There is evidence to suggest that the Charlson Index, an index used to predict one years rate of death from the study end point depending on how many medical conditions they present, is an inaccurate indicator to assess death today. We propose to develop a new valid index for categorising medical conditions that occur at the same time as another to predict one years rate of death for patients at the end of the study index period by using routinely collected health information from the Clinical Practice Research Datalink (CPRD). We will identify a sample of patients and investigate the burden of disease and how this will contribute to predicting their chance of death. In order to make sure that the Charlson Index is still valid and representable today, we will run a test to determine if the existing Charlson score, a score assigned to each burden of disease depending on how severe it is, is still the same as 30 years ago or whether the weighting of classifying these diseases has changed.
This study is a retrospective cohort study using patient data from the CPRD to gain insights on devising an alternative method to classify comorbid conditions and predict one years mortality, for patients with one or more comorbid conditions, for Read coded databases. The main objective is to develop a prognostic classification for comorbid conditions which individually or in combination might alter the risk of short term mortality for patients enrolled in clinical research. All individual diseases identified in the Charlson index will be applied to our patient cohort and then this list will be condensed down into the main comorbid conditions displayed in our population. A model will be run to determine whether the existing Charlson weighting for comorbid diseases is still valid or whether the severity of these diseases has changed over time. The potential effect that this may have on predicting mortality rate is yet to be studied. Descriptive statistics, univariate and multivariate analyses, Coxs proportional hazard model, will be used to perform data analysis associated with predicting mortality, The results from this study will be used to support scientific understanding of how this updated assessment of disease severity and predicting mortality may influence others.
Ruth Farmer - Chief Investigator - Boehringer-Ingelheim Pharmaceuticals, Inc
Urvee Karsanji - Corresponding Applicant - University of Leicester
Alicia Gayle - Collaborator - Imperial College London
Michael Marcus - Collaborator - University of Liverpool
Naj Rotheram - Collaborator - Boehringer-Ingelheim - UK
Urvee Karsanji - Collaborator - University of Leicester
Chris D Poole - Chief Investigator - Digital Health Labs Limited
ONS Death Registration Data