There is an ethical imperative to provide the best quality of care for all people, including people at the end of life. In order to provide such care, each intervention used in clinical practice must be tested objectively for net clinical benefit (response and toxicity). For pharmacological and non-pharmacological interventions, the gold standard of evidence remains a phase III randomized, controlled study. In palliative care, there is a gradually emerging evidence base for clinical practice derived from rigorously designed and conducted interventional studies. Such studies can effectively recruit, even in this clinical setting, to completion.There are design and implantation challenges in all controlled clinical trials. Some issues are highly likely in palliative care and will influence trial design and analysis including expected attrition unrelated to the intervention, wide variation in the diagnosis and timing of referral to individual palliative care services and responsible reporting of serious adverse events. Factors that are not unique to palliative care but are often magnified in such studies include concerns by referrers and participants about randomization and blinding, how and when to measure endpoints that are clinically relevant, how to minimize the burden of data collection, ensuring studies are adequately powered and how best to deal with primary outcomes that are subjective. Many studies in the palliative care population have successfully overcome these challenges by thoughtful study design. Factors associated with successful (palliative care) study completion include (i) a shared vision about what the research is designed to achieve; (ii) governance structures independent of the researchers; (iii) ways of defining standard practice for the control arm using best available evidence and consensus; (iv) face-to-face meetings especially for protocol development; (v) adequate funding given estimates that four to six palliative care patients will be screened for every person who completes the study; (vi) collaboration with a wide range of experts in clinical trials and other clinical disciplines of importance to the study question; (vii) adequate timelines given recruitment difficulties; (viii) adequate feasibility planning; (ix) standard operating procedures; (x) web-based data management in multi-site research; and (xi) site visits as a two-way learning and quality process, especially for services that are new to controlled clinical trials. Systematically, analyses need to include the characteristics of the sub-population(s) who will most predictably benefit from an intervention.Improving the quality of care for people at the end of life with rigorously designed, adequately powered studies that can inform clinical practice and policy directly, is an important health research agenda. Such studies are feasible and can help to inform practice improvement.