the pandemic. The health behavioral response by older adults to COVID-19 has varied over time with the geographic spread of the pandemic and affected access and utilization of medical services. The purpose of this paper is to examine changes in access to telemedicine in 2020 in response to the pandemic for the U.S. Medicare population. METHODS: The first two waves in June and October 2020 of the rapid response survey fielded by the Centers for Medicare and Medicaid Services (CMS) to track and monitor the effects of the pandemic within the U.S. Medicare population. With a panel sample size of 9686 Medicare beneficiaries, the calculated statistics use replicate weights to adjust for the complex survey sample design and balanced repeated replication using Fay’s adjustment of 0.3 for variance estimation. RESULTS: Nearly 45 percent of the Medicare population reported use of a telehealth appointment between June and October of 2020. The likelihood of using telemedicine increased for those with chronic conditions, such as depression, and for those with higher incomes and education. Medical practices were more likely to encourage telehealth visits for Medicare patients between March and June with 57 percent of the Medicare population reporting that their usual provider offered a telemedicine appointment to replace a regular office visit during the spring and 48% reporting the suggested telemedicine replacement from July through October 2020. Overall access to telehealth increased from 60% to 64% but varied by race/ethnicity, gender, Census regions, and rural status. CONCLUSIONS: Access to telemedicine services expanded for the U.S. Medicare population during the pandemic but usage varied by chronic disease status, socioeconomic and demographic factors, and geography. P23: Change in Healthcare Utilisation and Inpatient Mortality in Patients Hospitalised with Heart Failure during the Coronavirus Pandemic in England: A Retrospective Cross-Sectional Study Utilising HES
4:30PM - 4:45PM
, Barham L Jones D Learna Ltd in partnership with the University of South Wales, Royston, UK
OBJECTIVES: This study quantifies change in healthcare utilisation and inpatient mortality of all adult patients hospitalised with Heart Failure in England during three coronavirus national lockdowns compared to the same time period in the previous year. METHODS: A retrospective cross-sectional study using the Hospital Episode Statistics (HES) database was conducted. All adults admitted to an English hospital with a primary diagnosis of I110 Hypertensive heart disease with (congestive) heart failure, I255 Ischaemic cardiomyopathy, I420 Dilated cardiomyopathy, I429 Cardiomyopathy unspecified, I500 Congestive heart failure, I501 Left ventricular failure and I509 Heart failure unspecified between 1 st March 2019 and 28 th February 2021 were included. Admissions, bed days and inpatient deaths of patients admitted between 1 st March 2020 and 28 th February 2021 (during pandemic) was compared with patients admitted between 1 st March 2019 and 29 th February 2020 (prior to pandemic). The difference in event count was used to test national changes and a P-value of ≤0.05 was used to test significance RESULTS: There were 140,035 heart failure admissions in the observational period, 64,770 during the pandemic and 75,265 prior to the pandemic, all data were analysed. There were reductions in admissions (69,555 vs 80,715, P<0.0000000000), bed days (586,430 vs 753,985, P=0.0000000000) and inpatient deaths (7,650 vs 8,305, P=0.0000002154) during the pandemic. CONCLUSIONS: There were significantly fewer admissions, bed days and inpatient deaths for patients admitted with heart failure during the coronavirus pandemic. Interpretation of this change is challenging as this may reflect unmet health needs as patients ‘put off’ seeking care. Further research is required to analyse the change in out of hospital healthcare utilisation, deaths in other settings and to explore potential for excess and latent morbidity and mortality that may result from reduced access to hospital services during the pandemic. Methods and Controversies in the Evaluation of Oncology Products
The evaluation of innovative oncology products often depends on evidence from a small number of landmark clinical trials. These trials often present challenges of short follow-up and lack of power to measure differences in overall survival. This session discusses recent work to handle these challenges and explore the degree of uncertainty around benefits and costs. Among the work presented are novel methods for survival prediction that synthesize clinical data (from trials and external sources) with expert opinion, validation of extrapolation assumptions, comparison of structural modelling methods, and decision makers' willingness to accept surrogate endpoints.
Oriana Ciani, PhD
SDA Bocconi School of Management, Milan, MI, Italy
Oriana Ciani is Associate Professor of Practice at SDA Bocconi in Milan. She holds a MSc in Biomedical Engineering from Politecnico di Milano and postgraduate degree in Healthcare Management from Bocconi University. She received her PhD from the University of Exeter with a thesis focusing on the evaluation of surrogate end points. She has been 2020 Fulbright Research Scholar at Yale School of Medicine and Yale School of Public Health. Oriana's research interests are centred on the use of evidence synthesis techniques to inform policy decisions, health technology assessment (HTA) and healthcare policies evaluation.
P52: Surrogate Survival Endpoints: Are They Sufficient to Support Access?
4:45PM - 5:00PM
1, Dacosta RFD 2, Grosvenor A 3, Wang GD 4, Macaulay R 3 1Precision Advisors, London, LON, UK, 2PrecisionADVISORS, London, UK, 3Precision Advisors, London, UK, 4PrecisionAdvisors, London, UK
Immuno-oncology therapies (IOs) have revolutionised metastatic cancer management over the last decade. These are now being investigated in earlier stages of cancer, where surrogate survival endpoints such as relapse-free survival (RFS), disease-free survival (DFS), and pathological complete response (pCR) are needed to bridge the evidence gap prior to maturation of overall survival (OS) data. This research assesses surrogate survival endpoints in supporting access/reimbursement in early-stage cancers. Publicly accessible assessments of early-stage cancer therapies by seven HTA-bodies (HAS, G-BA, Medicinrådet, NICE, SMC, PBAC, CADTH) up to May 2021 were extracted. 52 assessments of 11 drug:indication pairings were identified, which utilised four surrogate primary endpoints (RFS, DFS, invasive DFS, pCR). 30 were fully reimbursed, 11 restricted reimbursement, 11 not reimbursed. Payer support was defined as being considered clinically/patient-relevant and/or a valid surrogate to OS. RFS was the best-supported surrogate endpoint by HTA bodies (7/7), followed by DFS (5/7) and IDFS (5/7), while pCR (2/7) was least well-supported. NICE and SMC were the most supporting of access using surrogate endpoints, followed by G-BA, while CADTH was least supporting. Assessments were most positive where the magnitude of the surrogate benefit were greater/showed a superior benefit-risk and where early data showed a likely long-term OS benefit. However, some assessments from NICE/G-BA were conditional and contingent on more mature follow-up data being provided, and PBAC imposed a flow-on restriction limiting retreatment with IOs as later line(s) of therapy. Several IOs already approved and reimbursed in the metastatic setting are being investigated in the early-stage setting using surrogate survival endpoints. These surrogate endpoints have been widely utilised for regulatory approval, and as shown here, HTA bodies are supportive of some surrogate endpoints based on their patient relevance and/or their validated OS correlation.
P49: Accuracy of Life Year Gains Predictions for CAR-T Therapy in the Long Term: An Analysis for Axicabtagene Ciloleucel in Refractory Large B-Cell Lymphoma
4:00PM - 4:15PM
Porteous A 1, Gregori D 1, Hilton B 2 1Costello Medical, London, UK, 2Costello Medical, Cambridge, UK
OBJECTIVES: Survival profiles for chimeric antigen receptor T cell (CAR-T) therapies commonly exhibit a plateau that may indicate a cure. This leads to challenges predicting long-term outcomes for novel CAR-T therapies when trial data are immature. This study retrospectively analysed the accuracy of overall survival (OS) extrapolations from interim data cuts in predicting realised long-term life years (LYs) for axicabtagene ciloleucel in patients with refractory large B-cell lymphoma. METHODS: Published OS data for axicabtagene ciloleucel from successive data cuts of ZUMA-1 (median follow-ups of 15.4, 27.1 and 51.1 months) were digitised. Standard parametric, spline (1–2 knots; normal, odds and hazard) and mixture cure models (MCMs) were fitted to the first two data cuts. Statistical fit was tested using Akaike and Bayesian information criteria (AIC and BIC). Cumulative LYs were estimated for each model over a 58-month time horizon, corresponding to the longest duration of published OS data. These projected LYs were then compared to realised LYs over this period. RESULTS: At the earliest data cut, MCMs provided the best predictions of realised LYs, with a mean absolute difference of 6.8% between predicted and realised LYs across MCM models (range: 0.5%–10.7%). Standard parametric extrapolations considerably underestimated realised LYs (mean absolute difference: 19.2%; range: 9.9%–28.0%), whilst spline models offered a mean absolute difference of 8.3% (range: 1.4%–13.2%). Similar findings were observed for extrapolations based on the second data cut (representing 11.7 months additional follow-up), but differences between model classes were less pronounced. CONCLUSIONS: MCMs may offer the best predictions of long-term survival for CAR-T therapies, particularly when only short-term data are available. Standard parametric models may be inappropriate to predict survival when extrapolating immature data, failing to capture the plateau in survival typical of CAR-T therapies. Further research is required to determine whether these findings are generalisable across CAR-T therapies and indications. P50: Competing Risk and Multistate Model Compared to Partitioned Survival Model in Metastatic Non-Small Cell Lung Cancer
4:15PM - 4:30PM
Le Mezo A 1, Kandel M 2, Caillon M 3, Chauny JV 4, Borget I 5 1Master 2 Market-Access et Evaluation Médico-Economique, Université Paris-Saclay, Sèvres, France, 2IQVIA France, La Défense Cedex, France, 3Amgen SAS, Boulogne Billancourt, 75, France, 4Amgen SAS, Boulogne Billancourt, France, 5Department of Biostatistics and Epidemiology, Gustave Roussy, Paris-Saclay University, Villejuif, France; Oncostat; GRADES, Paris-Saclay University, Châtenay-Malabry, France U1018, Inserm, Paris-Saclay University, “Ligue Contre le Cancer” labeled team, Chatenay-Malabry, 92, France
OBJECTIVES Partitioned survival analysis (PartSA) is commonly used for economic evaluations in oncology. We compare PartSA to a competing risk semi-markov multi-state model (MSM) and investigate differences in estimated cost-effectiveness in metastatic non-small cell lung cancer from a French perspective. : METHODS To populate both models, pooled data from Checkmate 017 and 057 was digitized to reconstruct patient-level data from overall survival, progression-free survival and post-progression survival Kaplan Meier curves. Models consisted of three states: progression-free, progressed disease and death. Statistical analysis was performed under : R to fit parametric survival models for the PartSA and to estimate state transitions for the MSM (package “mstate”). Costs and utilities were integrated in the model using values from nivolumab’s French HTA submission. Outcomes were discounted at 2,5% per annum over a 7-year time horizon. RESULTS Both models produced similar results during trial period (15 months) but showed discrepancies during extrapolation, inducing different costs and quality-adjusted life years (QALYs) over time. Incremental life years gained (LYG) of nivolumab versus docetaxel were 1,11 LYG (PartSA) and 0,88 LYG (MSM). However, both incremental cost effectiveness ratios (ICER) were similar, 88 979€/QALY (PartSA) and 90 148€/QALY (MSM). Scenarios assessing different parametric extrapolations produced an average of 112 496€/QALY (PartSA) and 113 085€/QALY (MSM) with coefficients of variation (CV) estimated at 19,0% (PartSA) and 16,3% (MSM). Probabilistic ICER results were similar between models with CV estimated at 21,8% (PartSA) and 23,2%(MSM). : CONCLUSIONS : This study provides new evidence on structural uncertainty. The MSM produced similar results to the PartSA for a 15 months follow-up. Ideally, both models should be tested to validate extrapolations and to ensure that the structural uncertainty is limited. However, when the time horizon is short (<10 years), the impact of the model choice on the ICER seems limited. P51: Blended Survival Curves: A New Approach to Extrapolation for Time-to-Event Clinical Trial Data in Health Technology Assessment
4:30PM - 4:45PM
Che Z 1, Baio G 2, Green N 2 1University College London, London, UK, 2University College London, London, LON, UK
OBJECTIVES: Survival extrapolation is generally required in the cost-effectiveness analysis to estimate the survival benefit of a new intervention, due to the limited duration of randomized controlled trials (RCTs). Current techniques of extrapolation often assume constant treatment effect beyond the observed period in the RCT, which is implausible and highly influential in survival estimates for resource allocation decisions. The objective of this study is to develop a novel methodology based on “blending” survival curves as a possible solution. METHODS: We mixed a flexible Cox semi-parametric model conducted in Bayesian setting to fit the observed data and a parametric model either by prior assumptions or external data on the long-term expected behavior of the underlying survival curves. The two are “blended” into a single survival curve that is equivalent to the Cox model over the follow-up intervals and gradually approaching to the parametric model over the extrapolation period based on a weight function. The weight function and mixing area of the blended curve control the way the internal and external data sources influence the estimated survival. RESULTS: A 4-year follow-up RCT of rituximab in combination with fludarabine and cyclophosphamide (RFC) v. fludarabine and cyclophosphamide alone (FC) for the first-line treatment of chronic lymphocytic leukemia is used to illustrate the method. Two kinds of prior information, registry data and summary of clinical knowledge were respectively used for the long-term estimate. CONCLUSIONS: Long-term extrapolation with various assumptions of treatment effect may give significantly different estimated mean survival gains. The blending process allows a consideration of plausible scenarios, abandoning the over-optimistic constant treatment effect and provides sufficient flexibility. Not only internal but also external validity could be carefully considered since a wide range of external evidence can be used to inform the long-term estimate, including hard data from real world and clinical expert opinion. Access to Care/Real World Evidence to Inform Decision Making
This session covers the use of real world evidence to inform decision making within innovative therapies, metastatic colorectal cancer, statin therapy and atopic dermatitis.
Olga Ovcinnikova, MSc
MSD UK, Chorleywood, United Kingdom
Over 12 years of working in pharmaceutical industry, health economics consultancies and public health centers across Europe. I have experience across a wide range of therapeutic areas including oncology, vaccines, GI, mental health and blood disorders.
