A patient-centered approach to benefit-harm assessment in treatment decision-making

Several structured and quantitative approaches have been developed to determine the benefit-harm balance of medical treatments in specific populations. These approaches summarize the key factors of a benefit-harm assessment that may include baseline outcome risks, treatment effects and relative importance for outcomes. The overarching goal of this dissertation was to develop a patient-centered approach to benefit-harm assessment in treatment decision-making. In part one of this dissertation, we reviewed the labels and medical reviews of 58 drugs approved by the US Food and Drug Administration for four diseases to learn how they dealt with surrogate outcomes when they were assessing the benefits and harms of drugs. Most drugs for chronic obstructive pulmonary disease, diabetes, and glaucoma were approved based only on surrogates but for osteoporosis, most drugs were also approved for patient-centered outcomes. The rationale for using surrogates was not often discussed (11 out of the 43 drug approvals based only on surrogates, 26%) in medical reviews. We accordingly proposed a framework for use of surrogate outcomes in doing patient-centered benefit-harm assessments. In part two of this dissertation, we conducted a survey of patients with non-infectious uveitis to elicit their preferences for six treatment outcomes associated with corticosteroid therapy. Eighty-two patients in the Multicenter Uveitis Steroid Treatment (MUST) Trial Follow-up Study and 100 patients treated at two academic medical centers (Johns Hopkins University and University of Pennsylvania) completed the best-worst scaling tasks, in which they repeatedly selected the most and least worrying from a list of three outcomes. Results showed that participants were more likely to select vision not meeting the requirement for driving, glaucoma, and needing eye surgery as the most worrying outcomes as compared against needing medicine for high blood pressure/cholesterol, cataracts or infection (e.g., sinusitis). In part three of this dissertation, we conducted a quantitative benefit-harm assessment using data from the MUST trial that compared corticosteroid implant versus systemic corticosteroids in non-infectious intermediate, posterior, and panuveitis. We calculated benefit-harm metrics to reflect the benefit-harm balance that took into account the treatment effects on different patient-centered outcomes and the patient preferences for these outcomes (derived from the preference-elicitation survey). The benefit-harm metrics at 6, 12, 18, and 24 months follow-up were all negative and the probabilities of the metric being positive were small or 0%. This implied that implant therapy had a worse benefit-harm balance than systemic therapy. In summary, using an example of corticosteroid therapy for treating non-infectious uveitis, we demonstrated a patient-centered approach to benefit-harm assessment where we focused on patient-centered outcomes and incorporated patient preferences to estimate treatment benefit-harm balance. Our approach can be applied in the future to different diseases and settings and help make evidence- and preference-based treatment decisions