Outpatient Management Practices Associated With Reduced Risk of Pediatric Asthma Hospitalization and Emergency Department Visits
Objectives. Effective outpatient care is believed to prevent hospitalization and emergency department (ED) visits resulting from childhood asthma. The aim of this study was to suggest priority areas for intervention by identifying outpatient management practices associated with the risk of these adverse outcomes in a large population.
Methods. This case-control study included children aged 0 to 14 years with asthma who were members of a regional health maintenance organization. Cases were children undergoing either a hospitalization or an ED visit for asthma during the study period. Control subjects were children with asthma without a hospitalization or an ED visit during the study period who were matched to patients on age, gender, and number of asthma-related hospitalizations in the past 24 months. Data on provider and parent asthma management practices were collected using chart review, closed-ended telephone interviews with parents, and computerized use databases. Multivariate analyses were conducted using conditional logistic regression models.
Results. Data were collected on 508 cases and 990 control subjects. A total of 43% of cases were reported by their parents to have moderately severe or severe asthma, compared with 20% of control subjects. Parents of cases with hospitalization were less likely than control subjects to have a written asthma management plan (44% vs 51%) and to report washing bedsheets in hot water at least twice a month (77% vs 86%). Cases with hospitalization were more likely to have a nebulizer (74% vs 56%).
In the final multivariate model, race/ethnicity was not associated with having had either a hospitalization or an ED visit, as was lower socioeconomic status. Having a written asthma management plan [odds ratio (OR): 0.54; 95% confidence interval (CI): 0.30, 0.99] and washing bedsheets in hot water at least twice a month (OR: 0.45; 95% CI: 0.21, 0.94) were associated with reduced odds of hospitalization. Having a written asthma management plan (OR: 0.45; 95% CI: 0.27, 0.76) and starting or increasing medications at the onset of a cold or flu were associated with reduced odds of making an ED visit.
Conclusions. Practices that support early intervention for asthma flare-ups by parents at home, particularly written management plans, are strongly associated with reduced risk of adverse outcomes among children with asthma.
Hospitalization for childhood asthma has increased in the United States over the past decade.1,2 Adverse outcomes such as hospitalization and emergency department (ED) visits account for almost three fourths of the direct costs of asthma and are believed preventable by effective outpatient care.3 For some children, lack of adequate health insurance may preclude optimal care and may contribute to adverse outcomes.4,5 Yet, even among insured children, there is wide variation in rates of hospitalization and ED visits.6,7
National guidelines issued in 1991 recommend specific outpatient asthma management practices, including peak flow monitoring, dust mite allergen control, use of spacers with inhalers, and use of written management plans.8 However, there is scant information to suggest which of these management practices needs more emphasis in actual populations.
The aim of this study was to suggest priority areas for intervention by identifying outpatient management practices associated with increased or decreased risk of adverse outcomes.
Identification of Case and Control Subjects
The study was conducted within the Kaiser Permanente Medical Care Program (KPMCP) of Northern California, a group model health maintenance organization (HMO) providing comprehensive care to ∼2.4 million members, 20% of whom are ≤14 years old. Most members receive all of their inpatient and outpatient care within the plan, although at the time of this study, 20% of children did not have KPMCP coverage for medications and may thus have used outside pharmacies.
We identified subjects by 1) using computerized data to identify a cohort of children with asthma-related use, and 2) using a telephone interview with the parent to confirm that the child had been diagnosed with asthma. The cohort included KPMCP members ≤14 years of age with asthma-related use at 13 facilities that had computerized pharmacy information. The following criteria were used: 1) a hospitalization for asthma (International Classification of Diseases, Ninth Edition code 493.0 to 493.9), (2) an ED visit for asthma, identified by text strings matched to the diagnosis text field, or (3) at least one asthma medication prescription filled and recorded in KPMCP's pharmacy information system. Pharmacy database criteria have been shown previously to identify patients with asthma with acceptable accuracy.9
Cases were children with asthma-related hospitalizations or ED visits between January 1995 and July 1995. We attempted to interview the parents of all patients with hospitalization and a random sample of parents of children with ED visits. For each case, we interviewed parents of two control subjects who did not have a hospitalization or an ED visit during the study period. Potential control subjects were sampled from the cohort with asthma-related use. For each case, control subjects were selected at random from among those patients matched on age, gender, and number of asthma-related hospitalizations during the 24 months before the study period. These variables were matched because previous studies have already established their association with adverse outcomes,10,11 and the goal of this study was to identify other variables amenable to intervention. Only children whose parents confirmed during the telephone interview that their child had been diagnosed with asthma by a doctor were included in the study. If used once as a case, a child was not used again either as a case or a control subject. A child used as a control subject who subsequently experienced a hospitalization or an ED visit was converted to a case and the parent reinterviewed. Converted control subjects were replaced with new control subjects.
