Electronic health records analytics to identify cancer patients with metabolic syndrome.

Abstract

e18649 Background: Metabolic syndrome, defined as the presence of at least 3 of 5 clinical factors including hypertension, elevated triglyceride levels, low high-density lipoprotein level, insulin resistance, and central obesity, increases the risk of heart disease, fatty liver, and multiple cancers. Metabolic syndrome in cancer patients has been associated with poor cancer-specific and overall survival. Lifestyle modification in patients with metabolic syndrome may reduce the risk of poor outcomes. In this quality improvement project, we aimed to determine the prevalence of metabolic syndrome among cancer patients and survivors seen in an outpatient general internal medicine (GIM) clinic and to determine the feasibility of using electronic health records (EHR) analytics to systematically identify such patients and refer them to lifestyle interventions and liver imaging. Methods: Study period was January-December 2021. During this period, an EHR algorithm was used to identify patients with metabolic syndrome based on the presence of ICD-10 diagnoses of metabolic conditions (diabetes, hypertension, lipid disease, and obesity). This algorithm was used to direct data from patient visits into an interactive dashboard to track metabolic syndrome prevalence and continuously monitor referrals to interventions. In September 2021, a best practice alert based on the EHR algorithm was created to identify patients with metabolic syndrome and prompt providers to refer them to nutrition counseling, liver ultrasound with elastography, and/or a community-based active-living support group for cancer survivors. GIM clinic nurses also reviewed medications and utilized an EPIC SmartPhrase that incorporated laboratory values (e.g., glucose, A1c, and lipids), blood pressure, and body mass index to confirm whether patients actually met the criteria for metabolic syndrome, and if so, they notified medical providers who then ordered the interventions. Patients confirmed to have metabolic syndrome received educational materials about lifestyle modifications. Data extracted from the dashboard were analyzed using Minitab 17 statistical software. Results: Among 1133 patients seen in the GIM clinic during 2021, 609 (54%) had metabolic syndrome. A total of 1045 patients (92%) had hypertension, 802 (71%) had hyperlipidemia, 571 (50%) had obesity, and 483 (43%) had diabetes. Among the 609 patients with metabolic syndrome, 148 (24%) were referred to liver ultrasound with elastography, 124 (20%) to nutrition counseling, and 21 (3%) to the support group. Beginning September 1, the best practice alert was triggered for 1131 clinical encounters meeting criteria for metabolic syndrome. Conclusions: The prevalence of metabolic syndrome among cancer patients seen in a GIM clinic was high. EHR analytics can lead to systematic identification and referral of patients with metabolic syndrome to lifestyle interventions and liver imaging.