AUGMENTed Real-World Data Enhances Comparative Efficacy Between Once-Weekly Insulin Icodec with Dosing Guide App Versus Once-Daily Insulin Glargine U300 in Insulin-Naive Type 2 Diabetes

Abstract

Introduction

ONWARDS 5 evaluated the effectiveness and safety of insulin icodec (icodec) titrated with a dosing guide app (icodec with app) versus once-daily insulin analogs in insulin-naive adults with type 2 diabetes. The insulin glargine U300 (glargine U300) stratum was too small to enable a robust post hoc efficacy comparison. Augmentation methodology was applied to increase the glargine U300 group size using real-world data (RWD), to facilitate efficacy comparisons of icodec with app versus glargine U300, and to demonstrate the potential of the augmentation methodology to strengthen underpowered treatment comparisons (AUGMENT study).

Methods

ONWARDS 5 data were augmented with RWD collected from the US Ambulatory Electronic Medical Records database. Randomized and augmented comparisons (propensity-score-matched) between icodec with app and glargine U300 were weighted to provide a fully augmented estimate of the primary outcome (change in glycated hemoglobin [HbA_1c] after 52 weeks). Data were adjusted for trial effects. Sensitivity analyses were conducted.

Results

The nonaugmented randomized estimated treatment difference (ETD; 95% CI) between icodec with app and glargine U300 (trial stratum) for change in HbA_1c was − 0.21 (− 0.70 to 0.28) percentage points. After adjusting for trial effects, the overall fully augmented ETD (95% CI) was − 0.33 (− 0.68 to 0.01) percentage points numerically in favor of icodec with app, although not statistically significant. Sensitivity analyses supported the findings.

Conclusions

Using augmented data, the precision of the change in HbA_1c estimate was increased compared with the trial stratum analysis alone. These findings help to validate the principle of utilizing augmentation to strengthen trial outcomes.

Trial Registration Number

The ONWARDS 5 trial is registered with ClinicalTrials.gov (NCT04760626).

Supplementary Information

The online version contains supplementary material available at 10.1007/s13300-024-01679-3.

Key Summary Points

Why carry out the study?

Using an augmented approach of ONWARDS 5 data, the precision of the estimated treatment difference for change in glycated hemoglobin between insulin icodec (in combination with a dosing guide app) and insulin glargine U300 was increased compared with that derived from a post hoc stratum analysis alone.

To our knowledge, this is the first diabetes study to use augmented data, which combine clinical trial data and propensity-score-matched real-world data to strengthen the assessment of comparative effectiveness between two treatments evaluated in a randomized clinical trial.

What was learned from the study?

Our findings demonstrate the potential for data augmentation to help to strengthen clinical trial results, to expand small treatment groups, and to bridge the gap between clinical trial efficacy and real-world effectiveness.

Introduction

Despite the introduction of several compounds into the market for the treatment of diabetes, insulin remains a therapeutic cornerstone. Insulin icodec (icodec) is a once-weekly basal insulin that was evaluated in the ONWARDS phase 3a clinical trial program. The six ONWARDS trials (ONWARDS 1–6) assessed the efficacy and safety of icodec compared with once-daily (OD) basal insulin analogs (OD analogs: insulin degludec [degludec], insulin glargine U100 [glargine U100], or insulin glargine U300 [glargine U300]) in adults with T2D or T1D [1–7].

ONWARDS 5 (ClinicalTrials.gov: NCT04760626) was a phase 3a clinical trial with real-world elements designed to evaluate the long-term effectiveness and safety of icodec titrated with a dosing guide app (icodec with app) compared with OD analogs dosed as per standard clinical practice in insulin-naive adults with T2D [1]. The icodec with app group showed a noninferior (P < 0.001) and superior (P = 0.009) change in glycated hemoglobin (HbA_1c) from baseline to week 52 (− 1.68 percentage points) compared with the OD analog group (− 1.31 percentage points). Furthermore, the rates of clinically significant or severe hypoglycemia were low and similar between icodec with app and OD analogs [1].