P2: Estimation of Metastatic Colorectal Cancer Patients Carrying BRAF Mutation Potentially Eligible to Targeted Therapy: A Real-World Evidence Study in Italy
4:15PM - 4:30PM
Degli Espositi L 1, Sangiorgi D 2, Andretta M 3, Bacca M 4, Barbieri A 5, Bartolini F 6, Cavaliere A 7, Chinellato A 8, Ciaccia A 9, Cillo MR 10, Citraro R 11, Costantini A 12, De Francesco A 11, Enieri N 8, Ferrante F 13, Gentile S 14, Lavalle A 14, Mancini D 4, Mensurati M 15, Moscogiuri R 16, Pastorello M 17, Procacci C 18, Re D 19, Santoleri F 12, Serao Creazzola S 20, Tegon M 8, Ubertazzo L 21, Vercellone A 22, Perrone V 23 1CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, Italy, 2CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, BO, Italy, 3Azienda ULSS 8 Berica, Vicenza, Italy, 4ASL Brindisi, Brindisi, Italy, 5ASL Vercelli, Vercelli, Italy, 6USL Umbria 2, Terni, Italy, 7ASL Viterbo, Viterbo, Italy, 8Azienda ULSS 3 Serenissima, Mestre (VE), Italy, 9ASL Foggia, Foggia, Italy, 10ASL Salerno, Salerno, Italy, 11Azienda ospedaliero-universitaria Mater Domini, Catanzaro, Italy, 12ASL Pescara, Pescara, Italy, 13ASL Frosinone, Frosinone, Italy, 14Direzione Generale per la Salute Regione Molise, Campobasso, Italy, 15ASL Roma 3, Roma, Italy, 16ASL Taranto, Taranto, Italy, 17ASP Palermo, Palermo, Italy, 18ASL BAT, Trani, Italy, 19ASL Teramo, Teramo, Italy, 20ASL Napoli 1, Napoli, Italy, 21ASL Roma 4, Civitavecchia (RM), Italy, 22ASL Napoli 3 SUD, Torre del Greco, Italy, 23CliCon S.r.l. Health, Economics & Outcomes Research, Ravenna, RA, Italy
OBJECTIVES: Cancer treatments represent one of the most expensive items for the National Health System. In a limited-resource system, the introduction of costly and innovative oncological therapies makes it necessary to balance the innovation and the access to treatments based on patient eligibility. The study aimed to evaluate the possibility of identifying metastatic colorectal cancer (mCRC) patients carrying BRAF-gene mutation, potentially eligible to targeted therapy, by linking administrative and pathological anatomy (PA) databases. METHODS: A retrospective study was conducted across 2013-2019 in a sample of Italian Entities, using the data-linkage between administrative and PA databases. Data were reported per million of health-assisted individuals. CRC and mCRC patients [diagnosed by at least one hospitalization for CRC or mCRC (ICD-9-CM codes 153-154 and 196-197-198, respectively)], were screened. Mutational status of mCRC patients was identified by BRAF genetic test (procedure codes: 91.30.3/91.36.5/91.29.3/91.29.4). Data-linkage of these data with those from the administrative databases allowed the identification of mCRC patients carrying BRAF mutation (BRAF +). RESULTS: Overall, 4,666 CRC patients were identified, with an incidence (2019) estimated of 0.7/1,000 of health-assisted individuals. Among them, mCRC accounted for the 39% (N=1,818) of patients. The 50% (N=915) of mCRC patients had an outpatient test for BRAF. After the data-linkage between administrative and PA databases, 83% (N=765) of them performed the BRAF test, and 107 patients (14% of patients with BRAF test reported) were BRAF +. CONCLUSIONS: These results reported an epidemiological scenario of CRC and mCRC-BRAF + Italian patients in line with published data, showing that our methodology could be a supportive tool to identify eligible patients for targeted therapy. Furthermore, the use of PA database would allow to quantify patients with a specific genetic profile required to access to innovative therapies, thus enabling to estimate health-costs and to plan the pharmaceutical expenditure in a perspective of economic sustainability. P4: Switching, Persistence and Adherence to Statin Therapy: A Retrospective Cohort Study Using the Australian National Pharmacy Data
4:30PM - 4:45PM
Talic S 1, Marquina C 1, Zomer E 1, Lybrand S 2, Liew D 3, Ademi Z 1 1Monash University, Melbourne, Australia, 2Amgen Australia, North Ryde, Australia, 3Monash University, Melbourne, VIC, Australia
OBJECTIVES: Statins are widely prescribed for the primary and secondary prevention of cardiovascular disease, but their effectiveness is dependent on the level of adherence and persistence. This study aimed to explore switching, adherence and persistence among the Australian general population with newly dispensed statins. METHODS: A retrospective cohort study was conducted using a random sample of data from the Australian national prescription claims data. Switching, adherence to and persistence with statins were assessed for people starting statins from 1 January 2015 to 31 December 2019. Cox proportional hazard models were used to compare outcomes between different statins. RESULTS: A cohort of 141,062 people dispensed statins and followed over a median duration of 2.5 years were included. Of the cohort, 29.3% switched statin intensity, 28.4% switched statin type, 3.7% switched to ezetimibe and in 2.7%, ezetimibe was added as combination therapy during the study period. Overall, 58.8% discontinued statins based on the 90-day gap criteria, of whom 55.2% restarted. The proportion of people non-adherent was 24.0% at 6 months to 49.0% at 5 years. People on low and moderate intensity statins were more likely to discontinue compared to those on high-intensity statins HR 1.20, 95% confidence interval [CI] 1.09–1.31), (HR 1.28, 95%CI 1.14–1.42), respectively. Compared to maintaining same statin type and intensity, switching statins, which includes up-titration (HR 0.77, 95%CI 0.70 to 0.86) was associated with less likelihood of discontinuation after reinitiation. CONCLUSIONS: Long-term persistence and adherence to statins remains generally poor among Australians, which limits the effectiveness of these medicines and the consequent health impact they may provide for individuals. Switching between statins is prevalent in one third of statin users, although any clinical benefit of the observed switching trend is unknown. This, combined with the high volume of statin prescriptions, highlights the need for better strategies to address poor persistence and adherence. P1: Data-Linkage Between Administrative and Pathological Anatomy Databases for the Identification of Lung Cancer Patients Eligible to Innovative Therapies
4:00PM - 4:15PM
Degli Espositi L
1, Sangiorgi D 2, Andretta M 3, Bacca M 4, Barbieri A 5, Bartolini F 6, Cavaliere A 7, Chinellato A 8, Ciaccia A 9, Cillo MR 10, Citraro R 11, Costantini A 12, De Francesco A 11, Ferrante F 13, Gentile S 14, Lavalle A 14, Mancini D 4, Mensurati M 15, Moscogiuri R 16, Pastorello M 17, Procacci C 18, Re D 19, Santoleri F 12, Serao Creazzola S 20, Ubertazzo L 21, Vercellone A 22, Perrone V 1 1CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, Italy, 2CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, BO, Italy, 3Azienda ULSS 8 Berica, Vicenza, Italy, 4ASL Brindisi, Brindisi, Italy, 5ASL Vercelli, Vercelli, Italy, 6USL Umbria 2, Terni, Italy, 7ASL Viterbo, Viterbo, Italy, 8Azienda ULSS 3 Serenissima, Mestre (VE), Italy, 9ASL Foggia, Foggia, Italy, 10ASL Salerno, Salerno, Italy, 11Azienda ospedaliero-universitaria Mater Domini, Catanzaro, Italy, 12ASL Pescara, Pescara, Italy, 13ASL Frosinone, Frosinone, Italy, 14Direzione Generale per la Salute Regione Molise, Campobasso, Italy, 15ASL Roma 3, Roma, Italy, 16ASL Taranto, Taranto, Italy, 17ASP Palermo, Palermo, Italy, 18ASL BAT, Andria (BT), Italy, 19ASL Teramo, Teramo, Italy, 20ASL Napoli 1, Napoli, Italy, 21ASL Roma 4, Roma, Italy, 22ASL Napoli 3 SUD, Torre del Greco, Italy
OBJECTIVES: The introduction of innovative and costly oncological therapies highlights the need for healthcare system to balance the innovation with the access to therapies based on patient eligibility. This study aimed to identify, through real-world data from administrative and pathological anatomy (PA) databases, patients with metastatic lung cancer carrying specific tumor markers, potentially eligible to innovative immunotherapy treatments. METHODS: The data-linkage between administrative and PA databases on a pool of Italian local Healthcare Entities (LHEs) was performed. Data were reported per million of health-assisted individuals. All patients diagnosed for lung cancer (ICD-9-CM code:162) from 2013-2019 were included. Metastatic disease has been diagnosed by using ICD-9-CM codes 196-197-198. The data integration with PA data-set was carried out to evaluate morphologic characteristics (M-80023; M-80413; M-80423) and the level of PD-L1 expression, required for the eligibility to specific immunotherapies. RESULTS: Overall, 4,387 lung cancer patients were included, with annual incidence of 0.7/1,000 health-assisted individuals for year 2019. Among them, 37% (N=1,625) presented metastasis, in line with published evidence indicating around 30-40% of lung cancer to present metastasis. Further analyses were performed in metastatic patients with a record in the PA database (N=365). The 87% (N=317) of them had non-small cell lung cancer, accordingly to the literature estimation, and 79% of patients were potentially eligible for immunotherapy since showed positivity to PD-L1 TPS ≥1%. CONCLUSIONS: The present study results are in line with epidemiological data reported by AIRTUM (incidence 0.7/1,000) and with international literature and showed how the applied methodology could represent a valuable tool for identifying patients eligible to new therapies. The use of PA data-set could allow the detection of patients with specific genetic profiles and their access to innovative medications. The quantification of potentially eligible patients would also allow budget impact estimation needed to plan the pharmaceutical spending by LHE. Evaluating Individuals and Patients Preferences: Discrete Choice Experiments and Beyond
This session illustrates how discrete choice experiments (DCE) can be used to better understand individuals' or patients' preferences and discusses how alternative methods such as the probabilistic threshold technique may be better to estimate maximum acceptable risk.
Anne E Spencer, BSc, MPhil, PhD
University of Exeter, Exeter, United Kingdom
P19: Assessing Heterogeneity in MAR: Methods and Models Beyond DCE
4:30PM - 4:45PM
Janssen E 1, DiSantostefano R 1, Falahee M 2, Simons G 2, Englbrecht M 3, Radawski C 4, Raza K 2, Hauber B 5, Veldwijk J 6 1Janssen R&D, Titusville, NJ, USA, 2University of Birmingham, Birmingham, UK, 3Freelance, Eckental, Germany, 4Eli Lily & Company, Indianapolis, IN, USA, 5Pfizer, New York, NY, USA, 6Erasmus University Rotterdam, Rotterdam, Netherlands
OBJECTIVES: Discrete choice experiments (DCEs) are robust stated-preference methods frequently used to estimate maximum acceptable risk (MAR) as a secondary outcome. However, DCEs provide sample-level estimates and explaining preference heterogeneity for MARs based on participant characteristics can be difficult. The study objective was to compare the capability of a DCE and a probabilistic threshold technique (PTT) to identify preference heterogeneity among MARs for preventive rheumatoid arthritis (RA) treatment. METHODS: Participants from 3 countries (United Kingdom (UK), Germany, and Romania, n = 2959) completed a DCE and PTT in random order. Participants made choices between treatments that reduced chance of developing RA but increased chance of three risks (mild and serious side effects, serious infection). For the PTT, interval regressions estimated MARs that accounted for age, education, numeracy, literacy, and RA family history. For the DCE, random parameters logit (RPL) models were used to calculate MARs for subgroups in which heterogeneity was identified in the PTT. RESULTS: The PTT identified preference heterogeneity for numeracy, literacy, and family history. Regarding these characteristics, the PTT identified statistically significantly different MARs (p<0.05) for at least one risk in at least two countries. The DCE identified preference heterogeneity for the chance of serious infection between UK participants with low vs. high numeracy (p<0.05). Using the DCE, no statistically different MARs were identified for other combinations of participant characteristics, risks, or countries. CONCLUSIONS: The PTT identified preference heterogeneity in MARs for more participant characteristics by directly incorporating participant characteristics in the regression model. When attempting to estimate MARs, PTT may partially overcome challenges with stratified DCE models, particularly if analyses such as latent class analysis are not feasible or desirable. Further research is needed to confirm the findings in this case studies and to explore which method most accurately identify true underlying preference heterogeneity are needed. P17: Preference of Rheumatoid Arthritis Patients for Tapering Biologics: A Discrete Choice Experiment
4:00PM - 4:15PM
Suz Jack C 1, Stamp L 2, Treharne G 1, Marra C 1 1University of Otago, Dunedin, New Zealand, 2University of Otago, Christchurch, New Zealand
OBJECTIVES: Tapering of biologics is a safe and feasible approach in the long-term management of rheumatoid arthritis (RA) patients who are in remission. However, the appeal of tapering strategies needs to be balanced against the risks of disrupting patients’ disease control. The aim of this study was to measure the preferences of RA patients and their risk-benefit trade-offs in relation to biologic tapering. METHODS: A web-based discrete choice experiment (DCE) was employed. Seven attributes (identified via focus groups and a systematic review) of varying levels describing three hypothetical choice were presented: frequency of treatment, chances of known adverse effects, chances of regaining disease control and healthcare service-related features. DCE data were analysed using mixed logit model to estimate the preference weights for key treatment features and to quantify trade-offs between the attributes. RESULTS: A total of 142 complete responses were analysed. Mean age was 60.3 years with an average disease duration of 20.8 years Frequency of biologic treatment was the most important attribute, followed by the chance of flare upon tapering. Time to see the rheumatology team after a flare was ranked the least important among the seven attributes. On average, participants were willing to accept between 25.3% to 50.2% increase in chance of disease flare in exchange for reducing the frequency of biologic treatment, chance of serious infection and chance of skin cancer. CONCLUSIONS: This study provides evidence that RA patients’ preference for tapering biologics are most influenced by the frequency of treatment and chance of flare. For these attributes, they are willing to accept a greater chance of flare in exchange for treatment benefits in the form of a reduction in biologic dosing and potential risk of serious infection and skin cancer associated with long-term biologic use. These findings have implications for clinical practice and policy making about tapering. P18: Patient Preferences for Attributes of a Multi-Cancer Early Detection Test: A Discrete Choice Experiment (DCE) Quantitative Pilot Study
4:15PM - 4:30PM
1, Ross M 1, Kansal AR 2, Fung E 2, Seiden M 3, Chung KC 2 1Evidera, Bethesda, MD, USA, 2GRAIL, Inc., Menlo Park, CA, USA, 3McKesson, The Woodlands, TX, USA
OBJECTIVES : Early cancer detection and intervention can significantly improve patient outcomes and reduce mortality rates. Evidence shows that emerging blood-based multi-cancer early detection (MCED) tests can detect a variety of cancer types across stages and provide a predicted cancer signal origin with high specificity. However, little is known about patients’ preferences for MCED tests. This study aimed to quantify preferences for attributes of blood-based MCED tests among the US general population aged 50-80 years. METHODS : A DCE consisting of five attributes (true positives, false negatives, false positives, likelihood of the cancer type unknown [e.g., inaccurate cancer signal origin], and number of cancers tested for) was administered online to US general population members to elicit preferences to quantitatively pilot test the DCE. Data were analyzed using an error-component multinomial logit model and relative attribute importance (RAI) was obtained. RESULTS : Participants (N=303) were 62.0% male (n=188), mean age 68.2 years (SD=6.4). RAI indicated that the rank order of attribute importance was false negatives (35.7%), true positives (27.6%), false positives (17.3%), number of cancers tested for (16.8%), and cancer type unknown (2.7%). Attributes related to improved test accuracy were important and participants strongly preferred screenings that tested for more cancer types (all p < 0.05). Preferences were non-significant for the likelihood of cancer type unknown attribute levels. Overall, 71.9% of participants reported that they would prefer to receive the MCED test in addition to their currently recommended cancer screenings. CONCLUSIONS : Participants’ preferences were strongly driven by the desire for a screening test with fewer false negatives and more true positives, with these 2 attributes comprising 63.3% of the RAI. False positive results and number of types of cancer tested for also impacted preferences but were of lower importance. The majority of participants preferred adding a MCED test to supplement current cancer screenings. P20: Evaluating Preferences and the Effect of Altruism on COVID-19 Vaccine Decisions: A Discrete Choice Experiment
4:45PM - 5:00PM
Kelley M 1, Drabo EF 2, Gong CL 3 1University of Southern California, Los Angeles, CA, USA, 2Johns Hopkins Bloomberg School of Public Health, Department of Health Policy and Management, Baltimore, MD, USA, 3Children's Hospital Los Angeles, Los Angeles, CA, USA
OBJECTIVES: To elucidate how Americans value COVID-19 vaccine characteristics, and determine whether their willingness to vaccinate is altered by the framing of the vaccination decision as altruistic or not. METHODS: We conducted a discrete choice experiment (DCE) with Amazon MTurk participants randomized into a control group with standard DCE questions, versus a treatment group with questions framed altruistically. The survey consisted of demographic questions, an altruism index, and a DCE of 12 choice tasks with 3 profiles (Vaccine A, Vaccine B, and No Vaccination). Vaccine attributes included number of doses, efficacy in preventing infection, risk of severe disease, severe side effect type, risk of severe side effect, and subsidy. We estimated preference weights using multinomial logit models, controlling for framing, sex, age, political party, health status, race/ethnicity, and altruism score. RESULTS: Sample included 2,014 respondents (control with no framing, n=1,037; altruism framing, n=977). Respondents preferred COVID-19 vaccines with allergic reactions vs neurological disorder as side effects (OR: 1.32; P <0.01), higher efficacy (OR: 1.03; P < 0.01), higher subsidies (OR: 1.00; P < 0.01), lower risk of side effects (OR: 0.99; P < 0.01), and lower risk of severe disease (OR: 0.99; P < 0.01). Preferences for single- vs double-dose formulations did not significantly differ (P > 0.01). Respondents with higher baseline altruism scores were more likely to prefer vaccination compared to those with lower altruism scores (RR: 1.83; P < 0.01). However, framing neither significantly affected preferences for vaccination nor modified the effect of baseline altruism on these preferences for vaccination. CONCLUSIONS: Preferences were strongest for vaccines with less severe side effects, suggesting that innovators should prioritize COVID-19 vaccines with these characteristics. More altruistic individuals were more likely to vaccinate, but framing did not modulate vaccination decisions, implying its limited nudging effects for vaccination. Using Real World Data to Assess Patient Outcomes
This session explores the use of real world data to assess patient outcomes drawing from examples in Herpes Zoster, Type-2 Diabetes, non-small cell lung cancer and psoriasis.
Laura Mccullagh, PhD
National Centre for Pharmacoeconomics, Dublin, Ireland
Dr Laura McCullagh is Head of Research, a Chief I Pharmacist and a member of the Senior Management Team within the National Centre for Pharmacoeconomics (NCPE), Ireland. Dr McCullagh is a lead health technology assessor within the NCPE. The NCPE is affiliated with Trinity College Dublin; roles here include various teaching and research commitments. Previous roles have included Associate Professor of Pharmacoeconomics & Health Technology Assessment, Trinity College Dublin, Senior Medicines Information Pharmacist (Oxford Radcliffe Trust, UK). Dr McCullagh holds a PhD (Trinity College Dublin) for research which focused on HTA within the Irish Healthcare Setting.
