[Secondary Data-based Trend Analysis of How Quality Differences in Two Forms of Healthcare Develop Over Time]

The evaluation of complex interventions such as different forms of healthcare are methodologically challenging. The aim of this study was to use generalized estimating equations (GEE) to investigate how quality differences between family practitioner-based healthcare (HZV) and standard care develop over time.

A longitudinal secondary data analysis for 2011-2018 was conducted using routine data from the AOK Baden Wuerttemberg health insurer. A dynamic cohort of insured persons that fulfilled the inclusion criteria were included in the analysis at the beginning of each year of observation, so the size and composition of the cohorts varied from year to year. Quality differences between HZV and standard care were investigated, whereby an autoregressive covariance structure (AR1) was assumed for multiple measurements. Under observation were a group variable for healthcare type, a variable for year of observation, and any interaction between the two. The resulting estimates provided information on the relationship between the two groups at the beginning of the observation period, and on how the groups developed both individually and in relation to one another over time.

The GEE were used exemplarily on the quality indicator prescription of potentially inappropriate medication (PIM) in elderly insured patients. At the beginning of the observation period, the chance of PIM in those over 65 years of age (year 2018; N=628,523) was significantly lower in the HZV group than in the group receiving standard care (odds ratio 0.978; 95% confidence interval: 0.968-0.987). The chance of a PIM in the following seven years declined in both groups, but faster in the HZV group than the group receiving standard care.

A secondary data-based trend analysis with GEE of quality differences in comparison groups over time has considerable potential in the evaluation of new and existing forms of healthcare.