I read with interest the article by Lu and colleagues1 which shows that quality indicators of pharmacotherapy are associated with increased odds of admission to hospital, but reduced odds of death. However, some problems in the analysis may explain the results. Specifically, based on Tables 1 and 3, there are 15 102 patients without use of drugs included in the analysis of the effects of potentially inappropriate medications (PIMs) and anticholinergic burden on the outcomes but not included in the analysis of polypharmacy. Because of the high mortality of these patients (91.6%), they may not be representative of general older adults. If these patients are excluded, the mortality rates in patients without the use of PIMs or anticholinergics are much lower than those in the original reports (24.6% v. 56.2% and 21.3% v. 41.9%, respectively).