Associations of Smoking With HospitalBased Care and Quality of Life in Patients With Obstructive Airway Disease* Jeffrey M. Sippel, MD; Kathryn L. Pedula, MS; William M. Vollmer, PhD; A. Sonia Buist, MD; and Molly L. Osborne, MD, PhD, FCCP

Study objectives: To investigate the relationship between direct or environmental tobacco smoke (ETS) exposure and both hospital-based care (HBC) and quality of life (QOL) among subjects with asthma. Study design: We report baseline cross-sectional data on 619 subjects with asthma, including direct or ETS exposure and QOL, and prospective longitudinal data on HBC using administrative databases for 30 months following baseline evaluation. Setting and patients: Participants were health maintenance organization members with physiciandiagnosed asthma involved in a longitudinal study of risk factors for hospital-based asthma care. Measurements: Demographic characteristics and QOL were assessed with administered questionnaires, including the Marks Asthma Quality-of-Life (AQLQ) and SF-36 questionnaires. HBC was defined as episodes per person-year of hospital-based asthma care, which included emergency department and urgency care visits, and hospitalizations for asthma. Results: Current smokers reported significantly worse QOL than never-smokers in two of five domains of the AQLQ (p < 0.05). Subjects with ETS exposure also reported significantly worse QOL than those without ETS exposure in two domains (p < 0.05). On the SF-36, current smokers reported significantly worse QOL than never-smokers in five of nine domains (p < 0.05). Subjects with ETS exposure reported significantly worse QOL than those without ETS exposure in three domains (p < 0.05). Current smokers used significantly more hospital-based asthma care than never-smokers (adjusted relative risk [RR], 1.40; 95% confidence interval [CI], 1.01 to 1.95) while ex-smokers did not exhibit increased risk compared with nonsmokers (adjusted RR, 0.94; 95% CI, 0.7 to 1.3). Also, subjects with ETS exposure used significantly more hospital-based asthma care than those without ETS exposure (RR, 2.34; 95% CI, 1.80 to 3.05). Conclusions: Direct or environmental tobacco exposure prospectively predicted increased health-care utilization for asthma and reduced QOL in patients with asthma. These findings add to our existing knowledge of the detrimental effects of tobacco smoke and are of relevance specifically to patients with asthma. (CHEST 1999; 115:691–696) Key words: asthma; health-care utilization; quality of life; smoking Abbreviations: AQLQ 5 Asthma Quality-of-Life Questionnaire; CI 5 confidence interval; ETS 5 environmental tobacco smoke; HCU 5 health-care utilization; HMO 5 health maintenance organization; LSD 5 least significant differences; QOL 5 quality of life; RR 5 relative risk

smoking is known to have deleterious A lthough effects on health-care utilization (HCU) and quality of life (QOL) in selected populations,1–5 its *From the Division of Pulmonary and Critical Care Medicine (Drs. Sippel and Osborne), Portland Veterans Administration Medical Center, Portland, OR; Kaiser Permanente Center for Health Research (Ms. Pedula and Dr. Vollmer), Portland, OR; and the Department of Medicine and the Division of Pulmonary and Critical Care Medicine (Dr. Buist), Oregon Health Sciences University, Portland, OR. This project supported by NIH grant HL 48237. Manuscript received March 18, 1998; revision accepted October 26, 1998. Correspondence to: Molly Osborne, MD, PhD, FCCP, P3 PULM Pulmonary/Critical Care, Veterans Administration Medical Center, 3710 SW US Veterans Hospital Rd, Portland, OR 97207; e-mail: [email protected]

effects on HCU and QOL in patients with asthma have not been well characterized. Such effects would be expected, since smoking has been linked to adverse health effects that could influence HCU and QOL in patients with asthma. For example, smoking has been associated with increased allergen sensitization,6 chronic bronchitis,7 and heightened bronchial reactivity.8 Furthermore, the increase in asthma morbidity over the past two decades9 has focused attention on other modifiable risk factors, such as patterns of medication use, allergen exposure, and patient management plans.10 –12 To investigate the relationship between tobacco exposure and both HCU and QOL among subjects with asthma, we studied 619 participants in a longitudinal study CHEST / 115 / 3 / MARCH, 1999

