ADA 2018 — Real-world insulin efficacy hampered by nonadherence

  • W. Todd Penberthy, Ph.D.
  • Conference Reports
Access to the full content of this site is available only to registered healthcare professionals. Access to the full content of this site is available only to registered healthcare professionals.


  • The reduction in HbA1c and weight observed from insulin administration to patients with type 2 diabetes is greater in randomized controlled trials (RCTs) than under real‐world conditions. 
  • Although RCTs remain the gold standard for regulatory bodies, real-world evidence (RWE) measures the gaps in real-world outcomes relative to outcomes observed in RCTs. 

Why this matters

  • We have moved from an era of poor-quality and unreliable real-world data to one of reliable real-world evidence based on robust methodologies and technologies.
  • RWE measures gaps in real-world practice treatment effectiveness compared with outcomes generated from RCTs. This is increasingly being used to modify guidelines.
  • RCTs focus on efficacy and are the gold standard of drug regulation, whereas RWE addresses the question of whether these drugs can work in real-world settings. 

Study design

  • RWE comes from everything beyond phase 3 clinical trials in terms of efficacy, collected mostly from electronic health records under real-life practice circumstances and from multiple studies. RWE can include tens of thousands of patients and huge data sets. Often the patients actually treated are not considered because they are ineligible for RCTs.
  • RCTs were compared with RWE to evaluate concordance for lowering of HbA1c and weight with basal-bolus or premixed insulin. 
  • 8 RCTs were analyzed comparing basal-bolus (n=1893) or premixed insulin (n=1517) regimens. Real-world data were extracted from UK primary care data sets for people receiving basal-bolus (n=7483) or premixed insulin regimens (n=10,744). 

Key results

  • Greater reductions were observed with RCTs than with RWE for HbA1c (real-world data, −0.28%; RCT data, −1.4%) and weight gain (real-world data, +0.27 kg; RCT data, +2.96 kg).
  • Further analysis revealed that the decreased effectiveness seen in RWE was the result of a lack of adherence 75% of the time. 


  • Advanced methodologies are required to overcome confounding factors with RWE.

Please confirm your acceptance

To gain full access to GPnotebook please confirm:

By submitting here you confirm that you have accepted Terms of Use and Privacy Policy of GPnotebook.