Elsevier

Bone

Volume 110, May 2018, Pages 170-176
Bone

Full Length Article
Development of a risk assessment tool for osteoporotic fracture prevention: A claims data approach

https://doi.org/10.1016/j.bone.2018.02.002Get rights and content

Highlights

  • Development of a fracture risk assessment tool based on administrative claims data.

  • Compared to a model with age, sex and prior fracture adding other risk factors involved small improvement in discrimination.

  • The model can be readily employed by healthcare decision-makers to identify and improve targeting of high risk populations.

Abstract

Background

In aging societies osteoporotic fractures are a major health problem with high economic costs. Targeting prevention at individuals at high risk is important to reduce the future burden of fractures. Available risk assessment tools (e.g., FRAX®, QFracture, the algorithm provided by the German Osteology Society (DVO-Tool)) rely on self-reported patient information to predict fracture risk. Time and resource constraints, limited access to clinical data, and (un)willingness to participate may hamper the use of these tools. To overcome such obstacles, the aim is to develop a fracture risk assessment tool based on claims data that may be directly used on an institutional level.

Methods

Administrative claims data of an elderly (≥65 years) population (N = 298,530) for the period from 2006 through 2014 was used. Major osteoporotic fractures (MOF) were identified based on hospital diagnoses. We applied Cox proportional hazard regression to determine the association of individual risk factors and fracture risk. Hazard ratios were used to construct a risk score. The discriminative ability of the score was evaluated using C-statistics.

Results

We identified 7864 MOF during follow-up. The median time to first fracture during follow-up was 371.5 days. Individuals with a MOF during follow-up had a higher mean and median risk score (mean: 4.53; median: 4) than individuals without MOF (mean: 3.07; median: 3). Adding drug-related risk factors slightly improved discrimination compared to a simple model with age, gender, and prior fracture.

Conclusion

We developed a fracture risk score model based on in-hospital treated subjects to predict MOF that can be used on an institutional level. The score included age, sex and prior fracture as risk factors. Adding other risk factors involved very small improvement in discrimination.

Introduction

Osteoporotic fractures represent a major health problem in aging societies, placing high health and economic burden on the individual and the public health system [1]. Osteoporotic fractures are associated with a loss of quality of life, a decrease in functional mobility, and an increase in mortality [[2], [3], [4], [5], [6]]. In developed countries the lifetime risk for osteoporotic fractures at fracture sites including hip, spine, and wrist has been estimated to amount to 30%–40% [7]. With the rapid rise in the number of older people osteoporotic fractures may become a prime future health concern as these fractures have been shown to correlate strongly with age [[8], [9], [10], [11]]. Findings for Germany indicate an increase in the total number of incident osteoporotic fractures by 238% between 2010 and 2050 if fracture rates remain constant. Unsurprisingly, total direct costs of incident osteoporotic fractures are projected to rise to 3.2 billion Euros in 2030 from approximately 898 million Euros in 2010 [12]. Therefore, preventive measures to reduce the risk of fractures are crucial; but such measures can be very expensive, adding to costs.

Various national and international fracture risk assessment tools exist to identify persons at increased risk of osteoporotic fractures. FRAX® [13], adapted in several countries, is probably the most prominent tool besides QFracture [14] that includes more risk factors and in greater detail. On national level there is the tool developed by the German Osteology organization (German DVO-Tool, [15]). These tools aim at assisting clinicians in the management of their patients through calculation of 5-year or 10-year fracture risk. To estimate individual fracture risk, these tools involve direct patient assessment and information about their risk factors. However, identifying potential persons for risk assessment may strongly depend on the physician's knowledge about the existence of available risk models, the physician's time budget, and patients' willingness to participate. A risk assessment tool based on readily available data may overcome such obstacles and could be used to identify persons at risk by an automatized algorithm. Such a case finding strategy would enable health care funds to offer their insured persons specific measures to improve bone health and reduce fall risk. Hence, the objective of this study was to develop a fracture risk assessment tool based on claims data that may be directly performed on health institutional and supplier level.

Section snippets

Study population and data source

The study was based on data provided by the German social insurance for agriculture, forestry and horticulture (Sozialversicherung für Landwirtschaft, Forsten und Gartenbau, SVLFG). The SVLFG is responsible for carrying out the agricultural accident insurance, the old-age provisions for farmers, the agricultural health insurance and the agricultural long-term care insurance. The extensive dataset contained data from 2006 through 2014. Information on age, gender, date and reason for termination

Results

Table 1 presents the descriptive characteristics of the derivation, the validation, and the entire study cohort. The entire study cohort (298,530 individuals) had a mean age of 75.43 years (SD 6.28) (at index date) and slightly less than half were women (48.78%). A prior major fracture within the 2 years preceding index date was recorded for 1.60% of the cohort. A total of 7864 MOF occurred during the 2 years follow-up and the median time to the first major fracture during follow-up was 371.5 days

Discussion

In this study we developed a risk assessment model for fracture prediction based on claims data that can be used for recommendations for preventive measures. We chose to follow a pragmatic approach considering risk factors that can be easily and reliably retrieved from claims databases and implemented in routine health care.