P67: Analysis of the Pharmaco-Utilization and Healthcare Costs in Pediatric Psoriatic Patients: A Real-World Retrospective Study Among the Italian Population
4:30PM - 4:45PM
1, Losi S 2, Sabatino S 3, Mezzetti M 4, Dovizio M 1, Veronesi C 1, Degli Espositi L 1 1CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, Italy, 2Eli Lilly Italy S.p.A, Sesto Fiorentino, Italy, 3Eli Lilly Italy S.p.A, Sesto Fiorentino (FI), Italy, 4Eli Lilly Italy S.p.A., Roma, Italy
OBJECTIVES: The objective of the study was to assess the profile of pediatric patients with psoriasis (PSO) in Italy, in terms of treatment patterns, pharmaco-utilization, and healthcare costs. METHODS: A retrospective study was conducted using the administrative databases of selected Italian health departments, covering around 22% of the population. Patients (<18 years) affected by PSO [diagnosed by at least one PSO hospitalization and/or an active PSO exemption code and/or topical antipsoriatics prescription] during 01/2010-10/2019, were included. The first occurrence of ≥1 inclusion criterion was considered index-date. During follow-up (≥12 months from index-date), the pharmaco-utilization (as treatment lines and discontinuation), and healthcare cost/patient [for drugs, outpatients specialistic services (OSS), and hospitalizations] were assessed in biological users under the age of 18 years during the follow-up. RESULTS: Overall, 7,077 patients were included (mean age 12.6 years, male 47%). During follow-up, 29.6% were prescribed systemic corticosteroids, 4.2% used conventional systemic therapies (cyclosporine/methotrexate/acitretin/dimethylfumarate), and 40.2% (index-date excluded) used topical antipsoriatics. Eighty-five patients (1.2%) had a biological prescription during follow-up, 61 of which were under 18 throughout follow-up. Among them, 42.6% were prescribed etanercept, 55.7% adalimumab, and the remaining ustekinumab. During available follow-up, 47.5% and 24.6% of biologic-treated patients maintained their therapy with adalimumab and etanercept, respectively. During the first-year follow-up, 39.7% of biological-treated patients interrupted (>90 days of treatment-gap) their treatment. Mean annual healthcare cost/patient for biological utilizers under 18 throughout follow-up(N=61) was 6,830€ (5,422€ for drugs, 986€ for OSS, and 422€ for hospitalizations). CONCLUSIONS: This Italian real-world study reported the pharmaco-utilization and costs for pediatric PSO patients in clinical practice. A low proportion of pediatric patients currently treated with biological therapies was found, with less than half of patients maintaining treatment. Furthermore, their management was associated with a high economic burden. The optimization of treatment for pediatric PSO patients is still an unmet need. P66: Epidemiology of Non-Small Cell Lung Cancer (NSCLC) by Histology and Disease Stage in Western Europe (WE): Population-Level Projections 2021-2026
4:15PM - 4:30PM
Kanas G 1, Kalilani L 2, Durbin L 3, Clark O 3, Nersesyan K 3, Keeven K 3, Giove TJ 4, Chao J 5, Aziez A 6, Stojadinovic A 7, Hogea C 8 1Kantar Health, Dublin, CA, USA, 2GlaxoSmithKline, Durham, NC, USA, 3Kantar Health, New York, NY, USA, 4GlaxoSmithKline, Mississauga, ON, Canada, 5GlaxoSmithKline, Collegeville, PA, USA, 6GlaxoSmithKline, Zug, Switzerland, 7GlaxoSmithKline, Upper Providence, PA, USA, 8GlaxoSmithKline, Philadelphia, PA, USA
OBJECTIVES: NSCLC accounts for approximately 85% of all lung cancers. Population-level projections are critical to further anticipate disease burden and address unmet needs. METHODS: We obtained the historic, annual age- and sex-specific incidence of lung cancer from country-specific registries in WE (France, Germany, Italy, Spain, and the United Kingdom). Histology data from the IARC’s Cancer Incidence in Five Continents (volume XI) were used to estimate total number of patients with NSCLC and by histology (non-squamous [NSQ] and squamous [SQ]). Assuming that observed trends will continue, the most recent country-, age-, and sex-specific incidence of NSCLC by histology were estimated and multiplied by the respective projected country populations to estimate annual number of cases from 2021 through 2026. Stage-specific distribution of NSCLC by histology according to the American Joint Committee on Cancer 8 th edition staging system from the Surveillance, Epidemiology, and End Results program for the year 2018 was applied. RESULTS: Total number of new NSQ NSCLC cases for the five countries is projected to increase from 162,789 in 2021 to 174,177 in 2026. NSQ stage IIIB/C increases from 9,735 to 10,417 and stage IV from 77,585 to 83,013 cases. The total number of new SQ NSCLC cases is projected to increase from 43,187 to 46,286 over the same period. SQ stage IIIB/C increases from 5,994 to 6,424 and stage IV from 13,423 to 14,386 cases. CONCLUSIONS: Newly diagnosed cases for advanced/metastatic NSCLC among all five WE countries is projected to rise 7.0% for NSQ and 7.2% for SQ NSCLC between 2021 and 2026, with most new cases being advanced/metastatic (stage IIIB/C-IV) patients. The 5-year overall survival rate for NSCLC remains ~30% for stage III and ~6% for stage IV. Coupled with projected increases in case numbers, this underlines the continued need for novel therapies and synergistic combinations to treat NSCLC. P68: Cardiovascular Events and Mortality in Type-2 Diabetic Patients under Second-Line Treatment with Hypoglycemic Agents: An Italian Real-World Study
4:45PM - 5:00PM
Degli Espositi L 1, Sangiorgi D 2, Nappi C 3, Andretta M 4, Barbieri A 5, Citraro R 6, Bartolini F 7, Cavaliere A 8, Ciaccia A 9, Chinellato A 10, Cillo MR 11, Dell'Orco S 12, Ferrante F 13, Gentile S 14, Procacci C 15, Re D 16, Ubertazzo L 17, Vercellone A 18, Perrone V 2 1CliCon S.r.l. Health, Economics & Outcomes Research, Ravenna, Italy, 2CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, BO, Italy, 3CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, Italy, 4Azienda ULSS 8 Berica, Vicenza, Italy, 5ASL Vercelli, Vercelli, Italy, 6Azienda ospedaliero-universitaria Mater Domini, Catanzaro, Italy, 7USL Umbria 2, Terni, Italy, 8ASL Viterbo, Viterbo, Italy, 9ASL Foggia, Foggia, Italy, 10Azienda ULSS 3 Serenissima, Mestre (VE), Italy, 11ASL Salerno, Salerno, Italy, 12ASL Roma 6, Albano Laziale, Italy, 13ASL Frosinone, Frosinone, Italy, 14Direzione Generale per la Salute Regione Molise, Campobasso, Italy, 15ASL BAT, Trani, Italy, 16ASL Teramo, Teramo, Italy, 17ASL Roma 4, Civitavecchia (RM), Italy, 18ASL Napoli 3 SUD, Torre del Greco, Italy
OBJECTIVES: To evaluate the incidence and risk of major adverse cardiovascular events (MACE) and all-cause mortality in type-2 diabetic patients with metformin failure in second-line treatment with hypoglycemic drugs, using real-world data in Italy. METHODS: A retrospective study was conducted using administrative databases from a sample of Italian Local Health Units. During 01/2015-12/2017, all adult type-2 diabetic patients who failed metformin therapy and in second-line with hypoglycemic drugs: sulfonylurea (SULF), dipeptidyl peptidase inhibitors (i-DPP-4), glucagon-like-peptide-1 (GLP-1), or inhibitors of the sodium/glucose-2 co-transporter (SGLT-2i), alone or combined with metformin, were included. Index-date corresponded to 1 stprescription date; patients were followed-up from index-date to end of 2019. The Propensity Score Matching (PSM) methodology was applied to abate covariates variability among cohorts. RESULTS: Among the 18,818 patients included (mean age 65.6 years, 57.5% male), 40.3% had SULF as second-line, 36.1% i-DPP-4, 15.3% SGLT-2i, 8.3% GLP-1. After PSM, the incidence rate per 100 person-years of MACE was 3.7 for SULF patients, 3.1 for i-DPP-4 patients, 2.7 for SGTL-2i patients, and 2.6 for GLP-1a patients. All-cause death incidence rate per 100 person-years was higher in SULF cohort (1.4) than DPP-4 (1.3), SGLT2-i (0.7) and GLP-1 (0.9) (p= 0.001). Compared to SULF patients, SGLT-2i and GLP-1 had a 26% (p-value=0.013) and 30%(p-value=0.003) lower risk of MACE events, respectively. SGLT-2i and GLP-1 patients were also associated to a lower risk (SGLT-2i:52%, p-value=0.001; GLP-1:40%, p-value=0.012) of all-cause death rather than SULF cohort. CONCLUSIONS: This study provides an evaluation from real-world data of the clinical profile of patients in second-line with hypoglycemic agents after metformin failure. Our findings showed that in matched cohorts, lower incidences and risks of MACE and all-cause death were observed among GLP-1 and SGLT2-i patients than in SULF ones. These data suggest that the evaluation of clinical outcomes with therapeutic interventions might support health-decision making, to also improve prescriptive appropriateness. P65: Health Care Resource Utilization and Potential Disease Deterioration After Herpes Zoster Incidence in Patients with Underlying Conditions: A Retrospective Cohort Study Based on German Claims Data, 2007-2018
4:00PM - 4:15PM
Witte J 1, Batram M 2, Schwarz M 3, Hain JJ 3, Ultsch B 4, Steinmann M 1, Bhavsar AB 5, Greiner W 1 1Department for Health Economics and Health Care Management, Bielefeld University, Bielefeld, Germany, 2Department for Econometrics, Bielefeld University, Bielefeld, Germany, 3GSK, Munich, Germany, 4GSK, Munich, BY, Germany, 5GSK- Vaccines, Wavre, WBR, Belgium
OBJECTIVES: This study aimed to investigate potential disease deterioration for underlying conditions (UCs: chronic obstructive pulmonary disease, diabetes mellitus type 1 or 2, depression, coronary heart disease, chronic heart failure (CHF), and rheumatoid arthritis) after an acute herpes zoster (HZ) using health care resource utilization (HCRU) and disease-specific worsening indicators. METHODS: Analyses were based on claims data, representing 13% of the German statutory health insurance population (corresponding to 87% of the entire German population). Patients aged ≥18 years with UCs were included when an incident HZ-diagnosis (defined by International Classification of Diseases and prescription of an antiviral drug) was observed between 2008-2016. Patients were matched to controls with the same UC but without HZ in the observational period using propensity scores. HCRU was analyzed for four quarters prior and eight quarters after acute HZ. Regression models were used to identify potential excess HCRU and disease-specific worsening indicators in the post-acute quarters. RESULTS: Over the observational period, HCRU data from 172,093 HZ patients, matched to 172,093 controls, were analyzed. Matching achieved almost perfect balance, as there was no significant difference for 12 out of 13 matching variables. Excess costs were observed in the index quarter of acute HZ for all UCs (range: €92-€259). A significant increase in post-acute excess costs was observed for patients with CHF only, amounting to 8.56% (152€) and 6.61% (114€) in the first and second quarter after HZ-incidence, respectively. Disease-specific worsening indicators were observed only for patients with depression who showed somatoform disorders excess diagnosis and excess antidepressant use in post-acute quarters. CONCLUSIONS: HZ-incidence is associated with a significant increase in excess HCRU in the index quarter of HZ-diagnosis for patients with UCs. The present results suggest that deterioration of UCs more than 6 months after incident HZ is relevant for patients with CHF and depression. Is Social Media Information Useful to Understand Patient Experiences and the Burden of Disease?
Social media is increasingly used as a source of information in health technology appraisals. This session illustrates how this data can be used to understand the patient journey, quality of life, unmet needs, and experiences with available treatments.
Lynda Doward, Master of Research
RTI Health Solutions, Manchester, United Kingdom
P34: Use of Social Media Listening (SML) Methods for Understanding the Patient Experience of Chronic Disease: A Scoping Review
4:15PM - 4:30PM
Trevisan F 1, Clifford S 2, Cooper V 2 1Sprout Health Solutions, Pinner, UK, 2Sprout Health Solutions, Los Angeles, CA, USA
OBJECTIVES: To conduct a scoping review of the literature related to the use of social media listening (SML) methods for understanding patients’ experiences of chronic disease. METHODS: Methodological guidance for conducting scoping reviews was followed. PubMed, Medline and EMBASE databases were searched using terms for social media listening, patient experience and quality of life. Primary research studies and systematic reviews that investigated the use of SML for understanding the patient experience of health or healthcare were included. Titles and abstracts were reviewed by two reviewers. One reviewer extracted data from full text articles. Quality of data extraction was assessed and verified by a second reviewer. A descriptive synthesis was performed with a focus on summarizing information related to the type of social media platforms, methods, data extraction tools and outcomes reported. RESULTS: The literature search identified 53 studies; 14 met the inclusion criteria and were included in the full-text review and data extraction. Social media platforms included Twitter, Facebook, HealthUnlocked, and disease-specific online communities. Methods used by researchers conducting SML included keyword frequency analysis, identification of patient reported terms for functional status, qualitative content analysis, identification of frequently asked questions and modelling of the patient experience. Outcomes included identification of symptoms experienced by and of importance to patients (including physical, psychological and cognitive symptoms), experience of medical treatment and procedures, social relationships and support, financial difficulties, information on the temporal patterns of symptoms experienced by patients and identification of racial and ethnic disparities in patient experience of disease. CONCLUSIONS: The findings highlight the potential value of SML as a method for capturing the patient voice and understanding patient experiences of chronic disease. SML has the potential to complement traditional methods, reduce patient burden and include the experience of people who may not take part in formal research studies. P33: Patient Experiences and Insights on Chronic Ocular Pain from a Social Media Listening Study
4:00PM - 4:15PM
1, Parashar N 2, Aasaithambi S 2, Verma H 2, O'Brien P 3, Sloesen B 4 1Novartis Healthcare Pvt. Ltd., Hyderabad, AP, India, 2Novartis Healthcare Pvt. Ltd., Hyderabad, India, 3Novartis Business Services Center, Limerick, Ireland, 4Novartis Pharma NV, Vilvoorde, Belgium
Objective: To identify the perceived causes for chronic ocular pain (COP) (≥3 months pain duration), its impact on quality of life (QoL) and understand the patient journey from social media posts. Methods: In this retrospective study, publicly available social media conversations were identified from searches triaged by a combination of automated relevancy keyword algorithm and manual review, and subsequently analyzed post anonymizing for COP content. Twitter, forums, and other (Facebook, Blogs, etc.) platforms were leveraged for the time period February 2020 to February 2021. Results: A total of 464 (UK=208, US=175, Canada=65 and Australia=16) patient/caregiver conversations on COP were identified. Top discussion points were symptoms (62%) and causes of COP (58%). Ocular factors (including dry eye disease, thyroid/Graves’ disease, and ocular surgeries) contributed to ~46% of causes identified, while non-ocular factors (including migraine, COVID, and side-effects/withdrawal of medications) contributed to ~54%. The most commonly mentioned symptoms (555) were headache/head pressures (96), dry/gritty eyes (67), light sensitivity (34), insomnia (29), and redness/pink eyes (28). Symptoms impacted all aspects of patients’ QoL: physical day-to-day activities such as reading, driving, and sleeping; emotional wellbeing such as depression/hopelessness, frustration/anger, fear, and suicidal thoughts; functional wellbeing such as difficulty at work/study place, reduced productivity or having to quit their job; social impacts such as being irritated around people, and having a less active social life. Eye drops (58/140 mentions) are the most commonly mentioned treatment option. Common coping strategies mentioned were blue-light filter glasses/eyeglasses (17), and hot compresses (11). Key unmet needs mentioned by patients were failed, improper, delayed diagnosis (62), and lack of effective treatments or appropriate management (30). Conclusion: Insights from this study reported patients’ experiences, concerns, and the adverse impact on overall QoL. The results can help in better understanding the patients’ perspective, which can be considered during drug development. P36: Unmet Needs of Caregivers in Locally Advanced or Metastatic Bladder Cancer from Social Media in the US
4:30PM - 4:45PM
1, Loussikian P 1, Marrel A 2, Barbier V 2, Foulquié P 1, Mebarki A 1, Schück S 1, Bharmal M 3 1Kap Code, Paris, France, 2Icon, Lyon, France, 3EMD Serono, Billerica, MA, USA
OBJECTIVES: Bladder cancer (BC) is the sixth most common cancer in the US; the prognosis for patients with locally advanced or metastatic BC is very poor. Few studies have assessed its burden on caregivers. This study aimed to characterize difficulties and unmet needs of caregivers for patients with locally advanced or metastatic BC as reported on social media. METHODS: US caregiver testimonials were collected from social media posts between January 2015 and April 2021 using specific terms for locally advanced or metastatic BC. These were qualitatively analyzed to identify caregiver difficulties and unmet need until saturation. RESULTS: Of 1214 testimonials from 679 caregivers on 72 social media sources, 423 were randomly selected and analyzed until saturation. From those that reported age-related data (<15% of testimonials), most caregivers were women (83.2%) with a mean age of 35.4 years, whereas the reported mean age of patients was 67.2 years. A total of 177 testimonials that expressed ≥1 caregiver- or patient-centered difficulty were identified and classified into a list of 36 types of challenges. The main difficulties related to the caregivers’ psychological impact throughout the patient journey (26%), the desire to share experiences among peers/support groups (15.8%), and the fear and management of patients’ adverse events (12.4%). Other major difficulties expressed included the specific psychological burden of end-of-life support or grief work (10.2%), the daily impact of being a caregiver (relocation, time consumption; 9.6%), stress due to screening and diagnostic delay (7.3%), and the change in relationships between patients and caregivers (5.1%). CONCLUSIONS: Qualitative analysis of social media testimonials from caregivers of patients with BC in the US provided insights on the substantial psychological impact and burden of care on them. Future research may explore BC caregiver well-being and quality of life as outcomes in quantitative studies. Informing the Decision-Making Process in Real Time
This session illustrates the potential of real world data information to inform the decision-making process.
Michael Barry, MB, FRCPI, PhD
National Centre for Pharmacoeconomics, Dublin, Ireland
P29: Reimbursement Outcomes for Combination Therapies vs Monotherapies in Lung Cancer and Multiple Myeloma in the Top Five European Markets
4:00PM - 4:15PM
Izmirlieva MA 1, Reinaud F 2, Taiyeb M 3, Ando G 1 1GlobalData, London, UK, 2GlobalData, Paris, France, 3GlobalData, Bangalore, India
Objectives: Theoretically, combination therapies would face greater difficulty in demonstrating cost effectiveness because the backbone therapy is often priced close to the relevant country’s cost-effectiveness threshold. Unless the backbone therapy’s cost is reduced, the combination may not be cost-effective even if the add-on therapy is priced at zero. We set out to verify it this is true in practice by assessing reimbursement outcomes for combination therapies vs monotherapies in lung cancer and multiple myeloma. Methods: Reimbursement status and level of reimbursement for all drugs in lung cancer and multiple myeloma, which were first priced between 1 January 2011 and 31 December 2020, were assessed in France, Germany, Italy, Spain and the United Kingdom using data from the IHS Markit POLI database. The reimbursement status review was supplemented by Amélioration du Service Médical Rendu (ASMR) ratings in France, Federal Joint Committee (G-BA) ratings in Germany and NICE guidance in the UK to assess the likely pressure on prices for those combination therapies that gained reimbursement. Results: In lung cancer combination therapies were more likely to be rejected for reimbursement compared to monotherapies. Across the five countries, 20 out of the 56 combination therapy presentations (equivalent to 35.7%) were rejected for reimbursement compared to 11.4% (31 out of 271) for monotherapy presentations. In multiple myeloma, 5.3% of combination therapy presentations (7 out of 132) were rejected for reimbursement, while every single monotherapy was approved for reimbursement. Combination therapies also had less favourable ASMR and G-BA ratings. Conclusions: This review of reimbursement decisions and cost-effectiveness assessment outcomes for drugs approved over a 10-year period in the top five European markets confirms that combination therapies in lung cancer and multiple myeloma face greater difficulty in demonstrating cost-effectiveness compared to monotherapies. Even when approved for reimbursement, combination therapies are subject to greater pressure on prices. P31: Impact of COVID-19 on HTA/PRMA of Medicinal Products in Europe: A Payer Perspective
4:30PM - 4:45PM
1, Dellamano R 2, Lobb W 1, Dalal N 3, Dellamano L 2, Pereira E 1 1Medical Marketing Economics LLC (MME), Montclair, NJ, USA, 2ValueVector, Milan, Italy, 3Medical Marketing Economics LLC (MME), St Albans, NJ, USA
Assess payer perceptions of COVID-19 pandemic’s impact on health systems, focusing on HTA, pricing, reimbursement and market access (PRMA) of new, branded medicines in the EU4 and UK.
In June 2021, MME Advisors conducted a virtual, national payer / advisor board with representatives from France (2), Germany (2), Italy (1), Spain (1), and the UK (2) - to discuss key topics within the pandemic’s context, such as:
Disruption to healthcare systems HTA impact: backlog, re-prioritization, framework PRMA impact: net price pressure, conditional pricing/RWE and time to market Differences and similarities within oncology, rare diseases, ATMPs and general medicines RESULTS:
Unlike the significant disruptions seen during the height of the pandemic in 2020, payers saw impact ranging from moderate (Italy) to high (Spain) as of June 2021. Disruption by disease state varied: oncology was highly disrupted everywhere but Germany. Most payers did not anticipate shifts in long term priorities or budget cuts to healthcare post pandemic. HTA impact was minimal, with no need to re-prioritize by therapy area or alter plans to adjust frameworks. Likelihood of stricter HTA criteria varied with payers in Italy anticipating more scrutiny for oncology and in Germany for rare diseases/ATMPs. While time to market was expected to remain mostly stable, delays anticipated in Spain. Majority of payers anticipated increasing pressure on drugs’ net prices; however, they were divided on increases in conditional pricing/RWE.