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21909/ on 01/22/2017

691

of risk factors for hospital-based asthma care. We compared current cigarette smokers with both ex-smokers and never-smokers with respect to HCU and QOL. We also compared subjects with and without environmental tobacco smoke (ETS) exposure with respect to these outcomes.

and a four-point oral steroid use scale (never, occasional burst use, frequent burst use, daily use). We then classified patients as having “less severe” conditions if the sum of these two indexes was 0 to 2 and as “more severe” if this sum was 3 to 5. This crude “severity” score has been shown to correlate with self-assessed severity and separates subjects into “more” and “less” severe groups.23 HCU: Hospital-Based Asthma Care

Materials and Methods Subjects and Research Setting Participants were members of a large health maintenance organization (HMO), Kaiser Permanente Northwest Division, who were either hospitalized for asthma during 2 years prior to recruitment or had at least two dispensings of antiasthma medications in the previous year. All participants reported having physician-diagnosed asthma and ongoing symptoms consistent with asthma at the time of recruitment. By design, participants ranged in age from 3 to 55 years. This analysis is restricted to the 619 participants aged 15 to 55 years. The study methods have been reported in detail elsewhere13 and are highlighted herein. Study Design We report baseline cross-sectional data and prospective longitudinal data collected as part of a longitudinal study to characterize risk factors for episodes of hospital-based asthma care within an HMO.13 The baseline assessment included a questionnaire and spirometry using standardized methods and equipment that met or exceeded American Thoracic Society requirements.14,15 Longitudinal HCU data were collected prospectively for 30 months following study entry. Questionnaires Relevant sections of the American Thoracic Society-DLD 1978 respiratory symptoms questionnaire,16 International Union Against Tuberculosis and Lung Disease bronchial symptoms questionnaire,17 and National Asthma Education Program Expert Panel Guidelines18 were incorporated into a questionnaire focusing on respiratory symptoms, characteristics of asthma, demographic factors, tobacco use, allergen exposures, and medication use. The questionnaire was administered to subjects by a trained technician, and took approximately 20 min. COPD was defined as self-reported chronic bronchitis, emphysema, or COPD. Nonasthma medication use was assessed by having participants bring all of their medications to the clinic, where they were then coded for analysis into the following classes: GI, hormonal, nonsteroidal anti-inflammatory, cardiovascular, b-receptor antagonist, psychotropic, diabetic, and thyroid medications. ETS exposure was defined as self-reported regular exposure (yes or no) to other people’s tobacco smoke either at home or at work. The questionnaire also included both generic and diseasespecific measures of health status: the SF-36 health status questionnaire19 and the Asthma Quality-of-Life Questionnaire (AQLQ) developed by Marks and colleagues.20 Scores from these questionnaires have been shown to significantly correlate with severity of asthma21,22 and have a high internal reliability. Asthma Severity To provide a rough stratification for severity, we classified participants on a three-point spirometry scale (FEV1 $ 80%, 60 to 80%, , 60%)

HCU data were obtained from administrative databases for the 30-month period following baseline evaluation. Using these data, we defined an episode of hospital-based asthma care as one or more emergency department visits, urgency care visits, or hospitalizations for asthma that were clustered in time, with no two adjoining contacts separated by . 2 days.24 Adjoining contacts that were separated by . 2 days were counted as belonging to separate episodes of care. As there were only 21 hospitalizations during the study period, hospital-based asthma care represents predominantly emergency department and urgency care clinic visits. Participant Follow-up Person-years of observation were calculated as the number of months of health plan membership during the 30-month follow-up period. On average, participants were followed up for 27.2 months, and 85% had at least 2 years of follow-up. Subjects with , 2 years of follow-up did not differ significantly in terms of age, gender, and severity of asthma when compared to those with $ 2 years of follow-up. Similarly, person-years of follow-up did not differ significantly by smoking status or by ETS exposure. We excluded 12 persons from analysis who had no eligibility during the follow-up period. Statistical Methods All analyses were performed using statistical software (SAS; Cary, NC). We used standard methods for analyzing contingency tables, with p values based on the Pearson x2 statistic and, for tests of trend, on the Mantel-Haenszel x2 statistic.25 We computed rates of hospital-based care as the total number of episodes of hospital-based care for any given subgroup divided by the total number of person-years of follow-up for that group. This was then multiplied by 100 and expressed as the number of episodes per 100 person-years of observation. We used Poisson regression analysis26 to compare rates of hospital-based episodes of care and multiple linear regression to examine the joint effects of multiple variables on the QOL scores. Poisson regression allowed us to distinguish not only between users and nonusers of health care, but also to detect associations of various factors with frequency of utilization. Poisson regression is the preferred analytic approach for the episode data because it is able to account for both the differential amount of follow-up among participants (via the person-years method) and also the recurrent nature of the outcome (many individuals had more than one episode of hospital-based asthma care). Adjustment variables used in these analyses were age, gender, nonasthma medication use, self-reported COPD, severity of asthma, and self-reported income. For QOL analyses, we used analysis of variance and Fisher’s LSD method of multiple comparisons for post hoc analysis. Unless otherwise stated, all p values are two sided and a p value of , 0.05 is significant.

