Parent-Reported Environmental Exposures and Environmental Control Measures for Children With Asthma FREE

THE INCREASE in childhood asthma prevalence and morbidity during the past 2 decades has been substantial, with the highest rates reported among poor, urban, and minority populations.^{1 - 2} Mounting evidence points to exposure to indoor allergens and irritants, sometimes associated with poor housing conditions, as a major contributor to childhood asthma.^{3 - 5} Exposure to environmental tobacco smoke has consistently been shown to increase asthma risk for children.^{6 - 9} Studies have implicated dust mites,¹⁰ cockroaches,^{11 - 12} mice and cats,¹³ and other furry pets as contributing to asthma morbidity. Exposures to mice and cockroaches have been shown to be especially common in asthmatic patients living in inner-city housing.^{11 ,14} Positive skin test results to specific allergens, including dog, cat, and indoor molds, have been associated with sensitivity to inhaled methacholine chloride.¹⁵ A review⁵ by the US Institute of Medicine recently found evidence of a causal relationship between exacerbations and dust mite, cockroach, and cat exposure in sensitized individuals and sufficient evidence of an association between exacerbations and exposure to dogs and fungi (molds).

National and international guidelines for the treatment of asthma, disseminated since 1991,^{16 - 18} highlight the importance of the reduction of environmental triggers as part of a comprehensive approach to asthma treatment. However, Lanphear et al¹⁹ recently reported that 40% of children with asthma had at least 1 residential exposure to a potential asthma trigger. Avoidance of triggers may be difficult, especially for families who are less socially mobile and are more likely to live in shared housing environments, such as apartment buildings. Most intervention trials testing methods to decrease allergen exposure have focused on dust mite control measures. While individual studies³ have shown benefit of such strategies, a recent meta-analysis²⁰ did not confirm evidence of benefit for all children with asthma based on a subset of studies meeting strict inclusion criteria.

As part of the baseline evaluation of children with mild to moderate asthma enrolled in a multicenter trial, the Pediatric Asthma Care Patient Outcomes Research Team II (PAC PORT II), we questioned families about current environmental triggers in the home and on actions they routinely take to help avoid exposures that might exacerbate asthma. Our goals were to assess the prevalence of exposure to potential environmental triggers, to identify risk factors for such exposures, and to determine whether instructions to parents about trigger avoidance were associated with fewer such exposures.

We analyzed baseline data from the 638 patients enrolled in the PAC PORT II, a multicenter randomized trial of asthma care improvement strategies in practices affiliated with managed care organizations (MCOs) in 3 geographic areas: Group Health Cooperative, greater Seattle; Rush Prudential Health Plan, Chicago; and Harvard Pilgrim Health Care and Blue Cross Blue Shield of Massachusetts, greater Boston. Potential subjects were children aged 3 to 15 years who had an asthma-related claim for a hospitalization, emergency department visit, or ambulatory encounter during a 1-year period. Children with a single ambulatory encounter for asthma also needed to have received 2 or more asthma medications during the same period to be eligible for screening. Telephone screening then identified those children who used daily medications for at least 2 months during a 1-year period as eligible for the trial, and excluded those with severe asthma or another major chronic illness. The goal of this enrollment process was to identify children with mild persistent or moderate persistent asthma, without serious comorbid conditions, for enrollment into the trial.

Data were collected using closed-ended questions in face-to-face interviews. Caregivers reported on their child's asthma symptoms (number of symptom days based on a 2-week recall), functional status (using the American Academy of Pediatrics' Children's Health Survey for Asthma),²¹ and exposure to various environmental triggers in the home. These included environmental tobacco smoke, dust, pets, and household pests (including cockroaches and rodents). Key processes of asthma care were examined, including whether a health care provider "ever gave you written instructions about what to do about triggers/things that start asthma" and if these were given or updated in the prior 12 months. A second item asked if a provider, in the past 6 months, "discussed with you changing things around your child—such as smoke, dust, or pets—to keep your child from getting asthma symptoms."

For each environmental exposure of interest, we examined differences by parent-reported patient race or ethnicity, annual family income (<$30 000, $30 000-$60 000, and >$60 000), and the maximum educational attainment of either parent in the household (high school education or less, some post–high school education, or college graduation). Analyses of housing type compared those in unattached single family homes with those in multifamily dwellings. We also assessed differences in relation to reported symptom frequency. Many subjects were receiving asthma controller medications at the time of the interview. Because there is no widely accepted standard for integrating symptom frequency with current medication use in assigning severity, we categorized patients by symptom days of the past 14 as follows: 0, 1 to 4, 5 to 12, and 13 to 14. These categories approximate the symptom frequencies of the categories suggested by the National Asthma Education and Prevention Program guidelines¹⁷ but do not alone classify these patients' "severity."

