[Valorization of health insurance medico-administrative databases to identify and follow the progression of a disease, study its management and estimate the impact of a health policy implementation using a health-economic model: application to type 2 diabetes in Luxembourg]. (Doctoral thesis)

Type 2 diabetes is a chronic disease associated with several life-threatening complications. Its prevalence increases worldwide. According to the international literature this disease is associated with a high impact in terms of costs and quality of life. In a context of budgetary constraint, there is a need to estimate the resources to allocate to the management of chronic diseases. This implies following the evolution of the epidemiology and the costs of these diseases, given their impact on public health and the healthcare budget. Observational cohort studies are costly and often hard to maintain during long periods. However, medicoadministrative (or claims) data, used by medical insurance for the reimbursement of insured patients, appears to be an interesting alternative source of data for public health authorities, although this data is often under-exploited.
The work hypothesis used in this thesis was that medico-administrative databases are a relatively exhaustive, reliable and cheap source of information, allowing useful studies for public decisionmakers. The research question concerned the situations in which such databases could be used and the potential results obtained by working on epidemiological studies or medical practices assessment, economic evaluations and building health-economic models.
Consequently, the objective of the research carried out in this thesis was to explore a medicoadministrative database, not initially dedicated to research, to study the fields of exploitation of this database and their possible applications on public health decisions concerning type 2 diabetes in Luxembourg.
This thesis was carried out under the aegis of the EA4069 “Epidemiology, Evaluation et Health Policies” of the university Paris-Descartes (France), the Centre for Health Studies of CRP-Santé (Luxembourg) and the doctoral network of the EHESP School of Public Health (France). It was funded by the National Fund for Research (Luxembourg) and was supported by the Inspection Générale de la Sécurité Sociale (Luxembourg) for the data collection used in this project, named DIABECOLUX.
This project was designed, in accordance with the Helsinki declaration. Collected data concerned the exhaustive population, living in Luxembourg and covered by the national health insurance (CNS), who had a reimbursement of a hypoglycemic treatment between 2000 and 2006.
A 22-digits identification number was attributed to each patient in order to insure secure anonymity. Since the identity of individuals cannot be retrieved by databases crossing, no ethical approval or data protection request was needed.
The first part of the thesis (Context and state of art) presents the geo-economic context and the healthcare system of the Grand-Duchy of Luxembourg and the health status of its population. Next, type 2 diabetes and its main complications are described, followed by a description of the state of play regarding the knowledge on diabetes in Luxembourg and existing prevention policies. The use of claims data is then compared to the use of clinical data. Finally objectives and research questions, which guided this work, are stated.
The second part of the thesis (database building process) details the steps of the process to build the database and describes the tables included.
The third part of the thesis (Identification of treated diabetic patients and associated complications) describes the DIABECOLUX algorithm, which allowed for the identification of treated type 2 diabetic patients in Luxembourg between 2000 and 2006 (in 2006, the prevalence was 3.79%; N= 17,070). Besides the DIABECOLUX algorithm . two complications of diabetes were investigated: diabetic nephropathy and retinopathy. . An algorithm determining the state of diabetic nephropathy for each individual and each year is presented (in 2006, diabetic nephropathy reached 3.77% of type 2 diabetes cases) in addition to the main obstacles to retinopathy identification.
The fourth part of the thesis (Patients follow-up and prevention of complications) describes a study on the adherence of physicians and patients to seven essential criteria extracted from European guidelines for the follow-up of type 2 diabetic patients. This study revealed a critical situation with respect to adherence and the attributes associated to inadequate adherence were studied (sex, treating physician, living region…). Moreover, it was found that patients suffering from nephropathy, whose risk to develop retinal and cardio-vascular complications is high, do not benefit from an adequate prevention care.
The fifth part of the thesis (economic evaluation) starts by describing the direct medical costs associated to type 2 diabetic condition and its different components, taking a third party payer perspective (national health insurance). Next, the cost associated to each state of diabetic nephropathy is presented together with the important economic impact of dialysis treatment (116647€/patient in 2006). Lastly, the use of medico-administrative data is described in an example of health-economic evaluation. A Markov model was built to assess an alternative health policy with the aim to improve the quality of life of patients in need of a dialysis. In accordance with our expectations, peritoneal dialysis was shown less costly and with a higher impact on patients’ quality of life as compared with haemodialysis.
The sixth and last part of the thesis discusses the general objectives of the analyses that form the bulk of this thesis, the main advantages and perspectives brought by a wide use of medicoadministrative databases and obstacles faced.
Despite difficulties in assessing the quality of medico-administrative data, it is a precious source of information for decision-makers and healthcare professionals, with the aim to improve patients’ care management. Exhaustivity permits to study representative populations and to generate good quality epidemiologic and economic evaluations. Moreover, the diversity of included data allows studying medical practices as well as building decision tools. Finally, the improvements considered at the national level will allow a faster access to health insurance data. However, as is the case with all sources of data, medico-administrative data cannot answer all questions. It can be seen as complementary to other sources of data and therefore does not impede the use and development of new research protocols including clinical data analyses. Results from the DIABECOLUX project contributed to the European basis EUBIROD and are the main source of the « national report on diabetes », solicited by the Ministry of Health of Luxembourg.
Presently, this thesis was the subject of 3 articles submitted or accepted in international scientific reviews, 6 oral communications and 5 posters in international congresses.
Presented findings were awarded 2 prizes (CALASS 2009 and Doctoral Network of EHESP 2011). Finally, 3 additional articles are in preparation.
Opinions expressed in this thesis are those of the author and not necessarily those of the Direction de la Santé of the Grand-Duchy of Luxembourg (Ministry of Health) or of the Centre for Health studies (CRP-Santé).