Perceived COVID-19 impact varied by country based on infrastructure and adaptability. Germany less impacted, whereas in other markets (e.g., Spain) COVID-19 seemed to have accelerated changes, rather than drive PRMA policy. Given the importance of healthcare, overall budget cuts were not anticipated, although the need to deploy funds to diverse areas (e.g., healthcare worker salaries, hospital capacity) could complicate future scenarios, especially for high-cost therapies. Therefore, continued monitoring is warranted.
P30: Potential Impacts of the New MHRA Policy for Biosimilar Approval for the Industry and Patients
4:15PM - 4:30PM
, Walker A, Walsh K Ribeiro A Lifescience Dynamics Ltd, London, UK
OBJECTIVES: To understand the potential impact of the new MHRA guidance for biosimilar licensing in the UK and other key markets, in terms of accelerating biosimilar development, time to market, prescribing limits (e.g. automatic substitution) and, patient access to biologics. METHODS: We reviewed the MHRA guidance on biosimilar licensing, alongside that from the FDA/EMA, and country-specific guidance on biosimilar use. We also analysed FDA/EMA approvals for biosimilars approved without confirmatory efficacy trials. Finally, a virtual iAdBoard was organised with payers/KOLs from France, the UK and US to capture different perceptions on the new policy and downstream impacts on biosimilar access. RESULTS: The MHRA has discontinued the requirement for biosimilars to undergo confirmatory efficacy trials as a licensing condition. Although the new policy was celebrated by the biosimilar industry, one must note the FDA and EMA do not explicitly state a Phase 3 trial requirement for biosimilar approval, and to date, two pegfilgrastim biosimilars have been approved without a Phase 3 trial (Udenyca, Nyvepria), given their robust chemical characterization and Phase 1 trials’ results. Additionally, the new MHRA guidance contemplates exceptions where comparative trials are required, leaving uncertainty around for how many biosimilars, particularly monoclonal antibodies, chemical characterization plus PK/PD trials will suffice. The virtual iAdboard revealed payer differences in opinion regarding impacts on biosimilar development timelines (vs agreement on economic viability), and in future policies encouraging biosimilar uptake, with EU payers more receptive to the change than US counterparts, but concerned with backlash from HCPs. CONCLUSIONS: If the FDA/EMA endorse the MHRA decision, in the future, a strong CMC/Phase 1 package could replace Phase 3 studies for biosimilar licensing. However, it is yet unclear whether this abbreviated data package will result in faster times to market or if it will have negative impacts on prescribing freedom and patient access to biosimilars. P32: Price Analysis of Cancer Therapies for the Treatment of Patients with Unresectable Hepatocellular Carcinoma
4:45PM - 5:00PM
Williams A 1, Anderson N 2, Gwon YG 3, Wifler W 2 1Boston Scientific, Marlborough, MA, USA, 2Boston Scientific, Maple Grove, MN, USA, 3Boston Scientific, Kuala Lumpur, MN, Malaysia
OBJECTIVES: The price of cancer therapy for the treatment of adult US patients with hepatocellular carcinoma (HCC) remains unknown. This study estimated the price of systemic therapies (ST) compared to selective internal radiotherapy (SIRT) for the treatment of HCC from the payer and provider perspectives. METHODS: The National Comprehensive Cancer Network (NCCN) Guidelines were used as a framework to model the treatment strategies for HCC. The associated drug prices as of May 2021 for ST and SIRT were obtained from the IBM Micromedex Redbook (Average Wholesale Price [AWP]), (Wholesale Acquisition Cost [WAC]), Medicare’s Average Sale Price (ASP), and Decision Resources Group (DRG) ASP. Clinical parameters, such as treatment duration and FDA-recommended daily dose (DDD), were obtained from randomized controlled trials and FDA-approved labels. The total price/DDD was calculated for each treatment therapy and treatment duration over a short-term (<12 months) horizon. Sensitivity analysis was conducted to explore the impact of treatment duration uncertainty on model results. Because drug rebates are unknown, these price estimates did not account for drug rebates or patient assistance programs negotiated directly with manufacturers by Pharmacy Benefit Managers. RESULTS: 11 STs and 3 SIRTs were included in our analysis. The median price/DDD of ST varied by perspective: Medicare ASP: $97,466 (IQR: $341-$205,393); Provider WAC: $123,322 (IQR: $18,475-$305,615); Provider AWP: $186,389 (IQR: $22,170-$366,738). The median price for SIRT was estimated; Medicare ASP: $21,877 (IQR: $21,877-$22,269) and Provider ASP: $21,873 (IQR: $21,316–$21,873). The price differences are greater than SIRT when considering patients who progress through first-line and second-line ST. CONCLUSIONS: The price of cancer therapy for HCC varies widely by payer-provider perspective. The availability of alternative cancer therapies, such as locoregional (non-surgical) approaches, may offer clinical meaningful benefit and reduce the total costs of HCC care from the payer-provider perspectives. Development and Measurement of Health Utilities
This session presents interesting new research in health utility estimation. Studies include the estimation of the EQ-5D-5L value set for Italy, an exploration of the association between EQ-5D utility and migraine frequency, and a vignette valuation study for eye disease. The association between health state utility and work productivity is also discussed, informing the debate about the extent to which wider societal benefits of health technologies are associated with health utilities.
Sorrel Wolowacz, PhD
RTI Health Solutions, Manchester, United Kingdom
Sorrel Wolowacz, PhD, is Head of European Health Economics at RTI-HS, with 20 years of experience in health economics research and consulting. Her research focuses primarily on economic modelling, health utility estimation, observational studies, and health technology appraisal submissions. Dr. Wolowacz is a member of the editorial board for the Journal of Comparative Effectiveness Research and was co-chair of the ISPOR Good Research Practices Task Force addressing Measurement of Health State Utility Values for Economic Models in Clinical Studies and for the ISPOR Working Group for Oncology Health Economic Modeling.
P10: Evaluating the Correlation between Monthly Migraine Days and Quality-of-Life: Utility Analyses to Inform a Japanese Cost-Effectiveness Model for Fremanezumab in Migraine
4:00PM - 4:15PM
Peterse E 1, Bennison C 2, Paris J 3, Chatterjee A 2, Wang X 4, Kojima Y 4, Yamato K 4 1OPEN Health, Rotterdam, Netherlands, 2OPEN Health, York, UK, 3OPEN Health, Kent, KEN, UK, 4Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan
OBJECTIVES : A cost-effectiveness model (CEM) for fremanezumab in migraine prevention from a Japanese public healthcare payer perspective has been developed. To inform health state specific utilities, we analyzed the correlation between the number of monthly migraine days (MMD) and a patient’s quality-of-life using data from Japanese-Korean trials. METHODS : The health states in the CEM are defined by the number of MMD, ranging from 0 to 28. Data from three Japanese-Korean clinical trials (406-102-00001, 406-102-00002, 406-102-00003) was analyzed. MSQoL (migraine specific quality of life) values measured in the trials were mapped to EQ-5D-3L utility values using a previously published mapping algorithm. To account for the repeated nature of the data, linear mixed effects models were fitted to the EQ-5D-3L values. MMD, MMD at baseline, treatment arm (monthly injection, quarterly injection, placebo), scheduled visits (month 1, month 2 etc.), age, sex, prior migraine medication and country were explored as covariates. The final model was selected based on the Akaike information criterion (AIC) value using forward and backward selection. RESULTS : In total, 3743 utility values from 970 patients were included in our analyses. The mean observed utility value was 0.83 for patients with 0 MMD and 0.51 for patients with 28 MMD. Fremanezumab decreased the number of MMD, thereby increasing a patient’s quality-of-life. The variables MMD, baseline MMD, scheduled visits and country were included in the final model. The regression coefficient for MMD was -0.01 (p<0.001), demonstrating that, after adjusting for baseline MMD, schedule visits and country, utility decreased by 0.01 for every day increase in MMD. CONCLUSIONS : There was a strong correlation between the number of MMD and quality-of-life in patients with migraine. Estimates derived from the linear mixed-effects model can be used to inform health-state specific utilities in the Japanese cost-effectiveness model for fremanezumab in migraine prevention. P12: Are Gains in Health Utility Associated with Gains in Work Productivity and Role Functioning in Chronic Diseases? A Systematic Literature Review
4:30PM - 4:45PM
Aggio D 1, Williams A 2, McNamara L 3, Lloyd A 1 1Acaster Lloyd Consulting Ltd., London, UK, 2Kyowa Kirin International, Marlow, BKM, UK, 3Kyowa Kirin Ltd., Galashiels, UK
OBJECTIVES: Disease experience for people living with chronic diseases has changed dramatically with improvements in health utility. It remains unclear, however, the extent to which improvements in health utility leads to gains in work productivity and role functioning. This systematic literature review aimed to explore the relationship between health utility and work productivity or role functioning across chronic diseases. METHODS: Diseases selected were chronic and severe (based on health utility weights in range 0.50 to 0.70). Records from a structured search conducted in MEDLINE, Embase and PsycINFO were reviewed against inclusion criteria and assessed for study quality/relevance. Articles published from 2000 – February 2021 and available in English were considered. Studies included a measure of health utility (e.g., EQ-5D) and productivity or role function (e.g., employment status, presenteeism and absenteeism). Study quality was assessed in terms of design, analysis approach, missing data and evidence of bias. RESULTS: The search identified 876 records; 244 underwent full review, and 34 of the highest quality studies were extracted. Only 4 longitudinal studies were identified. Studies included different diseases including multiple sclerosis, rheumatoid arthritis, and stroke. Weighted mean health utilities of 0.79 were observed for employed (full/part time) people with a chronic disease, compared with 0.71 for part time employed, 0.61 for those unemployed/not in work, and 0.62 for those incapable of working. These associations held in studies controlling for potential confounders (e.g., age, symptom severity etc). Values corresponded to approximately a 5% increase in employment per 0.1 unit increase in health utility value. CONCLUSIONS: There is limited longitudinal research among people with chronic diseases exploring how changes in health utility may lead to changes in work productivity and role functioning. However, the findings suggest that amongst people with a chronic and severe disease, better health states are expected to be associated with higher productivity. P11: Development of an EQ-5D-5L Value Set for Italy Using Videoconferencing Administered Personal Interviews: Reporting on the Feasibility of a New Mode of Administration for Valuation Studies
4:15PM - 4:30PM
Finch AP 1, Meregaglia M 2, Ciani O 2, Roudijk B 3, Fattore G 4, Jommi C 5 1EuroQol Research Foundation, Amsterdam, Netherlands, 2SDA Bocconi School of Management, Milan, Italy, 3EuroQoL Research Group, Rotterdam, Netherlands, 4Bocconi University, Milano, Italy, 5SDA Bocconi School of Management, Bocconi University, Milano, Italy
OBJECTIVES: To derive an Italian value set for the EQ-5D-5L using videoconferencing interviews and to determine the feasibility of this mode of administration. METHODS: Preferences were collected using the EQ-VT V2. Two valuation methods were employed, composite time trade-off (cTTO) and discrete choice experiment (DCE). The target sample size was 1,000 participants. Participants were recruited using a market research company with experience in quantitative and qualitative data collection. Videoconferencing administered interviews were conducted by 11 interviewers selected among PhD students, researchers, and other academic affiliates. A pilot of 199 interviews was employed to assess the technical, operational and protocol feasibility of videoconferencing interviews. Standard QC parameters were used to monitor interviewers’ performance during the data collection. To inform the modelling choices, GLS, Tobit, Logit, Probit and Hybrid models were fitted to the data, with different error specifications. Models were compared in terms of monotonicity of coefficients, statistical significance, and theoretical considerations. RESULTS: 1182 videoconferencing interviews were completed between October 2020 and February 2021, including 199 feasibility pilot interviews. Dropouts and technical problems occurred in less than 5% of the interviews, and all interviewers complied with the protocol as well as showing significant improvements in QC parameters. The results suggested videoconferencing was a feasible mode of administration. The final sample was representative of the Italian general population for age, gender, and education as recorded in 2019 by ISTAT. Among the models tested, the Hybrid Tobit heteroscedastic model without constant was selected for the derivation of the tariff. In the selected model, coefficients for all dimensions levels were statistically significant and monotonically decreasing. Values ranged from -0.571 for the PITS state to 1 for health state 11111. CONCLUSIONS: An Italian societal value set for the EQ-5D-5L was developed. This can be used for economic evaluations and decision making in Italy. Videoconferencing appeared feasible for valuation interviews. Two Years of COVID-19: What Has Been the Global Impact on HRQoL and Clinical Outcomes?
This session illustrates recent evidence about the impact of COVID19 on patients' health related quality of life and clinical outcomes.
Carla DeMuro, MS
RTI Health Solutions, Research Triangle Park, NC, USA
P61: COVID-19 Concerns Experienced by Pregnant and Postpartum Women and Their Influence on Health-Related Quality-of-Life
4:00PM - 4:15PM
Aytha Swathi P 1, Regan A 1, Grinshteyn E 1, Nosek M 1, Gu NY 2 1University of San Francisco, Sacramento, CA, USA, 2NYG Technologies, Santa Clarita, CA, USA
OBJECTIVES : Pregnant and postpartum women may be particularly susceptible to pandemic-induced anxiety/depression, which can adversely affect maternal and infant health. This study investigates the COVID-19 concerns experienced by pregnant and postpartum women and their Influence on health-related quality-of-life (HRQoL). METHODS: We conducted an online, national US survey (EuroQol grant: 260-2020RA) between May and June 2021. Respondents completed the EQ-5D-5L instrument and rated their level of concerns for their own health, their baby’s health, and their family’s health, being pregnant and giving birth during the pandemic. Women indicated whether they strongly agreed, agreed, neither agreed nor disagreed, disagreed or strongly disagreed with each concern. Respondents who indicated they “strongly agreed” were classified as having strong concerns. We used median regression to estimate the EQ-5D-5L utility and EQ-VAS scores by level of maternal concerns. RESULTS: Among 2,070 respondents, pregnant and postpartum women commonly expressed strong concerns about giving birth during the pandemic (44%; 95% CI 39%, 48%), the health of their baby (44%; 95% CI 40%, 48%), the health of their family (38%; 95% CI 34%, 42%), and being pregnant during the pandemic (38%; 95% CI 34%, 42%). Fewer respondents expressed strong concerns about their own health (27%; 95% CI 23%, 31%). Overall, there was no association between HRQoL measures and maternal concerns during COVID-19. Among women who gave birth during 2020 (n=536), each unit increase in concerns about being pregnant was associated a 0.02 decline in the EQ-5D-5L utility (95% CI -0.04, -0.01). No difference was observed in EQ5D-VAS scores (ß =0.00; 95% CI -0.01, 0.01). No other differences in HRQoL were observed. CONCLUSIONS: Although overall, there was no relationship between maternal concerns during the COVID-19 pandemic and HRQoL, we did observe small declines in HRQoL associated with concern of being pregnant during the pandemic among women pregnant early in the pandemic. P64: Evaluating the Impact of the COVID-19 Pandemic on Mortality after Myocardial Infarctions Hospitalization in Germany
4:45PM - 5:00PM
Krieger J 1, Hardtstock F 2, Wilke T 3, Maywald U 4 1Cytel Inc, Berlin, Germany, 2Cytel Inc, Wismar, Germany, 3IPAM e.V., Wismar, Germany, 4AOK PLUS, Dresden, Germany
OBJECTIVES: It was hypothesized that COVID-19 lockdown measures led to later admission of myocardial infarction (MI) patients to hospitals and, consequently, higher average case severity and mortality. The aim of this study was to compare MI-associated mortality between COVID-19 lockdown and pre-COVID-19 periods. METHODS: We used German claims data from continuously insured adults hospitalized with a MI (ICD-10 I21), and compared 30-days mortality for cases in March-May 2020 (first COVID lockdown in Germany) with March to May in 2017-2019. Multivariable logistic regression models were conducted to test for differences in mortality between pre-COVID and COVID months while controlling for patients’ age, sex, previous MIs (2-years baseline), and cardiovascular comorbidities. RESULTS: In 2020, we observed 758/612/712 MIs in March/April/May, which was fewer MIs than the average for the same months between 2017-2019 (March: 901; April: 716; May: 853). Over the observational years, there was a shift towards younger patients (average age 2017: 77; 2020: 74), and the proportion of women decreased (2017: 42.54%; 2020: 39.19%). The length of the index-hospitalization was significantly lower during the COVID-19 pandemic (March-May 2017-2019: 9.19 days; March-May 2020: 8.11 days; p<0.001). Furthermore, the number of deceased patients was lower during the COVID lockdown period (March 2017-2019: 16.53% vs. March 2020: 14.12%; April: 17.19% vs. 16.99%; May: 16.72% vs. 13.48%). However, regression models showed no significant difference between COVID and pre-COVID months except for May (OR [COVID vs. pre-COVID]: 0.73, p=0.014). CONCLUSIONS: Even if we cannot rule out the possibility of a higher MI-related mortality during COVID lockdown periods in non-hospitalized MI patients, we cannot confirm the hypothesis that hospitalized patients generally showed a higher mortality due to later admission to hospitals and thus more higher case severity at time of hospital admission. P62: The Impact of the COVID-19 Pandemic on the Quality of Life of Patients with Melanoma – Findings from a UK Melanoma Registry
4:15PM - 4:30PM
1, Mulgina D 1, Kudlac A 1, Ouyang C 1, Larkin M 2, Ofori A 3 1Vitaccess, Oxford, LON, UK, 2Vitaccess, London, LON, UK, 3Vitaccess, London, UK
OBJECTIVES: The Melanoma UK digital registry, launched in collaboration with Melanoma UK in 2017, captures real-world patient-reported quality of life (QoL), symptoms, and side effects of melanoma treatment using a mobile application. 729 UK patients with melanoma have since joined the registry, with approximately 60% of participants in Stage 3 and 4 of the disease. Due to the COVID-19 pandemic, a national lockdown was announced in the UK in March 2020 with an instruction to stay at home, particularly for vulnerable people with health conditions. The impact of the fundamental changes in day-to-day living and the delivery of healthcare on people diagnosed with melanoma needs to be better understood. This study explored if the COVID-19 pandemic was associated with changes in Melanoma UK registry participants’ QoL. METHODS: QoL data, measured using EQ-5D and QLQ-C30 from 423 participants, were analysed. Data submitted between 03/2019 and 02/2020 were deemed pre-COVID-19 data; data submitted between 03/2020 and 03/2021 were considered as peri-pandemic data. A monthly average score was calculated for each domain of EQ-5D and QLQ-C30. T-tests were conducted to compare the individual domain scores calculated for the pre- and peri-pandemic periods. RESULTS: EQ-5D: Mobility ( p =0.03), self-care ( p =0.03), usual activities ( p =0.003), pain symptoms ( p =0.03), and health in general ( p =0.0004) improved during the pandemic. QLQ-C30: There was an observed improvement in physical ( p =0.02), role ( p =0.01), and social functioning ( p =0.003). Fatigue ( p =0.002) and pain symptoms ( p =0.05) decreased during the pandemic. Participants also reported experiencing less financial difficulty ( p =0.02) in comparison with the pre-pandemic period. No change in emotional well-being was observed. CONCLUSIONS: The QoL of patients with melanoma improved during the COVID-19 pandemic. Future research should involve qualitative interviews with melanoma patients and their caregivers to explore the mechanisms of this change, the role of support networks, and the impact on caregivers. P63: Impact of COVID-19 on Health-Related Quality-of-Life in the United States, Sweden and Norway: A Cross-Country Comparison Using a Panel Survey
4:30PM - 4:45PM
Chen J 1, Gong CL 2, Jiao X 1, Zawadzki NK 1, Persson U 3, Hay JW 1, Gu NY 4 1University of Southern California, Los Angeles, CA, USA, 2Children's Hospital Los Angeles, Los Angeles, CA, USA, 3The Swedish Institute for Health Economics (IHE), Lund, Sweden, 4NYG Technologies, Santa Clarita, CA, USA
OBJECTIVES: To assess and compare the impact of the COVID-19 pandemic on health-related quality-of-life (HRQoL) in the United States, Sweden and Norway. METHODS: Two waves of web-based survey were conducted in April 2020 and January 2021 to collect demographic data, COVID-19 status, behavior and employment changes related to COVID-19 in each country (EuroQol Grant: 246-2020RA). EQ-5D-5L was used to assess health status of respondents. Results were compared between the two waves to measure changes in HRQoL. One-way ANOVA was used to detect significant differences between countries, and t-tests for differences between waves. RESULTS: We collected 2,734, 1,003 and 1,020 responses in Wave 1, and 2,252, 1,013 and 1,011 responses in Wave 2 for the US, Sweden, and Norway respectively. Corresponding mean (SD) EQ-VAS scores were 74.6 (±19.2), 68.7 (±21.4), and 69.2 (±20.8) in Wave 1 (p<0.001), and 76.4 (±18.6), 68.2 (±20.3), and 67.8 (±21.7) in Wave 2 (p<0.001). Between waves, only the VAS scores in the US were significantly different (p<0.001). Mean (SD) utility scores were 0.822 (±0.222), 0.768 (±0.260), and 0.808 (±0.248) in wave 1 (p<0.001), and 0.823 (±0.221), 0.783 (±0.237), and 0.777 (±0.271) in wave 2 (p<0.001); there were no significant differences between waves for all three countries. Anxiety/depression was consistently the most problematic EQ-5D-5L subdomain among Swedish and Americans (>50%), followed by pain/discomfort. >45% Norwegians also reported problems in anxiety/depression subdomain in both waves. The proportions reporting problems in anxiety/depression increased in wave 2 for Sweden and Norway, but decreased for the US. CONCLUSIONS: Population HRQoL in Sweden and Norway has been similar throughout the pandemic, while a rebound in population mean VAS was observed in the US. However, the large proportions reporting problems in anxiety/depression across waves in all 3 countries indicates that mental health issues resulting from the pandemic are a major concern. The EQ-5D-5L in Practice in Europe and Beyond: Advantages and Limitations
This session explores the pros and cons of assessing people's quality of life using the EQ-5D-5L scores. It shows to what extend it can provide valuable insights of the quality of life in the general population, but has limitations to assess quality of life in Alzheimer's disease. It shows how online platforms can be used for valuing health states, and its current use for clinical benefit assessment.
