BackgroundRising health care costs are a major public health issue. Thus, accurately predicting future costs and understanding which factors contribute to increases in health care expenditures are important. The objective of this project was to predict patients healthcare costs development in the subsequent year and to identify factors contributing to this prediction, with a particular focus on the role of pharmacotherapy.MethodsWe used 2014–2015 Swiss health insurance claims data on 373′264 adult patients to classify individuals’ changes in health care costs. We performed extensive feature generation and developed predictive models using logistic regression, boosted decision trees and neural networks. Based on the decision tree model, we performed a detailed feature importance analysis and subgroup analysis, with an emphasis on drug classes.ResultsThe boosted decision tree model achieved an overall accuracy of 67.6% and an area under the curve-score of 0.74; the neural network and logistic regression models performed 0.4 and 1.9% worse, respectively. Feature engineering played a key role in capturing temporal patterns in the data. The number of features was reduced from 747 to 36 with only a 0.5% loss in the accuracy. In addition to hospitalisation and outpatient physician visits, 6 drug classes and the mode of drug administration were among the most important features. Patient subgroups with a high probability of increase (up to 88%) and decrease (up to 92%) were identified.ConclusionsPharmacotherapy provides important information for predicting cost increases in the total population. Moreover, its relative importance increases in combination with other features, including health care utilisation.

Background: Potential drug-drug interactions (pDDIs) are described in various case reports, but few studies have evaluated the impact of specific combinations on a population level. Objective: To analyze the type and frequency of multiple contraindicated (X-pDDIs) and major interactions (D-pDDIs) and to subsequently assess the impact of the particular combination of tizanidine and ciprofloxacin on outpatient physician visits and hospitalizations. Methods: Anonymized Swiss claims data from 524 797 patients in 2014-2015 were analyzed. First, frequencies of X- and D-pDDIs were calculated. Next, a retrospective cohort study was conducted among patients prescribed tizanidine and ciprofloxacin (exposed, n = 199) or tizanidine and other antibiotics (unexposed, n = 960). Hospitalizations and outpatient physician visits within 7, 14, and 30 days after initiation of antibiotic therapy were evaluated using multiple binary logistic regression and multiple linear regression. Results: The relative frequencies of X- and D-pDDIs were 0.4% and 6.65%, respectively. In the cohort study, significant associations between exposure to tizanidine and ciprofloxacin and outpatient physician visits were identified for 14 and 30 days (odds ratio [OR] = 1.61 [95% CI = 1.17-2.24], P = 0.004, and OR = 1.59 [95% CI = 1.1-2.34], P = 0.016). A trend for increased risk of hospitalization was found for all evaluated time periods (OR = 1.68 [95% CI = 0.84-3.17], OR = 1.52 [95% CI = 0.63-3.33], and OR = 2.19 [95% CI = 0.88-5.02]). Conclusion and Relevance: The interaction between tizanidine and ciprofloxacin is not only relevant for individual patients, but also at the population level. Further investigation of the impact of other clinically relevant DDIs is necessary to improve patient safety and reduce avoidable health care utilization.

Purpose Heart failure is among the leading causes for hospitalization in Europe. In this study, we evaluate potential precipitating factors for hospitalization for heart failure and shock. Methods Using Swiss claims data (2014-2015), we evaluated the association between hospitalization for heart failure and shock, and prescription of oral potassium supplements, non-steroidal anti-inflammatory drugs (NSAIDs), and amoxicillin/clavulanic acid. We conducted case-crossover analyses, where exposure was compared for the hazard period and the primary control period (e.g., 1-30 days before hospitalization vs. 31-60 days, respectively). Conditional logistic regression was applied and subsequently adjusted for addressing potential confounding by disease progression. Sensitivity analyses were conducted and stratification for co-medication was performed. Results We identified 2185 patients hospitalized with heart failure or shock. Prescription of potassium supplements, NSAIDs, and amoxicillin/clavulanic acid was significantly associated with an increased risk for hospitalization for heart failure and shock with crude odds ratios (OR) of 2.04 for potassium (95% CI 1.24-3.36, p = 0.005, 30 days), OR 1.8 for NSAIDs (95% CI 1.39-2.33, p < 0.0001, 30 days), and OR 3.25 for amoxicillin/clavulanic acid (95% CI 2.06-5.14, p < 0.0001, 15 days), respectively. Adjustment attenuated odds ratios, while the significant positive association remained (potassium OR 1.70 (95% CI 1.01-2.86, p = 0.046), NSAIDs OR 1.50 (95% CI 1.14-1.97, p = 0.003), and amoxicillin/clavulanic acid OR 2.26 (95% CI 1.41-3.62, p = 0.001). Conclusion Prescription of potassium supplements, NSAIDs, and amoxicillin/clavulanic acid is associated with increased risk for hospitalization. Underlying conditions such as pain, electrolyte imbalances, and infections are likely contributing risk factors. Physicians may use this knowledge to better identify patients at risk and adapt patient management.