Re: The survival time of chocolates on hospital wards: covert observational study
In the observational study by Gajendragadkar et al.(1), the half-time survival for a chocolate (99 minutes) was reported for the overall population (N=258 chocolates), but no separate estimates for this end-point were provided for the two subgroups under examination (Quality Street vs Roses).
This comparison was however addressed by calculating another effect measure -the hazard ratio (HR)-, the results of which showed a significantly longer survival for the Quality Street chocolates (hazard ratio for survival of Roses vs Quality Street of 0.70, 95% confidence interval from 0.53 to 0.93, which corresponds to a HR for all-cause mortality for Roses vs Quality Street of 1.43, 95% confidence interval from 1.07 to 1.89).
Since absolute effect measures are known to be more informative than relative ones (2), I have re-analyzed the data published by Gajendragadkar et al.(1) and I have converted their result, originally expressed as a relative effect measure, into the corresponding result expressed as absolute effect measure.
Using simple arithmetic calculations (3) and weighing the effect magnitude according to the sample sizes of 135 chocolates for Quality Street and 123 for Roses, I estimated the half-time values of 116 minutes for Quality Street and 81 minutes for Roses, which were consistent with the above-mentioned HR values, and were also in keeping with the medians shown in Gajendragadkar’s Figure 1. Hence, when expressed as an absolute effect measure, the difference between these two half-times is 35 minutes.
In my view, apart from the undisputed statistical significance of this result, this finding is also highly relevant in practical terms, even though one could argue whether this increase of 35 minutes is advantageous or disadvantageous.
Finally, in the Results section, one statement (“the median survival time of a chocolate was 51 minutes, 95% confidence interval 39 to 63”) needs some clarification because the reported values differ from the findings presented in Figure 1.
1. Gajendragadkar PR, Moualed DJ, Nicolson PLR, Adjei FD, Cakebread HE, Duehmke RM,
Martin CA. , The survival time of chocolates on hospital wards: covert observational study. BMJ 2013;347:f7198
2. King NB, Harper S, Young ME. Use of relative and absolute effect measures in reporting health inequalities: structured review. BMJ 2012;345:e5774a
3. Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials. 2007;8:16.
Competing interests: I dislike surrogate end-points and surrogate chocolate