Mismatching among guidelines, providers, and parents on controller medication use in children with asthma

Abstract

Background

Underuse of controller medicines among children with asthma remains widespread despite national guidelines.

Objectives

To: (1) assess provider prescribing patterns for asthma controller medications; (2) assess how frequently parents’ reports of their child’s asthma controller medicine use were mismatched with their provider’s recommendations; and (3) evaluate parent attitudes and demographic characteristics associated with these mismatches.

Methods

In this cross-sectional study, we conducted linked surveys of parents and providers of children with probable persistent asthma in a Medicaid program and four commercial health plans in 2011. Probable persistent asthma was defined as a diagnosis of asthma and ≥1 controller medication dispensing.

Results

This study included 740 children (mean age 8.6 years). Providers for 50% of the children reported prescribing controller medications for daily year-round use, 41% for daily use during active asthma months, and 9% for intermittent use for relief. Among parents, 72% knew which class of controller medication the provider prescribed and 49% knew the administration frequency and the medication class. Parents were less likely to report the same controller medication type as the provider, irrespective of dose and frequency, if they were Latino [OR 0.23; CI 0.057–0.90], had a household smoker [OR 2.87; CI 0.42–19.6], or believed the controller medicine was not helping [OR 0.15; CI 0.048–0.45].

Conclusions

Mismatches between parent reports and providers’ intentions regarding how the child was supposed to use inhaled steroids occurred for half of children. Efforts should focus on ways to reduce mismatches between parent and provider intentions regarding controller medication use.

Introduction

Controller medicines for asthma such as inhaled corticosteroids (ICS), leukotriene antagonist (LTRA), or combination ICS-long acting beta-agonist (LABA) improve lung function, decrease the number of asthma-related hospitalizations, reduce emergency department (ED) visits, and limit the use of oral corticosteroids.^1–4 However, despite national guidelines recommending that all children with persistent asthma use an ICS, LTRA, or ICS/LABA,⁵ morbidity from asthma remains high. Underuse of controller medicines for persistent asthma is common,⁶ and contributes to severe exacerbations, including hospitalizations and ED visits.^7,8–12

Underuse of controller medicines may be due to provider under-diagnosis or under-treatment of asthma, parent non-adherence, or misunderstandings by parents of what medicine should be given and how frequently.¹³ Designing effective interventions to enhance controller medicine use will require more precise understanding of the relative influences of these possible barriers than is available from current studies. The objectives of this study were: (1) to assess provider prescribing patterns of asthma controller medications; (2) to assess how frequently parents’ understanding of their children’s controller medicine regimen matches their providers’; and (3) to assess parental characteristics associated with a mismatch between parent and provider understanding of the child’s medicine regimen.

Methods

Design and Study Setting

This was a cross-sectional, linked survey of parents and providers of children with asthma within The Population-Based Effectiveness in Asthma and Lung Diseases (PEAL) Network, a distributed data network that was created for research on lung diseases and built on currently available computerized datasets from the TennCare Medicaid population at Vanderbilt Medical Center and four Health Maintenance Organization (HMO) sites: Harvard Pilgrim Health Care (HPHC), Health Partners Research Foundation (HPRF), Kaiser Permanente of Georgia (KPGA) and Kaiser Permanente of Northern California (KPNC). The study was approved by the Institutional Review Board (IRB) of each site. This study merges parent-reported data, provider-reported data, and administrative claims data.

Data Collection

Computerized medical records or claims data from the four HMOs were used to identify children with asthma and their providers. More specifically, we used computerized data to identify asthmatic children, then located relevant providers either as identified by parents or indicated in the computerized data.

The target population included parents and providers of members ages 4–11 years on January 1, 2011, with ≥1 diagnosis of asthma (ICD-9-CM 493.xx) between the years 2004 and 2008, and ≥1 controller medicine dispensed in 2011 (i.e., ICS, LTRA, ICS/LABA). We excluded children receiving omalizumab or chronic oral corticosteroids (≥180 continuous days supply) in order to exclude children with the most severe asthma. We also excluded children with diagnoses of cystic fibrosis, bronchiectasis, pulmonary hypertension, pulmonary embolism, immunodeficiency, hereditary and degenerative diseases, psychoses, and mental retardation.

