|
 |
|
|||||||
|
|||||||||
© 2002 American Public Health Association
RESEARCH AND PRACTICE |
Ronald D. Deprez and Nancy L. Asdigian are with the Public Health Research Institute, Portland, Me. L. Christine Oliver is with the Occupational Health Institute, Boston, Mass. Norman Anderson, Edgar Caldwell, and Lee Ann Baggott are with the American Lung Association of Maine, Augusta.
Correspondence: Requests for reprints and for a copy of the survey instrument should be sent to Ronald D. Deprez, PhD, MPH, Public Health Research Institute, 120 Exchange St, Suite 200, Portland, ME 04101 (e-mail: rdeprez{at}phrg.com).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Objectives. We developed and evaluated a statewide and community-level asthma surveillance system.
Methods. Databases and measures included a community prevalence survey, hospital admissions data, emergency department/outpatient clinic visit records, and a physician survey of diagnosis and treatment practices. We evaluated the system in 5 Maine communities varying in population and income.
Results. Asthma hospitalizations were high in the rural/low-socioeconomic-status communities studied, although diagnosed asthma was low. Males were more likely than females to experience asthma symptoms, although they were less likely to have been diagnosed with asthma or to have used hospital-based asthma care.
Conclusions. Databases were useful for estimating asthma burden and identifying service needs as well as high-risk groups. They were less useful in estimating severity or in identifying environmental risks.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
The prevalence of asthma has risen precipitously over the past 2 decades, especially among children.1–3 Estimates from the 1998 National Health Interview Survey suggest that 26 million Americans have been diagnosed with asthma at some point in their life, and nearly 11 million suffer an acute asthma attack each year.4 Annually, asthma is responsible for more than 500 000 hospitalizations, 5000 deaths, and 134 million days of restricted activity.5 As might be expected, the economic costs are staggering, totaling nearly $11 billion in 1994, a 54% increase since 1984.6
Although most asthma prevention and management programs are implemented and evaluated at the local level, there is no comprehensive statewide system to monitor the burden of asthma and/or evaluate the impact of asthma-related interventions.1,7 The Centers for Disease Control and Prevention has urged state and territorial health departments to combine existing utilization data with survey-based prevalence estimates as part of a comprehensive asthma surveillance effort.8 Moreover, one of the nation’s Healthy People 2010 health objectives is for at least 25 states to establish a surveillance system for tracking asthma mortality, morbidity, access to care, and management.5
The purpose of our research was to design, pilot-test (in a sample of 5 communities), and evaluate a prototype system of asthma surveillance for the state of Maine. Our immediate goal was to develop a geographically linked system to monitor prevalence and severity of diagnosed asthma among adults and children, symptoms of undiagnosed disease, asthma risk factors, accessibility and use of asthma-related health care in communities across Maine, and provider practices of diagnosing and treating asthma. Our ultimate goal was to develop a data collection and analysis system that could yield information about the prevention and disease management implications of community-level data on asthma risk, disease burden, and treatment patterns. Such information will be useful to health planners and providers at both state and local levels as they attempt to articulate the scope and dimensions of the problem and as they develop and target prevention and treatment initiatives to address the problem.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Pilot-Test Communities
We pilot-tested the system in a sample of 5 Maine communities. For purposes of this project, we defined a community as a hospital service area (HSA). Originally created by the Maine Health Care Finance Commission, HSAs are organized around hospitals and include areas with a common service delivery system, as well as similar patterns of provider practices and health care–seeking behavior. HSAs represent distinct health care delivery systems, so the data obtained for them are useful for program planning, especially for conditions, such as asthma, that are treated mostly with outpatient care.
The 5 HSAs varied in population density and income. Two represented relatively urban and affluent communities. The remaining 3 represented relatively rural and lower-socioeconomic-status (SES) communities. HSAs were designated as high- or low-SES on the basis of a classification algorithm that takes into account average annual household income, poverty status, employment status, and educational attainment. Records in all databases were coded by HSA so that they could be linked. The system thus permitted us to create a comprehensive profile of asthma in each community, based on integrated information about risk, prevalence, severity, utilization, and practice patterns.
Prototype Databases and Measures
General population telephone survey.
The survey instrument (available upon request) was based on previously validated questionnaires and included questions about physiciandiagnosed asthma, use of prescription asthma medication, asthma symptoms, respiratory and cardiovascular disease history, asthma risk factors and triggers (e.g., tobacco use, exposure to pet dander, secondary smoke), occupational history, and sociodemographics.
