A benefit–harm analysis of adding basal insulin vs. sulfonylurea to metformin to manage type II diabetes mellitus in people with multiple chronic conditions

Wendy Bennett, Hélène E. Aschmann, Milo A. Puhan, Craig W. Robbins, Elizabeth A. Bayliss, Renee Wilson, Richard A. Mularski, Wiley V. Chan, Bruce A Leff, Orla Sheehan, Carol Glover, Katie Maslow, Karen Armacost, Suzanne Mintz, Cynthia Boyd

Research output: Contribution to journalArticle


Objectives: The benefits and harms of diabetes treatments need to be carefully weighed in people with type II diabetes mellitus (DM) and multiple chronic conditions (MCCs). Our objective was to quantitatively assess the benefits and harms of the addition of basal insulin (insulin) vs. sulfonylurea (SU) to metformin in people with DM and MCCs. Study Design and Setting: Data inputs into the benefit–harms analysis included (1) baseline risks of patient-centered outcomes (death, myocardial infarction, stroke, severe hypoglycemia, diarrhea, nausea) from cohorts and trials; (2) treatment effects for the addition of insulin vs. SU from a network meta-analysis; and (3) patient preference survey for outcome weights. Statistical analysis calculated the probability that adding insulin has greater benefits than harms, when compared with an SU, overall and by prespecified subgroups. Results: Including the six outcomes, the probability of net benefit for insulin compared with SU was similar, across subgroups by age and diabetes duration (probability range, using conditional logit weights: 0.44–0.56). Adding patient preferences for treatment burden associated with insulin injections shifted the probability to favor SU over insulin (probability range, using conditional logit weights: 0.01–0.12). Conclusion: In people with DM and MCCs, we demonstrated incomplete evidence to conclude if basal insulin or SU should be added in people with DM and MCCs on metformin alone. The benefit–harm balance was sensitive to treatment preferences, that is., perceived treatment burden, indicating the importance of shared-decision making in caring for people with MCCs who are at high risk for experiencing harms associated with diabetes management.

Original languageEnglish (US)
Pages (from-to)92-100
Number of pages9
JournalJournal of Clinical Epidemiology
StatePublished - Sep 1 2019



  • Diabetes/complications
  • Diabetes/treatment
  • Multiple chronic conditions
  • Older adults
  • Patient preferences
  • Risk assessment/methods

ASJC Scopus subject areas

  • Epidemiology

Cite this