Mapping chronic disease prevalence based on medication use and socio-demographic variables: an application of LASSO on administrative data sources in healthcare in the Netherlands

Background Policymakers generally lack sufficiently detailed health information to develop localized health policy plans. Chronic disease prevalence mapping is difficult as accurate direct sources are often lacking. Improvement is possible by adding extra information such as medication use and demographic information to identify disease. The aim of the current study was to obtain small geographic area prevalence estimates for four common chronic diseases by modelling based on medication use and socio-economic variables and next to investigate regional patterns of disease. Methods Administrative hospital records and general practitioner registry data were linked to medication use and socio-economic characteristics. The training set (n = 707,021) contained GP diagnosis and/or hospital admission diagnosis as the standard for disease prevalence. For the entire Dutch population (n = 16,777,888), all information except GP diagnosis and hospital admission was available. LASSO regression models for binary outcomes were used to select variables strongly associated with disease. Dutch municipality (non-)standardized prevalence estimates for stroke, CHD, COPD and diabetes were then based on averages of predicted probabilities for each individual inhabitant. Results Adding medication use data as a predictor substantially improved model performance. Estimates at the municipality level performed best for diabetes with a weighted percentage error (WPE) of 6.8%, and worst for COPD (WPE 14.5%)Disease prevalence showed clear regional patterns, also after standardization for age. Conclusion Adding medication use as an indicator of disease prevalence next to socio-economic variables substantially improved estimates at the municipality level. The resulting individual disease probabilities could be aggregated into any desired regional level and provide a useful tool to identify regional patterns and inform local policy.

comparison of Australian chronic disease prevalence estimates using administrative pharmaceutical dispensing data with international and community survey data

Introduction: Chronic disease (CD) is a leading cause of population mortality, illness and disability. Identification of CD using administrative data is increasingly used and may have utility in monitoring population health. Pharmaceutical administrative data using World Health Organization Anatomic Therapeutic Chemical Codification (ATC) assigned to prescribed medicines may offer an improved method to define persons with certain CD and enable the calculation of population prevalence. Objective: To assess the feasibility of Australian Pharmaceutical Benefits Scheme (PBS) dispensing data to provide realistic measures of chronic disease prevalence using ATC codification and compare values with international data using similar ATC methodology and Australian community surveys. Methods: Twenty-two chronic diseases were identified using World Health Organization (WHO) formulated ATC codes assigned to treatments received and recorded in a PBS database. Distinct treatment episodes prescribed to individuals were counted annually for prevalence estimates. Comparisons were then made with estimates from international studies using pharmaceutical data and published Australian community surveys. Results: PBS prevalence estimates for a range of chronic diseases listed in European studies and Australian community surveys demonstrated good correlation (r > .83, p < .001). PBS estimates of the prevalence of diabetes, cardiovascular disease and hypertension, dyslipidemia, and respiratory disease with comparable Australian National Health Survey data by age groupings (>45 years) showed correlations of between (r = 0.82 - 0.99, p < .001) and a range of percentage difference of -15% to 77%. However, other conditions such as psychological disease and migraine showed greater disparity and correlated less well. Conclusions: Although not without limitations, Australian administrative pharmaceutical dispensing data may provide an alternative perspective on population health and a useful resource to estimate the prevalence of a number of chronic diseases within the Australian population.

Correction to: Comparing health insurance data and health interview survey data for ascertaining chronic disease prevalence in Belgium

An amendment to this paper has been published and can be accessed via the original article.

Mapping Chronic Disease Prevalence based on Medication Use and Socio-demographic variables: an Application of LASSO in healthcare in the Netherlands

Objectives Policymakers generally lack sufficiently detailed health information to develop localized health policy plans. Chronic disease prevalence mapping is difficult as accurate direct sources are often lacking. Improvement is possible by adding extra information such as medication use and demographic information to identify disease. The aim of the current study was to use a LASSO (Least Absolute Shrinkage and Selection) model on a wide set of variables including medication use to obtain small geographic area prevalence estimates for four common chronic diseases and investigate regional patterns of disease. Methods Administrative hospital records and general practitioner registry data were linked to medication use and socio-economic characteristics. The training set (n=707021) contained GP diagnosis and/or hospital admission diagnosis as the standard for disease prevalence. For the entire Dutch population (n = 16,777,888), all information except GP and hospital admission was available. A LASSO operator regression model for binary outcomes was used to select variables strongly associated with disease. Dutch municipality (non-)standardized prevalence estimates for stroke, CHD, COPD and diabetes were then based on the average of individual predicted probabilities. Results Adding medication use data as a predictor substantially improves model performance. Estimates at the municipality level are best for diabetes with a weighted percentage error (WPE) of 6.8%, and worst WPE for COPD, with 14.5%. Disease prevalence has clear regional patterns, also after standardization for age. Conclusion Adding medication use as an indicator of disease prevalence next to socio-economic variables substantially improved estimates at the municipality level. The resulting individual disease probabilities can be aggregated into any desired regional level and provide a useful tool to identify regional patterns and subsequently inform local policy.