OBJECTIVES

We aimed to (1) identify the costs and outcomes between patient sub-groups, including age, gender, prior treatment, co-morbidities, US region and urbanicity by approved treatments and (2) demonstrate ability to rapidly attain insights through the use of interactive visual analytics.

METHODS

A disease-model for asthma was built using FDA labels, professional guidelines, literature, and subject matter expertise. Patients from IBM MarketScan Commercial and Medicare Supplemental databases between 1/1/2011 and 12/31/2017 with either ≥1 hospital/ER visit, ≥2 outpatient visits, or ≥1 outpatient visit and an asthma medication were identified. ICD-9 493.xx or ICD-10 J45.xx were used for asthma excluding those with emphysema, COPD or CF with at least one year of enrollment prior to index treatment. The data was visualized with the IBM Access and Value Connect solution and analyzed to identify variations in utilization outcomes (emergency room visits, outpatient visits, hospitalizations) and costs (per patient per month or PPPM) for patients using 17 different drugs.

RESULTS

There were 222,312 patients analyzed. For pediatric patients (5-17) with ≥1 asthma exacerbation in the 12 months following treatment initiation, mean PPPM costs ranged from $2075 for budesonide to $901 for montelukast, driven by differences in all-care costs. For adult patients (18+), the range was $6326 for omalizumab to $1689 for albuterol, driven primarily by an average asthma-related drug cost of $3,850 for omalizumab. For those without an exacerbation, the highest mean PPPM costs were $5749 for omalizumab in adults and $4172 for levalbuterol for pediatrics. Use of omalizumab skews heavily towards older populations, with only 12% being used in pediatric populations versus 20% of all asthma medications used in adults.

CONCLUSIONS

Visual analytics provide an efficient way to perform cost and comparative effectiveness and enable standardized comparisons by disease. further analysis is warranted to determine causal reasons for asthma differences by X population or Y treatment.