Aims and method We aimed to audit the documentation of decision-making capacity (DMC) assessments by our liaison psychiatry service against the legal criteria set out in the Mental Capacity Act 2005. We audited 3 months split over a 2-year period occurring before, during and after an educational intervention to staff. Results There were 21 assessments of DMC in month 1 (6.9% of all referrals), 27 (9.7%) in month 16, and 24 (6.6%) in month 21. Only during the intervention (month 16) did any meet our gold-standard (n = 2). Severity of consequences of the decision (odds ratio (OR) 24.4) and not agreeing to the intervention (OR = 21.8) were highly likely to result in lacking DMC. Clinical implications Our audit demonstrated that DMC assessments were infrequent and poorly documented, with no effect of our legally focused educational intervention demonstrated. Our findings of factors associated with the outcome of the assessment of DMC confirm the anecdotal beliefs in this area. Clinicians and service leads need to carefully consider how to make the legal model of DMC more meaningful to clinicians when striving to improve documentation of DMC assessments.