If improved population health outcomes and population medicine use are associated, we need to assess whether the association can be plausibly attributed to other factors. A widely used strategy for causal inference in this situation is quasi-experimentation [12,15]. In quasi-experimentation, the investigator first generates and then evaluates plausible alternative explanations of the relationships that have been observed with the aim of arriving, by a process of elimination, at a conclusion that medicines use in the population is the most plausible explanation of the improved population health outcomes [12,53].
Changes in measurement
The first alternative hypothesis that needs to be considered is that changes in population health outcomes reflect changes in the way that the health outcome has been measured. This could occur as a result of a change in the system of classification (e.g. the edition of the International Classification of Disease) that is used to code cause of death or morbidity, a change in diagnostic criteria, or increased attention to some causes of deaths that may increase the apparent rate of deaths attributable to that cause and reduce the rates of other causes of death [12,44,48]. These possibilities need to be excluded by evidence that coding systems, diagnostic criteria or medical scrutiny have remained stable over the study period. If a change in measurement remains a plausible hypothesis, its effects may be assessed by sensitivity analyses.
Confounding factors
Rarely is a change in medicine use the only change that has occurred during the period of interest that may affect the population health outcome we are studying. Concurrent historical events that affect the health outcome are another plausible hypothesis that needs to be excluded [48]. One plausible alternative explanation is that improved population health outcomes may reflect a declining prevalence in the population of major risk factors for a disease. Another possibility is that changes in population health outcomes may be due to the effects of the increased use of other medicines or medical interventions in the population.
Answers to all your Biology Questions