Results Table 1 presents selected characteristics of the study sample. Eleven percent (n 5 68) were current smok-

692

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21909/ on 01/22/2017

Clinical Investigations

Table 1—Subject Characteristics by Smoking Status* Smoking Status Characteristics Age, yr Gender, % female Mild asthma, % Pack-years Self-reported COPD,§ % Nonasthma medication use, % Annual household income, % , $30k $30k–$50k . $50k

ETS

Current (n 5 68)

Ex (n 5 172)

Never (n 5 376)

p Value†

Yes (n 5 237)

No (n 5 382)

37 6 11 76 66 20 6 20 29 43

43 6 8 55 67 17 6 20 14 37

35 6 11 58 75 0 11 20

, 0.001 0.008 0.081 0.35‡ , 0.001 , 0.001

37 6 12 66 70 — 16 31

38 6 11 56 73 — 12 25

0.053 0.008 0.420

18 45 37

25 42 32

19 40 41

26 45 30

17 39 44

0.002

0.287

p Value†

0.147 0.099

*Values given as mean 6 SD. n 5 616 under Smoking Status because three subjects did not respond to this question. All subjects answered Exposed to Tobacco Smoke question, so n 5 619. Current 5 current smokers; Ex 5 ex-smokers; Never 5 never-smokers. †Two-sided p values based on analysis of variance or two-sample t test for age and pack-years, and on x2 analysis for all other comparisons. ‡Represents comparison of current smokers vs ex-smokers only. §COPD defined as self-reported chronic bronchitis, emphysema, or COPD.

ers, 28% (172) were ex-smokers, and 61% (376) reported never having smoked. These three groups differed significantly with respect to age, gender, nonasthma medication use, and prevalence of selfreported COPD. Ex-smokers were, on average, about 7 years older than current and never-smokers. Current smokers were more likely to be female (76%) than either ex-smokers (55%) or never-smokers (58%). They were also about 21⁄2 times more likely than ex-smokers or never-smokers to report a diagnosis of COPD. The use of nonasthma medications was lower in neversmokers than in ever-smokers and, although not significant, never-smokers also tended to have less severe asthma. Self-reported annual income did not differ by smoking status, but exhibited a trend toward higher income among those not exposed to ETS. QOL: Cross-sectional Analysis Asthma-Specific QOL: The AQLQ ranges from 1 to 10, with 1 representing the best QOL. In practice, however, observed scores fall in a much narrower range. In our study, the AQLQ scores ranged from 1.2 to 3.0 across all domains. Current smokers reported significantly worse QOL than never-smokers in two of five domains, and ex-smokers reported significantly worse QOL than never-smokers in all domains (Table 2). Disease severity and self-reported COPD were associated with significantly worse QOL scores in all domains, while female gender was associated with significantly worse QOL scores in all domains except mood (data not shown). Smoking status (eg, never-smoker vs ex-smoker vs current smoker) remained an independent predictor of QOL in all domains after adjusting

for disease severity, nonasthma medication use, gender, age, and self-reported COPD using multiple linear regression. Subjects with ETS exposure reported significantly worse QOL than those with no ETS exposure in breathless and mood domains after adjusting for these variables also (Table 3).

