This article is more than 5 years old. Information may no longer be current.
Aggressive intervention improves glycemic control among non-adherent patients with diabetes
CHICAGO —Aggressive intervention with insulin in patients with type 2 diabetes who were at risk for being non-adherent resulted in improved glycemic control despite not affecting treatment persistence, according to research presented during the American Heart Association Scientific Sessions.
“Our takeaway from this is that targeting patient populations for more intensive interventions based on both the risk of non-adherence, as well as what their level of disease control is so we don’t have any adverse outcomes, could be more effective than delivering a blanket intervention to patients,” Julie C. Lauffenburger, PharmD, PhD, of Brigham Women’s Hospital and Harvard Medical School, said during a presentation of results from the TARGIT Diabetes at the American Heart Association Scientific Sessions.
In their three-armed, intention-to-treat randomized trial, researchers enrolled 6,000 adult patients with type 2 diabetes treated with insulin. All patients were insured through Horizon Blue Cross Blue Shield of New Jersey and had insulin prescriptions filled in the past 6 months.
The researchers sought to determine whether targeting smaller groups of patients who would most likely benefit from an intervention for treatment adherence would increase the efficacy of the intervention. They used predictive analytics to determine which patients were more likely to benefit from the intervention. In arm one of the study, patients were not targeted and received usual care. In arm two, 60% of patients were targeted based on their risk for non-persistence. Patients with moderate risk received a moderate-intensity intervention. In arm three, 40% of patients were targeted based on their risk for non-persistence and baseline glycemic control. Patients predicted to have poor adherence to their medication and a higher risk for glycemic control issues in this group received the high-intensity intervention. All non-targeted patients in each arm received usual care. Each study arm was designed to be cost neutral but varied according to the number of follow-up calls by pharmacists and providers, text messages and mailings.
Each arm included about 1,860 patients in the final analysis. Across each arm, patients were, on average, aged 55 years and had a mean HbA1c of 8.5% at baseline. Investigators assessed data up to 12 months after randomization.
For the primary outcome of insulin non-persistence, researchers found no significant difference between the groups. For each insulin prescription fill, 5.4% of patients in arm one were considered non-persistent compared with 4.7% of patients in arm two and 4.9% of patients in arm three.
Regarding glycemic control, patients in arm one had a mean reduction in HbA1c of 0.06% vs. 0.21% for patients in arm two and 0.31% for patients in arm three.
“In arm three, this was a significant difference in terms of the mean HbA1c change in the absolute difference between the two arms,” Lauffenburger said.
The absolute difference in mean HbA1c for arm two vs. arm one was 0.15%, and 0.25% for arm three vs. arm one.
Researchers found no significant differences between the groups for resource utilization, total health care spending and office visits.
“But what was unexpected was that there was an increase in all-cause ER visits and hospitalizations in the arm that was targeted purely on non-persistence compared with arm one,” Lauffenburger said. “We investigated this a little further. This was completely unexpected, and it appeared to be associated with having a fairly low HbA1c at baseline. So, the patients who had a HbA1c of less than 7% and were targeted for the adherence intervention were the ones who seemed to be attributable to this increase in hospitalizations.”
“We didn’t see a change on insulin persistence, but we did see a change in glycemic control, in arm three specifically,” she said.
Lauffenburger noted that other interventions attempting a similar goal may try to optimize predictive analytics to target further, particularly thinking more about incorporating disease control, and use different measurements. – by Gina Brockenbrough, MA
Reference:
Lauffenburger JC, et al. Abstract 164. Presented at: American Heart Association Scientific Sessions; Nov. 10-12, 2018; Chicago.
Disclosure: Lauffenburger reports receiving a research grant from Sanofi and AstraZeneca. The research was funded by Sanofi through an unrestricted grant.