SMART (study)

SMART (study) was a major clinical investigation that examined adaptive treatment strategies within a sequential, multiple-assignment randomized trial framework. The project connected concepts from behavioral science, biostatistics, and clinical practice and engaged researchers across institutions such as National Institutes of Health, Johns Hopkins University, Harvard University, Stanford University, and University of Pennsylvania. It informed subsequent work at centers including Massachusetts General Hospital, University of California, San Francisco, and University of Michigan and influenced guidelines from agencies such as the Food and Drug Administration and the Centers for Disease Control and Prevention.

Background

The study emerged from prior methodological advances exemplified by work at Northwestern University, University of Pittsburgh, and Columbia University on adaptive interventions and dynamic treatment regimens. It responded to calls for pragmatic designs advocated by groups at Agency for Healthcare Research and Quality and demonstration projects funded by the National Institute of Mental Health and the National Cancer Institute. Precedents included pilot trials at Yale University, Duke University, and Vanderbilt University that tested staged therapies, and conceptual foundations traceable to influential statisticians at University of Washington and Carnegie Mellon University. Collaborations involved clinical networks such as Kaiser Permanente and consortia like the Clinical and Translational Science Awards Program.

Study Design and Methods

The protocol used a sequential multiple-assignment randomized trial architecture pioneered in methodological papers from Northwestern University and formalized by investigators affiliated with Harvard University, University of Michigan, and Duke University. Participants were recruited from sites including Mayo Clinic, Cleveland Clinic, and community clinics partnered with Columbia University and were randomized at decision points to different treatment options drawn from portfolios developed at Stanford University and Johns Hopkins University. Outcomes were assessed using measures validated in studies at UCLA, Brown University, and University of Chicago. Statistical planning referenced approaches from University of California, Berkeley and Princeton University, and analytic methods included techniques from teams at Massachusetts Institute of Technology and University of Pennsylvania. The trial governance incorporated data safety monitoring recommended by advisors from World Health Organization consultations and ethics review modeled on frameworks used by Johns Hopkins Bloomberg School of Public Health.

Results

Primary outcome analyses drew on longitudinal methods advanced at Columbia University, University of Michigan, and Harvard T.H. Chan School of Public Health. The study reported differential response patterns consistent with findings earlier published by investigators at Yale University and Vanderbilt University. Subgroup and moderator analyses referenced approaches from University of Washington and Northwestern University; sensitivity checks paralleled methods used at Brown University and Duke University. Effect estimates and confidence intervals were presented in formats familiar to audiences of journals affiliated with American Medical Association, BMJ Group, and The Lancet. Secondary endpoints compared trajectories that echoed results in trials conducted at Mayo Clinic, Cleveland Clinic, and Kaiser Permanente networks.

Interpretation and Impact

Authors interpreted findings in the context of translational priorities emphasized by National Institutes of Health strategic plans and clinical practice guidelines by organizations such as American Psychiatric Association, American Heart Association, and American College of Physicians. The study influenced implementation projects at Massachusetts General Hospital and policy discussions at Centers for Disease Control and Prevention and Food and Drug Administration workshops. Educational programs at Harvard Medical School, Johns Hopkins School of Medicine, and Stanford School of Medicine incorporated the trial's methods into curricula, and subsequent grant solicitations from National Institute of Mental Health and National Institute on Drug Abuse cited its design. Methodological follow-ups appeared from groups at Northwestern University, University of Pennsylvania, and University of California, San Francisco.

Criticisms and Limitations

Critiques mirrored concerns raised in methodological debates at Princeton University, Columbia University, and University of Chicago about external validity, sample heterogeneity, and complexity of implementation. Commentators from Yale University and Duke University questioned generalizability to settings outside specialty centers like Mayo Clinic and Cleveland Clinic. Statistical limitations discussed by analysts at Massachusetts Institute of Technology and University of Washington included potential model misspecification and power calculations; ethical and regulatory commentators at Harvard University and Johns Hopkins University highlighted consent complexities for multi-stage randomization. Subsequent replication attempts at University of Michigan, Vanderbilt University, and Brown University sought to address these concerns and refine analytic frameworks.

Category:Clinical trials