Data were collected using chart review, telephone interviews with parents and KPMCP computerized databases. The Appendix shows the complete list of predictor variables in this study. Chart review was conducted for cases to describe medications recommended at outpatient clinic visits before the adverse event. A 10-minute telephone interview was conducted with each child's parent. For cases, the interview was conducted ∼3 weeks after the adverse outcome; the matched control subjects were interviewed as soon as possible afterward. The parent was first asked whether a doctor had ever diagnosed the child as having asthma; if not, the child was considered ineligible for study. Parents who confirmed that their child had asthma were asked to complete the interview, the questions of which were based in part on those used in a previous study.12
The interview was closed-ended, with the interviewer reading the response choices for most of those questions with ordinal categories, eg: “In general, how severe is your child's asthma? Is it mild, moderate, moderately severe, or severe?” Parents were asked about family demographics, the severity of the child's asthma, their asthma management practices, and their experiences with the health care system (Appendix). Parents of cases were asked questions about specific asthma management practices before as well as after the adverse outcome.
We used KPMCP computerized databases to collect information about insurance coverage and copayments, asthma medications prescribed, hospitalizations, and ED visits made during the 12 months before the study period. As shown in the Appendix, variables from both sources were classified in the following major groups: demographic and social; use and parent-reported measures of severity; access to care; and patient and physician practices.
Our goal was to identify practices amenable to health care interventions, such as access to care and parent and physician practices. In multivariate analyses, demographic and severity variables were treated as confounders (risk adjusters). It is important to note that disease severity is associated with many of the parent and physician practice predictors, and thus may confound the relationship between these predictors and the adverse outcomes. Interventions such as asthma education classes and care by asthma specialists, although indicative of patient and physician behaviors, are also indirect measures of severity, because children with more severe asthma problems are more likely to have received them.
Conditional logistic regression models were developed separately for hospitalizations and ED visits. A series of models was fitted as follows. Model I examined the role of demographic and social variables. Model II retained those variables with P ≤ .10 from model I and added parent assessment of asthma severity (ie, the parent's response to the question, “In general, how severe is your child's asthma?”) and use-based measures of severity (previous medications, outpatient visits, ED visits, and hospitalizations).
Model III evaluated the independent effects of outpatient practice variables (access to care and parent and physician practices). First, we examined the association between each practice variable and outcome (hospitalization and ED visit), adjusted for sociodemographic variables and asthma severity. Because of the large number of outpatient practice variables relative to sample size, we eliminated from additional consideration those variables that were not significant atP = .20. A stepwise procedure was then used to arrive at the final model, which retained factors with P ≤ .10 from model II and outpatient practice variables that reached statistical significance at P = .10.
Adjusted odds ratios (OR) and 95% confidence intervals (CI) are presented in this paper. Continuous variables and those in the form of discrete counts (eg, age, number of previous hospitalizations) were entered in the regression models as linear terms. The appropriateness of the assumption of linearity in the log odds was checked by examining the pattern of coefficients in the categorized versions of the variables in the final models.
During the study period, interviews were attempted with 563 parents of potential cases and 1581 parents of potential control subjects. Of the 563 potential cases, 10 (2%) were ineligible because of not having asthma, and 8 (1%) were ineligible for other reasons. Of the 545 parents of eligible cases, 508 (93%) completed interviews (133 had hospitalization and 375 had ED visits), 27 (5%) were passive refusals (repeated answering machines or no answer), and 10 (2%) declined to be interviewed. Of the 1581 potential control subjects, 453 (29%) were ineligible because of not having asthma by the parent's report, and 36 (2%) were ineligible for other reasons. Of the 1092 parents of eligible control subjects, 990 (90%) completed interviews, 73 (7%) were passive refusals, and 29 (3%) declined to be interviewed. In analyses, data on 990 control subjects were included, 257 matched to the cases who were hospitalized, 733 to the cases who were treated in an ED.
