Assessing Future Need for Acute Care in Adult Asthmatics

CHEST

Original Research ASTHMA

Assessing Future Need for Acute Care in Adult Asthmatics* The Profile of Asthma Risk Study: A Prospective Health Maintenance Organization-Based Study Molly L. Osborne, MD, PhD, FCCP; Kathryn L. Pedula, MS; Mark O’Hollaren, MD; Kenneth M. Ettinger, MD; Thomas Stibolt, MD, FCCP; A. Sonia Buist, MD; and William M. Vollmer, PhD

Study objectives: To develop simple clinical tools predictive of acute asthma care and to identify modifiable risk factors. Design: Prospective cohort study. Setting: A large health maintenance organization (430,000 members). Patients/participants: Adult members (18 to 55 years old) with asthma. Interventions: Data from a questionnaire, skin-prick testing for inhalant allergens, and spirometry were collected at the baseline visit. Acute care utilization data were obtained from administrative databases for a subsequent 30-month period. Methods: This two-phase study first identified and performed a split-sample validation on three clinical tools to determine their predictive ability by employing data from a questionnaire, questionnaire plus spirometry, and questionnaire plus spirometry and skin-prick testing. Second, it identified modifiable independent risk factors. Measurements and results: The 554 study participants generated 173 episodes of acute care over 1,258 person-years of follow-up (0.14 episodes per person per year). Of these, 101 participants had at least one episode, and one third of this group had two or more episodes. Clinical scoring into risk groups was done by reverse stepwise regression analyses. Using relative risks (RRs) as a guide, high-risk, moderate-risk, and low-risk groups were identified. The high-risk groups, 13 to 21% of the validation sample, had a 7- to 11-fold increased risk for hospital care compared to the low-risk groups. The moderate-risk groups, 46 to 50% of the validation sample, had a twofold- to fourfold-increased risk. FEV1 was the most significant predictor (RR, 4.33). Of the four potentially modifiable risk factors identified, current cigarette smoke exposure (RR, 1.6) and ownership and skin-prick test positivity to cat or dog (RR, 1.5) were the most significant. Conclusions: These models stratify asthma patients at risk for acute care. Patients with lower FEV1 values are at significantly higher risk, underscoring the importance of spirometry in asthma care. (CHEST 2007; 132:1151–1161) Key words: asthma; health-care utilization; health maintenance organization; spirometry; tobacco Abbreviations: ETS ⫽ exposure to environmental tobacco smoke; %FEV1 ⫽ percentage of predicted FEV1; HMO ⫽ health maintenance organization; ICS ⫽ inhaled corticosteroid; KPNW ⫽ Kaiser Permanente Northwest; PAR ⫽ predictors of asthma risk; RR ⫽ relative risk

of the pathophysiology and treatO urmentunderstanding of asthma has improved dramatically in recent years.1,2 Potent, effective asthma treatment is readily available, as are tools to objectively follow these patients.3,4 Properly managed asthma patients should rarely need emergency department treatment or hospitalization. Dewww.chestjournal.org

spite these advances, patients and health-care providers are often surprised by asthma exacerbations requiring emergent management.5 Our inability to more accurately identify patients at higher risk of acute asthma exacerbations has grave consequences for both patients and the increasingly burdened health-care system. CHEST / 132 / 4 / OCTOBER, 2007

1151

Nearly half the $7.4 billion spent annually on direct medical expenditures for asthma care is spent on hospitalizations, visits to the emergency department, and hospital outpatient care (ie, unscheduled outpatient care).6 – 8 Concerns regarding increasing For related article see page 1162 For editorial comment see page 1112 asthma morbidity and cost have led to a national objective of reducing health-care utilization for patients with asthma.9 Factors known to be associated with increased risk of future acute asthma health-care utilization include advanced age, female gender, a reduction in percentage of predicted FEV1 (%FEV1), and current cigarette smoking.10 –15 Simple risk indexes for asthma that include variables that can be readily administered and scored in the clinical setting are scarce,16 however, and do not include variables such as lung function. This article reports on the results of a prospective evaluation of risk factors for acute asthma care among adult members of a large health maintenance organization (HMO). We had two goals. The first was to develop clinical tools that clinicians and health-care organizations could use to identify patients at increased risk of future acute asthma exacerbations, and perform a split-sample validation to determine their predictive ability. The second was to identify modifiable risk factors. Our results can be used by clinicians to monitor patients more closely to determine the need for referral for case management or specialty care.

Table 1—Characteristics of Participants (n ⴝ 554)* Characteristics

Data

Mean age ⫾ SD, yr Female gender White race Household income, $ ⬍ 30,000 30–39,999 40–49,999 50–59,999 ⱖ 60,000 Smoking status Current smoker Former smoker Never-smoker ETS Acute episodes during 30-month follow-up 0 1 2 ⱖ3

39.6 (9.3) 337 (60.8) 520 (93.9) 104 (18.8) 116 (20.9) 93 (16.8) 100 (18.1) 140 (25.3) 61 (11.0) 165 (29.8) 325 (58.7) 223 (40.3) 453 (82) 66 (12) 14 (2.5) 21 (3.7)

*Data are presented as No. (%) unless otherwise indicated.

