Will Investigators Enroll Particular Subjects in a Randomized Controlled Trial?: A Mixed-Methods Study to Gauge Investigator Equipoise in a Trial of Surgery Versus Nonoperative Therapy in Subjects with Meniscal Tear and Persistent Pain Following Physical Therapy.

Background: Clinician equipoise-indifference between treatment arms-provides an ethical foundation for clinician participation in randomized controlled trials (RCTs). In preparation for an RCT comparing arthroscopic partial meniscectomy versus enhanced nonoperative therapy for patients with a meniscal tear and knee osteoarthritis (OA) who remain symptomatic after a course of physical therapy, we conducted a mixed-methods study to assess equipoise among potential enrolling clinicians.

Methods: Fifteen clinicians with experience managing meniscal tears assessed 29 vignettes of hypothetical patients who met trial eligibility criteria. We randomly varied 13 vignette features (e.g., age, sex, radiographic severity, tear morphology). Clinicians expressed their willingness to enroll each hypothetical patient. After polling, we recorded and transcribed a moderated discussion to document clinician thought processes. We performed a quantitative analysis to identify clinical features associated with the likelihood of enrollment and an exploratory thematic analysis of the transcribed discussion to explicate the quantitative findings.

Results: The 15 orthopaedic surgeons and physician assistants assessed 29 vignettes describing hypothetical patients. Eight votes were missing, leaving 427 vignettes, of which the clinicians were willing to enroll 302 (71%) (range, 24% to 100%) in the trial. Three clinicians were willing to enroll <50% of vignettes. Clinicians were willing to enroll just 39% of vignettes with bucket-handle tears. In logistic regression analyses, a bucket-handle tear (adjusted odds ratio [aOR], 0.12; 95% confidence interval [CI], 0.04 to 0.37) and Kellgren-Lawrence 3 radiographs (aOR, 0.54; 95% CI, 0.36 to 0.82) were independently associated with clinician unwillingness to randomize. The qualitative analysis confirmed that clinicians believed that bucket-handle tears should be managed operatively, whereas combinations of greater age, severe OA, inability to walk 200 yards, and higher body mass index (BMI) pushed clinicians toward nonoperative therapy.

Conclusions: This methodology can be used to identify clinicians who should be engaged in discussions and interventions to support equipoise. It can also inform development of exclusion criteria (e.g., exclude bucket-handle tears) to increase the proportion of eligible subjects referred for enrollment.

Clinical relevance: Orthopedic surgeons play crucial roles in randomized controlled trials, the foundation of clinical practice guidelines, by enrolling patients from their practices. This article examines clinician equipoise, a key determinant of the willingness of clinicians to enroll eligible patients in trials.