Table 2—QOL Scores by Smoking Status* Smoking Status

QOL Scores

Current Smokers (n 5 68)

AQLQ domain (from 1 5 best 3 7 5 worst) Breathless 2.7 6 1.8† Mood 2.7 6 2.1† Social 1.2 6 1.7‡ Concern 1.8 6 1.7 Overall 2.1 6 1.5 SF-36 domain ranges (from 0 5 worst 3 100 5 best) Physical functioning 72 6 24† Social functioning 75 6 19 Role-physical 73 6 37‡ Role-emotional 71 6 38 Mental health 69 6 21† Vitality 48 6 22† Bodily pain 61 6 27† General health 53 6 23†‡

NeverEx-Smokers Smokers (n 5 172) (n 5 376)

3.0 6 2.0† 2.5 6 2.0† 1.7 6 1.9† 2.0 6 1.9† 2.4 6 1.6†

2.2 6 1.7 1.9 6 1.7 1.3 6 1.7 1.5 6 1.5 1.8 6 1.4

73 6 24† 73 6 22† 61 6 40† 73 6 38 73 6 19† 51 6 22† 65 6 26† 61 6 23†

82 6 17 76 6 17 70 6 38 77 6 36 76 6 16 58 6 20 72 6 23 66 6 20

*Unadjusted scores expressed as mean 6 SD. n 5 616 under Smoking Status because three subjects did not respond to this question. †Two-sided p , 0.05 for pairwise comparison vs never-smokers based on Fisher’s LSD method of multiple comparisons. ‡Two-sided p , 0.05 for pairwise comparison vs ex-smokers based on Fisher’s LSD method of multiple comparisons. CHEST / 115 / 3 / MARCH, 1999

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21909/ on 01/22/2017

693

Table 3—QOL Scores by ETS Exposure* ETS Exposure? QOL Scores

Yes (n 5 237)

AQLQ domain (from 1 5 best 3 7 5 worst) Breathless 2.7 6 1.9 Mood 2.5 6 2.0 Social 1.4 6 1.7 Concern 1.7 6 1.7 Overall 2.1 6 1.5 SF-36 domain ranges (from 0 5 worst 3 100 5 best) Physical functioning 73 6 22 Social functioning 74 6 20 Role-physical 67 6 39 Role-emotional 73 6 37 Mental health 73 6 20 Vitality 53 6 23 Bodily pain 65 6 25 General health 59 6 22

No (n 5 382)

p Value†

2.3 6 1.8 2.0 6 1.7 1.4 6 1.8 1.6 6 1.7 1.9 6 1.5

0.01 0.003 0.56 0.70 0.42

81 6 19 76 6 19 68 6 39 77 6 37 75 6 16 56 6 20 71 6 24 65 6 21

, 0.001 0.81 0.92 0.32 0.57 0.35 0.02 0.04

*Unadjusted scores expressed as mean 6 SD. †Two-sided p value based on multiple linear regression adjusting for age, gender, severity of illness, self-reported COPD, nonasthma medication use, and self-reported income.

Generic QOL: The SF-36 domain scores range from 0 to 100, with 100 representing the best QOL. As with the AQLQ, actual SF-36 scores fall in a narrower range than is theoretically possible. In our study, the various subscales ranged from a low of 48 to a high of 82. Both current smokers and ex-smokers reported significantly worse QOL than never-smokers in physical functioning, mental health, vitality, and general health domains (Table 2). Ex-smokers also reported worse QOL than never-smokers in rolephysical, and ex-smokers reported better QOL than current smokers in change in health (reflecting improved health status over the past year, data not shown). Smoking status remained an independent predictor of QOL in these seven domains after adjusting for age, gender, nonasthma medication use,