Table 1 shows the demographic characteristics of cases and control subjects. The mean age of both cases and control subjects at the time of the cases' ED visit or hospitalization was 6.4 years. Approximately 63% of the subjects were male. The case and control subjects were similar with respect to household income, household type (two-parent family versus other), and the parents' level of education and full-time employment status. However, there was a slightly higher percentage of whites among the control subjects (P = .0096).
As shown in Table 2, parents' reports on several measures indicated that cases had more severe disease than did control subjects. Among cases, 42.6% were reported to have “moderately severe” or “severe” asthma, compared with 19.5% of control subjects. Cases had more severe asthma based on other parent-reported measures, including school days missed and activity limitations. Cases also had higher severity of illness based on previous use measures: they were more likely to have made previous ED visits (P = .0001) and to have been hospitalized (P = .0001). They also had higher use of β-agonists (P = .0001), cromolyn (P = .0019), inhaled steroids (P = .0002), and oral steroids (P = .0001).
Hospitalization: Asthma Management Practices
Charts were available for review for 95% of the hospitalized cases. Thirty-six percent of these patients had been seen in the outpatient clinic or ED ≤7 days before the hospitalization; 51% had been seen during the previous 30 days. Among the 48 patients who made visits for flare-ups in the month before the adverse event, 44% were prescribed cromolyn or inhaled steroids, 65% were prescribed short-course oral steroids, and 85% were prescribed inhaled β-agonists. At the time of hospitalization, 46% of patients had an oxygen saturation (O2 sat) of ≤90%, 31% had an O2 sat of 91% to 94%, and 23% had an O2 sat of ≥95%.
The proportions of hospitalized cases and their control subjects who reported specific asthma management practices in the telephone interview are shown in Table 3. The proportions for cases reflect their parents' retrospective reports of how the asthma was managed before the adverse outcome. Among parents who classified their child's asthma as severe, >80% had a nebulizer at home, and most reported having almost always or usually used preventive medicines in the past 2 months. Nearly half of cases with severe asthma had a written asthma plan, and most reported washing their bedsheets at least twice a month.
Hospitalization: Multivariate Models
Preliminary models identified income, parent assessment of severity, and the number of bronchodilator prescriptions in the past 6 months as the risk adjusters to be used in the final model. Overall assessment of severity was chosen as the primary parent-reported severity variable, because it was strongly correlated with other parent-reported severity indicators (such as frequency of symptoms, number of school days missed, exercise limitations) as well as with medication use.
The association of each asthma management practice with case versus control subject status was evaluated individually, after controlling for these risk adjusters, in conditional logistic regression models. As Table 3 shows, cases were more likely to report having a nebulizer and almost always or usually using preventive medicine than were control subjects. Patients were less likely to have a written management plan and to wash bedsheets at least twice a month.
The final multivariate model (Table 4) showed that children from higher income households were less likely to be hospitalized because of asthma (OR: 0.82 per income category of $10,000, up to $50,000; 95% CI: 0.69–0.98). After adjusting for income and severity measures, having a written asthma management plan (OR: 0.54; 95% CI: 0.30–0.99) and washing the child's bedsheets at least twice a month (OR: 0.45; 95% CI: 0.21, 0.94) were associated with reduced odds of hospitalization. Having a nebulizer was associated with a higher risk of hospitalization (OR: 2.9; 95% CI: 1.4, 6.2).
ED Visits: Asthma Management Practices
Charts were available for review for 97% of the cases with ED visits. Eighteen percent of patients had been seen in the outpatient clinic ≤7 days before the ED visit; 35% had been seen during the previous 30 days. Among the 78 patients who made visits for flare-ups in the month before the adverse event, 49% were prescribed cromolyn or inhaled steroids, 47% were prescribed short-course oral steroids, and 73% were prescribed inhaled β-agonists.