It is important to evaluate the comparative effectiveness of emerging insulin regimens (such as icodec) against a range of different comparators, including glargine U300; however, there are limited data comparing these two treatments in people with T2D. ONWARDS 5 was the only ONWARDS trial to include participants receiving glargine U300 as a comparator against icodec with app [1], but a robust post hoc analysis of these treatment groups in ONWARDS 5 is not feasible because the study was not powered for a comparison between icodec with app and glargine U300. To address this, we hypothesized that augmenting the ONWARDS 5 population with real-world data (RWD) from a large data set could help to strengthen the comparative evaluation between icodec with app and glargine U300. The use of RWD from clinical practice to augment clinical trial data is a relatively recent technique that can increase the sample sizes of small comparator groups and enable comparisons with other treatments [8–12]. The aim of this analysis was twofold: (1) to compare the efficacy of icodec with app with glargine U300 based on augmented ONWARDS 5 data, and (2) to demonstrate the feasibility of using RWD to augment randomized clinical trial (RCT) data to increase the precision of estimates when the sample size of RCT data alone may limit a sufficiently powered analysis.

Methods

ONWARDS 5 Overview

The trial design of ONWARDS 5 has previously been published [1, 5]. Briefly, ONWARDS 5 was a 52-week, open-label, randomized, multinational, phase 3a trial with real-world elements. Eligible participants were adults (≥ 18 years old) with T2D requiring insulin initiation, with an HbA_1c value over 7.0%, receiving any noninsulin glucose-lowering medications. Participants were assigned to degludec, glargine U100, or glargine U300 at the discretion of the investigator at the initial screening visit and were subsequently randomized 1:1 to receive either the assigned OD analog (degludec [n = 378], glargine U100 [n = 96], glargine U300 [n = 69]) or icodec with app (n = 542). Real-world elements were incorporated to mimic standard clinical practice including broad inclusion criteria, choice and dosing of the comparator OD analogs at the investigator’s discretion, and a low number of individual trial-site visits (approximately 3 months apart) for both treatment groups.

The ONWARDS 5 trial was conducted in accordance with the principles of the Declaration of Helsinki and the Good Clinical Practice guidelines of the International Conference for Harmonisation. The protocol, consent form, and other relevant documents were reviewed and approved by the appropriate institutional review boards or independent ethics committees. All participants provided written informed consent for participation and publication of data before trial entry.

Augmentation of ONWARDS 5 with RWD (AUGMENT Study)

To create a larger cohort of participants with glargine U300 for comparison with icodec with app, the ONWARDS 5 clinical trial data were augmented with RWD collected between 2015 and 2019 from the US Ambulatory Electronic Medical Records (AEMR) database (Fig. 1). The AEMR database is commercially available, contains information from outpatient clinics and physician’s offices, and provides a comprehensive and robust repository of patient data. The database captures a wide range of data, including demographics, diagnoses, laboratory results, procedures, and prescriptions.

Fig. 1.

Augmentation of ONWARDS 5 RCT data with RWD to enable a robust comparative analysis of icodec with app and glargine U300 in adults with T2D. glargine U300 insulin glargine U300, icodec with app insulin icodec titrated with a dosing guide app, RCT randomized controlled trial, RWD real-world data, T2D type 2 diabetes

When conducting an augmented study, the supplementary RWD should emulate the components of the RCT protocol as best as possible, particularly with regard to participant eligibility, treatment assignment compatibility, follow-up period, and statistical analyses [12]. Accordingly, the inclusion criteria used for extracting participant data from the AEMR database were: at least one prescription claim for basal insulin (degludec, glargine U100, glargine U300) during the index window (January 1, 2015 to August 31, 2020); at least two fills of the index treatment (including the fill on the index date itself); enrollment at least 360 days before the index date (the 1-year pre-index period); continuous enrollment for at least 360 days from (and including) the index date (the 1-year post-index period); at least 18 years of age at index; consistent and known sex; at least one diagnosis on a confirmatory medical claim for T2D in any period before the index date (including the index date itself) defined by diagnosis or by at least two noninsulin glucose-lowering medication claims; at least one measure of HbA_1c in the 90 days before the index date; and at least one measure of HbA_1c in the 365 days after the index date.

Outcomes

The primary outcome for this post hoc analysis was change in glycemic control (HbA_1c levels) from the index date to 52 weeks after the index date or to drop out (e.g., discontinuation). For ONWARDS 5 participants and AEMR controls, missing data were imputed using multiple imputation (e.g., for individuals who discontinued or were lost to follow-up). Discontinuation in ONWARDS 5 was defined per protocol regarding rescue medication allowance (noninsulin glucose-lowering medications were maintained or altered at the investigator’s discretion, whereas basal-bolus intensification was a reason for treatment discontinuation); for AEMR data, discontinuation was based on individual electronic medical record (EMR) prescription fill/claim start and stop dates. The index date (time 0) was set for AEMR participant data as the start of basal insulin treatment to mimic the ONWARDS 5 trial.

