Understanding Treatment Preferences in Highly Sensitized ESKD Patients: A Latent Profile Analysis of Discrete Choice Data
Author(s)
Geof D. Gray, Business Communications1, Dana Michelle Saavedra Roman, MPH1, Shiyang Su, Psychometrics1, Diana Clynes, Mass Communications, Public Relations2, Gina Ewy, MHA, JD3, Jennifer Leonard, Pharm.D.4, Jerome Bailey, B.A. Telecommunication5, Erin Kahle, CNP, MPA5, Hannah Saul, BA British Politics and Legislative Study6.
1Insights and Evidence Generation, Syneos Health, Bridgewater, NJ, USA, 2Executive Director, America Association of Kidney Patients, Tampa, FL, USA, 3Hansa Biopharma, Chapel Hill, NC, USA, 4Hansa Biopharma, Lund, Sweden, 5American Association of Kidney Patients, Tampa, FL, USA, 6Hansa Biopharma Inc, Lund, Sweden.
1Insights and Evidence Generation, Syneos Health, Bridgewater, NJ, USA, 2Executive Director, America Association of Kidney Patients, Tampa, FL, USA, 3Hansa Biopharma, Chapel Hill, NC, USA, 4Hansa Biopharma, Lund, Sweden, 5American Association of Kidney Patients, Tampa, FL, USA, 6Hansa Biopharma Inc, Lund, Sweden.
OBJECTIVES: Identify distinct patient subgroups based on treatment preferences and individual variables.
METHODS: In collaboration with the AAKP Center for Patient Research and Education, individual characteristics of highly sensitized end stage kidney disease (ESKD) patients who completed a Discrete Choice Experiment (DCE) survey (N=99) were combined with Mixed Logit Model (DCE analysis) to create predictive profiles through a Latent Profile Analysis (LPA). The patient profiles identify similar pattern responses to attribute preferences of highly sensitized ESKD patients deciding between two therapeutic choices (remaining on dialysis or proceeding with same-day desensitization and transplant) with individual’s characteristics serving as predictors. Model fit statistics including Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were used to determine the number of patient profiles for the LPA. 10 variables of individual characteristics from the DCE survey were selected to predict the LPA profiles, using multinomial regression analysis.
RESULTS: LPA identified 3 distinct subgroups of patients based on their preferred attributes tested in the DCE (hope, risk, life participation, kidney survival and support), which were further explained by predictive variables of proactivity, optimism, willingness to proceed with transplantation, dialysis experience, transplant and waitlist history, desensitization, and quality of life.
CONCLUSIONS: LPA is a well-established and utilized methodology for delving deeper into preferences. This LPA identified distinct patient subgroups based on key variables to facilitate targeted interventions or policies that cater to the specific needs and preferences of each patient subgroup.
METHODS: In collaboration with the AAKP Center for Patient Research and Education, individual characteristics of highly sensitized end stage kidney disease (ESKD) patients who completed a Discrete Choice Experiment (DCE) survey (N=99) were combined with Mixed Logit Model (DCE analysis) to create predictive profiles through a Latent Profile Analysis (LPA). The patient profiles identify similar pattern responses to attribute preferences of highly sensitized ESKD patients deciding between two therapeutic choices (remaining on dialysis or proceeding with same-day desensitization and transplant) with individual’s characteristics serving as predictors. Model fit statistics including Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were used to determine the number of patient profiles for the LPA. 10 variables of individual characteristics from the DCE survey were selected to predict the LPA profiles, using multinomial regression analysis.
RESULTS: LPA identified 3 distinct subgroups of patients based on their preferred attributes tested in the DCE (hope, risk, life participation, kidney survival and support), which were further explained by predictive variables of proactivity, optimism, willingness to proceed with transplantation, dialysis experience, transplant and waitlist history, desensitization, and quality of life.
CONCLUSIONS: LPA is a well-established and utilized methodology for delving deeper into preferences. This LPA identified distinct patient subgroups based on key variables to facilitate targeted interventions or policies that cater to the specific needs and preferences of each patient subgroup.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
RWD187
Topic
Methodological & Statistical Research, Patient-Centered Research, Real World Data & Information Systems
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity), Surgery, Urinary/Kidney Disorders