Study Population and Research Setting Persons in our study population were members of Kaiser Permanente Northwest (KPNW). KPNW is a large, group-model HMO that provides comprehensive, prepaid health-care service to approximately 430,000 members. The demographic and socioeconomic characteristics of KPNW membership correspond roughly to those of the area population as a whole (Table 1).18 To be eligible for inclusion in the study, KPNW members had to have been hospitalized for asthma during the 2 years before recruitment or have at least two dispensings of antiasthma medication in the year before recruitment. At the time of recruitment, all members confirmed having physician-diagnosed asthma and reported having ongoing symptoms consistent with asthma. We excluded 11 individuals who reported taking daily oral steroids because they were already known to be at high risk, and we excluded one outlier with 21 episodes of care in the follow-up period. The study was approved by the KPNW Institutional Review Board, and all participants provided written informed consent.

Methods and Materials The study methods and characteristics of the population have been described in detail elsewhere17 and are summarized here. *From the Oregon Health and Science University (Drs. Osborne and Buist); Kaiser Permanente Center for Health Research (Ms. Pedula and Dr. Vollmer), Kaiser Permanente Northwest (Drs. Stibolt and Ettinger); and The Allergy Clinic (Dr. O’Hollaren), Portland, OR. Supported by National Institutes of Health grant HL #48237; the American Lung Association of Oregon; VA Foundation acct #279999; and the Allergy Clinic, Portland, OR. This study was conducted at Kaiser Permanente Northwest, Portland OR. The authors have no conflicts on interest to disclose. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Molly Osborne, MD, PhD, FCCP, Mail code L102, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd, Portland, OR 97239; e-mail: osbornem@ ohsu.edu DOI: 10.1378/chest.05-3084 1152

Study Design In this prospective cohort study, participants were followed up for 30 months. Data from a questionnaire, skin-prick testing, and spirometry were collected at the baseline visit. Questionnaires The baseline questionnaire was based on the American Thoracic Society-Division of Lung Disease 1978 respiratory symptom questionnaire, the International Union Against Tuberculosis and Lung Disease bronchial symptom questionnaire,19,20 and the National Asthma Education and Prevention Program expert panel report.21 Items assessed included respiratory symptoms, characteristics of asthma, demographic factors, tobacco use, allergen exposure, medication use, and prior acute asthma care. Indicators of Exposure to Cigarette Smoke Our omnibus measure of “cigarette smoke exposure” included current smoking or secondhand smoke exposure, as described Original Research

below. “Current smoker” was defined as smoking as of a month ago. “Exposure to environmental tobacco smoke” (ETS) was defined as regular exposure to other people’s tobacco smoke in the last 12 months. “Cigarette exposure on the job” was defined as exposure to cigarette smoke of others most of the time while at work. “Ever-smoker” was defined as having smoked at least 20 packs of cigarettes in a lifetime or at least one cigarette per day for a year. Allergen/Irritant Exposure Participants were asked about the presence of dogs and cats at home, visible mold or mildew indoors, use of double-pane windows, and types of floor covering and upholstery. Specifically, “sensitive to indoor allergens” required a positive response to any of the following: “When near animals, feathers or dust, do you cough, wheeze, feel tightness in the chest, start to feel short of breath?” We asked about double-pane windows with the rationale that they might offer protection from asthma by decreasing condensation and possibly lowering indoor mold concentrations. Participants were also asked about occupational exposure to solvents, fumes, dusts, and gases. Medication Use Because information about medication use reflects medical management more than personal characteristics or exposures, we deliberately chose to not include reported medication use as a predictor variable in the analyses. We focused instead on information that might not otherwise be readily known by the physician. In the population studied, the prescription of inhaled corticosteroid (ICS) was itself expected to be a marker of disease severity. Since adherence with ICS treatment is known to be generally poor in patient populations and highly variable among individual patients, including ICS prescriptions as a marker of risk could confound the analysis. We also recognized that a randomized clinical trial design would have been the optimal way to document efficacy of specific medications, rather than an observational study such as this. In our study population, in the year prior to recruitment, 93% used a ␤-agonist, 54% used ICS, 42% had a course of “burst” corticosteroids, and ⬍ 25% needed an oral aminophylline preparation, an anticholinergic agent, or cromolyn. Spirometry and Use of Metered-Dose Inhaler Spirometry was performed at the baseline visit prior to and 5 min after inhalation of two puffs of isoproterenol using standard methods.22 Prediction equations of Knudson and colleagues23 were used to calculate %FEV1. Asthma severity was categorized as severe (%FEV1 ⬍ 60%), moderate (%FEV1, 60 to 80%), or mild (%FEV1 ⬎80%), according to National Asthma Education and Prevention Program guidelines.21 Postbronchodilator FEV1 measurements were included primarily to confirm asthma, whereas prebronchodilator FEV1 was used in the risk models. Skin-Prick Testing We conducted skin-prick testing using 13 inhalant allergens appropriate for the Pacific Northwest17,24: alder, birch, juniper, grass, western weed, cat, dog, mite (Dermatophagoides pteronyssinus and Dermatophagoides farinae), alternaria, cladosporium, aspergillus, and pencillium. Follow-up and Outcome Assessment Acute care utilization data were obtained from administrative databases for the 30-month period following baseline evaluation. Using these data, we defined episodes of acute care as one or www.chestjournal.org

