Background: Attention to the quality of glycemic management during the hospital stay, including care transitions, has grown with increasing evidence linking inpatient and post-discharge glycemic control to clinical outcomes, widespread use of glucometric benchmarking, and impending pay-for-performance measures. Clinical inertia (nonadherence to guidelines) and therapeutic inertia (failure to adjust medications as indicated) or CTI among providers may adversely affect inpatient glycemic management. Knowledge gaps, competing priorities, or belief the hospital encounter provides insufficient time to achieve glycemic targets may lead to CTI. Recent data on adverse outcomes in older adults following insulin initiation or intensification at hospital discharge also likely increases concern about iatrogenesis. In this review, we synthesized the existing empirical evidence on CTI in the management of diabetes and acute hyperglycemia in the hospital.
Methods: We searched PubMed, CINAHL, and EMBASE databases (Jan/Feb 2022) from inception to December 31, 2021, for peer-reviewed studies on CTI in inpatient and transitional glycemic management. We retrieved 1178 citations (733 duplicates), adding 2 studies from the gray literature and 23 studies from the bibliographies of identified works. We screened 471 nonduplicative citations for relevance by title and abstract, yielding 75 studies for full text review. Main outcomes were provider behaviors suggesting inappropriate inaction. Secondary outcomes were clinical events related to CTI. Evidence quality was appraised (John Hopkins Evidence Level & Quality Guide). We report findings by narrative synthesis given the methodological heterogeneity among studies.
Results: We identified 14 studies [publication years: 1997, 2006(2), 2007, 2008, 2009, 2011, 2012, 2013, 2014, 2018, 2019(2), 2020], including 1 qualitative and 13 observational cohort studies. Of these, 13 (93%) had good methodological quality. The sample included 12,342 patients from 49 academic medical centers and 4 VA hospitals. Glycemic status was at times not discussed during rounds and omitted (38%) or not trended (47%) in progress notes. A1c testing 90 days prior or during the stay was omitted in 55% to 69% of patients. Sliding scale insulin was used as monotherapy for 23% to 31% of all patients and 80% of postoperative patients with diabetes. Basal insulin was used for 20% to 42% of patients, mostly among chronic insulin users, and rarely adjusted for ongoing hyperglycemia despite low post-titration hypoglycemia risk. Two studies found evidence of negative therapeutic momentum (i.e., insulin dose reductions in the setting of hyperglycemia). Only 20% of discharge summaries had transitional diabetes plans; 30-day primary care or specialist follow-up visit was arranged for 46% of patients with uncontrolled diabetes. Prehospital regimens were intensified in 22% to 29% of indicated cases (per A1c and inpatient glucose control); antihyperglycemic drugs were inappropriately deprescribed in 5% of discharges. In 1 of 2 studies, Black patients were less likely than others to receive treatment intensification when indicated. Inpatient hyperglycemia affected 27% to 83% of patients; of these, 10% to 29% had ≥80% of glucose values above goal. Hypoglycemia occurred in 1.2% to 23% of patients. Lastly, CTI was associated with discharge to a care facility but not 30-day readmission.
Conclusions: In this systematic review, only a small fraction of inpatient glycemic management was guideline-concordant.