To identify patients with probable persistent asthma, we intentionally used criteria that erred on the side of inclusiveness because variability in asthma severity might affect adherence to provider recommendations. To ensure parents and providers of children with persistent asthma were included in the sample, we oversampled for asthmatic children with ≥1 asthma-related hospitalization or emergency department visit in 2011. All results are weighted by the sampling fraction, accounting for the design and non-response to represent the target population.

Each participating site administered our surveys to its parent and provider populations based on its own respective IRB requirements. As a result of regulations regarding the sharing of personal health information and levels of consent necessary for subject participation, the four HMO Research Network sites used a combination mail and telephone approach for the parent survey, while surveys of parents in the TennCare population at Vanderbilt Medical Center were completed in person. Parent surveys were primarily conducted in English, but were also offered in Spanish at HPHC and Vanderbilt Medical Center.

Surveys were conducted following written informed consent at Vanderbilt or verbal informed consent at the other sites. The parent survey consisted of closed-ended questions and required approximately 20 minutes to complete. Parents were asked for the name of the child’s asthma provider and for permission to contact the provider. Participating parents received a $25 gift card. All of the surveys were conducted by the same trained research assistant at Vanderbilt.

Provider surveys were administered by mail at HPHC, KPNC, and KPGA sites while providers of the TennCare population completed secure web-based surveys via Research Electronic Data Capture (REDCap; http://project-redcap.org). HPRF offered the provider survey by mail and REDCap. Provider surveys were completed up to 3 months after the parent surveys as we required the parent to consent to allow us to contact the provider. The provider survey consisted of a combination of closed- and open-ended questions that took approximately 10 minutes to complete. We made a total of three attempts to reach the provider by mail. In the survey, we suggested that the provider use the child’s medical record to help complete the study.

Of 2,341 parents we identified based on computerized data and attempted to contact, 4.7% (111) were unreachable due to non-functioning telephone numbers. One hundred and nine (4.6%) were ineligible either because they did not speak English or Spanish, had moved out of state, or did not have a child with asthma. Of the 2,121 eligible parents, 1,321 completed the survey for a completion rate of 62% based on the standard formula from the American Association of Public Opinion Research.²¹ Of the completed parent surveys, 42% (549) were returned by mail, 47% (626) were completed over the phone and 8% (112) were completed in clinic. Of the 1,321 completed parent surveys, 94% (1,242) provided consent for us to contact the child’s provider while 30 parents did not. Of the 1,242 providers contacted, 770 completed the survey (62%). After removing the providers for whom parents did not give consent to link the surveys to their child’s electronic data, a total of 740 parent-provider pairs were included in the analysis.

Measures

With questions used in a previous asthma study, we assessed the asthma controller medication regimen as reported by the prescribing provider and the parent, including the specific controller medicine the child was prescribed or taking, and the frequency with which each respondent (provider or parent) thought the medicine should be administered.¹⁴ In the parent survey, a research assistant verbally read a list of the brand and generic names of all controller medications for asthma and asked, “Please choose a medicine that your/the child is taking.” Multiple responses were allowed for children taking more than one controller medication. We aggregated controller medications into three classes: ICS, LTI, and ICS/LABA. We assessed two primary outcomes: agreement on medication type and agreement on both type and frequency. To assess administration frequency, providers and parents were asked: “How is [the] child supposed to use this medicine?” with choice responses of “Everyday, year round; Everyday, during active asthma months; or Not everyday. Only for relief.” Providers were encouraged to reference the child’s electronic medical record when responding to the survey.

