“We Just Don’t Take Delirium That Seriously”: A Mixed-Methods Behavioural Analysis of the Determinants and Outcomes of Delirium Recognition and Management among Hospitalized Older Adults in Ghana

Background Delirium is highly prevalent among hospitalised older adults in sub‑Saharan Africa, yet little is known about the determinants that influence its recognition and management by healthcare providers in low- and middle-income settings. We quantified delirium recognition, assessed its correlates and examined its association with mortality and length of stay among hospitalized older Ghanaians, and explored behavioural factors influencing delirium care by physicians in a Ghanaian tertiary hospital. Methods Using an explanatory sequential mixed‑methods design, we reviewed hospital records of 260 adults aged ≥ 60 years with researcher‑confirmed delirium using the Confusion Assessment Method (CAM). Delirium recognition was defined as chart documentation of “delirium” or altered mentation. Demographic, clinical and outcome (inpatient mortality and length of stay) variables were obtained from caregivers or abstracted from hospital records. Predictors of delirium recognition were evaluated correlations with multivariate logistic regression. Chi-square and Student’s T-tests examined the impact of delirium recognition on mortality and length of stay. We then conducted 13 semi‑structured interviews with emergency and internal medicine physicians. Transcripts were analysed using the Capability, Opportunity, Motivation and Behaviour (COM-B) Model to characterize determinants of delirium diagnosis and management. Results Delirium was recognised in 88/260 (33.8%) adults but formally diagnosed in only one (< 0.4%). Recognition was not associated with decreased mortality or length of stay. Capability (e.g. limited delirium knowledge, clinical skill, and geriatrics exposure), Opportunity (e.g. overcrowding, time pressure, operational inefficiencies, lack of staff sensitization, screening tools and protocols), and Motivation (e.g. low perceived importance, misconceptions) barriers explained both low diagnosis rates and the absence of outcome differences. Conclusion Delirium care in this setting is constrained by behavioural barriers across the domains of capability, opportunity, and motivation. Delirium recognition did not improve outcomes as physicians lacked the skills, environmental support, and motivation required to act on it. Behaviourally informed interventions like brief delirium screening tools, workforce training, caregiver engagement, and delirium‑friendly workflows are needed to strengthen delirium detection and management to improve outcomes in resource‑constrained settings.