Background: Although intensively studied in hospital and emergency settings, chest pain has remained largely unstudied in primary care, where it is associated with considerable diagnostic uncertainty and high utilization of medical resources.
Methods: We employed an established primary care research network to prospectively collect detailed information on episodes of care for chest pain. Over a 12-month period, Michigan Research Network (MIRNET) clinicians prospectively collected demographic, clinical, and clinician decision-making information for all patients seen in their offices with the complaint of chest pain.
Results: Three hundred ninety-nine complete episodes were collected and used for analysis. Episodes were well distributed among urban, rural, academic, and private sites. The average episode length was 1.53 visits. Musculoskeletal chest pain accounted for 20.4% of all diagnoses, followed by reflux esophagitis (13.4%) and costochondritis (13.1%). Stable angina pectoris was the primary diagnosis in only 10.3% of episodes, unstable angina or possible myocardial infarction in 1.5%. Most of the ancillary services used were directed toward finding or excluding cardiac disease.
Conclusion: A practice-based network can be used to study episodes of care. Resource use during episodes of chest pain in primary care are directed toward exclusion of cardiac disease, despite the surprisingly low frequency of cardiac diagnoses.