HOW TO GAIN HIGH DATA QUALITY IN A NON-INTERVENTIONAL TRIAL TAKING THE LIVE-COM STUDY AS EXAMPLE
Author(s)
Schön H1, Theobald K2, Schädlich PK31CSG mbH, Berlin, Germany, 2Sanofi-Aventis Germany, Frankfurt, Germany, 3IGES Institut GmbH, Berlin, Germany
Presentation Documents
OBJECTIVES: The increasing scientificness and complexity of non-interventional studies is facing new operational and quality management aspects to gain high level quality data and results. Therefore it is crucial to implement quality measurements as part of the overall study design. METHODS: A non-interventional, comparative multicenter, representative cross sectional cost evaluation study was conducted (LIVE-COM1). The study retrospectively documented and compared diabetes-specific resource utilisation and patient-reported outcomes (PRO) associated with insulin administration in patients with type-2 diabetes mellitus over six months. The study was performed in accordance with the recommendations to improve quality and transparency of non-interventional studies of the German Association of Research-Based Pharmaceutical Companies and according to Good Epidemiological Practice. A detailed study plan, containing scientific objectives and statistical methods was developed before start of study and approved by the responsible Ethic Committee. In addition a feasibility assessment of the case record forms and PRO-instruments was performed in advance. A total of 4000 randomly selected primary care units were invited to participate in LIVE-COM. Throughout the study a close telephone monitoring with the sites was implemented, also to ensure the collection of the PRO data from the patient. A detailed in-house review of the case record forms, double data entry, query resolutions and source data verification at 10% of the sites was implemented. RESULTS: In total 1731 patients with signed informed consent were included in the study at 138 study centres in Germany. Overall only 1.3% of the documented patient data showed missing values and the validity of the data was high. Only in rare cases imputation algorithms were applied or data were excluded from analyses. CONCLUSION: The implementation of data quality management procedures and close contact with the study sites can deliver high quality data comparable to clinical studies. 1Long Acting Insulin Glargine Versus Insulin Detemir Cost Evaluation COMparison
Conference/Value in Health Info
2009-10, ISPOR Europe 2009, Paris, France
Value in Health, Vol. 12, No. 7 (October 2009)
Code
PMC30
Topic
Real World Data & Information Systems
Topic Subcategory
Health & Insurance Records Systems
Disease
Diabetes/Endocrine/Metabolic Disorders, Multiple Diseases