self-reported COPD, and severity of asthma using multiple linear regression. Self-reported COPD and nonasthma medication use were associated with worse QOL in seven of nine domains, while female gender was associated with worse QOL in four domains, and severity in two domains. Subjects with ETS exposure reported worse QOL than those with no ETS exposure in physical functioning, bodily pain, and general health domains (Table 3). Hospital-Based Care: Longitudinal Analysis Current cigarette smokers had a greater rate of hospital-based episodes of asthma care than either never-smokers or ex-smokers (Table 4). Similarly, subjects who reported ETS exposure had more frequent episodes of hospital-based asthma care than those who reported no ETS exposure. These differences persisted even after adjusting for age, gender, disease severity, diagnosis of COPD, and nonasthma medication use. Based on the Poisson model, the relative risk (RR) for current smokers vs neversmokers was 1.40 (95% confidence interval [CI], 1.01 to 1.95), and the adjusted RR associated with ETS exposure was 2.34 (95% CI, 1.80 to 3.05). Current smoker status and ETS exposure were highly correlated. Seventy-eight percent of current smokers reported ETS exposure, and 22.5% of those with ETS exposure were current smokers. As a result, we believed that it was inappropriate to attempt to adjust the effects of any one variable for those of the other. However, after excluding current smokers from analysis (68 subjects), the adjusted RR associated with ETS exposure was 2.87 and still remained highly statistically significant (95% CI, 2.15 to 3.82). Discussion These results demonstrate that patients with asthma exposed to direct or ETS report worse QOL when compared to those without such exposure and are at

Table 4 —Rate of Hospital-Based Episodes of Asthma Care During 21⁄2-Year Follow-up

Baseline Status Baseline smoking status Current smokers Ex-smokers Never-smokers Baseline ETS exposure Yes No

Person-Years of Observation*

No. of Episodes

Episodes per 100 Person-Years of Observation

Adjusted RR†

95% CI

146 384 870

40 52 145

27 14 17

1.40 0.94 1.00

1.01–1.95 0.68–1.31 Reference group

528 878

148 89

28 10

2.34 1.00

1.80–3.05 Reference group

*Total person-years of observation varies slightly for smoking status and ETS exposure due to missing data. †Based on Poisson regression adjusting for age, gender, severity, self-reported COPD, nonasthma medication use, and self-reported income. 694

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21909/ on 01/22/2017

Clinical Investigations

increased risk for subsequent hospital-based asthma care. The 619 study subjects were adult members of an HMO who all had active asthma, completed spirometry and two QOL questionnaires at baseline, and were followed up prospectively for 30 months to assess rates of hospital-based asthma care. These findings add to our existing knowledge of the detrimental effects of tobacco smoke and are relevant specifically to patients with asthma. Modifying direct smoke and ETS could potentially lead to improved QOL in patients with asthma and to reduced HCU. Although it is likely that the increased hospitalbased care and lower QOL seen in this study are due to the negative health effects of smoke on patients with asthma, cause and effect cannot be established. Variables such as age, gender, nonasthma medication use, self-reported COPD, and severity of asthma also influence HCU and QOL and could be important confounders.21,24,26 –29 Nevertheless, direct smoke exposure and ETS exposure remained independent predictors of HCU and QOL even after adjusting for these variables. It is important to note that the diagnosis of asthma was made in these patients via their medication use (at least two dispensings of antiasthma medications) or previous hospitalization for asthma, as well as all participants reporting physician-diagnosed asthma prior to being enrolled in the study. Furthermore, on chart review of a subsample of participants for validation of our severity score (n 5 193), no one was found not to have asthma.23 There are statistical limitations to our QOL and hospital-based asthma care analyses worth noting. For QOL, we used analysis of variance and Fisher’s LSD method of multiple comparisons for post hoc analysis, and did not adjust further for the various domains within each QOL instrument. Accordingly, conclusions that are not consistent across all domains need to be interpreted with caution. For HCU, our outcome variable was hospital-based asthma care, defined as the sum of three specific components of HCU. Emergency department visits, urgency care clinic visits, and hospitalizations have important similarities in that they are largely unplanned and potentially preventable health care encounters. For this reason, combining them for analysis provides valuable information about modifiable components of HCU. However, individual analysis may not have had the statistical power necessary to reach certain conclusions. For example, there were only 21 hospitalizations during the study period, which was too infrequent to warrant separate analysis. The finding of lower QOL among smokers with asthma complements our existing knowledge of the detrimental effects of smoking on QOL that have been observed in nonasthmatic populations. Elderly patients who smoke report worse QOL than never-smokers.3