Categorical variables were analyzed using the χ² test. To assess associations of reported exposures with demographic factors, we used separate multivariate logistic regression models for each environmental trigger. Age, MCO, and the 3 available demographic variables (race, family income, and parental educational level) were included in all models. Housing type was tested in all models and retained in final models only if significant (P<.05). Odds ratios and 95% confidence intervals (CIs) are reported. Finally, reports of having received written environmental control instructions or having a discussion with a clinician about such issues were added to each final model to assess their association with the exposure of interest. All analyses were performed using SAS statistical software (SAS Institute Inc, Cary, NC). The study was undertaken with the approval of the institutional review boards at each of the study sites.

A total of 638 (64%) of the 1000 eligible children identified completed the baseline interview for the PAC PORT II trial and are described in Table 1. The mean age of enrollees was 9.4 years; 60% were male; and 32% were nonwhite. The families of these patients were, on average, highly educated, with 89% reporting some college education for at least 1 member. Patients in the sample varied in their report of current symptoms, with 33% reporting 5 or more symptom days during the past 14 days. Patients of the 3 health systems differed according to patient race (P<.001), family income (P = .005), and parental educational level (P = .01), but there were no differences in reported symptom frequency across the sites (P = .39). Symptoms were reported to be more frequent among lower-income children (P = .04) and varied by family educational level (P = .04), but did not differ significantly by patient race (P = .36). Enrollment rates differed modestly in Boston, Chicago, and Seattle (65%, 69%, and 59%, respectively; P = .01). Those patients meeting the eligibility criteria who did not enroll in the study were older (10.3 vs 9.3 years; P<.001), but had a similar illness severity as measured by the proportion with hospitalizations (P = .78) or emergency department visits (P = .56) and the number of weeks taking asthma medications in the past year (P = .57).

Exposures to environmental triggers are summarized in Table 2. In 30% of households, at least 1 family member smoked cigarettes (including 17% of mothers), and 71% of parents expressed worry about their child's exposure to tobacco smoke. The presence of household pests (including cockroaches and rodents) was a problem for 18% of the families, including 6% who reported exposure to cockroaches. Furry pets were present in 59% of households, including 32% with cats and 39% with dogs. Other exposures to potential environmental triggers included bedroom carpeting in 78% and forced-air heat in 58% of the households. Most children did not have a plastic cover on their mattress or pillow. Suboptimal housing attributes such as mildew or water damage were relatively common. Most patients did not follow recommendations to wash linen in hot water, and 18% slept with a natural feather pillow.

Bivariate analyses of selected environmental exposures are shown in Table 3. Smoking in the home was inversely associated with income (47% vs 19% in the lowest and highest income groups, respectively, P<.001). Smoking rates also differed (P<.001) among those not graduating from high school (51%), those with some college (41%), and college graduates (18%), but not with patient race or current symptoms. Smoking was more common among residents of multifamily dwellings (43% vs 27%, P<.001). Cat and dog ownership were least frequent among families of black patients, and dogs were most prevalent among families with annual incomes between $30 000 and $60 000 (P<.01). Dog ownership was more frequent in single family homes. Exposure to household pests was more commonly reported by black families and those in multifamily dwellings, with trends toward less exposure among those with a lower family income (P = .07) and lower parental educational attainment (P = .05). Conversely, bedroom carpeting was more frequent in the homes of white children compared with black and Hispanic or Spanish children. There were no associations between symptoms and these environmental exposures.

Consistent with membership in an MCO, virtually all patients reported having a primary care provider (Table 4). Some aspects of asthma care recommended by the National Asthma Education and Prevention Program were followed in most patients with asthma. For example, 70% had their peak expiratory flow measured in the physician's office at least once and 73% reported owning a spacer device for inhaled medications. Finally, 91% reported being shown how to use a metered-dose inhaler by a health professional. Most patients reported having received some written instructions for taking medications; however, the proportion of patients ever receiving a written action plan for the management of their asthma was only 21%. Forty-five percent reported ever receiving written instructions regarding trigger avoidance, but only 11% reported receiving such instructions in the past year. Forty-two percent reported a discussion with a health care professional about changing potential triggers in the home environment in the past 6 months.