The primary analysis of ONWARDS 5 adhered to the treatment policy estimand for the analysis of HbA_1c, and missing values were imputed based on trial participants who had a measurement at the week 52 visit but who discontinued treatment before the week 52 visit (retrieved dropouts). This type of multiple imputation differed from the sequential multiple imputation by Bayesian linear regression performed herein for AEMR missing values (see Supplementary Material for further details).

Propensity Score Matching

Propensity score matching between participant data from ONWARDS 5 and EMRs from the AEMR database was performed to ensure appropriate matching between RWD and RCT participant data. The 1:1 propensity score matching was performed based on age, HbA_1c, and noninsulin glucose-lowering medication use (glucagon-like peptide-1 receptor agonist [GLP-1 RA], sodium-glucose transport protein 2 inhibitor [SGLT2i], dipeptidyl peptidase 4 inhibitor [DPP4i]) at baseline. Table S1 shows the matches between ONWARDS 5 and AEMR data that facilitated the calculation of trial effects and the treatment comparison of icodec with app against glargine U300.

Statistical Analysis

The strategy for the analysis was to generate an augmented synthetic glargine U300 data set (by replacing the ONWARDS 5 degludec and glargine U100 control groups) suitable for comparison against ONWARDS 5 icodec with app data. This ‘augmented comparison’ would then supplement the original ‘randomized comparison’ between glargine U300 and icodec with app generated from ONWARDS 5 (Fig. S1). The effects of the two comparisons were weighted based on their relative sample sizes, and a weighted mean was reported.

A trial effect is considered to be a benefit that participants may experience when they are enrolled in RCTs simply because of the fact they are participating in an RCT [13]. These trial effects need to be accounted for when performing comparisons of augmented clinical data. The generation of the augmented data set was conceptually performed in two steps: (1) the estimation of the trial effect based on trial participants who received OD analogs, then (2) the application of the trial effect to real-world glargine U300 matched participants to create glargine U300 synthetic controls. Figure 2 presents a schematic illustrating these two steps. During the initial step, propensity score matching was used to identify individuals from the available AEMR database population who matched the characteristics of the participants from the degludec, glargine U100, and glargine U300 control arms of ONWARDS 5. The change in HbA_1c from baseline at 52 weeks between ONWARDS 5 and AEMR participants was estimated to identify the trial effect by stratum (i.e., degludec, glargine U100, and glargine U300). The estimated trial effects were then combined in a fixed effects meta-analysis to produce a single trial effect estimate and associated uncertainty. For the second step, ONWARDS 5 participants assigned to degludec or glargine U100 at screening but randomly assigned to receive icodec with app were matched with AEMR participants who received glargine U300. Subsequently, the change from baseline in HbA_1c between the icodec with app group and the glargine U300 synthetic control group was calculated, taking into account the overall trial effect estimated during the initial step.

Fig. 2.

Conceptual overview of the two-step process to augment the ONWARDS 5 data with RWD from the AEMR database. AEMR Ambulatory Electronic Medical Records, DPP4i dipeptidyl peptidase-4 inhibitor, glargine U300 insulin glargine U300, GLP-1 RA glucagon-like peptide-1 receptor agonist, HbA_1c glycated hemoglobin, icodec with app insulin icodec titrated with a dosing guide app, OD once-daily, RWD real-world data, SGLT2i sodium-glucose transport protein 2 inhibitor

For estimation, a joint statistical model for the propensity-score-matched ONWARDS 5 and AEMR data was developed to account for the covariance generated by the contribution of ONWARDS 5 glargine U300 data to both the trial effect calculations and the treatment comparison against icodec with app (see Supplementary Material—statistical appendix for further details).

The primary analysis assumed full augmentation. Acknowledging that clinical observers may differ in their belief about the relevance of augmenting clinical trial data with RWD, an additional parameter, ρ, was introduced to control for augmentation plausibility. The parameter 0 ≤ ρ ≤ 1 corresponded to the belief in the plausibility of augmentation (ρ = 0 corresponded to a belief that augmentation provided no additional information, whereas ρ = 1 corresponded to a belief that augmentation was fully relevant). At the extremes, ρ = 0 represented only the glargine U300 stratum-specific randomized comparison, whereas ρ = 1 represented the fully augmented (primary) analysis.