Several risk factors were identified to be associated with fracture risk for common osteoporotic fractures sites (i.e., hip, spine, forearm, humerus [17,18]). We found a

Conclusion

In summary, we developed a risk score to predict osteoporotic fracture risk in an elderly population. The score is based on a large administrative claims dataset and can be readily used by decision-making bodies at the institutional level or when self-reported patient information is not accessible. The discrimination statistics suggest that a model including drug-related risk factors improves predictive accuracy, but a simple model with age, gender and prior fracture may be used without

Ethical approval

Anonymised data were obtained by the insurance company. Insurants were not contacted personally for informed consent. The study was approved by the Ethical Committee of Ulm University.

Funding

The study was supported by the German Federal Ministry of Education and Research (grant no. 01EC1404D).

Conflicts of interest

Katrin Reber (KCR), Ivonne Lindlbauer (IL), Kilian Rapp (KR), Clemens Becker (CB), Gisela Büchele (GB), Sarah Mächler (SM), and Hans-Helmut König (HHK) declare that they have no conflict of interest.

Author contributorship

KCR, IL, KR, CB, GB and HHK planned the study. KCR, IL, KR, SM and HHK developed methodological approach. KCR, IL conducted statistical analyses. KCR wrote the manuscript. All authors participated in editing and revising the drafted version of the manuscript.

Acknowledgement

The authors thank the SVLFG and especially Daniel Stöger and Andrea Grunz from the SVLFG for granting access to the data and data support. The authors also wish to thank Andreas Meid from the Department of Clinical Pharmacology and Pharmacoepidemiology, University of Heidelberg for helpful advice and support of our analyses.

References (50)

  • E. Osnes et al.

    Consequences of hip fracture on activities of daily life and residential needs

    Osteoporos. Int.

    (2004)
  • M. Hiligsmann et al.

    Utility values associated with osteoporotic fracture: a systematic review of the literature

    Calcif. Tissue Int.

    (2008)
  • B. Abrahamsen et al.

    Excess mortality following hip fracture: a systematic epidemiological review

    Osteoporos. Int.

    (2009)
  • S. Morin et al.

    Mortality rates after incident non-traumatic fractures in older men and women

    Osteoporos. Int.

    (2011)
  • S.L. Hui et al.

    Age and bone mass as predictors of fracture in a prospective study

    J. Clin. Investig.

    (1988)
  • A. Konnopka et al.

    The health and economic consequences of osteopenia-and osteoporosis-attributable hip fractures in Germany: estimation for 2002 and projection until 2050

    Osteoporos. Int.

    (2009)
  • K.E. Ensrud

    Epidemiology of fracture risk with advancing age

    J. Gerontol. Ser. A Biol. Med. Sci.

    (2013)
  • F. Bleibler et al.

    The health burden and costs of incident fractures attributable to osteoporosis from 2010 to 2050 in Germany—a demographic simulation model

    Osteoporos. Int.

    (2013)
  • J. Kanis et al.

    FRAX™ and the assessment of fracture probability in men and women from the UK

    Osteoporos. Int.

    (2008)
  • J. Hippisley-Cox et al.

    Derivation and Validation of Updated QFracture Algorithm to Predict Risk of Osteoporotic Fracture in Primary Care in the United Kingdom: Prospective Open Cohort Study

    (2012)
  • DVO-Leitlinie D

    Zur Prophylaxe, Diagnostik und Therapie der Osteoporose bei Männern ab dem 60

  • A. Icks et al.

    Trend of hip fracture incidence in Germany 1995–2004: a population-based study

    Osteoporos. Int.

    (2008)
  • J. Kanis et al.

    Long-term risk of osteoporotic fracture in Malmö

    Osteoporos. Int.

    (2000)
  • D. Bliuc et al.

    Mortality risk associated with low-trauma osteoporotic fracture and subsequent fracture in men and women

    JAMA

    (2009)
  • H. Johansson et al.

    Imminent risk of fracture after fracture

    Osteoporos. Int.

    (2017)
  • Cited by (15)

    View all citing articles on Scopus
    View full text