Patients who have suffered myocardial infarction and continue smoking also report worse QOL than exsmokers or never-smokers.1 Plausible mechanisms exist through which smoking could affect QOL. Cigarette use has been associated with wheezing,30,31 increased sensitization to certain allergens,6 chronic bronchitis,7 increased bronchial hyperresponsiveness,8 and resistance to inhaled corticosteroids.32 Interpreting QOL data depends on the validity of the instruments used for a given disease state. The asthma QOL questionnaire used in this study is disease specific and has been used to characterize patients with asthma and occupational disabilities due to asthma.20,33,34 One limitation of the Marks AQLQ is the fact that a “minimal clinically significant change” has not been determined, as it has for certain other asthma-specific instruments such as the Juniper AQLQ.35 Therefore, it is difficult to know the clinical relevance of the differences we observed. The SF-36 instrument was developed as a general health status questionnaire and is well studied, has high internal reliability, and has been shown to have good discriminating properties across a wide range of disease states.36 It has also been used extensively in populations with asthma.21,28,29 At the same time, it has been argued that the SF-36 is not a particularly good discriminator for asthma, because people with asthma have scores relatively close to those of normal subjects.36 For example, the never-smokers reported in this study had QOL scores that equaled or exceeded those of 638 healthy adult patients in six of seven categories when compared with data published by McHorney and colleagues.36 The issue of whether the SF-36 discriminates among risk subgroups within asthmatics, for example current smokers vs nonsmokers, is less well studied. Again comparing with published normal populations, current smokers in this study reported worse QOL in six of seven categories.36 The finding of increased hospital-based care among smokers with asthma also complements our existing knowledge about the detrimental effects of smoke on HCU that have been observed in nonasthmatic populations. For example, among the general Swiss population, smokers have been shown to use outpatient and inpatient services more frequently than nonsmokers.5 Also, patients who have had successful coronary angioplasty but continue to smoke are twice as likely to suffer myocardial infarction and death as nonsmokers.37 Cigarette use in patients with asthma has also been strongly correlated with the need for subsequent hospitalization.38 Finally, it has also been shown that increased use of health-care resources among smokers ceases within 4 years of quitting.39 This is consistent with our finding that hospital-based care was similar for ex-smokers and never-smokers. In conclusion, this study found that exposure to direct smoke or ETS prospectively predicted increased HCU for asthma, and cross sectionally was CHEST / 115 / 3 / MARCH, 1999

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21909/ on 01/22/2017

695

associated with reduced QOL patients with asthma. These findings add to our existing knowledge of the detrimental effects of tobacco smoke, and are of relevance specifically to patients with asthma.

19 20

References 1 Ekeberg O, Klemsdal TO, Kjeldsen SE. Quality-of-life on enalapril after acute myocardial infarction. Eur Heart J 1994; 15:1135–1139 2 Brown JK. Gender, age, usual weight, and tobacco use as predictors of weight loss in patients with lung cancer. Oncol Nurs Forum 1993; 20:466 – 472 3 Maxwell CJ, Hirdes JP. The prevalence of smoking and implications for quality-of-life among the community-based elderly. Am J Prev Med 1993; 9:338 –345 4 Dales RE, Dionne G, Leech JA, et al. Preoperative prediction of pulmonary complications following thoracic surgery. Chest 1993; 104:155–159 5 Gutzwiller F, LaVecchia 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 6 Jarvis D, Luczynska C, Chinn S, et al. The association of age, gender and smoking with total IgE and specific IgE. Clin Exp Allergy 1995; 25:1083–1091 7 Troisi RJ, Speizer FE, Rosner B, et al. Cigarette smoking and incidence of chronic bronchitis and asthma in women. Chest 1995; 108:1557–1561 8 Sunyer J, Munoz A. Concentrations of methacholine for bronchial responsiveness according to symptoms, smoking, and immunoglobulin E in a population-based study in Spain. Am J Respir Crit Care Med 1996; 153:1273–1279 9 Weiss KB, Gergen PJ. Breathing better or wheezing worse? The changing epidemiology of asthma morbidity and mortality. Annu Rev Public Health 1993; 14:491–513 10 Mayo PH, Richman J, Harris HW. Results of a program to reduce admissions for adult asthma. Ann Intern Med 1990; 112:864 – 871 11 Hartert TV, Windom HH, Peebles RSJ, et al. Inadequate outpatient medical therapy for patients with asthma admitted to two urban hospitals. Am J Med 1996; 100:386 –394 12 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 13 Osborne ML, Vollmer WM, Linton KP, 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 14 Enright PL, Johnson LF, Connett JE. Spirometry in the Lung Health Study: I. Methods and quality control. Am Rev Respir Dis 1991; 143:1215–1223 15 Becklake M, Crapo RO. Lung function testing: selection of reference values and interpretive strategies. Am Rev Respir Dis 1991; 144:1202–1218 16 Burney PGJ, 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 (IUATLD). Int J Epidemiol 1989; 18:165–173 17 Burney PGJ, Laitinen LA, Perdizet S, et al. Validity and repeatability of the IUATLD (1984) bronchial symptoms questionnaire: an international comparison. Eur Respir J 1989; 2:940 –945 18 National Asthma Education Program. Guidelines for the