In multivariate models (Table 5), independent predictors of smoking in the household included an annual family income of less than $30 000 (OR, 1.9; 95% CI, 1.0-3.7) and between $30 000 and $60 000 (OR, 1.8; 95% CI, 1.1-2.8) compared with greater than $60 000, and parental educational attainment of high school or less (OR, 4.5; 95% CI, 2.4-8.2) or some college (OR, 2.8; 95% CI, 1.8-4.3) compared with college graduates. Dog ownership was more likely with parental educational attainment of high school or less (OR, 2.3; 95% CI, 1.2-4.3) or some college (OR, 1.6; 95% CI, 1.1-2.5) compared with college graduates. Black patients were less likely than others to own cats (OR, 0.26; 95% CI, 0.12-0.6) or dogs (OR, 0.4; 95% CI, 0.2-0.7). Reports of pests were significantly increased for black children compared with white children (OR, 2.3; 95% CI, 1.2-4.2). Finally, black race was associated with lower rates of exposure to carpeting (OR, 0.5; 95% CI, 0.3-0.9).

Receipt of written instructions regarding environmental exposures did not differ significantly by family income (P = .25) or parental educational attainment (P = .06), but was more frequently reported by black patients compared with white patients (56% vs 44%; P = .02). This difference did not persist after controlling for MCO and age. After controlling for demographic variables, parental report of having been given written environmental instructions in the past year was not associated with differential exposure to environmental triggers. A report of having discussed these issues was positively associated with having a smoker in the home, but not with the other environmental exposures when examined in multivariate models.

Recent asthma care improvement efforts have focused predominantly on appropriate pharmacotherapy, but guidelines¹⁷ also recommend patient education to minimize potential environmental triggers. We sought to determine the prevalence of exposures to potential environmental triggers for children in their homes, whether parents reported being given clear instructions about how to avoid them, and whether such instructions were associated with fewer exposures. We found concerning rates of exposure to many of the agents known to be potent triggers for some children with asthma. Of the children in our sample with asthma, 30% were exposed to tobacco smoke at home, many homes had dogs or cats, and 18% reported household pests, such as cockroaches. As important as the substantial rates of these exposures was the finding that only 11% had received written instructions about environmental trigger avoidance in the past year, although more (42%) reported some discussion with a physician on these issues in the prior 6 months.

These rates of specific exposures are higher than those recently reported by the National Health and Nutrition Examination Survey.¹⁹ For example, smoking and pet ownership were more common in our study than in the nationally representative sample. Demographic factors were associated with differential risk for some of these triggers. Lower family income and parental educational level were associated with exposure to environmental tobacco smoke, and black race was associated with lower rates of exposure to furry pets and bedroom carpeting. We believe strongly that physicians should not use these data to make assumptions about the home environment of patients they see. Rather, they highlight the prevalence of exposures to potential environmental triggers in all socioeconomic strata and among all racial and ethnic groups. Structured approaches to asking each family with asthma about their home environment are necessary to tailor educational efforts.

Some triggers more closely linked to housing may be difficult for families to change. The rate of exposure to cockroaches, 6% by parental report in our sample, was much lower than that in a study¹¹ of inner-city patients with asthma that measured antigens in the home. Previous work²² has highlighted the discrepancy between parental report of specific environmental exposures and antigen levels measured in the home. Cockroach antigen has been reported to be 33 times more prevalent in high-poverty areas.²³ Likewise, the rate of reported exposure to rodents was lower in our sample than in the National Cooperative Inner-City Asthma Study.²⁴ In that study, which collected household samples from enrolled subjects, detectable mouse antigen was found in at least 1 room in 95% of the homes.

In the dissemination of the National Asthma Education and Prevention Program recommendations, the primary focus has been on improving pharmacologic treatment of asthma by increasing the appropriate use of anti-inflammatory agents for all patients with persistent asthma.¹⁷ This reflects the change in our understanding of asthma as a disease of chronic inflammation.^{25 - 26} Less attention has been paid to helping families reduce exposure to environmental triggers. As physicians and patients treat asthma more aggressively, it will be increasingly important to reduce offending environmental agents to minimize the medications required.

The lack of association we found between specific exposures and current symptoms bears comment. First, we did not control for medication use by these patients, which may reduce some of the symptoms attributable to these exposures. However, it is also clear that we still have an incomplete understanding of the contribution to symptoms of environmental exposures among the general population of children with asthma. Lanphear et al¹⁹ suggest that the elimination of residential risk factors could lower asthma prevalence by 39%, if the associations they report are causally related to the development of disease. However, we believe it is not possible, given the cross-sectional studies that make up most of the literature, to accurately assess such contributions.

