Comparing Inhaled Medications Reported by Adolescents with Persistent Asthma and Their Caregivers

Abstract

Objectives:

To compare concordance (agreement) between teens with persistent asthma and their caregivers on the reported number of inhaled asthma medications used (rescue and controller); examine concordance specific to controller medications; and determine whether concordance over controller medications within caregiver/teen dyads is associated with demographics or clinical outcomes.

Methods:

We used baseline data from the School-Based Asthma Care for Teens (SB-ACT) trial in urban Rochester, NY. Caregivers and teens (12–16 yrs.) with poorly controlled persistent asthma separately reported the teens’ inhaled therapy, and could name up to two rescue and two controller medications. We compared the total number of medications and number of controller medications reported by each dyad member with Cohen’s Kappa, and assessed whether concordance over the number of controller medications was associated with demographics, symptoms, or healthcare utilization using chi-square and t-tests.

Results:

Of 210 dyads (79% public health insurance, 61% Black teens), 132 (63%) were disconcordant in reporting the overall number of inhaled medications. Teens or caregivers from 173 dyads (82%) reported any controller medication; however, a majority (61%) were discordant in the reporting of controller medications. Compared with concordant dyads, fewer caregivers from dyads with controller medication discordance reported education past high school (35% vs 51%, p=0.04); no other differences in demographics, symptoms, or healthcare utilization were identified based on controller medication concordance.

Conclusions:

Most dyads identified different numbers of inhaled medications, with substantial disagreement over controller medications. Working to ensure a basic understanding of treatment plans may promote successful self-management in persistent childhood asthma.

INTRODUCTION

Adolescence poses a unique challenge for the management of childhood asthma. This developmental period is characterized by a shift in the locus of responsibility for medication use from caregiver to patient.⁽¹⁾ While children as young as five years old may be tasked with independent use of asthma medications,^{(2, 3)} caregivers retain a majority of medication responsibility for children ≤10 years old.⁽¹⁾ Young adolescents have more responsibility than caregivers for using medications after age 13, with full responsibility achieved by the end of the teenage years.⁽¹⁾ Coincident with increasing responsibility is a decrease in teen adherence with daily controller medications, such as inhaled corticosteroids (ICS),⁽⁴⁾ which national asthma guidelines recommend for all patients with persistent asthma.⁽⁵⁾ Routine use of ICS can reduce asthma-related emergency room visits and hospitalizations.^{(6, 7)} Only 50–60% of prescribed controller medication doses are taken by younger children, however, and these rates are halved among older teens.⁽⁴⁾

Regimen complexity and an insufficient knowledge of medications are common barriers to adherence.^(8–10) All adolescents with persistent asthma should have a treatment plan that includes two different classes of medication: a daily controller medication, and a fast-acting rescue medication.⁽⁵⁾ In order to successfully manage a chronic disease, patients first need to be able to identify and understand how/when to use the medications in their treatment plan.⁽¹¹⁾ Prior studies have demonstrated that when adults with hypertension experience difficulty naming blood pressure medications, or when they identify medications only by physical characteristics (i.e., size, shape, color), they are more likely to have poorly controlled blood pressure and be subsequently hospitalized for hypertension.^{(12, 13)} Our team has also demonstrated that the inability of caregivers to accurately name available inhaled medications is associated with lower adherence to preventive therapy and a higher burden of asthma symptoms among younger children.⁽¹⁴⁾ Despite the potential importance of medication identification to routine management, we recently found that adolescents may develop full responsibility for using asthma medications without being able to name them or state correct indications for use; many caregivers also demonstrated gaps in knowledge about their teen’s inhaled medications.⁽¹⁵⁾

Recalling medication names and indications for use from memory can be difficult, however, and many pediatric patients and caregivers identify medications using physical descriptors (i.e., color or shape) rather than name.^(14–16) During this transitional period, it is possible that a more generalized knowledge about the treatment plan (such as the number of inhaled rescue and controller medications being used) may be more relevant to a patient’s daily experience with asthma management than the ability to recall specific medication details.

Although little is known about whether caregivers and teens share a generalized understanding of medications, there is good reason to consider whether such a shared understanding might be protective for teens with asthma. Discordance between caregivers and adolescents when describing responsibility for management tasks has been associated with lower adherence for both HIV⁽¹⁷⁾ and asthma^{(18, 19)} medications, and worse outcomes in asthma.^{(18, 19)} Conversely, adolescents with diabetes who agree with their caregivers about the division of management responsibilities have lower HbA1c levels.⁽²⁰⁾ As medication knowledge is necessary for chronic disease management,⁽¹¹⁾ it is unclear whether dyadic agreement over basic details of a treatment regimen such as the number of inhaled asthma medications, and the number of inhaled controller medications in particular, would be similarly associated with clinical outcomes.

