Factors affecting basal insulin injection adherence investigated using smart insulin pen - PowerPoint Presentation

1. Factors affecting basal insulin injection adherence investigated using smart insulin pen dataMarkus Menzen1; Niels Væver Hartvig2; Anne Kaas2; Nikoline Nygård Knudsen2; Johan Jendle31Division of Internal Medicine, Department of Gastroenterology and Diabetology, Community Hospital Bonn, Bonn, Germany2Novo Nordisk A/S, Bagsværd, Denmark3School of Medicine, Institute of Medical Sciences, Örebro University, Örebro, Sweden1

2. Disclosures and acknowledgementsThis study was sponsored by Novo NordiskThe authors acknowledge the medical writing assistance of Oxford PharmaGenesis, Oxford, UKINCLUDEDIN POSTER

3. T1D, type 1 diabetes1. Peyrot et al. Diabetes Care 2010;33(2):240–5; 2. Brod et al. Patient Intell 2014;6:43–60; 3. Sarbacker et al. Diabetes Spectr 2016;29(3):166–70; 4. Ekberg et al. J Diabetes Sci Technol 2022; doi: 10.1177/19322968221104142INCLUDEDIN POSTERIn a real-world setting, we investigated the probability of and influencing factors for missing basal insulin injections in smart pen users with T1DAimManaging multiple daily insulin injections around the demands of daily life can be challenging and burdensome for individuals with T1D, resulting in individuals missing or forgetting insulin injections1,2Missed daily insulin injections can lead to inadequate glycaemic control in individuals with T1D3,4The use of smart pens can provide detailed insights into treatment adherenceBackground

4. Adherence to insulin degludec injections in smart pen users ADDITIONALINFORMATIONA missed basal dose defined as a period of ≥40 hours between consecutive basal injectionsFirst injection status ‘unknown’U, unitsExample for one participant

5. Data were collected from smart pens (NovoPen 6) used to administer insulin degludec injections in individuals with T1D across 13 countriesThe daily probability of missing an insulin degludec dose (≥40 hours between injections) was estimated using a generalised linear mixed model with logistic link functionMethodsINCLUDEDIN POSTERage (spline effect with three degrees of freedom)gendermedian (typical) injection time of day (00:00–09:00, 09:01–18:00 and 18:01–24:00)median daily basal dose (≤30 units and >30 units)number of smart insulin penstime since smart pen initiationdevice types used (smart insulin pen with or without CGM using a CGM device and/or self-monitoring of blood glucose using a BG meter)smart pen engagement (characterised by the number of days with data uploads in the previous 14 days)day of the weekThe patient and treatment factors investigated were:BG, blood glucose; CGM, continuous glucose monitoring; T1D, type 1 diabetesAn interaction term was tentatively included to investigate if the day of the week effect depends on age

6. Patient characteristics (1/2)aIncludes Austria, Belgium, France, Ireland and any participants whose country of residence was unknownSD, standard deviation; T1D, type 1 diabetesIndividuals with T1D (n=1336)Mean (SD) age at consent, years31.8 (17.7)Female, n (%)614 (46.0)Country of residence, n (%)SwedenUKGermanyJapanDenmarkNorwayFinlandCzech RepublicSwitzerlandOthera547 (40.9)390 (29.2)110 (8.2)79 (5.9)72 (5.4)59 (4.4)43 (3.2)16 (1.2)5 (0.4)15 (1.1)Number of smart insulin pens, n (%)123≥4116 (8.7)974 (72.9)164 (12.3)82 (6.1)INCLUDEDIN POSTER

