Detection of clinical relapses in multiple sclerosis cohorts: construction and validation of a model based on administrative data.

Multiple sclerosis (MS) is the main cause of chronic disability in young people during their most productive years of life and therefore carries a high social and economic burden. The present study aimed to: (1) verify the capacity of an administrative data source to furnish data for constructing a model able to detect the occurrence of clinical relapses in MS patients and (2) validate the constructed theoretical model on a set of real-world data. Two MS experts identified some administrative variables as proxies of clinical relapses. Thereafter, the two MS experts analysed 889 events in 100 MS patients, considering only the administrative data relating to these patients, while a third neurologist independently analysed the real-world data (documented medical history) of the same patients in the same period. Absolute concordance between the theoretical model and the real-world data was found in 86 % of the events. The model we propose is easily and rapidly applicable, requiring the collection of just a few variables that are already present in local health authority administrative databases in Italy. It can be used to estimate, with a good level of reliability, the occurrence of relapses in various settings. Moreover, the model is also exportable to different and larger MS cohorts and could be useful for healthcare planning and for evaluating the efficacy of drugs in the real-world, thus favouring better resource allocation and management.

Reliability and validity assessment of administrative databases in measuring the quality of rectal cancer management.

PURPOSE: Measurement and monitoring of the quality of care using a core set of quality measures are increasing in health service research. Although administrative databases include limited clinical data, they offer an attractive source for quality measurement. The purpose of this study, therefore, was to evaluate the completeness of different administrative data sources compared to a clinical survey in evaluating rectal cancer cases. METHODS: Between May 2012 and November 2014, a clinical survey was done on 498 Lombardy patients who had rectal cancer and underwent surgical resection. These collected data were compared with the information extracted from administrative sources including Hospital Discharge Dataset, drug database, daycare activity data, fee-exemption database, and regional screening program database. The agreement evaluation was performed using a set of 12 quality indicators. RESULTS: Patient complexity was a difficult indicator to measure for lack of clinical data. Preoperative staging was another suboptimal indicator due to the frequent missing administrative registration of tests performed. The agreement between the 2 data sources regarding chemoradiotherapy treatments was high. Screening detection, minimally invasive techniques, length of stay, and unpreventable readmissions were detected as reliable quality indicators. Postoperative morbidity could be a useful indicator but its agreement was lower, as expected. CONCLUSIONS: Healthcare administrative databases are large and real-time collected repositories of data useful in measuring quality in a healthcare system. Our investigation reveals that the reliability of indicators varies between them. Ideally, a combination of data from both sources could be used in order to improve usefulness of less reliable indicators.