Factors associated with withdrawal from insulin pump therapy: A large-population-based study

Background Although, number of diabetic patients received insulin pump (IP) therapy is increasing; there are limited data regarding factors associated with IP withdrawal. Methods We conducted a cross-sectional study using data from an Israeli health maintenance organization. All patients, 21 or older, with type 1 (T1DM) or type 2 (T2DM) diabetes, who received IP therapy for a 7-year period were identified. Patients who did not purchase IP maintenance supplies for at least six consecutive months were defined as withdrawn (N = 355). Patients who purchased supplies were defined as adherent (N = 352). Results In both T1DM and T2DM patients, withdrawal from IP therapy was positively associated with duration of diabetes longer than 5 years (odds ratio [OR] = 13.26 [CI, 7.16-23.34; P < .001] and OR = 10.92 [CI, 5.64-21.14; P < .001], respectively), nonadherence to dietician follow-up (OR = 5.78 [CI, 3.65-9.14; P < .001] and OR = 3.41 [CI, 1.99-5.85; P < .001], respectively), and poor glycaemic control prior to IP treatment (OR = 4.04 [CI, 2.18-7.48; P < .001] and OR = 4.59 [CI, 2.71-7.81; P < .001], respectively]. Co-morbid neuro-psychiatric disorders were also risk factors for IP withdrawal: diagnosis of depression (OR = 2.22 [CI, 1.16-4.27; P = .017] and Attention Deficit Hyperactivity Disorder (ADHD) OR = 2.45 [CI, 1.003-5.087; P = .043]) among T1DM patients; and diagnosis of depression (OR = 1.85 [CI, 1.05-5.27; P = .046] and dementia OR = 4.03 [CI, 1.03-19.77; P = .048]) among T2DM patients. Conclusion In our large real-world population-based study, we found that smoking, obesity, poor glycaemic control, and co-morbid neuro-psychiatric disorders were associated with a high rate of withdrawal from IP therapy. Health care providers ought to familiarize themselves with patient characteristics predictive of nonadherence and should intensify patient follow-up when incorporating this new, costly, and challenging technology.