21

22

23

24 25 26 27

28 29 30 31 32

33 34 35 36

37 38 39

diagnosis and management of asthma. NIH publication No. 91-3042. Bethesda, MD: National Heart, Lung, and Blood Institutes, 1991 Stewart AL, Hays RD, Ware Jr JE. The MOS short-form general health survey. Med Care 1988; 26:724 –735 Marks GB, Dunn SM, Woolcock AJ. A scale for the measurement of quality-of-life in adults with asthma. J Clin Epidemiol 1992; 45:461– 472 Bousquet J, Knani J, Dhivert H, et al. Quality-of-life in asthma: I. Internal consistency and validity of the SF-36 questionnaire. Am J Respir Crit Care Med 1994; 149:371–375 Marks GB, Burney PGJ, Premaratne UN, et al. Asthma in Greenwich, UK: impact of the disease and current management practices. Eur Respir J 1997; 10:1224 –1229 Osborne ML, Pedula KL, Vollmer WM, et al. Lack of correlation of current asthma symptoms with specialist-assessed long-term asthma severity. Chest (in press) 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 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22:719 –748 McCullagh P, Nelder JA. Generalized linear models. 2nd ed. London: Chapman and Hall, 1989; 127–136 Dompeling E, vanGrunsven PM, Molema J, et al. Early detection of patients with fast progressive asthma or chronic bronchitis in general practice. Scand J Prim Health Care 1992; 10:143–150 Lyons RA, Lo SV, Littlepage BN. Comparative health status of patients with 11 common illnesses in Wales. J Epidemiol Community Health 1994; 48:388 –390 Viramontes JL, O’Brien B. Relationship between symptoms and health-related quality-of-life in chronic lung disease. J Gen Intern Med 1994; 9:46 – 48 David MM, Hanrahan JP, Carey V, et al. Respiratory symptoms in urban Hispanic and non-Hispanic white women. Am J Respir Crit Care Med 1996; 153:1285–1291 Flodin U, Jonsson P, Ziegler J, et al. An epidemiologic study of bronchial asthma and smoking. Epidemiology 1995; 6:503–505 Pedersen B, Dahl R, Karlstrom R, et al. Eosinophil and neutrophil activity in asthma in a one-year trial with inhaled budesonide: the impact of smoking. Am J Respir Crit Care Med 1996; 153:1519 –1529 Blanc PD, Jones M, Besson C, et al. Work disability among adults with asthma. Chest 1993; 104:1371–1377 Blanc PD, Cisternas M, Smith S, et al. Asthma, employment status, and disability among adults treated by pulmonary and allergy specialists. Chest 1996; 109:688 – 696 Juniper EF, Guyatt GH, Willan A, et al. Determining a minimal important change in a disease-specific quality-of-life questionnaire. J Clin Epidemiol 1994; 47:81– 87 McHorney CA, Ware JE Jr, Rogers W, et al. The validity and relative precision of MOS short- and long-form health status scales and Dartmouth COOP charts: results from the Medical Outcomes Study. Med Care 1992; 30(5 suppl):MS253–MS265 Hasdai D, Garratt KN, Grill DE, et al. Effect of smoking status on the long-term outcome after successful percutaneous coronary revascularization. N Engl J Med 1997; 336:755–761 Morris RD, Munasinghe RL. Geographic variability in hospital admission rates for respiratory disease among the elderly in the United States. Chest 1994; 106:1172–1181 Wagner EH, Curry SJ, Grothaus L, et al. The impact of smoking and quitting on health care use. Arch Intern Med 1995; 155:1789 –1795

696

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21909/ on 01/22/2017

Clinical Investigations