The survey included questions that were based on published instruments. Questions on concerns about controller medications were previously validated questions on concerns about medications from a study on medication adherence.¹⁵ Each question posed a statement on concerns about controller medications, e.g., “My child will become dependent on asthma medications if he/she takes them everyday,” and parents were asked to agree or disagree on a 4-point Likert scale.¹⁵ We also asked parents whether they were worried about the long-term effects of corticosteroid use with a question that has been used in a previous study.¹⁶ The survey included questions about parental expectations for symptoms that were developed by Yoos and colleagues.¹⁷ Each question posed a statement on expectations for asthma, e.g., “I expect that my child can fully participate in gym, normal physical activity, and playing.” In order to assess asthma control at the time of the interview, the Asthma Control Test (ACT) was administered.^18–20

Statistical Methods

Analyses were conducted in SAS version 9.3 (SAS Institute, Cary, NC) and R. We evaluated parent demographic factors and measures of concordance using univariate statistics. In bivariate analyses, we evaluated the association of each outcome with parent attitudes and beliefs about medicines and other independent variables using the chi-square test for categorical variables and t-test for continuous variables. Our main outcomes were: (1) parental and provider agreement on which controller medicine the child is supposed to be taking, irrespective of dose and frequency and (2) parental and provider agreement on the controller medicine and frequency that the medicine should be given. In the multivariable analysis, we conducted logistic regression analyses and included those variables that had a p<0.2 in the bivariate analysis. We also adjusted in multivariate analyses for the time between the last dispensing of controller medication and the date of the survey in order to account for recall bias. All results are weighted by the sampling fraction, accounting for the matched design and non-response to represent the target population.

Result

Population Characteristics

Our study population included 740 parents of children with probable persistent asthma and their providers (Table 1). A total of 1,321 parents of children completed the parent survey and 770 of their providers completed the provider survey. Thirty parents declined to have their child’s survey information linked to their electronic claims data. The mean age of the children was 8.6 years [SD 2.7 years] and 38% (278) were female. The parent sample included 34% (246) who self-described as white, 27% (197) Latino, 22% (162) black or African American, 12% (88) Asian, and 5% (38) other. As measured by the Asthma Control Test (ACT), 39% (282) of children had poorly controlled asthma at the time of the survey as defined by ACT ≤ 19.¹⁹

Table 1.

Baseline demographics.

N=740
Controller Medicine Group*
ICS	68% (679)
LTI	22% (224)
ICS/LABA	10% (99)
Age (mean)	8.6 years [SD 2.7]
Female	38% (278)
Race
Hispanic/Latino	27% (197)
Black or African American	22% (162)
White	34% (246)
Asian	12% (88)
Other/Native Hawaiian or Pacific Islander,	5% (38)
American Indian or Alaska Native
Parent’s highest grade/degree in school
Some high school or less	6% (46)
High school graduate	12% (87)
Some college or technical school	34% (246)
College graduate/Postgraduate training	48% (354)
Household income
Under $20,000	16% (112)
$20,001–$40,000	17% (113)
$40,001–$60,000	15% (103)
$60,001–$80,000	14% (94)
$80,001–$100,000	13% (90)
$100,001 or more	25% (167)
Number (mean) of days used controller medicine in past 14 days	3.9 (4.8)
Asthma Control Test (ACT) suggests poor control, ACT<=19	39% (282)
Number (mean) days used rescue medicine in the past 7 days	3.7 (3.3)
Number (median) days between survey and last dispensing	108 (interquartile range of 192)

^*Each child could be taking more than one controller medicine

We allowed providers to have a maximum of two patients with probable persistent asthma in the survey population in order to have a diverse representation of provider recommendations. The parents who were approached for survey were selected randomly if a provider had more than two patients eligible for the study; 57% of providers completed surveys for a single patient and the remaining 43% completed surveys for two children.

Provider Survey

The providers reported that 77% (572) of the children were supposed to be taking ICS, 23% (173) LTRA, and 10% (76) ICS/LABA (Table 2). Among those patients prescribed inhaled steroids, providers reported recommending daily year-round use in 50%, daily use during active asthma months in 41%, and intermittent use for relief in 9%. Prescribing less frequently than daily, year round was more common for ICS compared with LTI and ICS/LABA. According to the provider survey, 37% (211) of the children prescribed ICS were supposed to be using their ICS every day, year round; 50% (287) were supposed to be using it during active asthma months; and 13% (74) were supposed to be using it only for relief as needed, and not daily. Similarly, providers reported 77% (133) of the children on LTRA were supposed to be using the LTRA daily, year round, and 23% (39) reported the LTRA was to be used daily, during active asthma months.

Table 2.