Table 5 shows the proportions of cases with ED visits and their control subjects who reported specific asthma management practices in the telephone interview, stratified by disease severity. The proportions for patients reflect their parents' retrospective reports of how the asthma was managed before the adverse outcome. Among parents of children with severe asthma, slightly more than half of both patients and control subjects reported that their physician was “very involved” in the child's asthma care. Most children with severe asthma had nebulizers at home. However, among patients with severe asthma, only 25% had attended an asthma education class, 49% had a written asthma plan, and only 23% of those ≥7 years of age reported using a peak flow meter for symptoms. Approximately one third of parents of patients and control subjects said that at least one household member had smoked during the past year.
ED Visits: Multivariate Models
Preliminary models identified level of education, parent-reported severity, number of ED visits for asthma in the past 6 months, and number of inhaled steroid prescriptions in the past 12 months as the risk adjusters to be entered in the final model. The association of each asthma management practice with case versus control status was evaluated individually, after controlling for these risk adjusters, in conditional logistic regression models. As Table 5 shows, cases were less likely to have attended an asthma education class, to have a written asthma management plan, to use a peak flow meter for symptoms, or to have a nebulizer. Their parents were less likely to report having started or increased asthma and/or cold medicine at the onset of a cold or flu and to rate the child's personal physician as very involved in asthma care. Contrary to intuition, cases were less likely than were control subjects to report that a household member or regular caretaker smoked. Patients were more likely to be in day care, to use a peak flow meter regularly, and to have an asthma specialist.
The final multivariate model (Table 6) showed that a higher paternal level of education was associated with a lower risk of an ED visit attributable to asthma (OR: 0.55; 95% CI: 0.36–0.84). After adjusting for level of education and severity measures, two parent management practices had significant independent associations with ED visits: first, having a written asthma management plan was associated with decreased risk (OR: 0.45; 95% CI: 0.27–0.76). A second significant predictor was the parent's behavior when the child first developed cold or flu symptoms. Children whose parents gave them both asthma and cold medications had a lower risk than did children whose parents did not increase or start any medication (OR: 0.40; 95% CI: 0.19–0.86). However, administering cold medication alone appeared associated with a lower risk than administering asthma medication alone. Finally, children in regular group day care had a higher risk of an asthma-related ED visit (OR: 1.85; 95% CI: 0.90–3.78).
This study found that practices that promote early intervention for asthma flare-ups by parents at home were associated with reduced risk of adverse outcomes. In this HMO population, children who had a written asthma management plan were half as likely to have a hospitalization and or an ED visit than those who lacked a plan. In addition, those whose parents gave medication at the onset of cold symptoms were less likely to make ED visits.
Clinical and Research Context
This analysis was unique in comparing specific parent management practices between children with and without adverse asthma outcomes in a broad population. Our results are in accord with those of previous studies of selected populations. Early intervention practices for asthma flare-ups were identified as important in a study of an asthma education program for parents of young children.12Early augmentation of home therapy for flare-ups was associated with reduced length of hospitalization in another recent study.13
Initial management of asthma flare-ups by parents at home is a relatively recent phenomenon, the potential of which to reduce morbidity may not yet be maximized.14 The finding that fewer than half of patients in this population had written asthma management plans is similar to that of studies in other settings.15-17 Parents who lack specific criteria for seeking emergency care tend to have higher use.18
This study was relatively free of confounding by financial barriers to health care, because all subjects were insured by a fully integrated HMO. The results generalize primarily to children with good access to primary care. This population was relatively well educated and may thus have been better able to benefit from written asthma management plans than less literate populations. In this population, race/ethnicity was not a significant independent predictor of adverse outcomes, although education and income were. These findings support the belief that elevated asthma hospitalization risk among black children relative to white children is largely attributable to socioeconomic status and/or access to care.4,5,19
This study's results should be interpreted carefully, because its design could not completely control for biologic severity of illness. Severity was measured by parent report, previous use of medications, outpatient visits, ED visits, and hospitalizations. Our study could not optimally evaluate medication underprescribing as a risk factor for adverse outcomes because medication use served as a risk adjuster. Cases were in fact more likely than were control subjects to have recently received bronchodilator prescriptions or inhaled steroids. These findings likely occurred because children with more severe asthma receive more medications, and other severity measures did not completely control this confounding.
