The projected increase of vertebral osteomyelitis in Germany implies a demanding challenge for future healthcare management of aging populations

Abstract

Purpose: Since an increase in the occurrence of native vertebral osteomyelitis (VO) is expected and reliable projections are missing, it is urgent to provide a reliable forecast model and make it a part of future health care considerations.

Methods: Comprehensive nationwide data provided by the Federal Statistical Office of Germany were used to forecast total numbers and incidence rates (IR) of VO as a function of age and gender until 2040. Projections were done using autoregressive integrated moving average model on historical data from 2005 to 2019 in relation to official population projections from 2020 to 2040.

Results: The IR of VO is expected to increase from 12.4 in 2019 to 21.5 per 100,000 inhabitants [95% CI 20.9-22.1] in 2040. The highest increase is predicted in patients over 75 years of age for both men and women leading to a steep increase in absolute numbers, which is fourfold higher compared to patients younger than 75 years. While the IR per age group will not increase any further after 2035, the subsequent increase is due to a higher number of individuals aged 75 years or older.

Conclusions: Our data suggest that increasing IR of VO will seriously challenge healthcare systems, particularly due to demographic change and increasing proportions of populations turning 75 years and older. With respect to globally fast aging populations, future health care policies need to address this burden by anticipating limitations in financial and human resources and developing high-level evidence-based guidelines for prevention and interdisciplinary treatment.

Keywords: Forecast model; Osteomyelitis; Projection; Spondylodiscitis; Vertebral infection.

Fig. 1

Accuracy criteria (A) and projected total number of VO (B; confidence intervals in shades) until 2040 depending on the model used. ARIMA model (autoregressive integrated moving average), ETS model (exponential smoothing), Poisson (Quasi-Poisson regression), and Log (logarithmic regression) are established forecast models. According to key performance indicators (MAPE, RMSE, MAE, AIC) that measure the accuracy of the different forecast models, ARIMA model is the most robust and provides high accuracy data

Fig. 2

Projected total number of VO for women and men (A) in total: historical data (blue and red line) from 2005 to 2019 and projected numbers with confidence intervals in shades from 2022 to 2040 (B) and per age group from 2005 until 2040

Fig. 3

Projected incidence rates of VO per age group until 2040. The clinically relevant climax of incidence rate is projected between 2030 and 2035 for each age group. The largest incidence is shown for patients aged 75–79 years, 80–84 years, and ≥ 85 years

Fig. 4

Past and projected total numbers of VO for men and women per localization on the spine until 2040. Abscissa and ordinate are scaled identically to provide better comparability between different levels. The highest incidence rate of VO is projected in the area of the lumbar spine for both women and men