Parent and provider reports on which controller medicine the child was supposed to be taking and at what frequency the medicine was supposed to be taken.

N=740	ICS	LTRA	ICS/LABA	Total
Total number prescribed by provider*	572 (77%)	173 (23%)	76 (10%)	821
Medicine prescribed to be given:
Every day, year round	211 (37%)	133 (77%)	65 (86%)
Every day, during active asthma months	287 (50%)	39 (23%)	10 (13%)
Not every day; only for relief	74 (13%)	1 (0.5%)	1 (1%)
Number parents who reported same medication class provider prescribed	404 (71%)	127 (73%)	59 (78%)	590 (72%)
Knew medication class and frequency	283 (49%)	94 (54%)	49 (64%)	426 (52%)

^*Providers could prescribe more than one controller medicine.

Parent Survey

According to the parents, 68% (679) of children were supposed to be taking an ICS, 22% (224) a LTRA, and 10% (99) an ICS/LABA (a child could be taking more than one controller medicine). There was mismatch between parent and provider understanding of which medication should be taken in 29% (168) of children who were supposed to be taking ICS according to their providers, 10% (20) of children who were supposed to be taking LTRA, and 16% (14) of children who were supposed to be taking ICS/LABA. For the 202 children who had providers who reported the child was supposed to be using the ICS every day, year round, there was mismatch between parent and provider understanding in 27% (54). In the 263 children who were supposed to be using ICS everyday when asthma is active, there was mismatch between parent and provider understanding for 54% (142) of the children, including 36% (94) parents who thought the child was supposed to use ICS everyday, year round and 18% (48) parents who thought ICS was supposed to be used not every day, only for relief. For the 67 children for whom providers recommended ICS only for relief, 34% (23) parents had the same understanding that ICS was only for relief. There was mismatch between parent and provider understanding in frequency of ICS in 66% (44) children; this included 18% (12) parents who thought ICS should be used everyday, year round, and 15% (10) who thought ICS should be used every day during asthma active months.

Multivariate Analyses

As presented in Table 3, parents were less likely to report the same controller medication type as the provider, irrespective of dose and frequency, if they were Latino [OR 0.23; CI 0.057–0.90] and believed the controller medicine was not helping [OR 0.15; CI 0.048–0.45]. Parents were more likely to report the same controller medication type as the provider if there was a household smoker [OR 2.87; CI 0.42–19.6]. Table 4 shows parental and provider agreement on both the controller medicine and frequency that the medicine should be given if the parent believed that the medicine is not helping [OR 0.22; CI 0.068–0.71] or that the child doesn’t need as much medicine as prescribed [OR 0.35; CI 0.17–0.76]. Families that have a household smoker were more likely to have parental and provider agreement on the controller medicine and frequency that the medicine should be given [OR 6.71; CI 1.05–42.91].

Table 3.

Odds of parent and provider agreement on which controller medicine child is supposed to be taking (irrespective of dose and frequency), adjusting for age, race/ethnicity, household income, highest education attained by parents, asthma control, controller medicine prescribed, site, concerns about medicines, exposure to household smoking, and time between the last dispensing of controller medication and the date of the survey.

N=740	Odds Ratio [Confidence Interval]
Poor asthma control (based on ACT ≤ 19)	0.42 [0.16–1.09]
Controller medicine
ICS/LABA	1.02 [0.24–4.32]
LTI	0.42 [0.12–1.46]
ICS (referent)	*
HMO
Site 1	0.61 [0.13–2.76]
Site 2	3.05 [0.78–11.89]
Site 3	4.44 [1.39–14.14]
Site 4	8.39 [1.56–45.1]
Site 5 (referent)	*
Race
White (referent)	*
Black	0.37 [0.12–1.16]
Latino	0.23 [0.057–0.90]
Asian	0.73 [0.19–2.89]
Income
≤$40,000 (referent)	*
>$40,000	1.10 [0.44–2.75]
Worry about long-term effects of inhaled steroid medicines	2.34 [0.87–6.27]
In past 7 days, how often has child taken rescue medicines	1.36 [1.14–1.62]
Medicine is too expensive	1.11 [0.36–3.40]
Medicine is not helping	0.15 [0.048–0.45]
Taking more than 1 controller medicine	0.97 [0.31–2.95]
Household smoker	2.87 [0.42–19.6]

^*Referent group

Table 4.

