Regression approaches
Regression models can be used to model relationships between temporal trends in population health outcomes and population medicine use. The strongest examples of such analyses are those in which the design enables the researcher to demonstrate a dose-response relationship between medicine use within subsets of the population and health outcomes (e.g. [46]), or between different populations with different levels of exposure to the medicines (e.g. [47]).
Interrupted time series analysis
Aggregate data on mortality and medicine use each comprise a time series. The health outcomes time series may consist, for example, of monthly mortality from a specific cause in 5-year age groups in a population over a number of years while the medicine use time series may consist of monthly sales data or prescription numbers for a drug or a drug class over the same time period. Interrupted Time Series (ITS) analysis is a set of statistical methods that can be used to assess the impact of an intervention (such as the introduction of a new medicine) on a health outcome time series [48-50]. The onset of the intervention is usually specified as the date when a new drug was introduced into a market.
A family of statistical methods can be used to analyze the effects of an intervention on time series data while taking account of autocorrelation: correlations between data at different time points that invalidates conventional statistical methods such as ordinary linear regression [51,52]. ITS analyses may involve using segmented linear regression methods [48,52] or Auto-regressive Integrated Moving Average (ARIMA) models [51]. More recently, econometricians have used generalized least squares or partial likelihood methods to fit more complex models to time series data [49]. All such statistical methods enable the effect of the intervention to be separated from general trends and serial dependencies in time so that valid statistical inferences can be made about whether an intervention has had an effect on the time series [50].
Answers to all your Biology Questions