more emergency department visits, hospital-based “urgency care clinic” visits, or hospitalizations for asthma.12,25 (Acute care delivered at non-KPNW facilities is reimbursable and is recorded in a claims database.) Visits separated by ⬎ 2 days were counted as separate episodes of acute care. In total, 101 participants had at least one episode requiring acute care, and at least a third of these had two or more episodes. Specifically, 453 participants had no episodes, 66 had a single episode, 14 had two episodes, and 21 had three or more episodes. The total number of episodes served as the primary outcome variable. Person-years of observation were calculated as the length of health plan eligibility during the 30-month follow-up period. Follow-up times ranged from 1 to 30 months, with an average of 27.2 months and a median of 30.0 person-months. Eighty-two percent of the sample was followed up for the full 30 months, and only 7% had ⱕ 12 months of follow-up. Statistical Methods The clinical scoring rules were developed in a three-stage process. Initially, we used the entire sample to develop a series of multivariate “epidemiologic” models that predicted the probability of future hospital-based care as a function of baseline information. Starting with variables that had p values ⱕ 0.20 in univariate analyses, we performed reverse stepwise regression analyses to arrive at a final model in which all variables were significant at p ⱕ 0.05. All models were fit using Poisson regression analysis26 using statistical software (SAS, version 8.2; SAS Institute; Cary, NC). The Poisson model is ideally suited to the analysis of count data and has the added advantage of being able to incorporate varying follow-up among participants. It directly models the incidence per unit time and provides relative risk (RR) estimates. The literature27 suggests there should be minimal bias because there are 10 to 20 outcome events for each predictor variable (on the order of 14:1). In order to maximize the clinical utility of the results, we constructed three separate models, which we call profile of asthma risk (PAR) [shown in Appendix]. The first model, PAR A, uses questionnaire data as a potential predictor. The second model, PAR B, uses questionnaire and spirometry data. The third model, PAR C, uses questionnaire, spirometry, and skin-prick test data. These models reflect the types of information clinicians might have available to them, depending on their specialty and practice setting. Because these epidemiologic models do not immediately lend themselves to clinical use, we simplified them so they could be readily used in the clinical setting to discriminate between patients who are at low, moderate, and high risk for subsequent acute exacerbations. Using the RRs as a guide, we assigned integer scores to the various factors in the models (Table 2) and summed these to arrive at an overall score for which higher values denote greater risk. Finally, we developed cut points for the overall score for each model that can be used to classify patients into low-, medium-, and high-risk categories. For this stage, we first randomly classified subjects into a “test sample,” consisting of 60% of subjects (n ⫽ 332), and a “validation sample,” comprised of the remaining 40% (n ⫽ 222). Using the test sample, we determined the cut points for defining low-, medium-, and high-risk patients in order to maximize the separation between groups in terms of their subsequent risk of hospital-based care. We then used the validation sample to provide a more unbiased estimate of the true predictive value of these cut points for each of the models. In developing the epidemiologic and clinical models, we deliberately did not constrain them to be hierarchical to one CHEST / 132 / 4 / OCTOBER, 2007

1153

Table 2—Summary of Multivariate Poisson Regression Models* PAR Model A: Questionnaire Data Only

PAR Model B: Questionnaire and Spirometry Data

PAR Model C: Questionnaire, Spirometry, and Skin-Prick Test Data

Factors†

RR

95% CI

RR

95% CI

RR

95% CI

Age Education‡ Double-pane windows in bedroom Caffeine consumption Sensitive to indoor allergens§ Owns a cat or dog Owns and is skin-prick test positive for cat or dog Nightly nighttime symptoms Perennial (as opposed to seasonal) asthma Impact of asthma on work/school attendance储 Saw a physician for breathing problems in the past year Ever seen in urgent care or the ER for breathing problems Ever hospitalized for asthma %FEV1 60 to 80%¶ %FEV1 ⬍ 60%¶

