Characterizing Clinical Inertia in a Large, National Database

Objective: To characterize clinical inertia in the treatment of diabetes using a large, geographically diverse clinical database.

Study Design: A retrospective descriptive analysis was conducted in a clinical database containing 22 million patient records across 22 health care organizations (HCOs).

Population Studied: A total of 281,000 patients aged 18-75 were included during the 5.5-year study period (1/2012-6/2017). Patients had an outpatient visit in the last 12 months of the study period, an HbA1c in the last 24-30 months (index A1c), and a diagnosis of type 2 DM on a claim or EHR problem list at least 6 months prior to index A1c. A subset of 47,693 patients with an index A1c ≥8 and a prior A1c ≥8 or lack thereof, was observed for four 6-month follow-up periods for actions including a new class of diabetes medication prescribed or an A1c <8. The absence of observable action following index A1c suggests potential “clinical inertia.”

Principal Findings: Six months following an index A1c≥8, 55% of patients received no observable clinical action ranging from 45-65% across HCOs and 18-96% across individual providers. A new diabetes prescription was observed in 35% of patients (7.5% moved into glycemic control, A1c<8) and 10% moved into glycemic control without a new prescription. Patient characteristics associated with increased clinical inertia, i.e., no observable action, during the 6002D and 24-month follow-up periods included black race, low-income insurance, normal body mass index, and being on bolus insulin (all P<.01). Within 24 months, clinical inertia was reduced to 19%, ranging from 13-28% across HCOs.

Conclusions: Lack of clinical action in the 6 months following an A1c ≥8 suggests clinical inertia vs. ADA guidelines. Greater rates in low-income insurance and race/ethnic minority adults suggests potential populations to target to ensure adequate treatment for diabetes. The decline in clinical inertia within 24 months suggests either actions not seen in the data, or later interventions that were ultimately effective.