Sensitivity Analyses

Sensitivity analyses for trial effect, emulation criteria, and emulation timing factors were conducted to test the robustness of the evaluated outcome. The relevance of using degludec and glargine U100 trial effects to partly represent the trial effect for glargine U300 could be subject to scrutiny; therefore, a sensitivity analysis was conducted by incorporating an additional parameter into the model: 0 ≤ φ ≤ 1, which corresponded to the belief in the weighted mean trial effect estimation (φ = 0 corresponded to a belief that only the glargine U300-matched comparator group sufficiently contributed towards a glargine U300 trial effect, whereas φ = 1 corresponded to a belief that the full weighted trial effect was as calculated in the primary augmented analysis). To test the effect of the imputation method, another sensitivity analysis was conducted while adopting a last observation carried forward imputation strategy. Additionally, to explore how AEMR emulation criteria affected results, further emulations were conducted testing a range (from narrow to broad) of different emulation criteria (Table S2). Finally, timing factors of the AEMR emulation were also explored via a sensitivity analysis, in which AEMR participants were selected with time 0 closer in proximity to the recorded baseline results of participants in the ONWARDS 5 trial.

Results

Study Population and Baseline Characteristics

All 542 ONWARDS 5 participants receiving icodec with app were used for the propensity-score-matched analysis, and 542 participants receiving OD analogs were included in the calculation of the trial effect. From the AEMR database, 18 433 individuals were selected and matched to the ONWARDS 5 trial population on a 1:1 basis.

Baseline characteristics of the participants from ONWARDS 5 and AEMR after emulation but before propensity score matching are shown in Table 1. The mean age of participants was approximately 59 years for both groups, and duration of diabetes was 11.9 years and 5.6 years in the ONWARDS 5 and AEMR groups, respectively. The mean baseline HbA_1c in ONWARDS 5 participants and AEMR individuals was 8.9% and 10.0%, respectively, whereas the median HbA_1c was 8.5% and 9.6%, respectively. Good distribution of characteristics between the two sources was achieved following propensity score matching (Table 1).

Table 1.

Baseline characteristics in ONWARDS 5 and AEMR before and after propensity score matching for the trial effect and treatment contrasts analyses

	Before propensity score matching		After propensity score matchinga
			Trial effect analysis		Treatment contrasts analysis
	ONWARDS 5 (N = 1085)	AEMR (N = 18 433)	ONWARDS 5 control arm (N = 542)	AEMR (OD analogs)b (N = 542)	ONWARDS 5 icodec arm (degludec, glargine U100 strata) (N = 459)	AEMR (glargine U300)c (N = 542)
Sex, female, %	42.7	47.2	42.3	43.7	41.8	48.4
Age, years, mean (SD)	59.3 (10.5)	59.9 (12.0)	59.4 (10.2)	59.4 (11.6)	59.0 (10.7)	59.3 (11.6)
HbA1c, %
Mean (SD)	8.9 (1.6)	10.0 (1.9)	8.9 (1.5)	8.8 (1.4)	9.0 (1.6)	9.1 (1.5)
Median	8.5	9.6	8.5	8.5	8.6	8.8
Duration of diabetes, years, mean (SD)	11.9 (7.3)	5.6 (3.6)	12.0 (7.6)	5.9 (3.7)	11.9 (6.8)	5.5 (3.6)
BMI, kg/m2, mean (SD)	32.8 (7.0)	NA	33.0 (7.0)	NA	32.4 (7.1)	NA
eGFR, ml/min/1.73 m2, mean (SD)	88.1 (20.7)	NA	88.0 (20.3)	NA	88.1 (20.9)	NA
Ethnicity, n (%)
Hispanic or Latino	95 (8.8)	84 (0.5)	44 (8.1)	2 (0.4)	46 (10.0)	2 (0.4)
Not Hispanic or Latino	989 (91.2)	16 465 (89.3)	498 (91.9)	506 (93.4)	412 (89.8)	416 (90.6)
Not reported	1 (0.1)	1884 (10.2)	0	34 (6.3)	1 (0.2)	41 (8.9)
Race, n (%)
American Indian or Alaska Native	3 (0.3)	NA	1 (0.2)	NA	1 (0.2)	NA
Asian	47 (4.3)	394 (2.1)	19 (3.5)	9 (1.7)	27 (5.9)	7 (1.5)
Black or African American	52 (4.8)	1925 (10.4)	28 (5.2)	45 (8.3)	22 (4.8)	30 (6.5)
Native Hawaiian or other Pacific Islander	3 (0.3)	NA	1 (0.2)	NA	2 (0.4)	NA
Not reported	1 (0.1)	1884 (10.2)	0	34 (6.3)	0	41 (8.9)
White	971 (89.5)	13 519 (73.3)	492 (90.8)	434 (80.1)	401 (87.4)	367 (80.0)
Other	8 (0.7)	711 (3.9)	1 (0.2)	20 (3.7)	6 (1.3)	14 (3.1)
Use of noninsulin glucose-lowering medications at baseline, %
SGLT2i	43.7	23.3	44.3	39.9	51.0	42.7
GLP-1 RA	28.2	23.0	29.2	26.4	32.2	23.1
DPP4i	28.2	38.5	26.9	29.2	34.9	27.2