0.98 0.57 0.71 1.16 2.07 1.66

0.97–1.00 0.43–0.77 0.52–0.97 1.01–1.33 1.17–4.08 1.16–2.43

0.97 0.63

0.95–0.98 0.47–0.84

0.97 0.57

0.95–0.98 0.43–0.76

1.15 1.97 1.71

1.01–1.33 1.11–3.88 1.19–2.50

1.19 1.92

1.04–1.37 1.08–3.77

1.61

1.19–2.18

1.99 1.78 1.45 1.78

1.40–2.80 1.15–2.87 1.16–1.80 1.22–2.66

1.53 1.94

1.23–1.90 1.32–2.91

1.57 2.02

1.27–1.94 1.38–3.04

3.36

1.81–6.98

3.16

1.70–6.54

3.63

2.00–7.41

1.67

1.21–2.31

1.42 2.43 4.33

1.02–1.97 1.59–3.65 2.94–6.39

2.47 4.61

1.63–3.72 3.16–6.77

*RRs are adjusted for all other factors in the model. All parameter estimates significant at p 0.05. CI ⫽ confidence interval. †All factors coded yes/no unless otherwise specified. ‡Education coded as follows: (1) less than high school diploma; (2) more than high school diploma but less than 4 yr of college; and (3) 4 yr of college or more. §When near animals, feathers, or dust, do you cough, wheeze, feel tightness in the chest, or start to feel short of breath? 储Response to question: “Is your work or school attendance affected by asthma?” Original response categories (which were used in the research model) were as follows: (1) “not at all”; (2) “a little”; (3) “usually”; and (4) “always.” ¶Risks compared relative to those with %FEV1 ⬎ 80%.

another. Rather, our intent was to develop the best-fitting models given the three different sets of available information. Formal statistical comparison of goodness-of-fit between the models is therefore not possible. We also developed a “modifiable risk factors” model that could be used by clinicians to advise patients on how to reduce risk of an acute episode. We fit a model with questionnaire data and skin-prick test data and excluded prior health-care utilization variables. A two-tailed p value ⬍ 0.05 was used to define statistical significance in the analyses (Table 3). All models were fit using statistical software (SAS, version 8.2; SAS Institute; Cary, NC).

Results The 554 study participants were predominantly white and never-smokers, more than half were women Table 3—Summary of Modifiable Risk Factors Model* Factors Double-pane windows in bedroom Current cigarette smoke exposure Regular workplace exposure to solvents Owns and is skin-prick test positive for cat or dog

RR

95% CI

0.632 1.579 1.443 1.490

0.463–0.866 1.154–2.171 1.010–2.026 1.094–2.024

*RRs are adjusted for all other factors in the model. See Table 2 for expansion of abbreviation. 1154

(61%), and had a median annual income ⬍ $50,000 (Table 1). Median age of the participants was 41 years. We grouped education into grades 0 to 8 (n ⫽ 1, 0.2%), grades 9 to 12 (n ⫽ 102, 18.4%), some college (n ⫽ 225, 40.6%), and college graduate (n ⫽ 226, n ⫽ 40.8%). ETS during the past 12 months was common (40.3%). There were 173 episodes of acute care over 1,258 person-years of follow-up, for an overall rate of 13.8 episodes per 100 person-years of follow-up, or 0.14 episodes per person per year. Table 2 summarizes the best-fitting models for each of the three sets of predictor variables: PAR A, PAR B, and PAR C. Younger age, better lung function, and more education were independent predictors of lower risk of acute care. Relative to those with %FEV1 ⬎ 80%, those with %FEV1 of 60 to 80% were at roughly a 2.5-fold–increased risk for future acute episodes, and those with %FEV1 ⬍ 60% were at a more than fourfold-increased risk. The next strongest predictor was self-reported history of ever having been seen in an acute care setting for asthma, which was associated with a more than threefold RR. The extent to which breathing problems affected work or school attendance, whether the patient saw a physician for breathing problems in the past year, and prior Original Research

hospitalization for asthma were all independently associated with risk for future acute care episodes. In addition to the health-care utilization variables, self-reported sensitivity to indoor allergens was significant in all three models, with RRs ranging from 1.9 to 2.1. Owning a cat or dog was associated with about a 70% increased risk, similar to the risk for those who were also skin-prick test positive to whatever type of pet they owned. The presence of double-pane windows in the bedroom was protective, although the latter effect was no longer significant after adjusting for level of lung function. Similarly, reports of nightly nocturnal symptoms were associated with increased risk but dropped out of the models after adjusting for lung function. Finally, caffeine intake, defined as cups of caffeinated beverage consumed per day, was associated with a slightly increased risk in all three models. As described in “Material and Methods,” we used the results from Table 2 to develop simple clinical scoring rules based on each model. The score for PAR A uses nine questions and ranges from 0 to 10. The score for PAR B uses four questions plus prebronchodilator lung function and also ranges from 0 to 10. The score for PAR C uses five questions plus both prebronchodilator lung function and skin-prick testing and ranges from 0 to 11. In each case, we used data from the test sample to define cut points for these overall scores that classify people into low-, medium-, and high-risk categories. Figure 1 summarizes the predictive value of the resulting clinical models based on data from the validation sample. For each model, we see a striking gradation in risk of future hospital-based asthma care in going from low to medium to high risk. PAR A divides the validation population into a low-risk group, approximately 31% of the sample; a medium-risk group, approximately 48%; and a high-

