Quantitative modelling of hip fracture trends in 14 European countries: testing variations of a shared reversal over time

Abstract

Qualitative similarities between hip fracture trends in different countries suggests variations of the same epidemic. We tested a single statistical shape to describe time trends in Europe, while allowing for country-level variability. Using data from 14 countries, we modelled incidence rates over time using linear mixed-effects models, including the fixed effects of calendar year and age. Random effects were tested to quantify country-level variability in background rates, timing of trend reversal and tempo of reversal. Mixture models were applied to identify clusters of countries defined by common behavioural features. A quadratic function of time, with random effects for background rates and timing of trend reversal, adjusted well to the observed data. Predicted trend reversal occurred on average in 1999 in women (peak incidence about 600 per 100 000) and 2000 in men (about 300 per 100 000). Mixture modelling of country-level effects suggested three clusters for women and two for men. In both sexes, Scandinavia showed higher rates but earlier trend reversals, whereas later trend reversals but lower peak incidences were found in Southern Europe and most of Central Europe. Our finding of a similar overall reversal pattern suggests that different countries show variations of a shared hip fracture epidemic.