1074 — Facility Implementation of Inpatient Hypoglycemia Safety Interventions and Episodes of Hypoglycemia.

Lead/Presenter: Paul Heidenreich,  COIN - Palo Alto
All Authors: Heidenreich PA (VA Palo Alto Health Care System), Lin S VA Palo Alto Health Care System Gholami P VA Palo Alto Health Care System Sahay A VA Palo Alto Health Care System

Objectives:
Inpatient hypoglycemia is a common and preventable cause of morbidity. Accordingly, many VA facilities have implemented interventions to limit inpatient hypoglycemia. We sought to determine if implementation of specific anti-hypoglycemia interventions were associated with lower episodes of inpatient hypoglycemia.

Methods:
We used data from a 2019 VA facility survey of hypoglycemia interventions conducted by the National Center for Patient Safety (NCPS). A priori, we selected two processes of care: avoidance of mixing short and long acting (NPH) insulin and avoidance of sliding scale only insulin regimens. Survey responses were categorized by NCPS into full implementation, partial implementation, or no implementation. We linked facility response data to hospitalization data for patients with diabetes in the VA Health System in 2019 and 2020 for the facilities responding to the questions. If patients were hospitalized more than once the first hospitalization was used. We determined lab values for hypoglycemia during admission.

Results:
There were 240,390 patients with diabetes hospitalized at the 85 facilities (3-digit station code) that responded to the question on combining insulins. There were 236,519 patients with diabetes hospitalized at the 83 facilities that responded to the question on use of sliding scale only insulin regimens. 44% were over the age of 70 and 96% were male (for both cohorts). Severe hypoglycemia ( < 50 mg/dl) occurred in 1.9% and any hypoglycemia ( < 70 mg/dl) occurred in 8.2% of patients during admission (for both cohorts). Severe hypoglycemia was less common for patients at facilities avoiding mixed insulin (full implementation) 1.9% vs. 2.7% for no implementation, p < 0.0001 but not different for those avoiding sliding scale only insulin regimens (1.9% if full implementation vs. 1.7% if no implementation, p = 0.13). Similarly, any hypoglycemia was less common for patients at facilities avoiding mixed insulin (full implementation) 8.0% vs. 10.1% (no implementation, p < 0.0001), but not different for those avoiding sliding scale insulin (8.2% if full implementation vs. 7.5% if no implementation, p = 0.16). After adjustment for patient demographics, prior diagnoses and clustering by facility, the odds ratio (OR) of severe hypoglycemia (compared to those not avoiding mixed insulin) was 0.73 (95%CI 0.66-0.82) for those fully implementing avoidance, and 0.81 (95% CI 0.70-0.93) for those partially implementing avoidance of mixed insulin. Similar results were found for any hypoglycemia, OR 0.84 (95%CI 0.79-0.89) for those fully implementing avoidance, and 0.84 (95%CI 0.78-0.91) for those partially implementing avoidance of mixed insulin. After adjustment, there remained no significant difference for facilities avoiding or not avoiding sliding scale only insulin regimens.

Implications:
Facilities that have implemented avoidance of mixed insulin regimens (long and short acting) had lower rates of inpatient hypoglycemia. This was not observed for avoidance of sliding scale only insulin regimens.

Impacts:
These findings support efforts to reduce inpatient hypoglycemia events by avoiding mixing of long and short acting insulins.