Abstract 3020: Cluster Analysis of Women’s Prodromal and Acute Myocardial Infarction Symptoms
Purpose: To perform a secondary cluster analysis of a large data base to identify symptom clusters most likely to occur prior to and during myocardial infarction (MI) among subgroups of women.
Methods: We used an existing data base compiled on 1270 women (43% black, 42% white and 15% Hispanic) recruited from 15 sites across the U.S. All women were > 21 years of age and 4 – 6 months post-MI hospitalization. Initial data were collected using the McSweeney Acute and Prodromal Myocardial Infarction Survey that contains 33 prodromal and 37 acute symptom questions and descriptors, plus risk factors/comorbidities. We used cluster analysis to group women into similar, naturally occurring configurations of prodromal symptom and acute symptom scores. We evaluated several clustering algorithms, adopting the most meaningful separation of clusters. Using bivariate analyses, we linked patient characteristics and comorbidities to symptom score configurations by examining the association between clusters and each characteristic and risk factor/comorbidity separately and then by examining the association between clusters and characteristics/risk factors combined using multinomial logistic regression.
Results: Three clusters were identified; each containing women with decreasing frequency and severity of prodromal symptoms. African-American women < 50 years were more likely to complain of frequent and intense symptoms compared to other women. Older, Caucasian women without history of diabetes or smoking were less likely to experience prodromal symptoms. Chest pain/discomfort was not identified as a primary prodromal symptom. Two clusters were identified based on acute symptoms. Cluster 1 contained women with frequent and severe symptoms including shortness of breath (91.0%) and chest pain/discomfort (76.1%). Cluster 1 women tended to be Hispanic, obese, < 50 years, and smokers. Women in cluster 2 experienced less acute symptoms; 55.2% reported chest pain/discomfort and 49.8% reported shortness of breath.
Conclusion: Developing clusters of women’s CHD and MI symptoms should provide useful information to inform clinicians, facilitate timely diagnoses and early treatment, and potentially improve CHD outcomes.