Estimation of Disease Incidence in Claims Data Dependent on the Length of Follow‐Up: A Methodological Approach

Objective

To analyze the impact of the length of disease‐free intervals on incidence estimation.

Data Source

Statutory health insurance sample in Germany.

Study Design

Overestimation of the incidence in the first quarter of 2008 for three selected diseases, diabetes mellitus, colorectal cancer, and heart failure, depending on different lengths of preceding disease‐free intervals.

Data Collection/Extraction Methods

Continuously insured from 2000 until 2008 ≥18 years ( = 144,907).

Principal Findings

Compared with an 8‐year disease‐free period, incidence overestimations for diabetes, colorectal cancer, and heart failure were 40, 23, and 43 percent defining a 1‐year, and 5, 9, and 5 percent defining a 5‐year disease‐free period, respectively.

Conclusions

Depending on the specific disease, caution has to be taken while using short disease‐free periods because incidence estimates may be extremely overestimated.