Odds of parental and provider agreement on the controller medicine and frequency that the medicine should be given, adjusting for age, race/ethnicity, household income, highest education attained by parents, asthma control, controller medicine prescribed, site, concerns about medicines, exposure to household smoking, time between the last dispensing of controller medication and the date of the survey.

N=740	Odds Ratio [Confidence Interval]
Poor asthma control (based on ACT ≤ 19)	1.27 [0.55–2.92]
Controller medicine
ICS/LABA	1.62 [0.56–4.73]
LTRA	2.24 [0.41–12.21]
ICS (referent)	*
Taking more than one controller medicine	0.45 [0.13–1.61]
HMO
Site 1	0.76 [0.21–2.75]
Site 2	1.28 [0.37–4.45]
Site 3	1.80 [0.58–5.66]
Site 4	3.77 [0.62–23.1]
Site 5 (referent)	*
Race
White (referent)	*
Black	0.49 [0.18–1.32]
Latino	0.56 [0.21–1.54]
Asian	1.28 [0.53–3.14]
Missing	2.62 [0.41–16.92]
Income
≤$40,000 (referent)	*
>$40,000	0.31 [0.13–0.72]
Expect child can fully participate in gym, physical activity, playing	0.56 [0.18–1.73]
In past 14 days, how often has taken controller medicine	1.03 [0.96–1.10]
Medicine is too expensive	2.02 [0.93–4.36]
Medicine is not helping	0.22 [0.068–0.71]
Sometimes stop giving child medicine to give his/her body a rest	1.02 [0.25–4.21]
Child doesn’t need as much medicine as prescribed	0.35 [0.17–0.76]
Household smoker	6.71 [1.05–42.91]

^*Referent group

Discussion

Our study had four key findings. First, providers did not prescribe controller medicines for most patients with possible persistent asthma on a daily basis. Second, only 70% of parents knew which controller medicine their provider recommended, and half of parents did not know the medicine and frequency it was supposed to be given. Third, parent report of the child’s controller medication was less likely to match the provider’s if the parent self-identified as Latino or felt the medicine was not helping. Fourth, a mismatch between parent and provider report of the prescribed frequency of controller medication use was associated with the parent believing their child did not need as much medicine as prescribed.

Our finding that providers do not prescribe controller medicines on a daily basis for asthma is consistent with previous studies. The Expert Panel Report (EPR-3) Summary Report 2007: Guidelines for the Diagnosis and Management of Asthma was developed by the NAEPP to improve the quality of care and asthma outcomes based on evidenced-based medicine. Adherence to the NAEPP guidelines has been demonstrated to result in improved outcomes for asthma, including decreased hospitalizations, ED visits, and outpatient visits.²¹ Despite the promise of NAEPP guidelines, morbidity from asthma remains high.⁶ Poor adherence to national guidelines likely plays a role. A study of inner-city providers found that only 62% of providers adhere to NHLBI guidelines for ICS use,²² which is higher than our study that found that 50% of children received recommendations from their providers to take controller medications daily. Our study findings are consistent with a previous survey of pediatric providers that found that almost half of pediatric providers did not prescribe daily ICS to children with mild persistent asthma and almost all providers felt that seasonal or periodic use of ICS could be effective for some patients with asthma.²³ We found that almost 10% of the children received provider recommendations to use controller medications on an as needed basis.

A large number of parents – 50% -- gave answers that conflicted with the provider on both which controller medicine the provider recommended and with what frequency the medicine was to be given. Improving provider-patient communication during medical visits could help with caregiver-reported medicine adherence.²⁴ Research conducted prior to the routine use of LTRAs found that the level of caregiver-provider agreement regarding whether the child was prescribed an ICS or cromolyn was 78%, suggesting moderate agreement between providers and caregivers. Our study suggests that even with the introduction of LTRAs which have overall higher adherence than ICS and may be more likely to be recalled by the parent,²⁵ parents frequently do not know which controller medicine was prescribed by the provider. The frequent mismatch in understanding of which controller medications a child should be taking and how often implies that many providers may be unaware of their patient’s lack of adherence and suggests that there is room for improvement in communication about asthma medication regimens.

