Predicting the risk of becoming eligible for the disability pension: Machine learning methods applied to French health data / Prédire le passage en invalidité : les méthodes d'apprentissage automatique appliquées aux données de santé françaises

Introduction: Benefiting from the disability pension implies morbid (physical and psychological) and social (fall in income) implications for the person. It also has economic consequences for society, with increasing expenses since 2011 (+4.9% on average per year). Investing in preventive actions against the loss of the ability to work should limit these consequences, but it requires targeting people at risk. The development of artificial intelligence opens up prospects in this regard. Purpose of the Research: To target, using supervised machine learning methods, those people with a high probability of becoming eligible for the disability pension over the course of the year based on their socio-demographic and medical characteristics (pathologies, work stoppages, drugs taken, and medical procedures). Method: Among the beneficiaries of the French public welfare system aged 20­64 in 2017, we compared the socio-demographic and medical characteristics between 2014 and 2016 of those who received a disability pension in 2017 and not before, and those who did not receive a disability pension from 2014 to 2017. The determination of the boundary between these two groups was tested using logistic regression, decision trees, random forests, naive Bayes classifiers, and support vector machines. The models' performance was compared with respect to accuracy, precision, sensitivity, specificity, and AUC (area under the curve). Finally, the predictive power of each factor was measured by AUC too. Results: The boosted logistic regression had the best performance for three of the five criteria, but low sensitivity. The best sensitivity was obtained with the support vector machines, with an accuracy close to that of the boosted logistic regression, but a lower precision and specificity. Random forests offered the best discriminatory ability. The naive Bayes classifier had the worst performance. The most predictive factors in becoming eligible for the disability pension were having 30 days or more off sick in 2014, 2015, and 2016 and being aged 55 to 64. Conclusion: Supervised learning methods have appeared relevant for identifying people with the highest probability of becoming eligible for the disability pension and, more broadly, for steering public and social policies.

Introduction: Le recours à la pension d'invalidité a des implications morbides (physiques ou psychiques) et sociales (baisse du revenu). Il a aussi des conséquences économiques pour la société, avec des dépenses croissantes depuis 2011 (+4,9 % en moyenne par année). Prévenir la perte de la capacité à travailler devrait permettre de limiter ces conséquences, mais nécessite de cibler les personnes à risque. Le développement des méthodes d'intelligence artificielle ouvre des perspectives en ce sens. But de l'étude: Cibler les personnes ayant une « forte ¼ probabilité de devenir bénéficiaires d'une pension d'invalidité dans l'année au regard de leurs caractéristiques sociodémographiques et médicales (pathologies, arrêts de travail, médicaments et actes médicaux) à partir de méthodes d'apprentissage automatique supervisé. Méthodes: Parmi les bénéficiaires du régime général âgés de 21 à 64 ans en 2017, comparaison des caractéristiques de 2014 à 2016 entre les nouveaux bénéficiaires d'une pension d'invalidité en 2017 et ceux n'en bénéficiant pas. La détermination de la frontière entre ces deux groupes a été testée à l'aide de la régression logistique, des arbres de décision, des forêts aléatoires, de la classification naïve bayésienne et des séparateurs à vaste marge. Les performances des modèles ont été comparées au regard de la justesse, la précision, la sensibilité, la spécificité et l'AUC (Area Under the Curve). Le pouvoir prédictif de chaque facteur est estimé à partir de l'AUC. Résultats: La régression logistique boostée avait les meilleures performances sur trois des cinq critères retenus, mais une faible sensibilité. La meilleure sensibilité était obtenue avec les séparateurs à vaste marge, avec une justesse proche de la régression logistique boostée mais une précision et une spécificité inférieures. Les forêts aléatoires offraient la meilleure capacité discriminatoire. Les facteurs les plus prédictifs du risque de passer en invalidité étaient le bénéfice d'au moins 30 jours d'indemnités journalières pour maladie en 2014, 2015 et 2016 et le fait d'être âgé de 55 à 64 ans. Conclusion: Les méthodes d'apprentissage supervisé sont apparues pertinentes pour le ciblage des personnes les plus à risque de recourir à la pension d'invalidité et, plus largement, pour le pilotage d'autres prestations sociales.

Budget Impact Analysis of Expanding Gliflozin Coverage in the CKD Population: A French Perspective.