A previous study suggested that variation in asthma hospitalization rates in three cities in the Northeast was in part attributable to differences in outpatient prescribing patterns.6Compared with the previous study, higher proportions of our patients had been prescribed preventive medications before hospitalization. The publication of national guidelines in 1991 may have promoted these prescribing practices.8 The HMO's medical group adopted similar guidelines in January 1995, but these would have had limited time to affect practice during the study period.
The association we observed between home nebulizer use and increased hospitalization risk might also be attributable to confounding by severity of illness, but we cannot rule out a causal association. Home nebulizers are overused by some parents.20 Overreliance on nebulized β-agonists might delay other protective actions such as initiation of oral steroids and increase hospitalization risk. Thus, this association deserves additional study.
Any retrospective study based on interviews may have recall bias. In this study, parents of cases were interviewed after an adverse event and could have been biased toward overestimating their child's severity of illness or reporting that they followed desirable management practices before the adverse event. However, either of these biases would have made it more difficult for us to identify differences between cases and control subjects in asthma management practices. The potentially protective practices in this study were identified despite this bias rather than because of it.
This study's observational design a precludes drawing causal inferences between the predictor variables and adverse outcomes. The significant practices we identified could be markers or intermediaries for other factors not studied explicitly. For example, having a written asthma management plan may reflect a parent who is highly motivated to prevent asthma adverse outcomes or a clinician who has adopted other modern management practices. Washing of bedsheets in hot water twice a month is not considered by allergists to be the optimal intervention against dust mite allergens and instead may be a proxy for attentiveness to other environmental control or general adherence to physician recommendations.
Conversely, the lack of an independent association between other practices and adverse outcomes in this study does not mean these practices are ineffective. For example, asthma education programs have been shown to reduce morbidity in several randomized controlled trials.12,21-23 The lack of significance of having attended an asthma education class in our study was most likely attributable to unmeasured confounding. If having attended a class were associated with other important predictors, this would have decreased the precision of the analysis to estimate its independent effects. In fact, asthma education classes are an important means of promoting the practices we identified as potentially protective.
Improving parent asthma management practices will most likely require initial efforts by clinicians, who can provide written asthma management plans and recommend medications to be initiated or increased for flare-ups. Our results may overestimate the proportion of parents who have useful written plans, because some parents may have counted ED visit or hospital discharge instructions as written plans, as opposed to the structured written management plans recommended by national guidelines.8 However, several of the medical centers in this study used highly structured plans for the home management of asthma that were developed and preprinted by KPMCP regionally, then tailored by clinicians for individual patients. (Copies of these are available from the authors on request.) Our findings suggest that prospective, controlled evaluations of alternative interventions to enhance parents' ability to manage asthma using medications and environmental control (eg, case management, individualized or group teaching, or additional instruction by clinicians) are warranted.
This project was funded by the Kaiser Permanente Innovation Program. We thank Sandra Wilson, PhD, of the American Institutes for Research, for providing the questionnaire from the Wee Wheezers study.
We thank Harold Farber, MD, who designed the Northern California Kaiser Permanente Asthma Self-management Plan, and Charlie Homer, MD, for thoughtful reviews of the draft manuscript. We thank Steve Black, MD, Robert Fields, PharmD, and the members of the Northern California Kaiser Permanente Pediatric Asthma Best Practices Committee for helpful suggestions on study design. This study relied on the excellent interview work of Michelle Bunker, Cynthia Eagleton, and Lynn Simonson.
- Received December 4, 1996.
- Accepted February 17, 1997.
Reprint requests to (T.A.L.) Division of Research, The Permanente Medical Group, 3505 Broadway, Oakland, CA 94611.
- ED =
- emergency department •
- KPMCP =
- Kaiser Permanente Medical Care Program •
- HMO =
- health maintenance organization •
- OR =
- odds ratio •
- CI =
- confidence interval
- Halfon N,
- Newacheck PW
- Homer CJ,
- Szilagyi P,
- Rodewald L
- ↵National Asthma Education Program. Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: US Department of Health and Human Services, National Institutes of Health Publication No. 91–3042; 1991
- Conway SP,
- Littlewood JM
- Bendefy IM
- Fireman P,
- Friday GA,
- Gira C,
- et al.
- Lewis CE,
- Rachelefsky G,
- Lewis MA,
- et al.
- Copyright © 1997 American Academy of Pediatrics