AEMR Ambulatory Electronic Medical Records, BMI body mass index, DPP4i dipeptidyl peptidase-4 inhibitor, eGFR estimated glomerular filtration rate, glargine U100 insulin glargine U100, glargine U300 insulin glargine U300, GLP-1 RA glucagon-like peptide-1 receptor agonist, HbA_1c glycated hemoglobin, icodec insulin icodec, NA not available, OD once-daily, RWD real-world data, SD standard deviation, SGLT2i sodium-glucose transport protein 2 inhibitor

^aPropensity score matching was carried out using age, HbA_1c, and noninsulin glucose-lowering medication use (SGLT2i, GLP-1 RA, and DPP4i) at baseline

^bRWD from individuals who were matched 1:1 to ONWARDS 5 participants who received comparator OD analogs

^cRWD from individuals prescribed glargine U300 who were matched 1:1 to ONWARDS 5 participants who received icodec in the glargine U100 and degludec strata to perform icodec versus glargine U300 comparisons using all icodec-treated participants from ONWARDS 5

Trial Effect

The estimated trial effects (95% CI) based on change from baseline in HbA_1c between ONWARDS 5 and AEMR control strata were 0.19 (− 0.19 to 0.58) percentage points, − 0.31 (− 1.10 to 0.49) percentage points, and 0.29 (− 0.36 to 0.93) percentage points for degludec, glargine U100, and glargine U300, respectively (Fig. 3). Based on data from 1084 individuals, the overall meta-analyzed trial effect (95% CI) was 0.13 (− 0.25 to 0.49) percentage points (Fig. 3).

Fig. 3.

Trial effect and augmented treatment effect (icodec with app versus glargine U300) on change from baseline in HbA_1c (%) after 52 weeks. ^aRWD individuals were matched 1:1 to ONWARDS 5 individuals who received comparator OD analogs. ^bRWD individuals prescribed glargine U300 were matched 1:1 to ONWARDS 5 individuals who received icodec in the glargine U100 and degludec strata to perform icodec versus glargine U300 comparisons using all icodec-treated participants from ONWARDS 5. ^cN values correspond to the total contribution of individuals from the trial effect estimate (^a) and the treatment effect estimate (^b). degludec insulin degludec, glargine U100 insulin glargine U100, glargine U300 insulin glargine U300, HbA_1c glycated hemoglobin, icodec with app insulin icodec titrated with a dosing guide app, RCT randomized clinical trial, RWD real-world data

Change in HbA_1c

The randomized and the augmented (unadjusted and adjusted) estimated treatment differences (ETDs) for change in HbA_1c from baseline to week 52 between icodec with app and glargine U300 are shown in Fig. 3. Regarding the ONWARDS 5 post hoc randomized comparison, the ETD (95% CI) between icodec with app (n = 83) and glargine U300 (n = 69) for change in HbA_1c from baseline to week 52 was − 0.21 (− 0.70 to 0.28)  percentage points. Estimates for glargine U100 and degludec strata are shown in Fig. S2. For the unadjusted augmented comparison, the ETD (95% CI) between icodec with app (n = 542) and the propensity-score-matched AEMR glargine U300 individuals (n = 528) was − 0.44 (− 0.70 to − 0.18) percentage points. After adjusting for trial effect and combining the randomized ONWARDS 5 comparison and the augmented synthetic comparison, the overall weighted ETD (95% CI) between icodec with app (n = 542) and glargine U300 (n = 542) for change from baseline in HbA_1c at week 52 was − 0.33 (− 0.68 to 0.01) percentage points numerically in favor of icodec with app. This estimate accounted for data from the entire trial effect analysis population (n = 1084), which included the 69 glargine U300 participants from ONWARDS 5 who were also included in the overall treatment contrast.