INTRODUCTION: Gliflozins have historically been indicated for type 2 diabetes in France. However, their efficacy has recently been demonstrated in heart failure and chronic kidney disease (CKD), with positive recommendations by Haute Autorité de Santé for gliflozin therapies in these indications. The study objective was to investigate the 5-year budget impact associated with the introduction of gliflozins in addition to standard therapy in people with CKD and elevated albuminuria, regardless of diabetes status, from the perspective of the French healthcare system. METHODS: A budget impact model was developed to estimate the 5-year implications of incorporating gliflozins in the treatment of patients with CKD in France, using efficacy data from the Dapagliflozin and Prevention of Adverse Outcomes in Chronic Kidney Disease (DAPA-CKD) trial. Direct medical costs associated with drug acquisition and management, treatment-related adverse events, dialysis and kidney transplantation, and adverse clinical outcomes were considered. Market share forecasts were estimated from historical data and expert opinions. Event rates were derived from trial data, while cost data were sourced from published estimates. RESULTS: The introduction of gliflozins was estimated to be cost saving compared to the no gliflozins scenario, with an expected cumulative 5-year budget impact of -€650 million, driven by slowed disease progression in patients treated with gliflozins, with fewer patients cumulatively progressing to end-stage kidney disease (84,526 vs. 92,062). This, in addition to fewer hospitalisations for heart failure and deaths from any cause, led to substantial medical care cost offsets (kidney-related: - €894 million; hospitalisation for heart failure: - €14.3 million; end-of-life care: - €17.3 million) to the additional drug acquisition (€273 million) and treatment-related adverse events costs (€2.98 million). CONCLUSION: In concert with early diagnosis and proactive management of CKD, the expansion of the gliflozin indications into the French CKD population presents the opportunity to reduce the substantial burden associated with cardio-renal complications which outweighs the additional cost of the new treatment. INFOGRAPHIC.

Allocation aux adultes handicapés : pathologies et recours aux soins des bénéficiaires en 2017.

OBJECTIVE: We described the pathologies and health care utilization of beneficiaries of the general health insurance scheme via the Allocation Adulte Handicapé (AAH - Adult Disability Allowance) compared to the general population. METHOD: Mapping of pathologies and expenditures allowed the identification of 58 pathologies and chronic treatments in the SNDS, thanks to ICD-10 codes for long-term conditions or hospitalizations, specific drugs or medical procedures, among all beneficiaries of the general health insurance scheme aged 20 to 64 years with reimbursed care (>1€) in 2017. The prevalence and annual rates of care utilization among all beneficiaries of the general scheme via AAH (“AAH” group) and in the rest of the population (“non-AAH”) were standardized and described. RESULTS: Among the 793,934 (2.51% of the population) “AAH” persons, all the pathologies studied were more frequent than among the “non-AAH”, with 44% having psychiatric pathologies (compared with 3.2%), and 14% a neurological pathology (compared with 1%). AAH beneficiaries were more likely to use hospital care (63% versus 40%), but less likely to use specialist care (63% versus 68%) and dental care (37% versus 45%). CONCLUSION: The beneficiaries of the general scheme via the AAH had mainly psychiatric and neurological pathologies, but other pathologies were also much more frequent than in the general population. The lower use of dental and specialist care was probably related to a lack of access to care, potentially caused by the absence of 100% coverage of care.

Evolution of health care utilization and expenditure during the year before death in 2015 among people with cancer: French snds-based cohort study.

BACKGROUND: Cancer patients have one of the highest health care expenditures (HCE) at the end of life. However, the growth of HCE at the end of life remains poorly documented in the literature. OBJECTIVE: To describe monthly reimbursed expenditure during the last year of life among cancer patients, by performing detailed analysis according to type of expenditure and the person's age. METHOD: Data were derived from the Système national des données en santé (SNDS) [national health data system], which comprises information on ambulatory and hospital care. Analyses focused on general scheme beneficiaries (77% of the French population) treated for cancer who died in 2015. RESULTS: Average reimbursed expenditure during the last year of life was €34,300 per person in 2015, including €21,100 (62%) for hospital expenditure. "Short-stays hospital" and "rehabilitation units" stays expenditure were €14,700 and €2000, respectively. Monthly expenditure increased regularly towards the end of life, increasing from 12 months before death €2000 to €5200 1 month before death. The highest levels of expenditure did not concern the oldest people, as average reimbursed expenditure was €50,300 for people 18-59 years versus €25,600 for people 80-90 years. Out-of-pocket payments varied only slightly according to age, but increased towards the end of life. CONCLUSION: A marked growth of HCE was observed during the last 4 months of life, mainly driven by hospital expenditure, with a more marked growth for younger people.