The finding that so few patients reported receiving current written suggestions regarding environmental control is disappointing, but not surprising. These data reflect parental report of education by clinicians before initiation of any structured intervention by the PAC PORT II study. We have no information about the content of such education, which likely varied by clinician within these practices. Providing parental education is time-consuming for health care providers. Unless care delivery systems are designed specifically to provide high-quality patient education, such goals are hard to achieve in the context of usual primary care visits. A combination of educational materials (print, videotape, and computer based) may be used, but all require additional time and effort to ensure comprehension and tailor education to a particular child and family.

The patients interviewed were enrolled in MCOs with relatively diverse memberships in 3 geographic areas. These data represent an insured sample whose socioeconomic status was relatively high, and may not be generalizable to other regions of the country or to publicly insured or uninsured populations. Subjects from the 3 sites differed demographically, as did crude rates of exposure to several of the environmental triggers. We cannot determine whether this reflects geographic differences, urban vs suburban residence, or attributes of members of particular plans. Unadjusted reports of prior parental education also differed among the 3 sites. This study was not designed to compare care processes among the participating health plans; therefore, we have controlled for MCO site in multivariate analyses, but do not report effect estimates for each site.

These data were collected as baseline information for a 2-year randomized trial of care improvement strategies, which may further limit their generalizability. Families choosing to participate might differ from those who would not have enrolled in such a trial. The inclusion criteria aimed to select children with mild to moderate persistent disease, so these results may not be representative for children with less severe asthma. Finally, we have no confirmation of parental reports of either exposure or process-of-care variables. There may be reluctance for parents (especially those with a child who has asthma) to accurately report their smoking habits or the presence of household pests. However, our high rates of these reported exposures argue against such intentional underreporting. Also, parental recall about education they received may be incomplete.

Environmental exposures that may add substantially to the morbidity of children with asthma continue to be prevalent across diverse geographic and sociodemographic populations. The extent to which these exposures are responsible for the high morbidity of asthma in low-income areas and minority populations is unclear. The associations reported herein between particular environmental exposures and patient race or ethnicity do not imply that such demographic attributes cause increased exposure. Rather, race, family income, and parental educational level are crude proxies for a complex set of societal forces that may affect the prevalence and severity of asthma, the quality of care received for this condition, and the risk of exposure to potential asthma triggers. For example, Litonjua et al²⁷ recently showed that a large fraction of the racial and ethnic differences in asthma prevalence could be explained by area of residence, income, and level of education. In our data, varying combinations of race, income, and education were independently associated with exposures. We report ORs for each of these variables for completeness and consistency. It is a weakness of this study that we did not obtain more detailed measures of poverty, including housing attributes, or more detailed data on ethnicity. It is possible that such additional variables might have better explained the variation in exposure rates than the demographic variables reported herein.

Some of these environmental exposures, especially those associated with urban substandard housing, may be difficult for families to mitigate. Others, such as exposure to tobacco smoke in the home, are recognized threats by most parents. Although smoking cessation is a challenge in any circumstance, interventions have attempted to leverage a child's asthma as a motivating factor.²⁸ Programs for structured asthma education that include trigger avoidance and enhanced self-management have been developed, but have shown mixed results.^{29 - 32} In an era of ever-tightening health care finances, MCOs and practicing physicians may have difficulty implementing such programs without clear evidence of their effectiveness in various health care delivery systems. As we continue to develop new and better pharmacologic agents for asthma, we should also focus on testing behaviorally oriented strategies through which nonpharmacologic recommendations of national guidelines might be more widely implemented.

Accepted for publication November 14, 2001.

This study was supported by grant HS08368 from the US Agency for Healthcare Quality and Research, Rockville, Md (PAC PORT II); and grant HS08368 from the National Heart, Lung, and Blood Institute, Bethesda, Md.

We thank Scott Weiss, MD, and Vincent Carey, PhD, for their helpful comments on manuscript drafts; Nancy Laranjo for data analysis; and the families who participated in this study in Chicago, Boston, and Seattle.

Asthma research, policy statements, and guidelines all highlight the contribution of indoor environmental exposures to morbidity from childhood asthma. Environmental tobacco smoke has been shown to be particularly noxious to children with asthma.

This cross-sectional study reports on the range of exposures reported by families of children with asthma. Although the specific exposures varied by sociodemographic factors, children from all backgrounds and circumstances had exposures of significant concern. These results confirm the need to further develop effective patient education strategies to reduce indoor environmental exposures.

Corresponding author: Jonathan A. Finkelstein, MD, MPH, Department of Ambulatory Care and Prevention, Harvard Medical School and Harvard Pilgrim Health Care, 133 Brookline Ave, Sixth Floor, Boston, MA 02215 (e-mail: Jonathan_Finkelstein@harvardpilgrim.org).