Testing for augmentation plausibility showed a gradual improvement in ETD, with tightening 95% CI as ρ increased from 0 (randomized comparison only) to 1 (overall weighted augmented comparison, primary analysis) (Fig. S3). A similar trend was observed when testing for φ (belief in the weighted mean trial effect estimation), with tightening 95% CI around a slightly larger treatment effect as φ increased from 0 (only glargine U300 matched comparator group contributing to the glargine U300 trial effect) to 1 (fully weighted trial effect) (Fig. S4). Fully augmented HbA_1c estimates were comparable when tested for different HbA_1c levels and T2D diagnosis emulation criteria (Table S2).

Discussion

ONWARDS 5 evaluated the long-term effectiveness and safety of icodec with app compared with OD analogs over 52 weeks in insulin-naive adults with T2D [1]; however, the glargine U300 stratum was of insufficient size (n = 69) to enable a robust post hoc comparison with the icodec with app treatment group. Indeed, the post hoc randomization comparison (without any RWD augmentation) yielded an uncertain ETD with wide CIs in terms of change in HbA_1c from baseline at 52 weeks, which was not entirely surprising given the low number of participants in the ONWARDS 5 glargine U300 subgroup. Thus, it was hypothesized that augmenting the glargine U300 group with propensity-score-matched RWD may help to strengthen the evaluation of glycemic control between groups. After full augmentation of the glargine U300 subgroup (now consisting of RWD from 459 individuals and RCT data from 69 individuals) and adjustments for inherent trial effects, the precision of the ETD point estimate, although statistically not significant, was increased compared with that derived from the glargine U300 stratum subgroup analysis alone. This finding was consistently shown after several sensitivity analyses, thereby supporting the robustness of our approach and conclusions.

Exchangeability (i.e., the extent to which data sets may resemble and be combined with each other) between the two populations is an important factor to reduce different biases when incorporating RWD and RCT data in an augmented analysis [10–12]. Important exchangeability criteria include matching participant eligibility criteria, baseline characteristics distribution, same treatment approach, similar treatment evaluation and endpoint definition, and similar data collection periods. The exchangeability of the ONWARDS 5 population was particularly well suited for our augmented study because it had several real-world design elements and three treatment randomization strata (degludec, glargine U100, and glargine U300) that assisted when propensity-score–matching participant-level data from the trial with those from the AEMR RWD database.

Another important consideration for our analysis was the need to account for inherent trial effects (i.e., an effect participants may experience simply by partaking in a trial). RCTs are usually performed in strictly controlled environments and are characterized by limited inclusion criteria, frequent monitoring, trial-site visits, and contact with investigators, which all may result in unintended trial effects on participants. Moreover, these trial effects could be related to the treatment, protocol, care management, Hawthorne effect (a change in behavior because of the awareness of being observed), or indeed a study-induced placebo effect [13, 14]. These different components of the trial effect are thought to be present in all trial participants regardless of study arm (investigational treatment, control, or placebo) and may have resulted in improved outcomes when comparing ONWARDS 5 participants and AEMR individuals [13]. In our analysis, a trial effect for ONWARDS 5 in terms of change in HbA_1c from baseline to week 52 was estimated for each control stratum (degludec, glargine U100, and glargine U300), from which a weighted mean was then calculated and applied to the fully augmented treatment comparison to adjust for any potential differences driven by a trial effect. Of note, the weighted mean (95% CI) trial effect of 0.13 (− 0.25 to 0.49) percentage points reflects a modest difference between RWD and RCT data, in contrast to other analyses in T2D that have shown a wider discrepancy in change in HbA_1c between RWD and RCT data [15]. The modest difference in effect between RCT data and RWD observed in our analysis may be explained partly by the real-world design elements of ONWARDS 5.

This post hoc analysis has shown the feasibility of conducting an augmented analysis of an RCT by generating an expanded synthetic control group based on RWD. Although this type of study design is gaining traction [8–12], to our knowledge, this is the first diabetes study to use augmented data combining RCT data and propensity-score-matched RWD. Further exploration of this approach may help to strengthen clinical trial results, to expand small treatment groups, and to bridge the gap between RCT efficacy and real-world effectiveness. Future studies may also benefit from the flexibility of using different databases to include individuals from the same region or country as that of the matching clinical trial participants.

