The MePFAC study enrolled 162 men and women, with advanced cancer from 17 palliative care services across England, including hospital support teams, palliative care outpatient clinics, community palliative care services and hospices. Our study sample was reasonably balanced with regard to the sexes, with 86/159 (54%) of the sample being women. According to Cancer Research UK statistics,41 adults aged 50–74 account for more than half (54%) of all new cancer cases, and the four most common cancers are breast, prostate, lung and bowel. The mean age of participants in the MePFAC trial was 63.7 years, and the four most common cancers in our study population mirrored the frequency of these diagnoses in the general population. In keeping with what one would expect of a palliative care population, 41% of our study sample were on strong opioids and 78% had either recently received, were currently receiving or were scheduled to start palliative disease-modifying therapies at the time of study enrolment.
In an effort to keep data collection to a minimum, we did not collect data about the ethnicity of participants in the trial. The rationale for this was that we wished to only collect the minimum number of data to reduce participant burden. Indeed, in the first version of the protocol, to minimise data collection procedures, we did not even collect detailed diagnostic data about patients’ cancer diagnoses (we only recorded whether or not they fulfilled the study inclusion and exclusion criteria – advanced incurable cancer receiving palliative care). Moreover, we had no prior hypothesis (and there were no previous studies to suggest) that ethnicity would be a modifier of treatment response. However, in retrospect, the failure to collect ethnicity data was a limitation and has not allowed us to make any statements about the representativeness of included participants with regard to ethnic diversity.
The gender balance, age distribution, diagnostic spread, use of concomitant medication and wide geographical diversity of our recruitment strategy suggest that our findings are likely to be generalisable.