• Cancer
• Methodological research
• Supportive care

The famous quote from Albert Einstein that ‘Everything that can be counted does not necessarily count; everything that counts cannot necessarily be counted’1 can perhaps be paraphrased to ‘Every patient that can be included does not necessarily count, every patients that counts cannot necessarily be included’. The study by Stone et al2 published in this issue of BMJ Supportive & Palliative Care highlights this. Stone et al demonstrated that in a carefully planned, non-interventional study with rather simple assessments, less than one-tenth of the patients admitted to the relevant units were included. The actual result from this observation is that the study may reflect the situation for a minor group of the patients only. This raises obvious concerns about the generalisability of findings from the study to those patients met in everyday clinical practice.

Stone et al should be recommended for reporting these data. There is reason to believe that these difficulties are under-reported in the Results section of many publications. Difficulties in recruitment and access to patients, and poor compliance are factors that are most likely under-communicated in the planning and conduction of studies and in the interpretation of the research.

What are the obstacles related to inclusion and what can be done in order to increase the number of included patients, thereby improving the generalisability of study results? The study by Stone et al categorised the patients who failed to be included into three categories. The first category included those who were ineligible for inclusion. Some reasons were nothing to be done with, …