Adherence to multiple medications in the TODAY (Treatment Options for type 2 Diabetes in Adolescents and Youth) cohort: effect of additional medications on adherence to primary diabetes medication

Abstract

Background:

Non-adherence to diabetes medication leads to poor outcomes and increased healthcare costs. Multiple factors affecting adherence in adults with type 2 diabetes (T2D) have been identified, but pediatric data is sparse. We aimed to determine whether initiation of additional oral medications or insulin affects adherence to primary study medication (PSM) in the Treatment Options for type 2 Diabetes in Adolescents and Youth (TODAY) study.

Methods:

Six hundred and ninety-nine youth (aged 10–17 years) with recent-onset T2D were randomized in the TODAY study. Participants were categorized as adherent (≥80% taken by pill count) or non-adherent (<80%), and adherence was compared between those on additional medications or not. Subgroup analyses to assess influence of race/ethnicity, gender, medication type, or depression were performed.

Results:

At 36 months, 46.3% of participants were taking additional oral medications and 31.9% were on insulin. There was no difference in study medication adherence with additional oral medications (55.1%, 67.1%, and 56.7% at month 36 in those prescribed 0, 1, or 2+ additional medications; p=0.16). Girls on oral contraceptives (OC) had higher adherence (65.2% vs. 55.8% at month 36; p=0.0054). Participants on insulin had lower adherence (39.7% vs. 59.3% at 36 months; p<0.0001). There was decreased adherence in participants with baseline depression (p=0.008).

Conclusions:

Additional oral medications did not influence adherence to diabetes medications in TODAY. Addition of insulin led to reduced adherence. In subgroup analyses, OC use was associated with higher adherence in girls, while baseline depression was associated with lower adherence overall. Further studies examining potentially modifiable risk factors of adherence in pediatric T2D are needed.

Introduction

Non-adherence to prescribed anti-diabetic medications constitutes a major barrier to effective care of type 2 diabetes (T2D), and is associated with multiple factors including demographics, comorbidities, and lower socio-economic status [1]. Lack of adherence is associated with poor glycemic control, increased emergency room visits and hospitalizations, and increased medical cost [2]. Studies examining potentially modifiable factors that influence adherence in adults with T2D suggest that an increased number of prescribed medications may lead to decreased overall adherence [3–5]; no studies have been reported in pediatric T2D. These findings in adults may, nonetheless, lead to hesitation amongst healthcare providers in prescribing additional medications to adolescents.

Our group previously reported that adherence to study medication in the TODAY (Treatment Options for type 2 Diabetes in Adolescents and Youth) study decreased over time and was associated with baseline depression but not with gender, age, family income, parental education, or treatment group [6]. Exposure to stressful life events over the course of the study was also associated with non-adherence [7]. Participants with higher adherence had improved insulin sensitivity and oral disposition index [6]. Adherence was influenced by both family-supported and self-directed adherence strategies [8].

Here, we aim to determine whether additional oral medication or insulin usage adversely affects adherence to primary study medication (PSM) in pediatric participants from the TODAY cohort.

Participants and methods

TODAY participants and study design

Design and results of the TODAY study have been previously reported [9, 10]. The basic study design is depicted in Figure 1. Briefly, 699 participants aged 10–17 years, 64.7% female, 32.5% non-Hispanic Black, 39.8% Hispanic, 20.3% non-Hispanic White, and 7.4% of other race, with recent (<2 years) diagnosis of T2D by the American Diabetes Association (ADA) 2005 criteria, overweight or obese (body mass index [BMI] >85th and >95th percentile, respectively), and with islet cell antibody negative and C-peptide positive were enrolled in the study.

Figure 1:

The TODAY Study design.

After screening, eligible participants entered a 2- to 6-month run-in period to wean from non-study diabetes medications, tolerate metformin up to a dose of 1000 mg twice daily but no less than 1000 mg/day, attain hemoglobin A_1c (HbA_1c) <8.0% (<64 mmol/mol) for at least 2 months on metformin alone, and demonstrate adherence to study medications and visit attendance [9]. To be eligible, participants had to take ≥80% of study drug (metformin) during an 8-week run-in period.

