Diabetes teams provide timelier, more effective inpatient care with early intervention model
Electronically identifying hospitalized patients with diabetes or new hyperglycemia and allowing a diabetes team to provide immediate expert care without a referral can improve glycemic measures and protect against adverse events, according to findings published in Diabetes Care.
“The implications of this trial may be significant for revising approaches to models of care for managing diabetes in hospitals worldwide,” Spiros Fourlanos, MD, PHD, director of the department of diabetes and endocrinology at the The Royal Melbourne Hospital in Australia, told Endocrine Today. “Potentially, diabetes models of care in hospital might need to shift towards delivering more proactive care early in the admission course for people with diabetes.”
Using data from the Randomized Study of a Proactive Inpatient Diabetes Services (RAPIDS), Fourlanos and colleagues evaluated a treatment strategy designed around electronic blood glucose monitoring and compared it with usual care during 24 weeks at Royal Melbourne Hospital. After receiving information from electronic blood glucose meters signaling dysglycemia, a diabetes specialist did not require a referral to begin treatment and could initiate within 24 hours. Usual care required a referral based on medical charts before a specialist could intervene.
The researchers randomly assigned eight wards from the hospital to the intervention or to usual care. Before initiation of the intervention, all wards provided usual care and were observed for 10 weeks as a baseline. Each ward was then observed for 14 weeks once the intervention began. This period was categorized as “active” by the researchers.
From March to August 2016, 1,002 adults were consecutively admitted to the hospital wards. At baseline, 221 were treated in wards assigned to usual care (mean age, 70 years; 38.9% women), and 270 were treated in wards assigned to the intervention (mean age, 70 years; 38.5% women). During the active study period, 220 patients were treated in wards assigned to usual care (mean age, 70 years; 49% women) and 291 were treated in wards assigned to the intervention (mean age, 71 years; 46.4% women).
More patients were treated by a specialist diabetes team in the active intervention group compared with baseline (92% vs. 8%; P < .001); 64% of patients in the intervention group during the active period were treated by a specialist in less than 24 hours compared with only 4% at baseline (P < .001). A higher rate of patients without prior insulin treatment received insulin during active intervention compared with baseline (57% vs. 34%; P < .001).
Intervention led to a greater decrease in number of days with adverse glycemic levels (186 per 1,000 patient-days), based on a “liberal glycemic target” of between 4 mmol/L and 15 mmol/L, compared with usual care in the intervention group at baseline (243 per 1,000 patient-days; P < .001). Patients in the intervention group experienced adverse glycemic days 23% less often than those receiving usual care (P = .008).
The intervention group had a 0.4-mmol/L reduction in mean glucose compared with baseline (P = .003) and a 3.3% rate for days in which blood glucose averaged more than 15 mmol/L compared with 7.3% before intervention (P < .001). The usual care group did not have significantly different results from baseline in either measure. Additionally, a lower risk for hospital-acquired infections was observed in the intervention group compared with the usual care group (adjusted OR = 0.2; 95% CI, 0.07-0.58).
“This research is important as it suggests proactive models of diabetes care in hospital improves the safety of glucose management and could decrease the serious complication of hospital-acquired infection,” Fourlanos said. “We did not expect proactive care to have such a major impact on decreasing hospital-acquired infection. Of course infection was a secondary outcome measure, and ideally these findings need to be validated in future randomized clinical trials.” – by Phil Neuffer
Disclosures: The authors report no relevant financial disclosures.