Developing General Practice and Web-Based Toolkits for the Familial Hypercholesterolaemia Case Ascertainment Tool (FAMCAT)

Talk Code: 
4E.1
Presenter: 
Stephen Weng
Co-authors: 
Nadeem Qureshi, Joe Kai, Mike Taylor, Heather Wharrad
Author institutions: 
NIHR School for Primary Care Research (NQ, JK), Health E-Learning and Media Team (MT, HW), University of Nottingham

Problem

Familial Hypercholesterolaemia (FH) affects up to 1 in 500 individuals in the UK and, left untreated, can lead to premature heart disease. This can be prevented by statins but 80% of individuals still are not being identified in primary care. The FAMCAT study team has previously developed and validated a novel algorithm using a large primary care database (CPRD) based on routinely coded diagnostic indicators. We will demonstrate translating this algorithm into user-friendly toolkits, and present pilot data from testing the functionality of the toolkit in general practice.

Approach

The FAMCAT algorithm consists of several diagnostic indicators (cholesterol, age, triglycerides, lipid lowering drug prescribing, family history, diabetes, and chronic kidney disease). From analysing more than 2.9 million primary care records, FAMCAT was found to have a high discrimination, based on the area under the receiver operating curve (AUC 0.86, 95% CI 0.85–0.87) for predicting FH. For UK general practice, we have developed a novel data extraction toolkit (FAMCAT-UK) and tested this toolkit in three general practices. For international users, we have a developed a web-based version (FAMCAT-WEB) and a secure research database, with Google analytics which can capture user metadata.

Findings

FAMCAT-UK is a general practice data extraction tool which automatically extracts coded patient clinical data based on NHS READ/SNOMED codes. The tool then calculates a patient's probability of having FH based on the FAMCAT algorithm. Patients above a 1 in 500 population prevalence are flagged as having possible FH. These patients can then be ranked from highest to lowest probabilities. Treatment and management data are also presented on prescribing of lipid lowering therapies. Testing the tool in three general practices comprising of 27,349 total patients found that 11,499 adults over 16 years of age had a documented total or LDL-cholesterol recorded (42%). From these patients who have had their cholesterol assessed, 504 patients (1.8% of total list size; 4.4% of adults with cholesterol recorded) were above the 1 in 500 population prevalence of FH. FAMCAT-WEB is similar to FAMCAT-UK, except that users enter their own data on an online application which then presents the output of the probability along with guidance and interpretation. The corresponding online database captures all user data entry fields and output of the results, along with Google analytics metadata (i.e. usage statistics, location, time, date) based on user’s IP addresses.

Consequences

Previously, the FAMCAT algorithm has been developed as a research study and the algorithm was largely inaccessible to users. By developing both a user-friendly general practice and online-based toolkit, we can disseminate and translate our research to a broader audience. This approach offers the opportunity to enhance detection of FH in primary care by identifying individuals with greatest probability of having the condition. Such cases can be prioritised for further clinical assessment, appropriate referral and treatment to prevent premature heart disease.

Submitted by: 
Stephen Weng
Funding acknowledgement: 
The work is supported by the NIHR School for Primary Care Research