OBJECTIVES: To review design characteristics of T&M studies applied to healthcare, with a focus on choice of study design, statistical methodology, and handling of multi-centre data. METHODS: A PubMed search was performed using key search terms including “time and motion” (MeSH Term) AND any of the following: cost (analysis), (health) economics, observation(al), and prospective.  Articles (English; 2008 or later) were selected based on the following criteria: (1) observational study using T&M methodology; and (2) task-based data collection. Studies that measured broad aggregate health care professional tasks/hospital workflows, in the absence of task- or event-specific timings, were excluded.   RESULTS: Of 191 identified abstracts, 151 were excluded during screening; upon review, 21 of 40 remaining were retained for detailed assessment. Half (48%) were applicable to Europe, of which 2 were multi-country studies. Medical interventions studied were: drug (48%), diagnostic process (14%), medical procedure (24%), and IT systems to improve clinical management (e.g. EMR) (14%). The majority (86%) of studies were hospital-based, 86% were observational, and 14% employed hybrid methods, including chart review or survey. Only 20% used independent observers. Three quarters (76%) reported descriptive statistics. Of 9 multi-centre studies, one used a random effects regression model to account for “centre clustering”, and 8 reported pooled data (3 of which used a “mean of centre averages” approach). Eleven studies (52%) compared two groups, of which 3 applied an analytical design aiming to defect statistical differences, and 2 reported a sample size calculation. CONCLUSIONS: This review of T&M studies revealed that descriptive designs are most common (analytical designs using power calculations seem rare). Multi-centre comparator studies rarely use random effects regression models to account for “centre clustering”, though considered the method of choice to produce valid confidence intervals around point estimates. In general, statistical methodology is scarcely reported, affecting overall study credibility.