risk group, approximately 21%. Relative to the lowrisk group, the medium- and high-risk groups had RRs of acute care of 4.2 and 6.2, respectively, with an absolute risk of 3.6 episodes per 100 person-years for the low-risk group, and 26.4 episodes per 100 person-years for the high-risk group. For the validation sample, 7% of those classified as low risk had one or more episodes, as compared to 22% of those classified as moderate risk and 30% of those classified as high risk (Fig 1). For PAR B, the low-, medium-, and high-risk groups constitute 38%, 49%, and 13% of the validation sample, respectively, with the latter two groups having validated RRs of 3.0 and 10.9 compared to the low-risk group and an overall separation from low to high risk of 1.7 to 54.0 episodes per 100 person-years. For PAR C, the low-, medium-, and high-risk groups constitute 41%, 46%, and 13% of the validation sample, respectively, with the latter two groups having validated RRs of 2.1 and 9.5 compared to the low-risk group, and an overall separation from low to high risk of 5.1 to 54.0 episodes per 100 personyears. For groups B and C, for the validation sample, approximately 9% of those classified as low risk had one or more episodes, as compared to 17% of moderate-risk patients and 48% of high-risk patients (Fig 1). While the models summarized in Table 2 provide the best predictive power, they do not necessarily provide the clinician with insight into potentially modifiable risk factors for a given patient. We therefore also fit a model with questionnaire data only and excluded prior health-care utilization variables (Table 3). Four modifiable risk factors were identified. One risk factor, double-pane windows, was protective. The other three factors—“current cigarette smoke exposure,” “regular workplace exposure to solvents,” and “skin-prick test positivity with ownership of a cat or dog”— carried equal risk.

Discussion

Figure 1. Percentage of individuals with hospital-based care during follow-up as a function of risk strata for each of PAR models A, B, and C. Data are shown for individuals in the validation sample. www.chestjournal.org

We successfully developed three simple clinical models using independent risk factors suitable for use in the clinical setting that stratify adult asthma patients into risk groups. We identified and validated the models to determine their predictive ability. The high-risk groups, 13 to 21% of the validation sample populations, were at roughly 7- to 11-fold–increased risk for acute care compared to the low-risk groups. The moderate-risk groups, 46 to 50% of the validation sample, were at twofold to fourfold increased risk. Importantly, airflow obstruction (FEV1) was the most significant predictor of subsequent acute care. CHEST / 132 / 4 / OCTOBER, 2007

1155

This result underscores the importance of obtaining spirometry data to identify patients at risk. We separately analyzed the sample according to modifiable risk factors for acute asthma care and identified four independent risk factors. Current cigarette smoke exposure was identified as the strongest modifiable risk factor. This study builds on and extends previous studies that identify factors associated with acute asthma. The primary strengths of this prospective longitudinal study include the outpatient nature of the sample, the comprehensive baseline evaluation (including spirometry and skin-prick testing), the relatively large sample size, and the length and completeness of follow-up. The 544 study subjects contributed 1,258 person-years of follow-up and 173 distinct episodes of acute care during the 30-month follow-up period. Other strengths include the definition of asthma by doctor’s diagnosis, ability to test a large number of environmental and demographic features likely to influence risk of acute asthma, and the ability to document acute asthma in a large HMO. In addition, the large number of events allows us to identify independent risk factors that predict acute care. The PAR models are among only a few that enable clinicians to define the RR of acute care in a variety of clinical settings. This study emphasizes the strong predictive value of routine spirometry for characterizing asthma risk in an outpatient setting in adults, as has been demonstrated in children.28 This is highly relevant because a variety of national and international guidelines29 –31 recommend that lung function be regularly assessed as a part of routine asthma care. Our finding that a reduced FEV1 was associated with a significantly higher risk of subsequent acute asthma exacerbations reinforces the importance of routine lung function measurement as recommended by these guidelines and extends recent work by Kitch et al.13 Our PAR models have several advantages, including ease of use, available data on prospective properties, risk separation, and use of health-care utilization as the outcome. Several other models have been shown to be easy to use and practical in the clinical setting.32,33 Other models use administrative databases not easily available to clinicians11 or address physician-assessed control rather than health-care utilization as the primary end point.34,35 PAR A is useful when only questionnaire data are available, and PAR B (questionnaire plus spirometry) offers additional risk stratification, underscoring the importance of spirometry in asthma care. In light of the association of sensitization to specific allergens and life-threatening asthma, we had anticipated that specific allergen exposure and sen1156