Following the run-in phase, participants were randomized to one of three treatment groups (metformin, metformin+rosiglitazone, or metformin+intensive lifestyle program) and followed an average of 3.9±1.5 (2–6.5) years. Study medication arms were masked to investigators, study personnel, and participants. Half the cohort reached the primary endpoint (HbA_1c >8% [>64 mmol/mol] for 6 months), and the results demonstrated that adding rosiglitazone to metformin was associated with more durable glycemic control after an average follow-up of 3.9 years (range 2–6.5 years) [10]. Insulin was initiated at the time of primary outcome.

The protocol was approved by the Institutional Review Board for each of the participating institutions. All participants provided informed consent and minor child assent.

Categorization of adherence in the TODAY study

All participants were instructed to take four pills daily (study drug or placebo) provided as weekly blister packs, and pill counts were performed at each study visit (every 2 months in year 1 and then quarterly). Adherence was calculated as the percentage of pills missing from blister packs. If packs were not brought to the visit, adherence was noted as missing. Participants were categorized as “adherent” if ≥80% of pills were taken and “non-adherent” if <80% of pills were taken.

Definition and classification of additional oral medications

Participants reported the use of additional oral medications (prescribed by any healthcare provider) at each visit. Only data from oral medications prescribed on a minimum of two consecutive visits and on a regular schedule were included in this analysis (i.e. additional sustained medications or ASMs); inhalers, topical medications, and vitamins were ignored. Insulin was prescribed as part of the TODAY study design after loss of glycemic control (HbA_1c ≥8% [≥64 mmol/mol] for 6 months) on PSM.

Classification for subgroup analyses

Gender and race/ethnicity were determined by self-report at the baseline visit. Depressive symptoms were assessed using either the Children’s Depression Inventory (CDI) for participants <16 years or the Beck Depression Inventory II (BDI-II) for those ≥16 years. Total scores were calculated for each instrument; a cut-off score ≥13 on the CDI and ≥14 on the BDI-II indicated clinically significant levels of depressive symptoms [11, 12]. Depressive symptoms could be present or absent at each assessment.

Statistical methods

In order to account for the correlated nature of the data collected, generalized estimating equations (GEE) were used to compare percent medication adherence over time during the first 3 years of the study. Frequency distributions are shown in Tables 1–4 for select time points during the study (baseline, month 12, 24, and 36). Prescribed ASM information was available at each time point. Study medication adherence was available at each time point after baseline (i.e. after randomization of participants to one of the three intervention arms). Adherence to study medication was classified as non-adherent <80% vs. adherent ≤80% in the analyses. Similar GEE models were used to compare medication adherence across subgroups by adding an interaction term between the ASM and various subgroups including sex, race/ethnicity, and presence of depression. Pre-post analyses using generalized linear mixed models were used to compare differences in the proportion of participants adherent to medications (≥80% of pills taken) over time relative to time before and time after the start of the ASM, in models adjusted for time in the study. A random intercept model with a linear time trend was used. Adherence was analyzed for up to 3 years before and after the start date (first visit in which subject reported use) of additional medications. Apart from the testing specifically exploring the effect of insulin use on adherence, all other analyses were conducted using records before insulin use (i.e. before loss of glycemic control). All analyses were considered exploratory, with statistical significance defined as p<0.05. The SAS 9.4 software (SAS Institute Inc., Cary, NC, USA) was used for all analyses.

Table 1:

Percent of TODAY participants prescribed additional sustained medications (ASMs), cumulative ASM count, and insulin at select times during the study^a.

	Baseline (n = 699)	Month 12 (n = 595)	Month 24 (n = 541)	Month 36 (n = 376)
Cumulative pill countb
0	89.7%	71.4%	59.5%	53.7%
1	8.2%	22.2%	29.6%	32.2%
2+	2.1%	6.4%	10.9%	14.1%
Types of pills prescribedb
ACE or other hypertensive	3.9%	13.3%	21.8%	27.1%
Statin or other lipid lowering	0.9%	4.4%	6.3%	7.7%
Other diabetes	0.0%	0.3%	2.4%	2.1%
Psychiatric	2.2%	4.9%	5.7%	5.3%
Oral contraceptive (% of girls)	3.8%	8.3%	12.0%	15.8%
Other	1.7%	4.0%	4.4%	5.1%
Insulin prescribedc
Permanent insulin	0.0%	9.6%	23.7%	31.9%

^aOnly data from oral medications prescribed on a minimum of two consecutive visits (or three for antibiotics) are included as ASMs in this table.

