More on the weekend effect – this time concerning the CLAHRC WM Director’s speciality, obstetrics. This was another large database study, with an impressive 1.3 million deliveries from the English National Health Service database. The weekend effect is the difference in outcomes from weekend admissions versus weekday admissions. However, the main study finding was that “the perinatal mortality rate was 7.3 per 1000 babies delivered at weekends, 0.9 per 1000 higher than for weekdays (adjusted odds ratio 1.07, 95% confidence interval 1.02 to 1.13)”. But look at the figure and the rate was even higher on a Thursday, and Wednesday and Thursday combined would yield a similar increase over the mean. Granted, weekend was an a priori hypothesis and Thursday was not, but that does not mean we should ignore the observation – the babies who died were not aware of this ‘rule of the game’ in frequentist statistics. The authors make the usual mistake of enumerating preventable deaths on the basis of ‘cause and effect’ assumption. They came up with 770 per year.
By the time the CLAHRC WM Director received the print version of the BMJ, 36 critical comments had already been posted online.
Reinhart and Rogoff made their data of factors associated with economic depressions available to other scientists  – the CLAHRC WM Director assumes that Palmer and colleagues will also make their data available for re-analysis?
— Richard Lilford, CLAHRC WM Director
- Palmer WL, Bottle A, Aylin P. Association between day of delivery and obstetric outcomes: observational study. BMJ. 2015; 351: h5774.
- Reinhart CM & Rogoff K. This Time is Different: Eight Centuries of Financial Folly. Princeton, NJ: Princeton University Press, 2011.