7. Patient characteristics (2/2)aIncludes only participants administering bolus insulin with a smart penAspart, insulin aspart; BG, blood glucose; CGM, continuous glucose monitoring; faster aspart, fast-acting aspart; SD, standard deviation; T1D, type 1 diabetes; U, units INCLUDEDIN POSTERIndividuals with T1D (n=1336)Device type(s) used, n (%)Smart insulin pen Smart insulin pen with BG meterSmart insulin pen with CGM deviceSmart insulin pen with CGM device and BG meter398 (29.8)186 (13.9)640 (47.9)112 (8.4)Bolus insulin used,a n (%)Faster aspartAspartFaster aspart and aspart580 (43.4)569 (42.6)34 (2.5)Mean (SD) basal dose, U/day26.5 (16.0)Mean (SD) frequency of missed basal doses, %5.7 (8.5)Mean (SD) number of data uploads per day, n0.3 (0.7)

8. Study characteristicsSD, standard deviation; T1D, type 1 diabetesIndividuals with T1D (n=1336)Number of days in which degludec injections occurred, n171 378Mean (SD) observation period (days in which degludec injections occurred), days128.3 (117.4)Mean (SD) observation period (period from first to last injection), days156.0 (138.8)Study periodFrom 6 March 2021 to 5 December 2022INCLUDEDIN POSTER

9. Based on analysis of 171 378 injection days, the mean daily probability of missing a basal insulin dose was 3.5% (95% CI: 3.1–4.1%)Factors significantly associated with treatment adherence were:INCLUDEDIN POSTERKey resultstypical injection timebasal insulin dosedata upload frequencyday of the weekageCI, confidence interval

10. Effect of age, day of the week and typical injection time of day on basal insulin treatment adherenceData are estimated least-squares means (95% confidence interval) from 1336 individuals with type 1 diabetes and include 171 378 days in which insulin degludec injections occurredINCLUDEDIN POSTERIndividuals aged 20–25 years had the highest probability of missing a basal doseThe estimated probability of missing a basal injection was higher on Saturdays than on other days of the weekNo interaction between day of the week and age was found (p=0.148)There was a higher probability of missing a basal injection if the individual typically injected during daytime hours (09:01–18:00) than if they typically injected between 00:00 and 9:00 and between 18:01 and 24:00 Key results

11. ADDITIONALINFORMATIONData are estimated least-squares means (95% confidence interval) from 1336 individuals with type 1 diabetes and include 171 378 days in which insulin degludec injections occurredIndividuals with more frequent data uploads were less likely to miss a basal dose than those with less frequent data uploadsIndividuals with a median daily basal dose of ≤30 units were more adherent than those with doses of >30 unitsEffect of data upload frequency and basal insulin dose on basal insulin treatment adherence

12. ADDITIONALINFORMATIONNo significant effects on treatment adherence were observed for device type (CGM device or BG meter), gender and time since smart pen initiation (p>0.05)Data are estimated least-squares means (95% confidence interval) from 1336 individuals with type 1 diabetes and include 171 378 days in which insulin degludec injections occurredBG, blood glucose; CGM, continuous glucose monitoringEffect of device type, gender and time since smart pen initiation on basal insulin treatment adherence

13. Data are estimated least-squares means (95% confidence interval) from 1336 individuals with type 1 diabetes and include 171 378 days in which insulin degludec injections occurredThe number of smart insulin pens used by individuals revealed a significant effect: individuals with two pens had the best estimated adherenceADDITIONALINFORMATIONEffect of number of smart pens on basal insulin treatment adherence

14. ConclusionsThe mean daily probability of missing a basal insulin dose was 3.5% (95% CI: 3.1–4.1%)Basal insulin injection treatment adherence was significantly associated with the age of the individual, day of the week, typical injection time of day, level of smart pen engagement (data upload frequency) and daily basal insulin doseTreatment adherence was not significantly affected by device type, gender and time since smart pen initiation Encouraging engagement with smart pen data, advising individuals to inject early in the morning or later in the evening and providing guidance on how to manage injections across the week may increase basal insulin adherence and improve clinical outcomes in T1DA similar analysis in individuals with type 2 diabetes is underwayINCLUDEDIN POSTERCI, confidence interval; T1D, type 1 diabetes