^bCumulative ASM count is defined as the total number of ASMs prescribed to a given participant and include all types of medications listed above (with the exception of insulin).

^cTODAY participants were placed on insulin therapy upon loss of glycemic control. ACE, angiotensin-converting enzyme; TODAY, Treatment Options for type 2 Diabetes in Adolescents and Youth.

Table 4:

Percent of TODAY study participants that are never on insulin, pre insulin, or on insulin by study medication adherence status (using 80% cut-off) at select points in the study.

	Month 12 (n = 595)			Month 24 (n = 541)			Month 36 (n = 376)			p-Valuea
	Never insulin	Pre insulin	On insulin	Never insulin	Pre insulin	On insulin	Never insulin	Pre insulin	On insulin
Non-adherent <80%	31.0%	35.2%	52.9%	39.3%	37.0%	52.5%	40.7%	46.7%	60.3%
Adherent ≥80%	69.0%	64.8%	47.1%	60.7%	63.0%	47.5%	59.3%	53.3%	39.7%	p<0.0001

^ap-Value from unadjusted generalized estimated equation models examining differences in medication adherence by use of insulin over time. TODAY, Treatment Options for type 2 Diabetes in Adolescents and Youth.

Results

Use of ASM and insulin increased over time in the TODAY cohort

Table 1 describes use of additional medications throughout the study. The number of ASMs taken ranged from 0 to 5. At baseline, 8.2% of the 699 participants were prescribed one ASM and 2.1% were prescribed two or more (2+) ASMs. These percentages increased gradually over the study; at month 36, 32.3% of the participants were prescribed one ASM and 14.1% were prescribed 2+ ASMs. The most commonly prescribed ASMs at month 36 were: angiotensin-converting enzyme (ACE) inhibitors and other anti-hypertensives (27.1%), statins and other lipid- lowering medications (7.7%), non-study anti-hyperglycemic agents (2.1%), psychiatric drugs (5.3%; of which 93% were anti-depressants), and oral contraceptives (OC) in girls (15.8% of girls). No participants were on insulin at baseline per study design. At month 36, 31.7% of participants had lost glycemic control on PSM and were prescribed daily insulin.

Addition of ASM did not impact adherence in TODAY

As published, overall adherence in the TODAY study (using an 80% cut-off) declined as a function of time in the study (p<0.0001) but was not different between treatment groups or related to glycemic control [5]. Before insulin therapy, participants had a mean±standard deviation (SD) percent medication adherence of 81.1±24.8% at month 12, which decreased to 73.6±30.3% at month 36. Stated categorically, 67.1% of the participants had adherence ≥80% at month 12 as opposed to 59.0% at month 36.

When the cohort was divided into those prescribed 0, 1, or 2+ ASMs, there was no difference in adherence at any time point: 65.9%, 67.0%, and 81.8% at month 12; 65.5%, 57.1%, and 57.1% at month 24; and 55.1%, 67.1%, and 56.7% at month 36 in each group, respectively (p=0.16; Table 2, Figure 2A). The difference remained non-significant when ASMs were collapsed into 0 vs. 1+ medications (p=0.055).

Table 2:

Adherence of TODAY participants to primary study medication at select time points by additional sustained medication (ASM) count at select time points.

	Month 12 (n = 595)			Month 24 (n = 541)			Month 36 (n = 376)			p-Valuea
	0 med	1 med	2+ med	0 med	1 med	2+ med	0 med	1 med	2+ med
Non-adherent <80%	34.1%	33.0%	18.2%	34.5%	42.9%	42.9%	44.9%	32.9%	43.3%
Adherent ≥80%	65.9%	67.0%	81.8%	65.5%	57.1%	57.1%	55.1%	67.1%	56.7%	p = 0.16

^ap-Value from unadjusted generalized estimated equation models examining differences in medication adherence by ASM count over time. TODAY, Treatment Options for type 2 Diabetes in Adolescents and Youth.