We found that parents of Latino children were less likely to report the child was using the same controller medicine, irrespective of frequency, as the provider prescribed. This finding is important because some Latino subgroups (e.g. Puerto Ricans) experience higher morbidity and mortality due to asthma than white children. This mismatch of patient-provider expectations seems like a likely cause of the underuse of preventive medicines that has been observed in Latino children compared with white children, even among insured populations.²⁶ However, it is possible that this finding is a reflection of site-specific differences in collection of data, as only two sites were able to enroll Latino participants.

Other findings offer insight into ways that providers might be able to identify families at elevated risk for mismatched understandings of how to use controller medicines. For example, our finding that parents were less likely to agree on the controller medicine if the parents felt the medicine was not helping suggests that providers may be able to identify families at risk of non-adherence by asking parents for their impressions about whether the medication is helping.

Likewise, parents and providers were less likely to agree on the controller medicine and frequency if the parent thought the child didn’t need as much medicine as prescribed, which suggests that providers could specifically probe parents’ perceptions of this issue as a means of identifying patients who are using less medication than providers intended.

Strengths of our study include a diverse population of children from five geographic locations around the country, the completion of surveys from both providers and parents in order to compare understanding of controller medicine regimens, and the ability to link information from parent and provider surveys to claims data. Limitations of our study include the inability to distinguish between Latino subgroups and the fact that only two locations offered participation in Spanish. HPHC and Vanderbilt offered the survey in Spanish, but the other three sites excluded non-English speaking respondents, which meant we could not access this population. This is an important demographic to access because these parents may experience greater difficulties communicating with their providers, resulting in increased misalignment between provider recommendations and parental action. Additionally, parent gender was not evaluated, though it is possible that differences in asthma medicine adherence emerge between mothers and fathers. Furthermore, we were unable to be sure that the parents and providers were referring to the same controller medication , although we believe there is high probability that the parents and providers referred to the same controller medication prescription because we interviewed the providers whom the parents identified as the asthma provider. Moreover, we identified and asked about the most recent prescription. The vast majority of controller medications reported by parents and providers matched, supporting the likelihood that parents and providers were referring to the same prescription. Moreover, in our observational study, we were unable to assess causality. For example, when we found that a mismatch between parent and provider report of the prescribed frequency of controller medication use was associated with the parent believing their child did not need as much medicine as prescribed, reverse causation is possible as mismatches and subsequent medication non-adherence could lead to more poorly controlled asthma in children.

In conclusion, a substantial number of providers prescribe asthma controller medicines less frequently than recommended by evidence-based guidelines, particularly for ICS. In addition, half of parents of children with asthma reported controller medicine regimens that were mismatched with their providers’ reported recommendations. Efforts to increase use of controller medicines for asthma may need to focus on provider, parent and patient communications and understanding of controller medicine regimens.

Highlight Box.

What is already known about this topic?

Underuse of controller medicines among children with persistent asthma remains widespread despite clear national guidelines and quality measures.

What does this article add to our knowledge?

Providers do not recommend that inhaled steroids be used according to national guidelines. Mismatches between parent reports and providers’ intentions regarding how the child was supposed to use inhaled steroids occurred in half of children.

How does this study impact current management guidelines?

Efforts to increase use of controller medicines for asthma should focus on ways to align provider decision-making with national guidelines and ways to reduce mismatches between parent and provider intentions regarding controller medication use.

Abbreviations

ED

emergency department, HMO, Health Maintenance Organization

HMO

Health Maintenance Organization

HPHC

Harvard Pilgrim Health Care

HPRF

Health Partners Research Foundation

ICS

inhaled corticosteroids

ICS/LABA

inhaled corticosteroids/long-acting beta agonist

IRB

Institutional Review Board

KPGA

Kaiser Permanente of Georgia

KPNC

Kaiser Permanente of Northern California

LTRA

leukotriene antagonist

PEAL

Population-Based Effectiveness in Asthma and Lung Diseases