Objective: The aims of the present study were to establish rates of resuscitation order documentation of patients aged ≥65 years from both psychogeriatric and general medical units and to compare patients on predictors of resuscitation status, particularly examining the effect of depression. Methods: A retrospective case note audit of psychogeriatric (n = 162) and general medical (n = 135) unit admissions within a tertiary teaching hospital was performed. Multivariate logistic regression was used to determine significant clinical and demographic predictors of resuscitation status. Results: Resuscitation orders were documented in more psychogeriatric (94.4%) than general medical (48.1%) files. Depression did not significantly predict resuscitation status in either group. Having undergone competency assessment significantly predicted resuscitation status for the total sample and separately for psychogeriatric and medical patients. Older age (overall sample), poorer prognosis (overall sample), living in residential care (overall sample and medical group) and self-consenting to resuscitation status (overall sample and medical group) significantly predicted resuscitation status. Conclusions: Resuscitation orders were more frequently documented on the psychogeriatric unit. Further prospective analysis is needed of how resuscitation orders are made before depression is discounted as a predictor of end-of-life decision-making. What is known about the topic?: Despite increased community, media and research attention to end-of-life decision-making, resuscitation preferences of older patients are often poorly documented. Existing research into patient clinical and demographic factors that influence end-of-life decision-making have largely focused on general medical rather than psychogeriatric settings. There is a need to investigate rates of resuscitation documentation in psychogeriatric and general medical units and specific factors associated with having a 'do not attempt resuscitation' order in place, particularly the effect of current depression on decision-making. What does this paper add?: Resuscitation orders were more frequently documented on the psychogeriatric than medical unit. Depression was not a significant predictor of resuscitation status in either group of patients. Although having undergone a competency assessment, older age and poorer prognosis predicted not being for resuscitation for the total sample, living in residential care and self-consenting to resuscitation status predicted not being for resuscitation for the overall sample and the medical group specifically. What are the implications for practitioners?: This paper suggests that the need for clinicians to ensure documentation of preferences is a focus of day-to-day work with older patients. Clinicians should consider patient competency in end-of-life decision-making and how factors associated with depression, such as helplessness, may be more closely related to resuscitation decision-making in older patients.