Cancer diagnosis and outcomes in Michigan emergency departments versus other setting

The Behavioral Model of Health Services Utilization (1968) is used to investigate the association of individual and community level, predisposing, enabling, and need factors on cancer diagnosis location and its correlation with stage of cancer at diagnosis and receipt of guideline cancer treatment. This retrospective cohort study is based on a sample of patients 66 years and older from the Michigan Tumor Registry who had either a colorectal or lung cancer diagnosis.;Variables from the Michigan Tumor Registry, Medicaid Enrollment File, Medicare inapatient, outpatient, and physician claim files, Area Resource File, and 1990/2000 Census Summary File were used to operationalize the constructs in the Andersen model for a total of nine hypotheses relating to predisposing, enabling, and need factors and outcomes (i.e., stage of cancer at diagnosis and receipt of guideline cancer treatment).;Colorectal and lung cancer patients with an ED diagnosis were more likely to be diagnosed at a later stage as opposed to patients diagnosed in other settings. Patients diagnosed with NSCLC in the ED were significantly less likely to have a resection or have radiotherapy initiated compared to patients diagnosed in other settings. A statistically significant relationship did not exist between an ED colon cancer diagnosis and receipt of guideline colon cancer treatment.;In terms of predisposing characteristics, African American patients were more likely to be diagnosed with lung cancer in the ED relative to White patients. Patients who appeared to believe in the value of health and health maintenance with regular cancer screening prior to diagnosis were less likely to be diagnosed with colorectal and lung cancer in the ED compared to patients who did not receive prior screenings.;In terms of personal enabling resources, census tract median household income was an insignificant variable among colorectal and lung cancer patients. Dual eligibility was associated with an increased likelihood of a colorectal cancer diagnosis in the ED compared to patients who were on Medicare alone.;Community enabling resources were operationalized by the average number of general practitioners and specialists per 100,000 residents. This was not a significant predictor of diagnosis location among colorectal cancer patients. However, residing in an area with a higher average concentration of radiation oncologists was associated with a decreased likelihood of learning of a lung cancer diagnosis in the ED.;Perceived need and its association with diagnosis location varied based on the measure. Having a PCP or outpatient visit 2-12 months prior to colorectal and lung cancer diagnosis was associated with a lower likelihood of an ED cancer diagnosis. A higher evaluated need (i.e., increased comorbidity burden and at least one inpatient admission prior to diagnosis) was associated with an increased likelihood of being diagnosed with colorectal and lung cancer in the ED. Colorectal cancer patients with three or more comorbidities were four times more likely to be diagnosed with cancer in the ED while lung cancer patients with the same comorbidity burden were over twelve times more likely to have an ED cancer diagnosis relative to patients with no comorbidities at diagnosis. An inpatient admission 2-12 months prior to diagnosis increased the likelihood of being diagnosed with colorectal and lung cancer in the ED.;Among the demographic characteristics, older age was associated with an increased likelihood of an ED diagnosis and later stage cancer at diagnosis and a decreased likelihood of chemotherapy and radiotherapy initiation relative to their younger counterparts in both colorectal and lung cancer patients. Females were more likely to be diagnosed with colorectal and lung cancer in the ED and at later stages of colorectal cancer at diagnosis.