As far as study limitations are concerned, the set of covariates used to match control RWD individuals to ONWARDS 5 participants was limited; therefore, differences in important unmeasured factors between populations may be a potential source of bias. However, any systematic differences would have been captured by the estimated trial effect. For the analysis, it was assumed that the true trial effect was identical across OD analogs strata in ONWARDS 5, which enabled the determination of a common estimate. Exchangeability of the trial effect across strata may have potentially biased the precision of the augmented comparison; however, sensitivity analyses were conducted to address this limitation. Furthermore, the inherent observational nature of RWD may have introduced residual confounding factors subsumed in the trial effect estimation; therefore, sensitivity analyses were conducted to address these limitations. Although the primary analysis adjusted for region by accounting for the regional differences observed in ONWARDS 5, the AEMR database contains exclusively US data, whereas the ONWARDS 5 cohort represented individuals from both North America and Europe; consequently, differences in treatment and health-care systems may have been present. Methodologically, the difference in imputation method was a limiting factor of the analysis because the concept of retrieved dropout used in ONWARDS 5 was difficult to capture for AEMR data. Finally, we were not able to provide augmented evidence on the incidence and rates of hypoglycemia, owing to the nature of the AEMR database. However, rates of clinically significant or severe hypoglycemia were low and similar between icodec with app and OD analogs in the ONWARDS 5 trial [1]. In addition, broadly comparable hypoglycemia rates are expected across the OD analog strata [16] and, therefore, a clinically significant difference in hypoglycemia rates between icodec and glargine U300 is not anticipated; however, this hypothesis warrants further evaluation.

Conclusion

In conclusion, this augmented data analysis showed that the treatment effect of icodec with app compared with the augmented effect of glargine U300 on the change in HbA_1c was numerically in favor of icodec with app, with narrower CIs than for the randomized comparison using RCT data alone. In addition, the methodology and findings help to validate the use of augmented analyses to further strengthen, complement, and contextualize findings from RCTs.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Martin Bøg, Simon Clancy, and Christian Kruse contributed to the study design, study conduct and data collection. All authors contributed to the analysis and interpretation of the study data, the preparation of the manuscript with the support of medical writing services, and approved the version of the manuscript submitted for publication.

Funding

This study and the journal’s Rapid Service Fee were funded by Novo Nordisk A/S, Søborg, Denmark.

Data Availability

US AEMR data were sourced from IQVIA Inc., US (IQVIA Real-World Data Adjudicated Claims database) and are commercially available. Individual participant data from ONWARDS 5 will be shared in datasets in a de-identified/anonymized format. Shared data will include datasets from clinical research sponsored by Novo Nordisk that was completed after 2001 for product indications approved in both the EU and the USA. The study protocol and the redacted clinical study report will be made available according to Novo Nordisk data sharing commitments. These data will be available permanently after research completion and approval of product and product use in both the EU and the USA (no end date). Data will be shared with bona fide researchers who submit a research proposal requesting access to data for use as approved by the Independent Review Board according to its charter (see www.novonordisk-trials.com). These data can be accessed via an access request proposal form; the access criteria can be found at www.novonordisk-trials.com. The data will be made available on a specialized SAS data platform.

Declarations

Conflict of Interest

Liana K. Billings has received research support and consultant fees from, or has served on advisory panels for, Bayer, Dexcom, Eli Lilly, Endogenex, Novo Nordisk, Pfizer, and Sanofi. Ernesto Maddaloni has received research support and consultant fees from, or has served on advisory panels for, Abbott, Eli Lilly, Merck Serono, Merck Sharp & Dohme, MTD, Novo Nordisk, and Pikdare. Marisse Asong, Martin Bøg, Simon Clancy, Christian Kruse, and Elisabeth de Laguiche are employees of Novo Nordisk A/S and hold stock options.

Ethical Approval

The trial was conducted in compliance with the principles of the Declaration of Helsinki and in accordance with International Conference for Harmonization Good Clinical Practice guidelines. Relevant documents, including the protocol and consent forms, were approved by institutional review boards or independent ethics committees. All participants provided written informed consent for participation and publication of data before trial entry and could withdraw their consent at any time.