An evaluation of the Hawthorne effect in a clinical trial of trauma care in Ghana

Abstract

IntroductionHawthorne effect refers to changes in clinician behavior in response to being observed. It has been minimally evaluated in trauma and surgical research, especially in low‐ and middle‐income countries (LMICs).MethodsWe used data from the first (control) phase of a randomized clinical trial in Ghana (2020‐21). Trained observers were stationed in emergency units in eight non‐tertiary hospitals to observe the achievement of key performance indicators (KPI) (KPIs) of trauma care. We hypothesized that the observers' presence would create a Hawthorne effect, resulting in clinicians (doctors, nurses) performing above their usual level, with performance highest immediately after the observers were introduced, and with performance then declining over time toward the usual baseline.ResultsWe enrolled 2067 injured patients. Achievement of 13 out of 16 KPIs declined significantly over the 14 months after the introduction of observers. For example, airway assessment declined from 77% initially to 60% at the end of the phase (p < 0.001). Chest examination declined from 66% to 31% (p < 0.001). Distal pulses checked declined from 45% to 7% (p < 0.001). The slope of decline in KPI achievement ranged from 0.8%/month (mobility assessed) to 15.9%/month (distal pulses checked).ConclusionsThe decline in KPI achievement over time is consistent with the hypothesis that clinicians performed above their usual level just after the observers were introduced, with performance then declining toward the usual baseline. This study is among the first to provide quantitative data on the Hawthorne effect for trauma research in LMICs. These data are useful for the design of future observational studies and clinical trials.