Aureliano Paolo Finch, PhD
EuroQol Research Foundation, Amsterdam, Netherlands
Aureliano Finch is a scientist at the EuroQol Research Foundation, a member of the Descriptive System working group and coordinator of the EQ family of instruments valuation studies. His research mainly relates to the development and testing of preference based instruments of health and wellbeing. He applies both quantitative and qualitative methods for this purpose.
P59: To What Extent Is EQ-5D Used as a Tool for Clinical Outcome Assessment?
4:30PM - 4:45PM
Shaw C 1, Longworth L 1, Bennett B 2, Ruane P 3, Watson C 1, Francis LE 1, Shaw J 4 1PHMR Ltd., London, UK, 2Bristol-Myers Squibb, Uxbridge, UK, 3PHMR Ltd., Newcastle upon Tyne, NBL, UK, 4Bristol-Myers Squibb, Lawrenceville, NJ, USA
OBJECTIVES: The EQ-5D is widely used to inform economic evaluations of health technologies. However, the extent of its use for clinical outcome assessment (COA) is unclear. This review identified the prevalence with which EQ-5D data are evaluated by health authorities in clinical benefit assessments. METHODS: Drug technology assessments (TAs) published by HTA agencies in England, France, Germany and the US during the last 2 years were identified. Product labelling for drugs approved by the European Medicines Agency (EMA) and US Food and Drug Administration (FDA) over the last 5 years were also reviewed. Only documents reporting EQ-5D as a COA measure were included. RESULTS: EQ-5D data were reported for COA in 139 TAs with the majority reported for Germany (n=78) and the remainder for England (n=46), France (n=12) and the US (n=3). Visual analogue scale (VAS) scores were presented most frequently (n=111) followed by utility index scores (n=48) and dimension levels (n=1). The VAS accounted for 99% of EQ-5D reports in Germany. Minimally important differences (MIDs) were discussed in 51 TAs: 34% and 24% of VAS and index score reports, respectively. Three-hundred twenty drugs were approved by the EMA and 735 by the FDA, and among these 15 and 35, respectively, presented EQ-5D data for COA. All EQ-5D data submitted to the FDA were reported in supporting documentation. Index scores, VAS scores and dimension levels were cited for 32, 26 and 5 drugs, respectively. Discussion of MIDs was more frequent in EMA documents (35%) than FDA documents (11%). CONCLUSIONS: The EQ-5D has been used for COA in HTA submissions; most frequently the VAS in German TAs. EQ-5D was also used to support labelling claims in a minority of EMA and FDA decisions. No EQ-5D data were reported in FDA product labelling, which suggests that the data were not considered material. P57: Mapping Quality of Life - Alzheimer's Disease (QOL-AD) Scores to EQ-5D-5L Utilities in the Ambar Trial Population
4:00PM - 4:15PM
1, Gomez-Ulloa D 2, Runken MC 3, Serrano D 4, Barnes B 4, Grifols C 2, Barcelo M 2, Podger L 5 1OPEN Health Group, York, NYK, UK, 2Grifols, Sant Cugat del Vallès, Spain, 3Grifols SSNA, Research Triangle Park, NC, USA, 4OPEN Health Group, Bethesda, MD, USA, 5OPEN Health Group, London, UK
OBJECTIVES: Health state utilities are required to model the cost/benefit of novel treatments in accordance with Health Technology Assessment guidelines. The AMBAR phase 2b/3 trial (NCT01561053) evaluated the efficacy and safety of plasma exchange with albumin replacement in mild-to-moderate Alzheimer’s disease (AD) patients. Health-related quality of life was measured through the disease-specific Quality of Life - Alzheimer’s disease (QoL-AD) questionnaire. EQ-5D-5L utilities were estimated from AMBAR QoL-AD data using a validated mapping algorithm developed by Rombach et al. (2020). METHODS: A multinomial logistic regression model was employed to generate mapped utility values for the placebo and pooled treatment arms for both patient and proxy-rated QoL-AD scores. The model included age and sex as covariates and was run including QoL-AD item 7 (marriage). The UK value set (crosswalk to EQ-5D-3L value set by van Hout et al. 2012) was applied. RESULTS: The analysis sample included 298 patients and 298 caregivers with complete baseline QoL-AD data (i.e., all 13-items completed). The mean mapped patient-rated EQ-5D-5L utility at final visit was 0.8132 for placebo and 0.8312 for pooled treatment. For proxy-raters, the mean mapped EQ-5D-5L utility at final visit was 0.7333 for placebo and 0.7235 for pooled treatment. The correlation between observed QoL-AD scores and mapped EQ-5D-5L utilities was 0.526 for patients and 0.552 for proxies, in line with the findings in Rombach et al. (2020). CONCLUSIONS: Mapped EQ-5D-5L utilities were estimated for the AMBAR population. The estimates were relatively consistent at final visit across treatment arms. However, the correlation between observed QoL-AD scores and mapped EQ-5D-5L scores remained moderate. This may indicate a general lack of conceptual overlap in the domains covered by the QoL-AD and the EQ-5D-5L, limiting the precision of a mapping approach and reinforcing the benefit of directly collecting utilities in AD trials for increased accuracy of estimates. P60: A New Online Tool for Valuing Health States: Eliciting Personal Utility Functions for the EQ-5D-5L
4:45PM - 5:00PM
Schneider P 1, van Hout B 1, Heisen M 2, Brazier JE 1, Devlin N 3 1University of Sheffield, Sheffield, UK, 2Pharmerit International, Rotterdam, Netherlands, 3University of Melbourne, Melbourne, Australia
OBJECTIVES: Standard health valuation methods, such as TTO or DCE are inefficient. They require data from hundreds if not thousands of participants to generate utility values (=value sets) for health descriptive systems. Here, we present the Online elicitation of Personal Utility Functions (OPUF) tool; a new type of online survey for valuing EQ-5D-5L health states using more efficient, compositional preference elicitation methods, which allows estimating value sets on the individual level. The objectives of this study are to report on the development of the tool, and to test the feasibility of using it to obtain individual-level value sets for the EQ-5D-5L.
METHODS: We used an iterative design approach to adapt the PUF method, previously proposed by Devlin et al., for the EQ-5D-5L and for use as a standalone online tool. The valuation consists of three steps: level rating, dimension weighting, and anchoring. We conducted three iterative rounds of qualitative interviews to get feedback on the tasks and then piloted the tool in a sample of 50 participants from the UK.
RESULTS: On average, it took participants about 10 minutes to go through all tasks. Their responses indicated a good level of engagement. The two most important EQ-5D dimensions were Pain/Discomfort and Mobility. For 46 (92%) participants, we were able to construct a personal utility function. The results revealed that health state preferences differ considerably between individuals. CONCLUSIONS: We successfully piloted the OPUF tool and showed that it can be used to derive a social as well as personal utility functions for the EQ-5D-5L. Even though the development of the online tool is in an early stage, there are potential avenues for further research. With some further abstraction, the OPUF tool could provide a modular software platform for creating valuation tools for any health descriptive system. P58: EQ-5D-5L Population Norms for Italy
4:15PM - 4:30PM
Meregaglia M 1, Finch AP 2, Malandrini F 1, Ciani O 1, Jommi C 3 1SDA Bocconi School of Management, Milan, Italy, 2EuroQol Research Foundation, Amsterdam, Netherlands, 3SDA Bocconi School of Management, Bocconi University, Milano, Italy
OBJECTIVES: The EQ-5D is a widely used instrument to measure patient-reported outcomes and health state utility values. This study aimed to report normative data for the EQ-5D-5L questionnaire in Italy based on a nationally representative sample. METHODS: This study is part of the EQ-5D-5L Italian valuation study. The target participants were a sample of the Italian adult population (aged 18 and above) and were recruited through a market research company. Eleven trained interviewers conducted one-to-one interviews using a videoconferencing software (Zoom) and a survey web application (LimeSurvey) between November 2020 and February 2021. The distribution of answers was estimated for the descriptive system of the EQ-5D-5L, and descriptive statistics were calculated for the visual analogue scale (EQ-VAS). Data analysis was performed using Stata (StataCorp). RESULTS: The sample was composed of 1,182 people and fully represented the Italian population (2020) in terms of age, gender, and geographical area. Mean age was 48.3; men were 48.7%. Half of the participants selected the two most common health states (i.e., ‘11111’ and ‘11112’). Mean VAS was 81.8, and steadily decreased with increasing age (from 87.0 in the 18-24 group to 75.1 among the over 75). In participants affected by chronic illness (39%), mean VAS dropped to 75.5. At least one problem (from slight to extreme) was reported by 12.1% of respondents for mobility, 4.2% for selfcare, 11.6% for usual activity, 43.3% for pain/discomfort and 41.2% for anxiety/depression. The frequency of problems generally increased with age, except for the last dimension, where 56% among the youngest reported complaints versus 30% of participants aged 75 and above. Moreover, self-reported anxiety/depression was far more common in women (49.7%) than in men (32.3%). CONCLUSIONS: EQ-5D-5L population norms provide useful insights into the health status of the Italian population and can be used as reference values for future surveys using the same tool. Decision Modeling and Simulation
This session covers studies using decision modeling and simulation to investigate the impact of different strategies to inform decision making.
Linus Jönsson, MD, PhD
H. Lundbeck A/S, Copenhagen, Denmark
P5: Real-World Validation of the Implementation of Healthcare Capacity Optimization Measures Guided by the Simpli Tool: An Ophthalmology Proof of Concept in Portugal
4:00PM - 4:15PM
1, Oliveira F 1, Moita Fidalgo R 1, Oliveira A 1, Franco C 1, Fernandes C 1, Martins I 1, Brito de Sá M 1, Silva JP 1, Fonte S 1, Perpetua P 2, Freitas R 3, Bandeiras C 4 1Novartis Farma, Porto Salvo, Portugal, 2Novartis Pharma Region Europe, Basel, Switzerland, 3Novartis Pharma AG, Porto Salvo, Portugal, 4Novartis Global Service Center, Dublin 4, D, Ireland
OBJECTIVES: Hospital services in ophthalmology face significant capacity constraints in Portugal. In 2020, SimPLI – Simulating Capacity Performance Leading to Impact was introduced to assist services in simulating the impact of measures to decrease the backlog of outpatient consultations and procedures. The aim of this work is to introduce the validation of the simulator outputs with the real-world production in a reference ophthalmology center in Portugal. METHODS: A spreadsheet-based simulator was developed for testing measures to accelerate the realization of delayed external consultations and outpatient surgeries. The baseline backlog is represented by the number of delayed procedures. The time to eliminate this backlog was calculated and compared with the output for scenarios where one or more measures were implemented. The number of consultations and procedures simulated at 2020 year-end (YE) was compared with real hospital production data for model validation. RESULTS: The implementation of 15-minute telemedicine consultations for follow-up appointments and the reduction in 10 minutes in the time of face-to-face consultations was predicted to increase the number of 1 st consultations by 68% and the number of follow-up consultations by 77% in comparison with a scenario without any optimization measures. The time to solve outstanding consultations would be reduced by 7 months. In real practice in the same service, the application of the aforementioned measures was successful, with an additional 58% of 1 st consultations and 85% follow-up appointments until 2020 year-end. In outpatient surgeries, the application of measures for capacity optimization reduced by 22% the number of patient lost to private hospitals, with considerable resulting savings. CONCLUSIONS: SimPLI is invaluable for planning the investment in efficient actions towards optimization of hospital capacity. The proof of concept demonstrates that the proposed measures were validated in real practice with improved provision of care in ophthalmology, eventually resulting in improved patient outcomes. P6: Projecting COVID-19 Hospitalizations and Deaths Under Scenarios of Vaccination in Jefferson County, Kentucky
4:15PM - 4:30PM
Patel N University of Louisville, LOUISVILLE, KY, USA
OBJECTIVES : This report investigated the simulated effect of several vaccination scenarios on COVID hospitalizations and deaths in Jefferson County, Kentucky. Study Design: Eight scenarios were considered. First, it was assumed that the status quo scenario (~30,000 doses of Pfizer and Moderna vaccines distributed and administered every week) would continue without Johnson & Johnson’s vaccine. Then, three scenarios of the addition of Johnson & Johnson’s vaccine (10,000, 15,000, and 20,000 weakly) were considered. Next, an expansion over the status quo scenario (~40,000 doses of Pfizer per week) was considered with and without Johnson & Johnson’s vaccine scenarios. METHODS : An epidemic dynamics model (namely, a Susceptible-Exposed-Infectious Recovered (SEIR) model) is adopted and estimated in this study. In the model, transmission through different phases of the COVID-19 epidemic (susceptible, exposed, infectious, hospitalized, vaccinated, recovered, and dead) is regulated with transmission and clinical dynamics parameters. Key transmission parameters are the population, basic and effective reproduction factors, lengths of incubation periods, pre-infectiousness, infectiousness with and without symptoms, and vaccines’ efficacy rates and coverage. Key clinical dynamics parameters are hospitalization rate among the symptomatic, time from onset of severe symptoms to hospitalization, length of hospital stay, fatality rate among the hospitalized, recovery time among the hospitalized, the time from hospitalization to death. RESULTS : More intense vaccination than the status quo is expected to decrease hospitalizations and deaths in the next three months. However, the magnitude of the decrease in deaths is small, < 3 dozen. Importantly, it is expected that the COVID-19 infection continues to spread. Therefore, social distancing and other COVID-19 protection measures (for example, mask-wearing) must continue – should they be relaxed, a “during vaccination surge” will occur and should be expected in the late April-early May period. CONCLUSIONS : Implications for Policy or Practice are Social distancing and other COVID-19 protection measures (for example, mask-wearing) must continue. P8: New Onset Cardiovascular Disease in Australia by Socioeconomic Groups: A Modelling Study
4:45PM - 5:00PM
1, Marquina C 2, Talic S 2, Zomer E 2, Morton J 2, Liew D 1, Ademi Z 2 1Monash University, Melbourne, VIC, Australia, 2Monash University, Melbourne, Australia
OBJECTIVES : To project incident cardiovascular disease (CVD) and related health economic outcomes in Australia by socioeconomic status between 2020 and 2029. METHODS : A dynamic population model was built to project the annual incidence new-onset CVD by quintile of socioeconomic disadvantage in Australians aged 40-90 years between 2020 and 2029 using the Pooled Cohort Equation (PCE). The model projected years of life lived, quality adjusted life years (QALYs), direct healthcare medical costs, and productivity losses due to new-onset CVD. All outcomes were discounted by 5% annually. RESULTS : Cardiovascular risk profiling using the PCE showed that 20% of the most disadvantaged quintile were considered at high risk of CVD, compared to 12% in the least disadvantaged quintile. From 2020 to 2029, the model projected 211,901 incident cardiovascular events would occur in the most disadvantaged quintile compared to 184,846 in the least disadvantaged. Acute healthcare costs in the most socioeconomically disadvantaged group were AU$ 206 million higher than in the least disadvantaged group, while the difference in societal costs was AU$ 820 million. Scenario analyses estimated that a 17% risk reduction in CVD would be needed in disadvantage quintiles 1-4 to achieve the same outcomes as the least disadvantaged quintile (quintile 5). CONCLUSIONS : The number of CV events and associated costs highlight the urgent need to implement scalable primary prevention interventions targeted at disadvantaged groups. This model provides a platform to assess which interventions are likely to yield more benefits in each socioeconomic group at the population level. P7: Evaluating Impact of Universal Varicella Vaccination Strategies on Clinical Burden of Varicella and Herpes Zosterin England and Wales
4:30PM - 4:45PM
1, Sharomi O 2, Xausa I 3, Nachbar R 3, Matthews I 4, Pawaskar M 5 1Merck & Co., Inc., Rahway, NJ, USA, 2Merck & Co., Inc., West Point, PA, USA, 3Wolfram Solutions, Champaign, IL, USA, 4Merck Sharp & Dohme Ltd., Wokingham, WOK, UK, 5Merck & Co., Inc., Kenilworth, NJ, USA
OBJECTIVES: England and Wales have not implemented universal varicella vaccination (UVV) primarily due to its hypothesized impact on herpes zoster (HZ) incidence. Our study evaluated long-term clinical impact of UVV and exogenous boosting on varicella and HZ in England and Wales. METHODS: An age-structured, deterministic, dynamic transmission model using a dynamic population was adapted to England and Wales to assess varicella cases, associated hospitalizations and HZ cases. Ten (one- and two-doses, with/ without catch-up) vaccination strategies at short (12m/18m) and medium (18m/40m) dose intervals with and without catchup for 2 doses at 14 and 15 years of age were compared to no vaccination over a 50-year time horizon. Four varicella vaccines were considered with monovalent and quadrivalent formulations [Varivax ®, ProQuad ® (V/MMRV-MSD) Varilrix ® and Priorix-Tetra ®(V/MMRV-GSK)]. Vaccination coverage was assumed to be 91% for first doses and 89% for 2 nd doses and 87% for catch-up. The model accounted for the impact of exogenous boosting on HZ cases. RESULTS: All vaccine strategies substantially reduced the clinical burden of varicella over no vaccination: with 82-97% reduction in total varicella cases and 78%-86% reduction in the number of hospitalizations. One-dose strategies without catchup resulted in the smallest reduction in cases, hospitalizations, while the greatest reduction was seen with the 2-dose short-interval (12m and 18m) strategy. Incidence of HZ is estimated to be reduced by 7-11%. Strategies with V/MMRV-MSD vaccines were more effective in averting all four outcomes than with V/MMRV-GSK vaccines with similar intervals. CONCLUSIONS: Our model estimated that all one and two-dose UVV strategies significantly reducted the clinical burden of varicella including reduction in varicella related incidence, and hospitalization as well as reduction in HZ incidence compared to no vaccination in England and Wales. Policymakers should consider including UVV in their childhood immunization program to reduce disease burden. Impact of the COVID-19 Pandemic: Investigations in Populations of Interest
This session covers investigations on different populations of interest in regard to COVID-19 - pregnant and postpartum mothers, young adults and mental health, and infants.