Figure 2:

Effect of additional study medication (ASM) on adherence to primary diabetes medication in the TODAY study*.

*ASM, additional sustained medication; OC, oral contraceptive use in girls. (A) Use of ASM did not reduce adherence to study drug (p=0.16); (B) Girls on OC had overall higher adherence rates than those not on OC (p=0.005); (C) While baseline depression was associated with reduced adherence (p=0.008), addition of ASM in depressed patients did not have an impact (p=0.56); (D) Participants prescribed insulin had lower adherence rates compared to the pre-insulin and never-on-insulin groups (p<0.0001). Number of participants and percent medication adherence (≥80%) by subgroup at months 12, 24, and 36 are shown under each figure.

Girls on OC have improved adherence throughout the study

There was no relationship between any of the commonly prescribed ASMs and adherence except that use of OC in girls was consistently associated with increased adherence at all time points: 78.6% vs. 63.4% at month 12, 70.6% vs. 59.3% at month 24, and 65.2% vs. 55.8% at month 36 (p=0.0054; Table 3, Figure 2B). However, subgroup analysis of the 64 girls who were prescribed OC during the study showed no difference in adherence in the 36-month period before and after initiation of OC (p=0.62; data not shown).

Table 3:

Percentage of female participants prescribed oral contraceptive (OC) or not by primary study medication adherence status (using 80% cut-off) at select points in the study.

	Month 12 (n = 345)		Month 24 (n=270)		Month 36 (n = 161)		p-Valuea
	Not on OC	On OC	Not on OC	On OC	Not on OC	On OC
Non-adherent <80%	36.6%	21.4%	40.7%	29.4%	44.2%	34.8%
Adherent ≥80%	63.4%	78.6%	59.3%	70.6%	55.8%	65.2%	p = 0.0054

^ap-Value from unadjusted generalized estimated equation models examining differences in medication adherence by use of OC in TODAY female participants over time. TODAY, Treatment Options for type 2 Diabetes in Adolescents and Youth.

Baseline depression is adversely correlated to adherence, and psychiatric medication stabilized adherence decline

Subgroup analysis was then performed to determine how gender, race/ethnicity, or depression effected adherence in those on ASM. There was no effect of gender or race/ ethnicity. Baseline depression adversely correlated with adherence over time (p=0.008) as previously reported [5], but in participants with depression, use of ASMs did not impact adherence to PSM (Figure 2C). We were not able to determine specifically whether treatment of depression in those with symptoms altered adherence, but the decline in PSM adherence seen over the 36 months before initiation of psychiatric medication in the 43 participants prescribed sustained psychiatric medications stabilized for the 36 months following treatment initiation (p=0.02).

Insulin use leads to decreased adherence

Participants in the TODAY study were prescribed insulin when they failed glycemic control on PSM [6–7]. To determine whether addition of insulin influenced adherence to PSM, adherence was compared among three categories: all time points for participants never on insulin, pre-insulin time points for those eventually prescribed insulin, and on-insulin time points. There was a significantly lower rate of adherence in the on-insulin time points compared with the pre-insulin and never-on insulin: 47.1% vs. 64.8% and 69.0% at month 12; 47.5% vs. 63.0% and 60.7% at month 24; and 39.7% vs. 53.3% and 59.3% at month 36, respectively (p<0.0001; Table 4, Figure 2D). Subgroup analysis showed no effect of race/ethnicity (p=0.22) or baseline depression (p=0.16). However, there was a significant interaction between the insulin groups (collapsed into not-on-insulin vs. on-insulin) and gender (p=0.004). Adherence continued to decline after insulin initiation in girls, whereas adherence rate stabilized in boys. Additionally, for those on insulin, use of ASMs was also associated with decreased adherence to PSM (p=0.03).

