Abstract
Background
The National Health Service in England “Long Term Plan” aims to prevent 150,000 strokes and myocardial infarctions over the next 10 years. To achieve this, resources are being allocated to improve early detection of conditions strongly associated with cardiovascular disease. This includes working towards people routinely knowing their “ABC” risk factors (“A”: atrial fibrillation (AF), “B': hypertension and “C”: high cholesterol) (1).
Purpose
The aims of this study were to: 1) determine the proportion of participants with “A”, “B”, and “C” criteria; and 2) to identify risk factors for patients fulfilling any of these criteria.
Methods
LIPIDOGRAM2015 was a nationwide cross-sectional survey for adults in Poland. Adults were recruited in 2015 and 2016 by 438 family physicians. For the ABC criteria, “A” was defined as AF identified in the medical records of the participant, “B” was defined as either systolic blood pressure greater than 140mmHg or diastolic blood pressure greater than 90mmHg or both, and “C” was defined as total cholesterol greater than 200mg/dL (5.17mmol/L). The scaled and centred dataset underwent principal component analysis using singular value decomposition to achieve dimensionality reduction. K-means clustering was used to stratify patients with Hartigan's rule being used to identify optimal K number (2–4). The p-value for statistical significance used in this study was p<0.01 unless otherwise specified.
Results
13,724 patients were included in the study. 71.0% (n=9,747) of participants fulfilled the criteria for one or more of the “A”, “B” or “C” components (Fig. 1). 26 variables were used in this analysis with Principal Component Analysis showing 7 principal components explaining over 50% of the variance with 20 components explaining over 90%. K-means clustering was also performed, finding 39 separate clusters. Correlations and statistical significance tests showed a high degree of variability between variables. Participants with AF were older (mean (SD) 67.7 (9.5) vs 55.7 (13.7), p<0.0001), with higher prevalence of concomitant coronary heart disease (CHD) (OR 6.73, 95% CL 5.75, 7.87) and ischaemic stroke (OR 13.45, 95% CL 7.66, 23.6). Participants with hypertension were older (mean (SD) 60.1 (SD 12.4) vs 53.8 (14.0), p<0.0001), with a higher BMI (mean (SD) 29.9 (5.1) vs 27.5 (4.8), p<0.0001) and resting heart rate (mean (SD) 75.7 (10.7) vs 72.7 (8.9), p<0.0001), more likely to be male (OR 1.42, 95% CL 1.32, 1.53) and have diabetes (OR 1.61, 95% CL 1.46, 1.78). Participants with high cholesterol showed an inverse correlation with prevalence of both concomitant diabetes (OR 0.85, 95% CL 0.77, 0.94) and CHD (OR 0.85, 95% CL 0.76, 0.94) (Fig. 2).
Conclusion
Simple demographic and clinical variables could be used to guide targeted screening to increase population awareness of “ABC” status, allowing for a greater proportion of the population to be appropriately managed with cardiovascular prevention strategies.
FUNDunding Acknowledgement
Type of funding sources: None. “ABC” Venn diagram Correlogram and significance plot
Dealing with many correlated covariates in capture-recapture models