Filipa Sampaio, MSc, PhD
Uppsala University, Uppsala, Sweden
P26: The Practice of Face Masking Among Young Adults in South India: An Online Cross-Sectional Survey During Second Wave of COVID-19
4:15PM - 4:30PM
Kochuparambil J 1, Issac A 2 1Mary Queen's Mission Hospital, Kanjirappally, India, 2Mary Queen's Mission Hospital, Adoor, India
OBJECTIVES: COVID-19 pandemic urges the need for respiratory protective equipment like face masks as a public health measure to control the spread of infection. This study aimed to investigate the trends followed in the practice of mask-wearing by the South Indian population amid the second wave of COVID-19 outbreak in 2021. METHODS: A web-based, online cross-sectional survey was conducted among the young adult population in India in late April 2021. An eight-item questionnaire was designed to assess the social perceptions and attitudes regarding wearing a face mask as a part of universal safety precautions. The social perceptions towards wearing masks were categorized as excellent, good average and poor on a scale (Social Perception Scale -SPS) scored out of 8. The details collected using a predesigned google form are statistically analyzed using the Chi-square test with a p-value of < 0.05 is considered statistically significant. RESULTS: Among the 1283 participants who completed the questionnaire, 57% wore cloth masks followed by 26% wearing N95 masks and 12% wearing surgical masks. Even though the age of the study population varied from 19 – 76 years and with a male preponderance of 56.3% (n = 723), students and recent graduates participated largely in the study (71.8%, n = 922). A mean SPS score of 5.67±1.07 (out of 8) indicates that the social perception of the study population is good. A statistically significant association is observed between the SPS score and the age (p = 0.003), type of mask used (p < 0.001), and economic background of the study population (p <0.001). Breathing difficulty, communication problems, additional cost incurred and dermatologic issues were commonly reported barriers against mask-wearing. CONCLUSIONS: Adjunctive public health measures such as mask-wearing are crucial in curbing the COVID-19 transmission. By shaping an appropriate public attitude, policymakers can ensure compliance towards mask-wearing. P27: Vaccination Coverage Trends for Hepatitis B in Infants from the Brazilian and Colombian Expanded Immunization Program: A Real-World Analysis of COVID-19 Pandemic Impact
4:30PM - 4:45PM
1, Abreu A 2, Hernández F 3, Martins J 2, Kano B 4, Kashiura D 2, Julian G 1 1IQVIA Real World Insights, São Paulo, SP, Brazil, 2IQVIA Brasil, São Paulo, SP, Brazil, 3IQVIA Colombia, Bogotá, Colombia, 4IQVIA, São Paulo, SP, Brazil
OBJECTIVES: COVID-19 pandemic has posed major challenges for healthcare systems and societies worldwide. Mitigation measures and the fear of exposure to COVID-19 might have negatively impacted local health policies, such as pediatric immunization programs strategies. This observational study aims to analyze the vaccination coverage (VC) for hepatitis B in infants in Brazil and Colombia between 2015 to 2020. METHODS: This is a descriptive analysis using real-world data from the Expanded Immunization Program System from Brazil (SI-PNI) and the Epidemiological Surveillance System from Colombia (SIVIGILA). We calculated the annual variation of VC for hepatitis B in infants from 2015 to 2020 for both countries. RESULTS: Overall, Brazilian VC had an average annual decline of 3.6% in the pre-COVID-19 period (2015-2019), reaching the lowest coverage in 2019 (78.57%), while the Colombian VC had an increasing pattern for the same period (0.4% annually), reaching the highest coverage in 2017 (89.3%). In 2020, VC decreased by 19.8% in Brazil, compared with 2019. In Colombia, VC decrease was notably lower (1.0%). CONCLUSIONS: In Colombia, VC increase might be explained by the implementation of the national plan for hepatitis B elimination in infants during this period. In Brazil, VC coverages for several other infectious diseases have also faced a decrease during the last years, but no formal mitigation activity or plan was yet established. Although both countries showed a reduction of the VC coverage in 2020, the impact was considerably higher in Brazil. These trends could be explained by the distinct health strategies linked to the Expanded Immunization Programs for each country in preparation for the COVID-19 pandemic. P28: Impact of COVID-19 on Mental Health in Young Adults in the United States
4:45PM - 5:00PM
Intermill T 1, Gong CL 2, Gu NY 3 1University of San Francisco, Sacramento, CA, USA, 2Children's Hospital Los Angeles, Los Angeles, CA, USA, 3NYG Technologies, Santa Clarita, CA, USA
OBJECTIVES: To assess the impact of COVID-19 pandemic on mental health in young adults in the US. METHODS: Three waves of online surveys were designed to capture mental health status in the US (EuroQol grant: 84-2020RA): Wave1 (Apr 1 st – May 6 th, 2020 (n=2,734)), Wave2 (July 4 th – Sept 4 th, 2020 (n=2,454)), and Wave3 (Jan 10 th - Mar 15 th, 2021 (n=2,252)) using the EQ-5D-5L to evaluate respondent’s health-related quality-of-life (HRQoL) and the Patient Health Questionnaire (PHQ-4) to assess anxiety and depression. The EQ-5D-5L utility, VAS scores and 5 domains were stratified by age, gender, and race/ethnicity. Binary Logistic regressions were used to estimate the associations between anxiety/depression and various covariates. Chi-square tests were conducted for significant differences in mental health outcomes between age groups. RESULTS: Most participants were white (68.7%) non-Hispanic (89 %). On average, participants were 42 (±13) years old, 47% being female. In all 3 waves, self-reported anxiety and depression were significantly higher in young adults (18-34) compared with older adults (35+) (p<0.01). Anxiety scores were 42%, 53%, and 33% in waves 1-3 respectively for young adults, whereas 33%, 40%, and 22% were reported by adults 35-64 and 19%, 20%, and 12% were reported by adults 65+. Similar trends were observed for depression, with younger adults reporting 39%, 54% and 35%, compared with 27%, 38% and 22% for those aged 35-64 years and 14.5%, 16% and 14.85% for 65+. EQ-5D-5L utility in waves 1-3 were 0.82, 0.75, and 0.82 (P<0.01) and 74.7, 78.7, and 76.4 for EQ-VAS (P<0.01). Age and employment status were significant predictors for anxiety and depression outcomes. CONCLUSIONS: Mental health deterioration during COVID-19 was pronounced among young adults for all waves, especially in wave2. Findings suggest although people adapt over time, the US was ill-prepared for a mental health crisis, especially among young adults. P25: Impact of COVID-19 on the Health-Related Quality-of-Life of Pregnant and Postpartum Persons
4:00PM - 4:15PM
1, Aytha Swathi P 2, Nosek M 1, Gu NY 3 1University of San Francisco, Sacramento, CA, USA, 2University of San Francisco, Signal Hill, CA, USA, 3NYG Technologies, Santa Clarita, CA, USA
OBJECTIVES: To assess the impact of COVID-19 on health-related quality-of-life (HRQoL) of those who were pregnant or recently pregnant during the pandemic. METHODS: Individuals who were pregnant any time since January 2020, the beginning of the pandemic, were invited to participate in an online, national US survey (EuroQol grant: 260-2020RA). Respondents were asked to self-report their experiences with COVID-19, to complete the EQ-5D-5L, and other measurements of HRQoL. To estimate the association between COVID-19 infection with the EQ-5D-5L outcomes, we used median regression for the EQ-5D utility and EQ-VAS scores, and ordinal logistic regressions for the EQ-5D-5L health items. Post-stratification weights were used to ensure representation by age, race and US census region. RESULTS: Among pregnant or postpartum persons, the median EQ-5D-5L utility score was 0.87 and EQ-VAS was 0.80. The median EQ-5D-5L utility score increased by 0.0058 (95% CI 0.0026, 0.009) for each additional year of age of the respondent. We observed no change in EQ-5D-VAS utility measures by maternal age (ß = 0.00; 95% CI -0.09, 0.09). On average, comparing Black pregnant persons to White, EQ-5D-5L utility values were 0.44 points lower, and EQ-5D-VAS scores were 0.31 points lower. Although median EQ-5D-5L utility values were similar for those with and without a diagnosis of COVID-19 (0.87 and 0.88), utility values declined by 0.022 (95% CI -0.040, -0.010) for each unit increase in perceived COVID-19 severity. Similar results were observed for the EQ-5D-VAS scores. When we evaluated EQ-5D-5L items individually, respondents diagnosed with COVID-19 reported more problems related to anxiety/depression compared with those who did not (OR 2.43; 95% CI 1.35, 4.40). No other items were significantly associated with COVID-19. CONCLUSIONS: We observed lower HRQoL measures associated with severe COVID-19 infection during pregnancy. In particular, problems with anxiety and depression contributed most strongly to lowered HRQoL during pregnancy. Methodological Developments in Survival Analytic Methods to Inform Cost-Effectiveness Models
This session will explore advances in methods for developing appropriate parametric survival models to inform cost-effectiveness models. The session will include methods for fitting parametric models in the context of multistate models, methods for jointly analysing PFS and OS, estimating subgroup-specific effects and using external historical data to inform model development.
Jeroen P Jansen, PhD
School of Pharmacy, University of California San Francisco, San Francisco, CA, USA, and PRECISIONheor, Oakland, CA, USA
P46: Effective Use of Reconstructed Survival and Comparative Effectiveness Data: A Case Study from Estimating Unreported Subgroup Survival in Advanced Stage Gastrointestinal Cancers
4:15PM - 4:30PM
Alagoz O 1, Xiao H 2, Singh P 2, Gricar J 2, Dixon M 2, Kim I 2, Kurt M 3 1University of Wisconsin-Madison, Madison, WI, USA, 2Bristol Myers Squibb, Lawrenceville, NJ, USA, 3Bristol Myers Squibb, Princeton, NJ, USA
OBJECTIVES : This study devises a systematic approach that can utilize aggregate level survival and comparative effectiveness data published from randomized controlled trials (RCT) to assist subgroup-specific health economic and meta-analyses. METHODS : We developed a soft-constrained optimization model, which approximates the restricted mean survival time (RMST) for the overall population in each arm via weighted sum of the RMSTs of two subgroups of interest. Survivals of both subgroups in each arm were assumed to follow Weibull or log-logistic distribution. The constraint ensured that cumulative hazards between the arms were proportional for each subgroup at a sufficiently long pre-specified time point. Estimated subgroup-specific survival functions for the control arm were direct outputs of the model and were shifted by applying the reported hazard ratios from the forest plots to generate their counterparts for the intervention arm assuming proportional hazards between the arms. For validation, we tested our approach in a case study consisting of 10 distinct RCTs with reported subgroup-specific Kaplan-Meier (KM) curves from advanced stage gastrointestinal tumors. RESULTS : Across all 48 subgroups, on average, loglogistic model performed equally or better than Weibull model in performance criteria comparing overall survival (OS) rates, median OS and RMSTs. Predicted survival curves laid within the 95% confidence intervals (CIs) of reported KM-curves in 75% and 81% of the time for Weibull and loglogistic models, respectively. Predicted median survivals were within the 95% CIs of the reported medians in 34 and 40 subgroups for Weibull and loglogistic models, respectively. Average relative gap between the predicted and reported RMSTs was 10% in both models. Predicted RMSTs were within the 95% CI of reported RMSTs in 34 and 37 subgroups for Weibull and loglogistic models, respectively. CONCLUSIONS : Our elicitation approach is effective and demonstrably reliable in deriving unreported subgroup survival with flexible time-varying hazard functions. P47: A Bayesian Hierarchical Mixture Cure Modelling (MCM) Framework for the Joint Utilization of Progression Free Survival (PFS) and Overall Survival (OS) in Estimating Long-Term Survivorship Rates in Previously Untreated Metastatic Melanoma: A ...
4:30PM - 4:45PM
Green N 1, Paly V 2, Youn JH 3, Kurt M 4, Moshyk A 5, Baio G 1 1University College London, London, LON, UK, 2ICON plc, New York, NY, USA, 3ICON plc, Marlow, Bucks, UK, 4Bristol Myers Squibb, Lawrenceville, NJ, USA, 5Bristol Myers Squibb, Princeton, NJ, USA
OBJECTIVES: Differences in emergent survival plateaus between PFS and OS may imply clinically unintuitive dichotomy between the resulting proportions of long-term survivors (LTS) when they are analyzed separately via mixture cure models (MCM). We present a novel Bayesian hierarchical (BH) MCM framework assuming a dependency between PFS and OS to estimate LTS rates in CheckMate 067 and demonstrate its practical utility over frequentist MCMs in long-term QALY estimations. METHODS: Frequentist and BH MCMs were fitted to PFS and OS data from the trial with minimum 60-months of follow-up. In the frequentist MCMs, PFS and OS were modelled separately whereas in BH MCMs both endpoints were modelled jointly with a shared LTS rate. In both approaches, background mortality rates were taken from World Health Organisation’s age, gender and country-specific lifetables and time-to-event outcomes for the non-LTS were modeled using a range of standard parametric distributions. Estimated incremental QALYs gains for nivolumab containing therapies versus ipilimumab under both approaches were compared using local tariffs from US, Canada, UK, France, Sweden, Belgium, Netherlands, Portugal and Australia. RESULTS: Among all combinations of distributions considered for the BH MCM, the exponential-exponential fit adequately captured the observed survival trends for both endpoints with reasonable goodness-of-fit measures and shared LTS rates which were (95% credible intervals) 46.3% (32.8%, 62.6%) for nivolumab+ipilimumab, 37.8% (21.6%, 55.7%) for nivolumab, and 15.1% (6.8%, 26.0%) for ipilimumab. For each arm, shared LTS-rates were in between individually-estimated LTS rates from the OS and PFS data in the frequentist MCMs. Compared to frequentist MCMs, over 20-years BH MCMs produced higher incremental QALY gains for nivolumab+ipilimumab and nivolumab versus ipilimumab with differences ranging from 0.30-0.48 and 0.17-0.26, respectively. CONCLUSIONS: Our BH MCM framework can alleviate the disparity between individually estimated OS- and PFS-based LTS rates, and allow for more robust clinical inference and extrapolations. P45: The Use of Historical Clinical Trial Data to Inform Survival Extrapolation
4:00PM - 4:15PM
Pham HA 1, Smalbrugge D 2, Kroi F 2, Heeg B 1, Ouwens M 3 1Ingress-health, Rotterdam, ZH, Netherlands, 2Ingress-health, Rotterdam, Netherlands, 3AstraZeneca, Mölndal, O, Sweden
OBJECTIVES : Standard parametric distributions are commonly used for the extrapolation of survival data in cost-effectiveness analyses. However, survival data is often immature and uncertainty remains around the survival extrapolations. Mature historical data can be used to better predict survival beyond trial data. This study assessed two methods to incorporate historical data in the extrapolation of immature survival data. METHODS : Immature data of a breast cancer trial comparing pertuzumab+trastuzumab+docetaxel versus trastuzumab+docetaxel (follow-up time 38 months; data-cut 2015) was extrapolated and mature survival data (follow-up time 120 months; data-cut 2020) from the same trial was used to validate the extrapolations. The historical data was from a previous breast cancer trial including mature survival data of trastuzumab+docetaxel (follow-up time 50 months; data-cut 2005). Two methods to quantitatively inform the extrapolation of immature survival data with historical data were compared to standard parametric distributions: 1) historical shape parameter as informative prior for the shape of the immature data; 2) historical data as a third arm. Predictions were assessed with delta area under the curve (AUC) values based on the mature survival data. RESULTS : Without priors, the delta AUC was 7.59, 1.62, 13.15, 8.32, 25.15, and with the historical arm the delta AUCs were 9.65, 4.38, 6.79, 8.26, 21.81, for Weibull, loglogistic, lognormal, exponential, and Gompertz, respectively. With priors, the delta AUC were 8.43, 3.37, 9.11, 23.68, for Weibull, loglogistic, lognormal, and Gompertz, respectively (as for exponential there is no shape parameter). The loglogistic distribution without priors predicted the immature data the best. For three out of five distributions, the extrapolations with a historical arm resulted in better predictions compared to the extrapolations without prior. CONCLUSIONS : The impact of external data on clinically plausible survival extrapolations can further be improved by using historical data with longer follow-up with treatment patterns similar to the current standard of care. Emerging Opportunities for the Use of Real World Data in Comparative Effectiveness Research
This session illustrates applications around the use of real world data and electronic health records to conduct comparative effectiveness research studies or reduce uncertainty on long term outcomes in HTA evaluations.
