BackgroundAs in other areas of health delivery, there is a need to ensure that end-of-life care is guided by patient centred research. A systematic review was undertaken to examine the quantity and quality of data-based research aimed at improving the (a) processes and (b) outcomes associated with delivering end-of-life care in hospital settings.MethodsMedline, EMBASE and Cochrane databases were searched between 1995 and 2015 for data-based papers. Eligible papers were classified as descriptive, measurement or intervention studies. Intervention studies were categorised according to whether the primary aim was to improve: (a) end of life processes (i.e. end-of-life documentation and discussions, referrals); or (b) end-of-life outcomes (i.e. perceived quality of life, health status, health care use, costs). Intervention studies were assessed against the Effective Practice and Organisation of Care methodological criteria for research design, and their effectiveness examined.ResultsA total of 416 papers met eligibility criteria. The number increased by 13% each year (p < 0.001). Most studies were descriptive (n = 351, 85%), with fewer measurement (n = 17) and intervention studies (n = 48; 10%). Only 18 intervention studies (4%) met EPOC design criteria. Most reported benefits for end-of-life processes including end-of-life discussions and documentation (9/11). Impact on end-of-life outcomes was mixed, with some benefit for psychosocial distress, satisfaction and concordance in care (3/7).ConclusionMore methodologically robust studies are needed to evaluate the impact of interventions on end-of-life processes, including whether changes in processes translate to improved end-of-life outcomes. Interventions which target both the patient and substitute decision maker in an effort to achieve these changes would be beneficial.Electronic supplementary materialThe online version of this article (doi:10.1186/s12904-017-0204-1) contains supplementary material, which is available to authorized users.