sitivity would be associated with health-care utilization, and it was associated in the unvariate and multivariate analyses. However, in estimating the risk of health-care utilization in individuals, PAR B and C give similar risk estimates for low-, moderate-, and high-risk individuals (Fig 1). These findings do not minimize the importance of identification of aeroallergens that may trigger symptoms in patients with asthma. The utility of skin-prick testing is well established in patients with asthma. Another asthma risk assessment model, the Asthma Therapy Assessment Questionnaire model developed by Vollmer et al,12 (a simple index of number of asthma control problems) was prospectively validated in a large cohort and found to correlate with clinically significant impairment. It does not incorporate data potentially available to a clinician, however, such as spirometry and skin-prick testing. One advantage of our clinical models is better risk separation. The Asthma Therapy Assessment Questionnaire identifies a group at a threefold to fourfold risk, whereas our models have significantly stronger predictive power, particularly when spirometric data are incorporated. Although not yet validated, another model, the Asthma Control Test, is a useful five-item questionnaire that has been shown to correlate with FEV1 as well as specialist-assessed asthma severity and specialist-assessed need for a change in asthma therapy. Identification of modifiable risk factors that independently contribute to acute care in asthma is also clinically relevant. Our study is one of many that demonstrate the risks of smoking.36,37 Cigarette smoking was the major independent modifiable risk factor PAR associated with acute asthma exacerbations. This finding is consistent with studies demonstrating a high prevalence (35%) of current smoking in adults presenting to emergency departments with acute asthma,14 compared with the 24% prevalence rate in United States adults.14 Although the mechanism is not known, smoking may well modify the immunologic response in asthma,38 as well as reduce the response to corticosteroids,39 increasing symptoms and disability. Of note, smoking was not a significant variable in the clinical models, perhaps because the effects of smoking were taken into account by health-care utilization and lung function variables. Our study also specifically identifies exposure and sensitivity to cat or dog as an independent risk factor. These findings are consistent with results from other studies40 – 42 demonstrating that the combination of allergen exposure and sensitivity predicts hospitalization for asthma. Of interest, this association has not been clearly demonstrated in children.43 Our finding regarding solvents is consistent with studies44 – 46 demonstrating the risk of exposure in the workplace. Original Research

We do not know the source of the protective effect of double-pane windows, but it may arise from decreased exposure to fungal spores by decreasing window condensation and mold growth,47 or it may be a marker for the age of the home. Although an inverse association of double-pane windows (double glazing) has been found with asthma symptoms,46 this was not significant. All of these potentially modifiable risk factors can be discussed with at-risk patients to their benefit. Indeed, a one study48 demonstrated a reduction in asthma-associated morbidity with use of an individualized, home-based environmental intervention. Although we did not analyze FEV1 as a modifiable risk factor, it is important to mention that FEV1 can be at least indirectly modified (eg, it can improve after use of an antiinflammatory medication). Study limitations include study demographics and setting, potentially reducing generalizability of results. Minorities accounted for only 6% of the study population, reflecting the demographics of the Portland, OR, metropolitan region. Our age range of 18 to 55 years may mean that findings may not extrapolate to children and older adults. We restricted the upper age limit to 55 years to avoid misclassification with COPD. Additionally, HMO care is prepaid, so economic factors would likely not play a significant role in decision making. Notwithstanding limitations, we believe it is likely that the major risk factors for acute care identified in this study—reduced lung function and prior treatment in the acute care setting—are robust predictors. We recognize that a relatively small proportion of individuals tends to drive the majority of health-care costs for asthma, as for many chronic diseases. In the statistical analysis, the relative predictive value of individual risk factors was naturally more heavily influenced by those with more-frequent asthma episodes. The model selectively identifying such individuals may be, if true, advantageous for ensuring that they receive needed care. Furthermore, modifiable risk factors, such as exposure to cigarette smoke and pets, are likely to be generalizable.

www.chestjournal.org

Although we used a split-sample approach to validate the predictive ability of the clinical models, the fact that the epidemiologic models from which they were derived were constructed using the full sample may mean that the estimates of predictive value in Table 3 may be somewhat optimistic. Further studies are needed to independently validate these results in other populations. We deliberately chose not to include reported medication use as a predictor in the PAR models, for several reasons. Excluding medication use helped us develop a model independent of changes in guidelines and patient compliance. For instance, when faced with a high-risk patient who was already prescribed ICS, a provider might choose to increase the dose of medication, add an additional medication, or discuss possible adherence issues with the patient. A high-risk patient who was not already receiving these medications might be prescribed inhaled steroids. Our models can serve as tools to alert providers to patients who may need closer scrutiny during the clinic visit. They can also serve as the basis for beginning a discussion with the patient about his/her asthma (eg, “I see you are waking up at night from your asthma”). Future studies in which strict adherence to asthma medications is monitored could further evaluate the therapeutic value of the PAR models. In summary, we have identified important modifiable risk factors for asthma exacerbations, including current cigarette smoke exposure. We also have developed three clinical PAR models that can easily be used by the clinician to identify patients at risk, and we confirmed the predictive ability of the model. Importantly, we have demonstrated the independent contribution of baseline lung function in predicting future asthma exacerbations. Further testing of the PAR models is needed to determine if preemptive intervention in high-risk patients will reduce emergent health-care utilization. Our hope is that the PAR models may be widely used to identify patients who may need additional attention in order to prevent a serious exacerbation of asthma.