Isaac Corro Ramos, PhD
institute for Medical Technology Assessment, Eindhoven, NB, Netherlands
Isaac Corro Ramos, PhD obtained his Master’s degree in Mathematics (option Statistics and Operations Research) from the University of Sevilla in June 2001. Between June 2001 and July 2005 Isaac had several jobs in different working areas and countries. He worked as a high-school Mathematics teacher in Sevilla (Spain), as a software programmer in Madrid (Spain) and Vienna (Austria), and as a researcher in Rome (Italy). In July 2005 he started his Ph.D. at the Department of Mathematics and Computer Science of the Eindhoven University of Technology working on the STRESS (Statistical Testing and Reliability Estimation of Software Systems) project. He defended his thesis on December 15, 2009. Since August 2009 Isaac works as a scientific researcher at the Institute for Medical Technology Assessment (iMTA) in Rotterdam, The Netherlands. Since he enrolled iMTA he has worked in several research projects whose subjects include probabilistic modelling, cost-effectiveness analysis of health care technologies, value of information analysis and discrete event simulation.
P15: Transcatheter versus Surgical Aortic Valve Replacement: A Real-World Comparison of Clinical Outcomes Based on a German Claims Dataset
4:30PM - 4:45PM
Hardtstock F 1, Wilke T 2, Maywald U 3, Spitzer S 4 1Cytel Inc, Wismar, Germany, 2IPAM e.V., Wismar, Germany, 3AOK PLUS, Dresden, Germany, 4Praxisklinik Herz und Gefäße, Dresden, Germany
OBJECTIVES: This study aimed to describe clinical outcomes after transcatheter aortic valve implantation (TAVI) and surgical aortic valve implantation (SAVR). METHODS: This study consisted of a retrospective analysis of German health insurance claims data from 01/01/2013-30/06/2019. Continuously insured adults with either TAVI (OPS 5-35a.0) or SAVR (OPS 5-351.0) between 01/01/2014 and 30/06/2018, who had aortic valve stenosis (ICD-10 I35.0, I35.2) were included. Patients with previous TAVI or SAVR were excluded. Both cohorts were described with regards to their baseline characteristics (one-year baseline) and the incidence rate (IR) of events during the follow-up period for death, transient ischemic attack (TIA), stroke, major bleeding event, periprocedural complications, and myocardial infarction (MI). RESULTS: Overall, 2,932 TAVI and 826 SAVR patients were identified. Compared to SAVR patients, TAVI patients were on average older (81.75 years vs. 69.18 years), more often female (56.92% vs. 42.37%), more comorbid (CCI 5.86 vs. 3.82; CHA2DS2-VASc-Score 3.17 vs. 2.47), and they had a higher probability of previous TIAs (3.07% vs. 1.33%), strokes (8.29% vs. 4.00%), and MIs (10.57% vs. 3.87%). 3.07%/1.21% of TAVI/SAVR patients died during the index hospitalization. Outcomes were observed during a follow-up period of 2.43 years (TAVI) / 3.02 years (SAVR). The following IR have been observed for TAVI/SAVR: death (0.17 vs. 0.04; p<0.001), TIA (0.00 vs. 0.01, p=0.046), stroke (0.03 vs. 0.01, p<0.001), major bleeding event (0.08 vs. 0.04, p<0.001), periprocedural complications during index hospital stay (1.87 vs. 1.13, p<0.001), and MIs (0.02 vs. 0.00; p<0.01). CONCLUSIONS: TAVI has become the new standard of care in recent years and has replaced the classic aortic valve replacement, specifically in more fragile patients. The above results confirm that TAVI procedures are widely used in clinical practice, and that in line with current guidelines, physicians assess which patients should receive a TAVI or a SAVR procedure. P14: Exploring the Potential for EHR-Derived Real-World Data to Reduce Uncertainty in HTA Decision-Making: A Case Study of Long-Term Survival Outcomes
4:15PM - 4:30PM
Pittell H 1, Kent S 2, Groves B 2, Mpofu P 3, Baxi S 3, Copeland A 3, Bargo D 4, Adamson B 3, Jonsson P 5 1Flatiron Health, Great Neck, NY, USA, 2National Institute for Health and Care Excellence (NICE), London, LON, UK, 3Flatiron Health, New York, NY, USA, 4Flatiron Health, Brooklyn, NY, USA, 5National Institute for Health and Care Excellence (NICE), Manchester, UK
OBJECTIVES: Clinical trials are an important source of evidence for health technology appraisals (HTA). However, a key concern is uncertainty in survival due to immature data. This study investigates whether electronic health record (EHR)-derived data from the US may have the potential to reduce uncertainty in long-term outcomes, using NICE technology appraisal (TA) 531 as a case study. METHODS: We selected patients with previously untreated, Stage IV NSCLC, with positive or unknown PDL1 status, who initiated first-line pembrolizumab monotherapy between October 2016 and December 2020 from the nationwide de-identified EHR-derived Flatiron Health database. We applied additional lab and ECOG eligibility criteria. Outcomes were overall survival from treatment start and treatment duration. Sensitivity analyses assessed a sub-group with known PDL1 status and a time horizon ending at NICE TA publication. RESULTS: The study included 1109 patients (median age 72, 50% female). Real-world duration of therapy was median 5.0 months (95% Confidence Interval [CI]: 4.2-5.7). Median overall survival was 13.8 months (95% CI: 11.8-16.2) over the full study period. The sensitivity analysis excluding patients with missing PDL1 status found median overall survival of 14.9 months (95% CI: 12.5-17.6). The shorter time horizon (October 2016 to June 2018) estimated median overall survival of 13.1 months (95% CI: 10.8-NR). CONCLUSIONS: In this case, EHR-derived data offered longer follow-up time (max 49 months) than the trial follow-up (max 22 months) used for extrapolation. This cohort had a median overall survival of 13.8 months while the trial (n=154) estimated 30.0 months and a similar Medicare claims analysis (n=3079) estimated 11.4 months. Real-world median age was 7-9 years older than the trial. Our study demonstrates that EHRs can be a source of mature data on specific cohorts of interest with potential to contextualize trial evidence and inform HTA-decision making. P13: Use of Real-World Big Data to Assess the Effectiveness on Overall Survival Among Chemotherapy or Immunotherapy in First Line Metastatic Non-Small Cell Lung Carcinoma Patients in an Italian Setting
4:00PM - 4:15PM
Degli Espositi L
1, Sangiorgi D 2, Ancona DD 3, Andretta M 4, Barbieri A 5, Bartolini F 6, Cavaliere A 7, Chinellato A 8, Ciaccia A 9, Cillo MR 10, Citraro R 11, Costantini A 12, De Francesco A 11, Dell'Orco S 13, Di Manno G 13, Ferrante F 14, Gentile S 15, Lavalle A 15, Moscogiuri R 16, Pastorello M 17, Procacci C 3, Re D 18, Santoleri F 12, Ubertazzo L 19, Vercellone A 20, Perrone V 21, Dovizio M 2 1CliCon S.r.l. Health, Economics & Outcomes Research, Ravenna, Italy, 2CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, BO, Italy, 3ASL BAT, Trani, Italy, 4Azienda ULSS 8 Berica, Vicenza, Italy, 5ASL Vercelli, Vercelli, Italy, 6USL Umbria 2, Terni, Italy, 7ASL Viterbo, Viterbo, Italy, 8Azienda ULSS 3 Serenissima, Mestre (VE), Italy, 9ASL Foggia, Foggia, Italy, 10ASL Salerno, Salerno, Italy, 11Azienda ospedaliero-universitaria Mater Domini, Catanzaro, Italy, 12ASL Pescara, Pescara, Italy, 13ASL Roma 6, Albano Laziale, Italy, 14ASL Frosinone, Frosinone, Italy, 15Direzione Generale per la Salute Regione Molise, Campobasso, Italy, 16ASL Taranto, Taranto, Italy, 17ASP Palermo, Palermo, Italy, 18ASL Teramo, Teramo, Italy, 19ASL Roma 4, Civitavecchia (RM), Italy, 20ASL Napoli 3 SUD, Torre del Greco, Italy, 21CliCon S.r.l. Health, Economics & Outcomes Research, Bologna, Italy
OBJECTIVES: The use of big data to assess the effectiveness of oncological treatments in clinical practice is gaining increasing interest. This analysis aimed to assess the overall survival of metastatic non-small cell lung (met-NSCLC) patients receiving chemotherapy (CT) or immunotherapy (I/O) as 1 st line by using real-world data in a sample population in Italy. METHODS: A retrospective observational analysis based on administrative data from a sample of Italian Local Health Units was conducted. Met-NSCLC patients starting a 1 st line therapy with CT or I/O between 2017-2018 were identified. Stopping inclusion period up to 2018 enabled at least a two-years follow-up period for each included patient. Kaplan Meier overall survival analysis considered time (months) from therapy initiation to death. Multivariable analysis was performed to adjust for cofounders such as age, gender, metastasis, BRAF test prescription and pharmacological treatments. RESULTS: A total of 3,126 (mean age±SD 68.6±9.8 years, 68.2% male) and 316 (mean age±SD 68.6±9.7 years, 74.4% male) patients initiated treatment with CT and I/O respectively. In both groups, the more frequent metastases detected were related to lymph nodes (42.1% CT, 24.1% I/O), bone (25.8% CT, 14.9% I/O) and brain (18.3% CT, 10.1% I/O). Median [95%CI] survival was 8.0 [7.4-8.6] and 14.6 [12.2-18.9] months for CT and I/O patients, respectively. Death was not reported in 31.2% of CT and in 44.3% of I/O cohorts. Multivariable analysis showed the risk of death to be significantly lower in patients treated with I/O compared to CT (HR [95%CI] 0.796 [0.681-0.930]). CONCLUSIONS: Results from our study showed among met-NSCL patients in 1 stline a better overall survival of the I/O compared to CT patients and a reduced risk of death of I/O vs CT-treated patients. Our findings suggest real-world data could produce valuable insights into treatments and their outcomes in routine daily oncology practice, thus integrating the evidence from clinical trials. P16: Recent Estimates of Survival in Patients with Advanced Non-Small Cell Lung Cancer (NSCLC) in the US (2010-2020)
4:45PM - 5:00PM
1, Chao J 2, Hogea C 3, Stojadinovic A 4, Giove TJ 5, Sun X 1, Aziez A 6, Velcheti V 7 1GlaxoSmithKline, Durham, NC, USA, 2GlaxoSmithKline, Collegeville, PA, USA, 3GlaxoSmithKline, Philadelphia, PA, USA, 4GlaxoSmithKline, Upper Providence, PA, USA, 5GlaxoSmithKline, Mississauga, ON, Canada, 6GlaxoSmithKline, Zug, Switzerland, 7New York University Langone, New York, NY, USA
OBJECTIVES: Despite availability of new treatments, the prognosis of lung cancer remains poor. This study aims to provide recent estimates of survival in patients with advanced non-small cell lung cancer (NSCLC) in the US. METHODS: The survival of patients with advanced NSCLC was estimated using two US databases together covering 2010–2020. The study included patients with stage III or IV NSCLC diagnosed between 2010–2016 in the Surveillance, Epidemiology, and End Results Program (SEER) database, and patients with stage IIIB, IIIC or IV NSCLC, diagnosed between 2017–2020, without known oncogenic driver mutations who had completed ≥4 cycles of 1L treatment (restricted to platinum-based combinations, immuno-oncology monotherapy, or ipilimumab/nivolumab) in the Flatiron Health database, a US Oncology Electronic Medical Record database. Overall survival (OS) was defined as time from diagnosis of stage III or IV NSCLC to death or to date of last confirmed activity. RESULTS: A total of 49,298 and 133,395 patients with stage III and IV diagnosis respectively were identified in SEER. The 1-, 3- and 5-year OS for patients with Stage III disease were 55.1%, 26.3% and 17.5%, and for stage IV disease were 25.8%, 7.4% and 4.0%, respectively. The Flatiron database had 1,045 patients with stage IIIB, 130 patients with stage IIIC and 3,210 patients with stage IV disease at diagnosis. The 1- and 3-year OS for stage IIIB/IIIC disease were 72.5% and 36.4%, and for patients with stage IV disease were 65.9% and 24.6%, respectively. CONCLUSIONS: Despite differences in study population characteristics between the two databases, the study shows that mortality in patients with advanced NSCLC remains high, underscoring the need for continued efforts to identify novel treatments and synergetic treatment combinations to improve patient outcomes. Using Real World Evidence to Predict Risks of Adverse Outcomes and Disease Progression
Understanding patients risks of adverse outcomes and their rate of disease progression, and how this varies according to patient characteristics is essential in understanding the benefits of intervention. In this session we will explore how large real world data sets are being used to quantify these outcomes using examples from oncology and rheumatology. We will also hear how novel approaches to accessing and analysing data may facilitate use of datasets that must be stored in different geographical locations.
Praveen Thokala, MASc, PhD
Sheffield University, Sheffield, United Kingdom
Praveen Thokala joined the School of Health and Related Research (ScHARR) at the University of Sheffield after completing an MASc from the University of Toronto and a PhD from the University of Southampton.
During this time, he has completed several health economics projects including single technology appraisals (STA), multiple technology appraisals (MTA), and diagnostic assessment reports (DARs) for NICE. He has also recently led a modeling project for Institute for Clinical and Economic Review (ICER) in the US. His research interests include health economic modeling, multi-criteria decision analysis (MCDA), discrete event simulation modeling, and optimization. He co-supervised six PhD students to completion and currently co-supervises four PhD students.
In terms of MCDA, he has worked with the National Institute of Health and Care Excellence decision support unit (NICE DSU) in the UK on exploring the applicability of MCDA in HTA. He co-chaired the ISPOR Task Force on the use of MCDA in healthcare decision making and has been involved in several MCDA studies including supporting priority setting and benefit-risk analysis. He also co-edited a book titled, Multi-Criteria Decision Analysis to Support Healthcare Decisions.