The goals of this study were to compare the total number of inhaled therapies reported by teens with persistent asthma and their caregivers for concordance (agreement); examine concordance within dyads specific to controller medications; and determine whether dyadic concordance over controller medications is associated with patient demographics or clinical outcomes such as symptom burden or acute healthcare utilization. We hypothesized that caregiver/teen concordance on the number of inhaled controller medications being used for asthma management would signal a shared understanding of the treatment regimen, and would therefore be associated with fewer asthma symptoms and less acute healthcare use.

METHODS

Settings and Participants

The School-Based Asthma Care for Teens (SB-ACT) study is an ongoing, five-year long randomized controlled trial in the Rochester, NY metropolitan area. Young adolescents with physician-diagnosed asthma between the ages of 12–16 years old were eligible for enrollment into SB-ACT if: (a) they attended school at a partner site; (b) they could speak English; and (c) a caregiver reported symptoms over the previous month consistent with persistent or poorly- controlled asthma as defined by national management guidelines.⁽⁵⁾ Teens were excluded if they carried additional medical diagnoses such as congenital heart disease or cystic fibrosis which could impact asthma assessment. Only one child per family was enrolled, even if multiple children in the family satisfied inclusion criteria. This analysis includes data from baseline surveys completed separately by caregivers and teens throughout the first 3 years of the study (2014–2016).

We obtained written informed consent from all caregivers and assent from all children in accordance with institutional research guidelines. We compensated family participation time with a $50 gift card. The University of Rochester institutional review board approved the study protocol.

Measures

In the baseline survey, we asked caregivers and teens to separately name all of the inhaled asthma medications used by the teen. A research assistant asked respondents if they were taking any rescue medications or daily controller medications, and read aloud examples of medication names from both classes of medication. Up to two rescue and two inhaled controller medications could be identified by either generic or trade names. Respondents also reported non- inhaled asthma medications, such as leukotriene receptor antagonists (LTRA) and adjuvant therapies for allergic rhinitis; answers to these questions are not included in the present analysis due to our intended focus on inhaled therapy. We additionally asked caregivers to report demographic information (i.e., caregiver age, gender, level of education completed, marital status; child age, gender, race, ethnicity, insurance coverage status); asthma symptoms over the previous month; and asthma-related healthcare utilization (i.e. emergency room visits, hospitalizations, treatment with systemic corticosteroids) over the previous year using structured surveys.

Analysis

We used descriptive analysis to compare the total number of inhaled medications reported by caregivers and teens for concordance or discordance within caregiver/teen dyads (range: 0–4 medications). Both generic and trade names were accepted; a response of “albuterol” or “ProAir™” would have been equally counted as 1 rescue medication. We calculated a Cohen’s Kappa to compare observed concordance with the level of agreement expected with random response. Within discordant dyads, we examined how often each class of medication contributed to the observed discordance.

For dyads in which at least one member reported an inhaled controller medication, we used descriptive analysis to compare the number of reported controller medications and calculated a separate Cohen’s Kappa for controller medication concordance. These analyses were informed by an a priori assumption that discordance in controller medication reporting may be more clinically meaningful than discordance limited to rescue medications. The standard rescue medication for asthma (i.e., a short acting beta agonist) is intended for relief of acute symptoms; whether administered via nebulizer or metered-dose inhaler with a spacer, patients should expect comparable effectiveness with correct technique. Differential recall about the number of rescue medications by caregivers and teens may not be clinically important provided that at least one rescue medication is available to the teen when needed. Discordant reporting of controller medications between dyad members, however, may reflect regimen confusion, poor adherence, or potential asthma management dysfunction within dyads.

Finally, we examined whether dyadic concordance over controller medications was associated with patient characteristics, recent asthma symptoms, or acute healthcare utilization (i.e., asthma-related hospitalizations, emergency department visits, or courses of treatment with oral corticosteroids). We used Pearson chi-square and t-tests to compare categorical variables (i.e., caregiver gender, education, and marital status; child gender, race, ethnicity, insurance coverage) and continuous variables (i.e., child and caregiver age), respectively. A 2-sided α <0.05 was considered statistically significant. All analyses were performed using STATA (version 12.0).

RESULTS

Participants

We enrolled 222 caregiver/teen dyads in the SB-ACT trial between November 2014 and December 2016 (enrollment rate = 81% of eligible participants); complete answers to medication questions were available for 210 dyads (95%). The average teen included in our analysis was 13.5 years old (SD 1.0), male (56%), Black (61%), non-Hispanic (69%), and covered by public health insurance (79%). Caregivers were 40.6 years old on average (SD 7.7), and more likely to be female (90%) and single (76%). Educational attainment beyond high school was reported by 42% of caregivers.