CHEST / 132 / 4 / OCTOBER, 2007

1157

Appendix

1158

Original Research

www.chestjournal.org

CHEST / 132 / 4 / OCTOBER, 2007

1159

References 1 Sears MR, Greene JM, Willan AR, et al. A longitudinal, population-based, cohort study of childhood asthma followed to adulthood. N Engl J Med 2003; 349:1414 –1422 2 Tobin MJ. Asthma, airway biology, and nasal disorders in AJRCCM 2003. Am J Respir Crit Care Med 2004; 169:265–276 3 NAEP Expert Panel. NAEP Expert Panel report 2: guidelines for the diagnosis and management of asthma; update on selected topics 2002. Bethesda, MD: National Institutes of Health, 2002; publication 02-5075 4 Bateman ED, Boushey HA, Bousquet J, et al. Can guidelinedefined asthma control be achieved? The Gaining Optimal Asthma Control Study. Am J Respir Crit Care Med 2004; 170:836 – 844 5 Zeiger RS, Heller S, Mellon MH, et al. Facilitated referral to asthma specialist reduces relapses in asthma emergency room visits. J Allergy Clin Immunol 1991; 87:1160 –1168 6 Weiss KB, Sullivan SD, Lyttle CS. Trends in the cost of illness for asthma in the United States, 1985–1994. J Allergy Clin Immunol 2000; 106:493– 499 7 Weiss KB, Gergen PJ, Hodgson TA. An economic evaluation of asthma in the United States. N Engl J Med 1992; 326:862– 866 1160

8 Weiss KB, Sullivan SD. The health economics of asthma and rhinitis: i. Assessing the economic impact. J Allergy Clin Immunol 2001; 107:3– 8 9 US Public Health Service. Healthy people 2000: national promotion and disease prevention objectives; full report, with commentary. Washington, DC: US Department of Health and Human Services, Public Health Service, 1991; DHHS Publication (PHS) 91-50212. 10 Skobeloff EM, Spivey WH, St Clair SS, et al. The influence of age and sex on asthma admissions. JAMA 1992; 268:3437–3440 11 Lieu TA, Quesenberry CP, Sorel ME, et al. Computer-based models to identify high-risk children with asthma. Am J Respir Crit Care Med 1998; 157:1173–1180 12 Vollmer WM, Markson LE, O’Connor E, et al. Association of asthma control with health care utilization and quality of life. Am J Respir Crit Care Med 1999; 160:1647–1652 13 Kitch BT, Paltiel AD, Kuntz KM, et al. A single measure of FEV1 is associated with risk of asthma attacks in long-term follow-up. Chest 2004; 126:1875–1882 14 Silverman RA, Boudreaux ED, Woodruff PG, et al. Cigarette smoking among asthmatic adults presenting to 64 emergency departments. Chest 2003; 123:1472–1479 Original Research