P74: Treatment Patterns and Outcomes in Patients with Acute Myeloid Leukaemia (AML) in England: A Cancer Analysis System (CAS) Registry Retrospective Cohort Study
4:15PM - 4:30PM
Caseby S 1, Cranmer H 2, Ohlmeyer V 1, Groucott J 1, Eaton J 1, Podkonjak T 3, Lambova A 4, Adamson E 5, Kearns I 1 1Takeda UK Ltd, London, UK, 2Takeda Pharmaceuticals International Co., London, UK, 3Takeda Pharmaceuticals International AG, London, UK, 4IQVIA Ltd, Sofia, Bulgaria, 5IQVIA Ltd, LONDON, UK
OBJECTIVES : Myelodysplastic syndromes (MDS) and AML are rare blood cancers existing on a continuum, with poor clinical outcomes. This retrospective cohort study obtained real-world data on rates of transformation from MDS to AML, treatment patterns in AML, and survival outcomes for patients with AML in England. METHODS : This study analysed retrospective data from the English national CAS registry and included all adult patients with MDS or AML treated with systemic anti-cancer therapy (SACT) from January 2013 to December 2018, with follow-up to December 2020. Three study cohorts were generated: non-transformed patients with MDS of any risk category (cohort A), patients with AML who transformed from MDS of any risk category (cohort B), and all other patients with de novo and secondary AML (cohort C). Patient characteristics, rates of transformation to AML and SACT treatments received after azacitidine were described; overall survival (OS) was estimated using Kaplan-Meier methods. RESULTS : Cohorts A, B and C included 6,549, 368 and 6,936 patients (total N=13,853), with median age at diagnosis 67, 71, 65 years, respectively. The majority of patients were male (58.54%) and of white ethnicity (87.58%). Transformation rates of patients with MDS of any risk category (total cohorts A and B) were 4.48-11.23% depending on year of diagnosis and available follow-up; median time to transformation was 13 months. Treatment sequences following azacitidine were similar in cohorts B (N=56) and C (N=1,572); the majority of patients did not receive further SACT (66.07%; N=37 and 68.58%; N=1,078, respectively). Median OS was 2.7 months from AML transformation for cohort B and 16 months from SACT initiation for cohort C. CONCLUSIONS : This nationwide study highlights the poor survival for patients with AML in England, particularly for patients who have transformed from MDS to AML. Treatment options are limited, with significant unmet need for new effective therapies. P73: Risk of Hospitalization and Emergency Room Visits Among Community Oncology Patients
4:00PM - 4:15PM
Namasivayam G 1, Rahman MM 2, Mohammad N 3, Chang B 3, Karhade M 3, Robert N 3, Wu N 4, Heller B 5, Hoang S 6, Alwardt S 3, Neubauer M 3, Staggs S 4, Moore L 7, Smith H 3 1Ontada, Livermore, USA, 2Ontada, magnolia, USA, 3Ontada, The Woodlands, TX, USA, 4US Oncology Network, The Woodlands, TX, USA, 5Southern Cancer Center, Mobile, AL, USA, 6US Oncology, Austin, TX, USA, 7Ontada, Shaker Heights, OH, USA
OBJECTIVES : Machine Learning (ML) solutions can be used to bring insights to providers at the point of care. This research aimed to establish an explainable ML model to predict patient risk of emergency room (ER) visits and hospitalizations within 30 days after an oncology practice visit. METHODS : A retrospective cohort of 98,686 patients within McKesson Specialty Health | The US Oncology Network’s iKnowMed electronic health records (EHR) system between 05/01/16 and 06/30/20 was identified for research purposes. Inclusion criteria included Oncology Care Model patients aged >65 years and on cancer treatment. Approximately 300 clinical metrics from the EHR system were considered including diagnosis, staging, labs, vitals, treatments, comorbidities, drugs, and performance. Data were split into training (90%) and testing (10%). Model was selected using performance metrics (AUC, sensitivity, specificity, and accuracy). Model interpretations were reviewed by a team of oncologists. RESULTS : The extreme gradient boosting model (XGBoost) achieved the following testing results: AUC 72%, balanced accuracy 66%, sensitivity 65%, and specificity 67%. At threshold 0.5. The model correctly identified 65% of hospitalized and 67% of non-hospitalized patients. The top 5 attributes in order of SHapley Additive exPlanations (SHAP) impact values included albumin value (0.18), hemoglobin value (0.14), pain score (0.12), days since first chemotherapy (0.09) and lymphocytes (0.08). SHAP plots illustrate all associations between features and 30-day hospital admissions and ER risk. CONCLUSIONS : Real-world data were harnessed and applied in an ML approach to establish a high-performing patient ER visit and hospitalization prediction model. The next phase will include model deployment to several US Oncology Network practices to validate effectiveness in the real world. Results will inform providers when a patient is at risk for an ER visit or hospitalization with the aim of improving the overall quality of oncology care and reducing admissions and ER visits. P76: Federated Analysis of Multi-Centric Real-World Data: A Feasibility Study
4:45PM - 5:00PM
, Schwarz S, Blum M Andreux M Owkin, Paris, France
OBJECTIVES: Real-World Data (RWD) brings evidence about potential benefits and risks of a medical product that complements evidence generated during clinical trials. An important aspect of RWD is to incorporate data from multiple cohorts so as to minimize bias. However, data sharing in a central repository is not always possible due to data regulations and operational hurdles, thereby impeding the adoption of Real-World Evidence. Federated analysis (FA) is a novel approach to execute analyses on non co-located datasets, where data analysis is executed locally to preserve privacy and ensure compliance. In this work, we investigate the potential of FA to perform statistical analysis over large, decentralized datasets. The main purpose of the study was to evaluate if common statistical operations can be performed in a federated fashion. METHODS: We performed a proof-of-concept federated analysis implementation of several common statistical analysis methods on the public cancer TCGA dataset, split in three centers to mimic a decentralized setting. We quantitatively tested the difference between results obtained with a Federated Analysis and with a standard pooled implementation. We also qualitatively assessed the difficulty to perform an FA analysis. RESULTS: Our findings indicate that it is possible to perform federated analysis for descriptive statistics (mean, variance), survival analysis (Kaplan-Meier curves and Cox Proportional Hazards Model), and hypothesis testing (log-rank test) that provides results equivalent to a pooled implementation. Furthermore, the implementation is qualitatively simple for experienced data scientists. CONCLUSIONS: FA holds the potential to perform RWD analysis in a private and compliant fashion over decentralized data. This is especially relevant for rare diseases where local data scarcity is commonplace. Future works will investigate the evolution of the gap between FA and pooled methods when additional privacy-preserving techniques, such as differential privacy, are used. P75: Use of Electronic Health Records to Understand the Disease Progression in Lupus Nephritis (LN)
4:30PM - 4:45PM
, Mazumder D, Singh D Dhangar I Optum Global Solutions, Noida, UP, India
LN, a common manifestations of Lupus, often progresses to Chronic Kidney Disease (CKD) or End Stage Renal disease (ESRD) within 5 years of LN diagnosis. Early identification and monitoring could possibly delay the progression to CKD/ESRD in LN patients. The objective is to evaluate disease progression among LN patients in the real-world.
Patients with ≥2 records for lupus (ICD10) between 01/01/2016 and 31/12/2020 were identified using Optum de-identified electronic health records. The first record for lupus was set as index. Patients with continuous activity 6 months pre and 24 months post-index, with no lupus diagnosis during the pre-index were included. Patients were followed up to 24 months to assess evidence of LN (second index; based on ICD codes for proteinuria, nephritis, glomerular diseases, and renal failure). Patients were further followed for 12 months from LN diagnosis date to assess the progression to CKD/ESRD. We also evaluated the use of guideline recommended therapies and progression among the subset of patients with LN.
A total of 37,163 patients with lupus were identified after applying the study selection criterion. Nearly 8% of patients presented with LN at the time of lupus diagnosis. Approximately, 14% and 19% of patients developed LN in the first 12 months and 24 months of lupus diagnosis, respectively. Among patients who developed LN, nearly 41% progressed to CKD/ESRD within a year of diagnosis irrespective of their treatment status. Approximately, 60% patients progressed to CKD/ESRD during the 12 months follow-up period even after receiving the guideline recommended therapy.
Our analysis suggests that lupus patients may develop LN as early as 1 year of diagnosis, and subsequently to CKD/ESRD in a year. Well designed studies using EHR may provide greater insights into the factors associated with early progression, which will help design population health initiatives targeting early detection and intervention.
Methods for Decision Modelling and Economic Evaluation
This session includes studies that address challenges in economic evaluation in an innovative way or offer possible examples of good practice that will be of interest to analysts working in diverse fields. Studies address estimation of cost parameters (for cardiovascular disease), discrete event simulation, impact of changing price over the lifecycle of the product, and fiscal health modelling.
David Epstein, PhD
University of Granada, Granada, Spain
DAVID EPSTEIN is associate professor at the Department of Applied Economics, University of Granada, Spain. He has an MSc from University of York and a PhD from the University of Granada. His research interests include health technology appraisal, meta-analysis and cost-effectiveness analysis. David Epstein has participated in major clinical trials and other projects at national and international level. He is associate editor of Value in Health and Gaceta Sanitaria, co-coordinator of the Economic Evaluation group of the Spanish Health Economists’ Association, and member of the ISPOR, the Chartered Institute of Management Accountants and Royal Statistical Society.
P54: Cost-Effectiveness Analysis of Pharmacokinetic-Guided Prophylaxis versus Standard Prophylaxis in Adult Patients with Severe Hemophilia Α in China
4:15PM - 4:30PM
Gu C 1, Jiao Z 2, Huang H 1, Han Y 3 1Takeda China, Beijing, China, 2Shanghai Chest Hospital, Shanghai Jiao Tong University, Shanghai, China, 3Sun Yat-sen University, Guangzhou, China
OBJECTIVES: The objective of the study was to assess the cost-effectiveness of pharmacokinetic (PK)-guided prophylaxis with recombinant human coagulation factor VIII (FVIII) against standard prophylaxis in Chinese adult patients with severe hemophilia A. METHODS: A Discrete Event Simulation (DES) model was developed to evaluate the cost-effectiveness of PK-guided individualized prophylaxis versus standard prophylaxis from the healthcare system perspective which simulated 10,000 patients over a 1-year time horizon. Standard prophylaxis FVIII dose was 30 IU/kg by intravenous injection. PK-guided prophylaxis dosage was adjusted for each patient to maintain the FVIII trough level at 1-5 IU/dL to reduce the risk of bleeding. Dosing interval for both approaches was set to be 48h. Population characteristics, clinical data and utilities were collected from published literature and expert survey. The health outcomes included Annual Joint Bleed Rate (AJBR) and Quality-Adjusted Life Years (QALYs). Model considered the prophylaxis drug costs and bleeding treatment costs in the evaluation. Incremental Cost Effectiveness Ratio (ICER) was estimated to support decision. Probabilistic sensitivity analysis was automatically incorporated throughout DES. RESULTS: A total of 94.3% of patients receiving PK-guided individualized prophylaxis achieved the goal of maintaining the trough concentration at a level of 1 to 5 IU/dL compared with 62.7% on standard prophylaxis. AJBR in PK-guided and standard prophylaxis were 1.527 vs 1.601 and QALYs gained were 0.8384 vs 0.8383 respectively. The average FVIII dose for prophylaxis in PK-guided prophylaxis was 28 IU/kg. Prophylaxis drug costs and bleeding treatment costs in PK-guided prophylaxis (RMB 999,034; RMB 30,557) were both lower than those in standard prophylaxis (RMB 1,072,828; RMB 31,948). A total average savings of RMB 75,185 was obtained by the PK-guided approach. CONCLUSIONS: PK-guided individualized prophylaxis is a dominant treatment compared with standard prophylaxis for adult patients with severe hemophilia A in China, with slightly higher QALYs but lower total costs. P55: Excess Annual Hospital Costs Due to Cardiovascular Events in a Contemporary UK Population to Inform Health Technology Assessments
4:30PM - 4:45PM
Zhou J 1, Wu R 2, Williams C 1, Mihaylova B 1 1University of Oxford, Oxford, UK, 2Queen Mary University of London, LONDON, LON, UK
OBJECTIVES : Impacts of cardiovascular disease (CVD) events on hospital cost in contemporary populations are required to support health technology assessments. We developed models of annual hospital costs associated with a range of events of interest in CVD using the individual participant data from UK Biobank. METHODS : All hospital admissions of the 502,493 participants in UK Biobank were costed using the UK Healthcare Resource Groups reference costs (2019 UK£). Separate annual hospital costs models were indicated for participants with (n= 444,177) and without history of CVD (n= 57,556). We compared one- and two-part generalized linear regression models (GLMs) (part 1: logistic regression for probability of incurring cost, part 2- GLM with Gaussian, Poisson or Gamma distributions using identity, log or squared root links for costs, conditional on incurring any), adjusting for participants’ characteristics at entry (socio-demographic, clinical, prior diseases) and time-updated adverse events (myocardial infarction, stroke, coronary revascularization, cancer, diabetes, vascular death and non-vascular death) (p-value <0.01 in stepwise covariate selection). Standard errors were adjusted for clustering of annual periods by participant. Selected models were used to estimate the events’ impacts on annual hospital costs. RESULTS : For both participants with or without prior CVD, the two-part model with gamma distribution and identity link provided the best fit in modelling annual hospital costs. Overall, adverse events were associated with higher cost in participant with prior CVD than without prior CVD. In both populations, the highest excess costs were associated with years of incident cancer [mean and 95% confidence intervals for participants without and with prior CVD: 6095(5999-6191), 6232(5962, 6502)], coronary revascularization [6077(5920, 6234), 6693(6464, 6922)], stroke [5366(5129, 5603), 5856(5478,6234)], and non-vascular death [5317(5117, 5517), 6056(5652, 6460)]. CONCLUSIONS : These models could inform costs in health technology assessments in cardiovascular disease, such as cost-effectiveness analysis of interventions to reduce cardiovascular disease risks. P53: The Early Lessons of COVID-19: The Need for a Broader Health-Economic Perspective
4:00PM - 4:15PM
1, van der Schans S 2, Postma M 3, Boersma C 4 1Health-Ecore B.V., Amersfoort, UT, Netherlands, 2University Medical Center Groningen, University of Groningen, Groningen, GR, Netherlands, 3University of Groningen, Groningen, Netherlands, 4University Medical Center Groningen, University of Groningen, Groningen, the Netherlands; Department of Management Sciences, Open University, Heerlen, the Netherlands; Health-Ecore, Zeist, UT, Netherlands
OBJECTIVES: The current COVID-19 pandemic caused ~20,000 deaths and ~50,000 hospital admissions in the Netherlands. Efforts to manage this communicable disease and its impact on the health-care system without prior development of specific vaccines have put a strain on the fiscal budget. This study aims to indicatively quantify the impact of COVID-19 on the Dutch government’s fiscal position, simultaneously indicating the value of preventive vaccines from a payer perspective. METHODS: Dutch COVID-19 specific population data on laboratory-confirmed infections, hospital admissions and mortality, was collected from the domestic start of the COVID-19 pandemic on 27 February 2020 until the first administered vaccine on 6 January 2021. A fiscal health modelling approach was used to estimate the loss in tax revenues. Occurred productivity losses were added as an indicator for the future burden on the social security system. Tax revenue losses were caused by premature mortality, whereas the productivity losses occurred through mortality as well as morbidity. Outcomes were expressed in total monetary impact (€, 2020). RESULTS: The impact of the pandemic in the analysed time-period was estimated to amount to a total of €920.7 million. Tax loss due to premature mortality amounted to €58.8 million with 50% attributed to patients >60 years. Productivity loss due to morbidity summed up to €862 million with 46% due to patients 40-59 years. CONCLUSIONS: The fiscal impact of the current pandemic highlights the importance of a broader approach to health-economic analysis. A fiscal health framework, optimally linked to a disease simulation model, is a better instrument to inform decision-making in the context of communicable diseases. The reported fiscal estimates also highlight the benefit of investments in communicable disease prevention such as anticipative development of vaccines. In the decision-making process around pandemic preparedness measures, investment funding and real-options can consequently be informed by a fiscal health framework. P56: Early Health Technology Assessment to Guide Introduction of Novel Technology into the Market: The Vostars System Case Study
4:45PM - 5:00PM
Lorenzoni V 1, Trieste L 2, Turchetti G 1 1Institute of Management, Scuola Superiore Sant’Anna, Pisa, Italy, Pisa, Italy, 2Scuola Superiore Sant'Anna, Pisa, Italy
OBJECTIVES : The aim of the study is to show potential approaches and usefulness of early Health Technology Assessment exemplifying the case study of the Video and Optical See-Through Augmented Reality Surgical System (VOSTARS), an advanced navigation tool for non-endoscopic surgeries, recently developed, also thanks to funding from H2020 programme. METHODS : Surgeons in potential fields of application were surveyed about current practice, potential application of the system, perceived usefulness, intention to use, organizational issues, etc.. Those data, data from in vitro and in vivo tests and inputs from literature were used to inform early evaluation of different HTA dimensions comprising also cost-effectiveness (CEA) and budget impact analysis (BIA). RESULTS : A total of 99 surgeons were surveyed. About the possibility to use the VOSTARS system, 39 surgeons (41%) declared they will absolutely or probably use it (45%). Positive feedbacks were also obtained for both the potential usefulness and the possibility of the system to improve performance of work (about 80% of surgeons). High degree of variability emerged about the opportunity to enhance job effectiveness and improve operating time, anyway surgeons offered insight into potential limits and benefits of the system, related to the possibility of personalized approach. Surgeons also informed the identification of specific surgical fields and interventions for which the system could allow satisfy unmet need (ie, neurosurgery and maxillofacial surgery). Although variability of results, CEA showed potential savings for the Italian National Health System (INHS) in both neurosurgery and maxillofacial surgery, when compared to freehand approach. The BIA highlighted 3-year savings of -1,269€ for a local HS performing about 50 procedures per year. CONCLUSIONS : Despite the evaluation underlined high degree of uncertainty about the economic implications, end users involvement helped highlighting technical issues to be solved, offered insight to better target the technology, assess fact related to acceptability and also design the potential market Methodological Developments in Network Meta-Analysis and Comparative Effectiveness Research
This session explores advances in evidence synthesis including a range of topical issues: managing randomised and non-randomised designs, single arm studies, unanchored indirect comparisons and appropriate methods for analysing survival endpoints.
Matthijs Versteegh, PhD
Erasmus University Rotterdam, Rotterdam, Netherlands
Matthijs Versteegh, PhD, is director at the institute for Medical Technology Assessment (iMTA) of Erasmus University Rotterdam
P42: Comparison of Estimation Methods for Single-Arm Trials in Rare Diseases with Historical Control Groups
4:15PM - 4:30PM
Barlev A 1, Brookhart MA 2, Xun P 3, Thirumalai D 3, Sadetsky N 1, Suissa S 4 1Atara Biotherapeutics, South San Francisco, CA, USA, 2Duke University, Durham, NC, USA, 3Atara Biotherapeutics, Thousand Oaks, CA, USA, 4McGill University and Jewish General Hospital, Montreal, QC, Canada
OBJECTIVES : Randomized controlled trials are the gold standard for estimating treatment effects. However, in rare diseases with high unmet need, single-arm trials are used when randomization of patients to placebo or standard of care is infeasible or unethical. We evaluated various methods to control for confounding in estimating treatment effects in a small single-arm trial with a historical comparator group. METHODS : We used simulation to evaluate different techniques to estimate the “true” treatment effect on overall survival (OS) and objective response rate (ORR) in a specific target population using an external comparator design. We varied effect size, sample size, confounders, and correlation between confounders. We assessed two broad categories of methods: i) requiring specification for treatment allocation [propensity score (PS)-based inverse probability of treatment weighting (IPTW), standardized mortality/morbidity ratio (SMR), stabilized IPTW (SIPTW), overlap weighting (OW), stratification, and matching], and ii) adding outcome information (g-computation). Their precisions and accuracies were evaluated by a combination of 95% confidence interval (CI) coverage, power, bias, and mean square error (MSE). RESULTS : G-computation resulted in the most accurate and precise estimator of OS (95% CI coverage: 93.5%, power: 69.3%, bias: -0.001, MSE: 0.055) in a small sample size scenario of 30 treated subjects compared with 120 comparator subjects. Similar results were observed for ORR. In comparison, results for OS were: 95% CI coverage: 72.8%, 65.6%; power: 75.9%, 62.6%; bias: -0.026, 0.072; MSE: 0.114, 0.167 for SMR and IPTW, respectively. CONCLUSIONS : In our simulated example, the g-computation estimator performed best to control confounding in a small single-arm trial with an external comparator group. PS based methods (e.g., SMR & IPTW) may be suitable as an initial step in the creation of the comparator arm when researchers are blind to the outcome, while g-computation can be subsequently used to estimate the efficacy of treatment. P41: Crosnma: A New R Package to Synthesize Cross-Design Evidence and Cross-Format Data
4:00PM - 4:15PM
, Salanti G Hamza T University of Bern, Bern, BE, Switzerland