Concordance within dyads: total number of inhaled medications (rescue and controller)

Across the full group of respondents, the total number of inhaled medications reported by caregivers and teens shared a similar distribution. Despite this apparent overlap, a majority of caregiver/teen dyads (63%) were discordant on the overall number of inhaled medications reported (Table 1). Half of the discordant dyads included one member reporting 1–2 medications while the other member reported 3–4. Most discordant dyads (67%) disagreed about the number of controller medications in use. Among concordant dyads reporting the same total number of inhaled medications, 24% of caregiver/teen pairs reported different numbers of rescue and controller medications. For example, the caregiver and teen may have both reported 2 medications, but the caregiver reported 2 rescue and no controller medications while the teen reported no rescue and 2 controller medications. The Kappa score for overall concordance was 0.066 (P = 0.13), indicative of poor agreement that might be expected if caregivers and teens were answering randomly.⁽²¹⁾

Table 1.

A comparison of the total number of inhaled medications (rescue and controller) reported by caregivers and teens within dyads

		Caregiver Report					Total
		Zero	One	Two	Three	Four
Teen Report	Zero	0	1	3	2	0	6
	One	1	18	27	9	4	59
	Two	3	22	49	18	3	95
	Three	1	15	13	11	2	42
	Four	0	2	4	2	0	8
Total		5	58	96	42	9	210
Discordant (Total)		132 (63%)
Concordant (Total)		78 (37%)

Concordance within dyads: number of inhaled controller medications

The caregiver, teen, or both members from among 173 dyads (82% of enrolled) reported teen use of at least one inhaled controller medication. Most of these dyads (61%) were discordant with respect to the number of controller medications reported (Table 2). Almost all discordance in controller medication reporting (92%) resulted from one dyad member not reporting any controller medication. The Kappa score for concordance in controller medications was −0.231 (P < 0.001), indicating a level of slight disagreement that is significantly worse than would be expected from random guessing.⁽²¹⁾

Table 2.

A comparison of the number of inhaled controller medications reported from dyads reporting at least 1 controller medication

			Caregiver Report			Total
			Zero	One	Two
Teen Report	Zero			48	5	53
	One		38	65	3	106
	Two		7	5	2	14
Total			45	118	10	173
Discordant (Controller)		106 (61%)		Both respondents identified ≥ 1 controller medication(s)			8 (8%)
Concordant (Controller)		67 (39%)		One respondent did not identify any controller medication			98 (92%)

Controller medication concordance: demographics, recent symptoms, and healthcare use

Compared with discordant dyads, caregivers from dyads with concordance over controller medications were significantly more likely to have completed any education past high school (51% vs. 35%, p = 0.04) (Table 3). Concordance was more likely among caregivers who were female (96% vs. 87%, p = 0.06) and teens who used more rescue medication (48% vs. 34%, p = 0.07), however these differences did not reach statistical significance. No other differences in demographics, symptoms over the past month, or acute healthcare utilization over the past year were identified between dyads based on controller medication concordance.

Table 3.

Baseline demographics, symptom report, and healthcare utilization within dyads reporting at least 1 inhaled controller medication

	Concordant	Discordant	P- value
	N =67	N= 106
Demographics
Caregiver
Age, years (mean, SD)	40.9 (7.1)	40.7 (8.6)	0.89
Gender: female	96%	87%	0.06
Education: >high school	51%	35%	0.04*
Marital status: married / domestic partner	27%	17%	0.12
Teen
Age, years (mean, SD)	13.6 (1.1)	13.5 (1.0)	0.50
Gender: male	51%	55%	0.61
Race: Black	72%	58%	0.17
Ethnicity: Hispanic	27%	31%	0.55
Insurance plan: public	79%	80%	0.90
Symptoms (past month)
Daytime symptoms: ≥2/week	58%	51%	0.35
Nighttime symptoms: >2/month	50%	41%	0.97
Rescue medication use: >2/week	48%	34%	0.07
Health Care Utilization (past year)
Oral corticosteroids courses: ≥2	18%	16%	0.75
ED visits: ≥1	39%	41%	0.82
Hospitalizations: ≥1	10%	5%	0.15

^*P < 0.05 considered statistically significant

DISCUSSION

In this cross-sectional analysis of urban teens with poorly controlled/symptomatic persistent asthma, we found that most caregivers and teens within dyads reported a different overall number of inhaled medications. Discordance specifically in reporting the number of guideline-recommended controller medications was also highly prevalent, and almost all dyads that were discordant over controller medications included one respondent who did not name any controller medications. According to NHLBI asthma management guidelines,⁽⁵⁾ all adolescent subjects that were enrolled in the SB-ACT trial should be using a daily controller medication. Prior studies have demonstrated that many pediatric patients with persistent asthma are not prescribed an appropriate level of controller therapy;⁽²²⁾ our findings suggest that even when controller therapy is reportedly being used, there is little consensus between caregiver and teen reports.