Discussion

This is the first study to examine whether prescription of additional oral medication and/or insulin alters adherence to PSM in adolescents and youth with type 2 diabetes mellitus (T2DM). We found that adding additional oral medications did not reduce adherence rates in our participants. Studies of the effect of multiple prescribed medications to adherence in adults with T2DM have shown variable results, with some reporting no difference in adherence [13–15] and others showing increased medication prescriptions as a risk factor for reduced adherence [1, 3–5, 14]. These disparate results may be explained by differences in study populations and research methods employed, and highlight the complexities involved in predicting adherence to therapeutic regimens.

Looking deeper into various phases of medication adherence, a large Danish study found that polypharmacy was a risk factor for slower initiation of diabetes medication, but did not effect ongoing adherence and in fact decreased risk of discontinuation [16]; as our participants had already initiated diabetes medication prior to ASM, we could not address these phases of adherence, but our results corroborate that ongoing medication adherence was unaffected. This suggests that adherence to diabetes medication may be influenced by unique factors not applicable to other classes of therapeutics.

We found that adherence to diabetes medication was higher in girls on OC; the mechanisms for this cannot be elicited from our study design but may be related to inherent engagement with medical therapy in those that sought such preventive care or effective counseling received by these participants about the importance of daily medication. There are no other studies addressing the effects of OC on adherence to other medications for any condition.

The addition of insulin injections, which were initiated per study protocol after 6 months of HbA_1c >8% on PSM, did lead to decreased PSM adherence. Notably, while adherence initially declined in all participants, continued decline over time was seen only in girls. We could not investigate whether this was due to the complicated therapeutic regimen or underlying participant characteristics in those that failed oral medication. While there is substantial literature on adherence to insulin in patients with diabetes, we found no published studies examining adherence to oral diabetes medications after initiation of insulin. One report of a large Polish and German cohort did find improved adherence to insulin in males, and in those prescribed additional oral medications [17], while another German study similarly found that additional oral medication use was associated with higher adherence to ongoing use of insulin [18].

As previously published, in the TODAY study cohort, 14.8% of 689 screened participants scored above the threshold for clinically significant depressive symptoms at baseline. We found that adherence to PSM was lower in those participants screening positive for depression, as has been shown consistently in adult populations [1, 19–23]. Prevalence of depression in adults with T2D has been reported in multiple large population-based studies as nearly double that of the general population in adults with T2D [24], though data in childhood-onset T2D is limited. Two small retrospective studies of pediatric T2D cohorts found high rates (19.4% and 26.2%) of neuropsychiatric disease, including depression [25, 26]. In adolescents and young adults with type 1 diabetes, several studies have shown an association between depressive symptoms and poor adherence and glucose control [27–29]. Future studies should examine whether treatment of depression improves diabetes medication adherence in the pediatric age group.

Studies of medication adherence, including ours, are limited by the inability to accurately determine whether participants truly ingested the study medication. Our methodology of using pill counts is validated and widely used due to its low cost and simplicity; other methods such as biochemical monitoring, pharmacy records, or observer report may improve accuracy [30]. While our cohort is demographically similar to the general population of youth with T2DM, there were aspects of the controlled clinical trial (i.e. no cost for visits or medication, self-selection of participants) that are not generalizable and unlikely reproducible in a real-world setting.

Conclusions

Prescription of ASM did not influence adherence to PSM in the TODAY cohort of adolescents and youth with T2D. In girls, OC use was associated with higher rates of adherence. Baseline symptoms of depression were associated with lower adherence overall. Addition of insulin led to reduced PSM adherence, particularly in girls. These results suggest that pharmacologic treatment of comorbid conditions should not be withheld in youth with T2D due to concerns of impacting adherence to the diabetes regimen. Further studies are needed to determine whether treatment of depression improves adherence in youth with diabetes.

Novelty statement.

• Our study shows that in youth with type 2 diabetes (T2D), additional oral medications do not reduce adherence to primary diabetes medication.

• Pharmacologic treatment of comorbid conditions should not be withheld in youth with T2D due to concerns of impacting adherence to the diabetes regimen.