ABSTRACT
Objective In this retrospective cohort study we aimed to assess, in 202,247 people who started a statin therapy between 2007-2014, the factors that led to the initiation of the drug. To do this we explored CVD risk factors singularly and in combinations as recorded in electronic health records in the year before they receive their first prescription and we compared the risk scores with that suggested by the NICE guideline at that time.
Methods We summarised demographic characteristics and proportions of people with a risk score below the threshold. Regression-based analyses are performed to evaluate the association between the missingness of the risk score and relevant risk score components.
Results 45,364 individuals (22.4%) were prescribed statins without a record of a risk score being available in the year prior to the prescription date. When the risk score was available, 68,174 out of 156,883 patients were prescribed statins even with a score below the 20% threshold. Smoking status was the most frequently recorded variable (74.9% of the instances), followed by systolic blood pressure (71.6%) and total cholesterol (70%), while HDL cholesterol was the least recorded (34.1%). Cholesterol levels are positively associated with the missingness of the risk score, while systolic blood pressure shows a negative association.
Conclusions GPs often start statins on people with no risk score recorded in their clinical records or in those with risk scores below the recommended threshold. Higher cholesterol values may result in a GP starting statin therapy without recording the other relevant components required to calculate a risk score.
STRENGTHS AND LIMITATIONS OF THIS STUDY
Our cohort contains a large number of individuals: the study provide a representative picture of initiation of statins in UK primary care.
We specifically focus on which variables and factors GPs record in electronic databases in the year prior statin treatment initiation: this is the first study to directly tackle the issue of statins prescribing in the absence of all the information required by the NICE guideline
We are not able to verify if GPs actually used the records of the individual health indicators, when these were available, to calculate the risk score.