This lack of shared understanding over the basic details of a patient’s medication regimen is consistent with previous work indicating a lack of shared asthma knowledge between caregivers and children.⁽²³⁾ These findings also provide context for our companion analysis examining medication identification and teen responsibility, in which we found that the caregiver-to-teen transition of responsibility for using medications was not associated with the teen’s knowledge of the names of medications and the indications for their use.⁽¹⁵⁾ The method we used to elicit medication knowledge in the current analysis is distinct from the teen responsibility study. In the prior work, respondents were asked to name their asthma medications entirely from memory. The medication questions used in this analysis were posed later in the baseline survey, were structured to specifically identify rescue and controller medications, and included examples of medication names in each class read aloud by the research assistant in order to facilitate easy recall. It is clear that medication recall and understanding is an underreported challenge for this population.

Dyadic agreement about the number of controller medications within this population was found to be associated with the level of caregiver education, with greater concordance observed when caregivers reported having more than a high school level of education. Dyads with controller medication concordance were not significantly different from discordant dyads on the basis of demographic characteristics, recent asthma symptoms, or acute healthcare utilization. On further consideration of these results and our original hypothesis, it is perhaps understandable why concordant/discordant dyads appear similar with respect to these variables. Our analyses use baseline data from a larger randomized controlled trial conducted within an urban school district; all caregivers reported symptoms consistent with poorly controlled or symptomatic persistent asthma at enrollment, and the lack of observed differences may reflect clearly defined enrollment criteria contributing to sample homogeneity around asthma morbidity.

Moreover, while patient knowledge of asthma⁽²³⁾ and asthma medications⁽¹⁰⁾ has been associated with adherence, agreement on the number of medications in a regimen does not imply that concordant dyads are more likely to either have or adhere with controller therapy. This is of particular importance in an adolescent population, for whom adherence is known to fall with increasing age.⁽⁴⁾ We were unable to compare caregiver/teen reports of adherence in this study, because adherence questions in the baseline survey were triggered by respondent identification of at least one controller medication. Most dyads with discordant reporting of controller medications had one member who did not identify any controller medications. Although adherence was not examined in this study, discordance in describing regimens may be a marker of dysfunctional dyadic management, a known risk factor for poor adherence and disease control.^{(18, 19)} Additional work will be needed to determine if concordance is associated with increasing patient age or adherence. Concordant reporting on the number of controller medications also does not ensure that caregivers and teens are reporting the same medications. Future studies should consider whether a family history of asthma, medication sharing between family members, or recent changes in asthma prescriptions influences either concordant reporting or clinical outcomes.

It should be noted that the dyads with controller medication discordance (Table 2) may comprise two subgroups operating under different dynamics. Teens did not report any controller medications in half of the discordant dyads, while caregivers did not report any controller medications in 42% of discordant dyads. Future work should examine discordant dyads further to better understand the basis for these discrepancies and to determine whether clinical outcomes differ depending on who reports the controller medication.

There are several limitations to consider. We do not compare reported medications with medical record or pharmacy-based prescription data. We did not have access to such data, but more importantly it was not our intent to determine whether the caregiver or teen was more accurate in reporting inhaled asthma medications. Rather, we wanted to explore whether patients and caregivers share an understanding about the basic details of asthma regimens based on lived experience. The discordance in medication reporting observed between our population of inner-city adolescents with uncontrolled asthma and their caregivers may not be generalizable to other populations managing childhood asthma. This study is part of a larger ongoing trial which was powered to identify longitudinal differences in symptom-free days between randomized groups. We were limited to the sample size available from the trial at the time of analysis, and thus our study may not be powered to identify true differences in symptoms or healthcare use between caregiver/teen dyads on the basis of controller medication concordance. We did not include non-inhaled controller therapy (i.e., LTRA tablets) in our assessment of concordance; however, this analytic decision likely created a conservative bias within our findings. Finally, social desirability bias may be informing our results: when respondents are prompted to report controller medications, they may feel compelled to provide an answer even if the teen is not using any routine medications.

It is concerning that caregiver/teen dyads disagree on asthma regimen information as fundamental as the number of inhaled medications being used by the teen. It is particularly worrisome that such disagreement is being observed during a critical period in which many adolescents develop increasing responsibility for medication use. Although the clinical impact of our findings is not clear, medication knowledge is a necessary (if not sufficient) requirement for home asthma management. Providers can start to address dyadic asthma management and regimen confusion by ensuring a shared understanding of treatment plans among both patients and caregivers.

CONCLUSIONS

Teens with uncontrolled persistent asthma and their caregivers often disagree about the number of inhaled medications used by the teen. Dyadic agreement about the number of controller medications within this population is not associated with clinical outcomes such as recent asthma symptoms or acute healthcare use.