Re: Statistical methods to compare functional outcomes in randomized controlled trials with high mortality
Colantuoni et al. are to be commended for their review on methods to compare functional outcomes in trials with high mortality1. This is important in individuals with advanced life-limiting illnesses requiring palliative care2 where 18% of participants have been reported to have truncated data - for example, on quality of life - due to death3.
A survivors analysis in palliative care trials would be a reasonable approach to managing data truncated due to death if numbers and characteristics of those dying per arm were equal. A recent meta-analysis of palliative care trials found no evidence of differential attrition due to death4. However, a survivors analysis would only give an effect size for those well enough to survive to the primary endpoint, and would not provide generalisable findings for those who did not. This is significant in palliative care research where patients may be offered participation in a clinical trial with the knowledge that a proportion will not reach the end point, but these individuals cannot be identified at the time of enrolment. It is important to understand whether the treatment benefits or harms those who withdraw early due to deterioration.
We believe that palliative care trials, where the aim is to improve quality of life rather than survival, require the development of new methods of analysis that include outcome data from participants who die, reflect the outcome at all time points, take account of survival and are clinically interpretable. This could involve a new composite endpoint approach where mortality and the values of the outcome at all times before death are combined into a single variable. Importantly such variables could account for states worse than death, which is pertinent in palliative care where death is not necessarily a negative outcome. However, the composite endpoint approach may be too sensitive to random differences in survival. An alternative may be to perform a survivors analysis at each time point and then average the estimated treatment effects over time.
1. Colantuoni E, Scharfstein DO, Wang C, Hashem MD, Leroux A, Needham DM and Girard TD. Statistical methods to compare functional outcomes in randomized controlled trials with high mortality. BMJ 2018; 360: j5748. DOI: 10.1136/bmj.j5748.
2. Currow DC, Plummer JL, Kutner JS, Samsa GP and Abernethy AP. Analyzing phase III studies in hospice/palliative care. a solution that sits between intention-to-treat and per protocol analyses: the palliative-modified ITT analysis. J Pain Symptom Manage 2012; 44: 595-603. DOI: 10.1016/j.jpainsymman.2011.10.028.
3. Hussain JA, Bland M, Langan D, Johnson MJ, Currow DC and White IR. Quality of missing data reporting and handling in palliative care trials demonstrates that further development of the CONSORT statement is required: a systematic review. J Clin Epidemiol 2017; 88: 81-91. DOI: 10.1016/j.jclinepi.2017.05.009.
4. Hussain JA, White IR, Langan D, Johnson MJ, Currow DC, Torgerson DJ and Bland M. Missing data in randomized controlled trials testing palliative interventions pose a significant risk of bias and loss of power: a systematic review and meta-analyses. J Clin Epidemiol 2016; 74: 57-65. DOI: 10.1016/j.jclinepi.2015.12.003.
Competing interests: No competing interests