15 Thomson NC, Chaudhuri R, Livingston E. Asthma and cigarette smoking. Eur Respir J 2004; 24:822– 833 16 Nathan RA, Sorkness CA, Kosinski M, et al. Development of the asthma control test: a survey for assessing asthma control. J Allergy Clin Immunol 2004; 113:59 – 65 17 Osborne ML, Vollmer WM, Linton KL, et al. Characteristics of patients with asthma within a large HMO: a comparison by age and gender. Am J Respir Crit Care Med 1998; 157:123– 128 18 Freeborn DK, Pope CR. Promise and performance in managed care. Baltimore, MD: Johns Hopkins University Press, 1994 19 Burney PG, Chinn S, Britton JR, et al. What symptoms predict the bronchial response to histamine? Evaluation in a community survey of the bronchial symptoms questionnaire (1984) of the International Union Against Tuberculosis and Lung Disease. Int J Epidemiol 1989; 18:165–173 20 Burney PG, Laitinen LA, Perdrizet S, et al. Validity and repeatability of the IUATLD (1984) Bronchial Symptoms Questionnaire: an international comparison. Eur Respir J 1989; 2:940 –945 21 National Heart, Lung, and Blood Institute, National Asthma Education Program. Guidelines for the diagnosis and management of asthma: expert panel report. J Allergy Clin Immunol 1991; 88:425–534 22 American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995; 152:1107–1136 23 Knudson RJ, Lebowitz MD, Holberg CJ, et al. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis 1983; 127:725–734 24 Ten RM, Klein JS, Frigas E. Allergy skin testing. Mayo Clin Proc 1995; 70:783–784 25 Vollmer WM, Osborne ML, Buist AS. Temporal trends in hospital-based episodes of asthma care in a health maintenance organization. Am Rev Respir Dis 1993; 147:347– 353 26 McCullagh P, Nelder JA. Generalized linear models. London, UK: Chapman and Hall, 1983 27 Stokes ME, Davis CS, Koch GG. Categorical data analysis using the SAS system. 2nd ed. Cary, NC: SAS Institute, 2000 28 Fuhlbrigge AL, Kitch BT, Paltiel AD, et al. FEV1 is associated with risk of asthma attacks in a pediatric population. J Allergy Clin Immunol 2001; 107:61– 67 29 Celli BR. The importance of spirometry in COPD and asthma: effect on approach to management. Chest 2000; 117(suppl):15S–19S 30 Global Initiative for Asthma. GINA report: global strategy for asthma management and prevention. Available at: http:// www.ginasthma.com/Guidelineitem.asp??|1⫽2&|2⫽1&intID⫽60. Accessed September 26, 2007 31 National Asthma Education and Prevention Program. Expert panel report 2: guidelines for the diagnosis and management of asthma. Betheada, MD: National Institutes of Health, 1997; publication 97-4051; 1– 4 32 Schatz M, Nakahiro R, Jones CH, et al. Asthma population management: development and validation of a practical 3-level risk stratification scheme. Am J Manag Care 2004; 10:25–32

www.chestjournal.org

33 Li D, German D, Lulla S, et al. Prospective study of hospitalization for asthma: a preliminary risk factor model. Am J Respir Crit Care Med 1995; 151:647– 655 34 Nathan RA, Sorkness CA, Kosinski M, et al. Development of the asthma control test: a survey for assessing asthma control. J Allergy Clin Immunol 2004; 113:59 – 65 35 Juniper EF, O’Byrne PM, Guyatt GH, et al. Development and validation of a questionnaire to measure asthma control. Eur Respir J 1999; 14:902–907 36 Gutzwiller F, La Vecchia C, Levi F, et al. Smoking, prevalence of disease and health service utilization among the Swiss population. Rev Epidemiol Sante Publique 1989; 37:137–142 37 Dales RE, Dionne G, Leech JA, et al. Preoperative prediction of pulmonary complications following thoracic surgery. Chest 1993; 104:155–159 38 Sunyer J, Basagana X, Roca J, et al. Relations between respiratory symptoms and spirometric values in young adults: the European Community Respiratory Health Study. Respir Med 2004; 98:1025–1033 39 Chaudhuri R, Livingston E, McMahon AD, et al. Cigarette smoking impairs the therapeutic response to oral corticosteroids in chronic asthma. Am J Respir Crit Care Med 2003; 168:1308 –1311 40 Gelber LE, Seltzer LH, Bouzoukis JK, et al. Sensitization and exposure to indoor allergens as risk factors for asthma among patients presenting to hospital. Am Rev Respir Dis 1993; 147:573–578 41 Sarpong SB, Karrison T. Sensitization to indoor allergens and the risk for asthma hospitalization in children. Ann Allergy Asthma Immunol 1997; 79:455– 459 42 Rosenstreich DL, Eggleston P, Kattan M, et al. The role of cockroach allergy and exposure to cockroach allergen in causing morbidity among inner-city children with asthma. N Engl J Med 1997; 336:1356 –1363 43 Gruchalla RS, Pongracic J, Plaut M, et al. Inner City Asthma Study: relationships among sensitivity, allergen exposure, and asthma morbidity. J Allergy Clin Immunol 2005; 115:478 – 485 44 Kilburn KH. Chronic disease in the workplace and the environment: lung disease: lessons from the past: keys to the future. Arch Environ Health 1984; 39:139 –143 45 Petsonk EL. Work-related asthma and implications for the general public. Environ Health Perspect 2002; 110(Suppl): 569 –572 46 Zock JP, Kogevinas M, Sunyer J, et al. Asthma characteristics in cleaning workers, workers in other risk jobs and office workers. Eur Respir J 2002; 20:679 – 685 47 Garrett MH, Rayment PR, Hooper MA, et al. Indoor airborne fungal spores, house dampness and associations with environmental factors and respiratory health in children. Clin Exp Allergy 1998; 28:459 – 467 48 Morgan WJ, Crain EF, Gruchalla RS, et al. Results of a home-based environmental intervention among urban children with asthma. N Engl J Med 2004; 351:1068 –1080

CHEST / 132 / 4 / OCTOBER, 2007

1161