353. Quantifying severity of chronic conditions in English Primary Care using the Clinical Practice Research Datalink
- Principal Investigator: Evan Kontopantelis
- 1 April 2017 to 31 March 2018
- Project No: 353
- Funding round: FR 13
Severity in English Primary Care
Aim: to longitudinally characterise two exemplar conditions, type-2 diabetes (T2DM) and coronary heart disease (CHD), focusing on developing clinical decision algorithms that will grade the severity and respective health care needs of a patient with T2DM or CHD.
Importance: Analyses of routinely collected data tend to characterise a condition in a simplistic binary fashion. However, there are numerous grades of the clinical severity of these conditions which could be informative to clinicians, researchers and policy makers.
Design: Cohort study of Electronic Health Records from the Clinical Practice Research Datalink. We will extract all data from over 650 general practices, covering 11 million people. Data will be organised and analysed in annual data bins, 2006- 2015.
Over time and by region, age-group, sex and socio-economic deprivation, we will:
- Generate algorithms to estimate severity grades for T2DM and CHD, using within-condition diagnoses (e.g. complications), co-morbidities (focusing on 18 well-recorded chronic conditions), treatments, referrals to secondary or community care and emergency hospitalisations.
- Validate the algorithms using survival analysis models and refine them using the regression weights and estimate years lost by severity stratum.
- Describe the disease progression journey, time needed to progress from one stage to the next, what is damaged and what is at risk in the short/long term.
- Estimate the prevalence of all 18 chronic conditions, the referrals associated with each, and their multimorbidity patterns.
Outputs will include academic papers, presentations and briefing papers directed at policy makers, the public and our PPIE partners. This work will demonstrate the value of the NIHR School for Primary Care Research and will lead to a larger programme of work in relation to longitudinal risk-prediction modelling and appropriate management and intervention (e.g. when is a review/intervention needed, what are appropriate treatments by grade and when a condition is well controlled or not).
Amount awarded: £48,259