Research Proposal
Background: Randomized clinical trials (RCT) provide rich data related to safety and efficacy of pharmacological agents. This is often at variance with real-world data since patients in the real-world i.e. (real-world data (RWD)) are not constrained by RCT design. Even when the results of RCT and RWD are consistent, they often do not apply directly to an individual patient who presents before a physician.
Objectives: To examine the utility of Desirability of Outcome Ranking (DOOR) in measuring the risk-benefit profile of antipsychotic treatment in schizophrenia.
Study Design: Randomized control trial data will be reanalyzed using Desirability of Outcome Ranking (DOOR), a measure that combines clinically-relevant risks and benefits.
Participants: Subject-level data from randomized, placebo-controlled studies of risperidone and paliperidone in schizophrenia patients.
Main Outcome Measures: Utility of DOOR in characterizing treatment responders/non-responders and overall concordance with results of trial using traditional outcome measures.
Statistical Analysis: We will re-analyze the trial data using DOOR for finer gradations of composite outcomes. DOOR probability, proportion in favor of treatment, and win ratio will be used to compare treatments. Their 95% confidence intervals will be estimated by bootstrap.
Randomized clinical trials (RCT) provide rich data related to safety and efficacy of pharmacological agents. This is often at variance with real-world data since patients in the real-world i.e. (real-world data (RWD)) are not constrained by RCT design. Even when the results of RCT and RWD are consistent, they often do not apply directly to an individual patient who presents before a physician. We attempt to overcome these challenges by use of a novel statistical technique, Desirability of Outcome Ranking (DOOR), to better inform clinical-decision making in the treatment of schizophrenia. DOOR has proven utility in antimicrobial therapy in conjunction with partial credit scoring to allow for quantitative comparisons of the clinical desirability of treatment decisions (Claeys et al, 2021, Evans et al, 2020). In this proposal, we extend this novel statistical technique to clinical decision making in antipsychotic treatment of schizophrenia.
The specific aims of the project are:
a) Develop a Desirability of Outcome Ranking (DOOR) for schizophrenia trials
b) Examine the utility of DOOR in characterizing treatment responders/non-responders and overall concordance with results of trial using traditional outcome measures
Individual-level data from RCTs of risperidone and paliperidone in schizophrenia
Desirability of Outcome Ranking (DOOR) will be defined a-priori from input from expert psychiatrists who have experience in treating patients with schizophrenia. The outcome is a composite outcome, combining the efficiency and safety measures together.
The predictor / independent variable is the treatment assignment (i.e drug vs placebo). The analysis will include the treatment assignment, age, gender, race/ethnicity, and other potential risk factors.
We will evaluate the heterogeneity of the composite outcomes in the subgroups of age, gender and race/ethnicity.
We will generate tables that summarize the distribution and extent of missingness of potential risk factors, for example, age, gender, and race/ethnicity, overall and by treatment arm to assess for random baseline imbalances for the trial data.
Our interest is to compare the DOOR between two treatments. The DOOR probability, proportion in favor of treatment and win ratio will be estimated by making all possible pairwise comparisons between two treatment arms according to DOOR, and their 95% CI will be constructed using bootstrap.
We will also evaluate heterogeneity of the composite outcome in several pre-specified subgroups, for example, age, gender, and race/ethnicity. These are hypothesis generating analyses. We will use the method described above to evaluate the effect of treatments stratified by subgroup. All subgroup analyses will be clearly reported, including how subgroups will be defined, outcomes examined, and how both the point and interval estimates of treatment effects.
Randomized clinical trials (RCT) provide rich data related to safety and efficacy of pharmacological agents. This is often at variance with real-world data since patients in the real-world i.e. (real-world data (RWD)) are not constrained by RCT design. Even when the results of RCT and RWD are consistent, they often do not apply directly to an individual patient who presents before a physician. We attempt to overcome these challenges by use of a novel statistical technique, Desirability of Outcome Ranking (DOOR), to better inform clinical-decision making in the treatment of schizophrenia.
Estimation of key milestone dates for the proposed study:
1) Anticipated project start date = Nov-2021
2) Data analysis completion = May 2022
3) Dissemination of results= Nov 2022
Dissemination plan includes presentation at conferences (such as ASCP Annual conference, ISCTM annual meeting) and publication in peer-reviewed journals (such as Journal of Clinical Psychiatry).
1. Claeys CK, Hopkins TL, Schlaffer K, Hitchcock S, Jiang Y, Evans S, Johnson JK, Leekha S. Comparing the Clinical Utility of Rapid Diagnostics for Treatment of Bloodstream Infections Using Desirability of Outcome Ranking Approach for the Management of Antibiotic Therapy (DOOR-MAT). Antimicrob Agents Chemother. 2021;65(9):e0044121.
2. Evans SR, Knutsson M, Amarenco P, Albers GW, Bath PM, et al. Methodologies for pragmatic and efficient assessment of benefits and harms: Application to the SOCRATES trial. Clin Trials. 2020 Dec;17(6):617-626.