We computed the prevalence of diagnosed and undiagnosed asthma and associated 95% confidence intervals (CIs) among male and female adults aged 18 to 44 years, 45 to 64 years, and 65 years and older and among children aged 3 to 17 years in the total sample and in the urban/high-SES and rural/low-SES HSAs. We did not break down estimates for children by sex, because child sex was not included on the survey.
We defined diagnosed asthma for both adults and children as a physician diagnosis of asthma and current use of prescription asthma medication. We defined undiagnosed asthma for adults as no physician diagnosis of asthma, combined with a report of experiencing 4 or more of the following symptoms consistent with a diagnosis of asthma in the previous year: cough with exercise, wheeze with exercise, chest tightening with exercise, sleep broken by wheeze, sleep broken by difficulty breathing, wheeze upon waking, difficulty breathing upon wakening, wheeze in smoke, or wheeze in dust.9 We defined undiagnosed asthma for children as no physician diagnosis of asthma, combined with a report of wheeze in the previous year and either wheeze with exercise or cough at night in the previous year.10
Using the 1997 National Heart, Lung, and Blood Institute’s (NHLBI) Expert Panel Report 2,11 we also developed a medication-based algorithm to estimate the prevalence of severe persistent, moderate persistent, mild persistent, and mild intermittent asthma among adults and children with current physiciandiagnosed asthma. The algorithm took into account the type, mode, and frequency of self-reported prescription asthma medication use among adults and children.
Trained interviewers administered the survey in 1997 to a stratified random sample of 18- to 64-year-old adults residing in each of the 5 pilot-test communities. We randomly sampled households with telephones, with a probability of selection proportional to the size of the community to which they belonged. We asked 627 randomly selected adults (62% of those eligible to participate) about themselves and the 456 children (aged 3–17) living in their households. Of these, 352 adult respondents and 277 children came from the 2 urban/high-SES communities, and 275 adults and 179 children came from the 3 rural/low-SES communities.
We used poststratification weights (based on the age and sex distribution in the study communities) to adjust for nonresponse and gaps in telephone coverage. Just under half (49%) of the sample were male, and 64% were married. The average age was 40 years. Just under half (49%) of all respondents had some college education, and 49% reported an annual household income of $35 000 or more. Approximately 75% were employed either full- or part-time; the largest percentage of these (35%) reported working in a professional/technical occupation. Compared with the general population of the 5 pilot-test communities, the sample included a larger proportion of respondents with an annual household income of $35 000 or more (34% vs 49%) and more professional/technical workers (25% vs 35%). All other sample demographics were comparable to general population characteristics.
Hospital admissions. We used 1994/1995 data from the Maine Uniform Hospital Discharge Database to compute annual average rates of asthma admissions per 1000 population. To compute rates and associated 95% confidence intervals, we used asthma as the principal diagnosis (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] code 49312) separately for males and females aged 0 to 4 years, 5 to 17 years, 18 to 44 years, and 45 to 64 years in the total study area and in the urban/high-SES and rural/low-SES HSAs.
Emergency department/hospital outpatient visits. We used 1994 data (the only year of data available at that time) from the Maine Health Care Finance Commission to compute annual rates of asthma-related emergency department (ED)/outpatient clinic visits per 1000 population. Because many rural hospitals in Maine do not have scheduled outpatient clinics, outpatient visits are comparable to, and therefore combined with, ED visits. A visit is classified as outpatient if an individual presents to the ED either for a nonemergent problem or in the evening, when many clinics are closed. ED visits would have to be analyzed separately, however, in a statewide system that includes the several larger hospitals in Maine that do have scheduled outpatient clinics. We computed rates and associated 95% confidence intervals based on asthma as the principal diagnosis (ICD-9-CM code 49312) separately for males and females aged 0 to 4 years, 5 to 17 years, 18 to 44 years, and 45 to 64 years in the total study area and in the urban/high-SES and rural/low-SES HSAs.
Physician survey. We also developed a postal survey to collect information from pulmonologists/internists, family practitioners, and pediatricians about their practices of diagnosing and treating adult and childhood asthma. The objective was to assess the contribution of provider diagnostic and treatment practices to community patterns of asthma prevalence and related health care utilization. The instrument presented 4 case studies that required diagnostic judgments of "asthma likely," "asthma suspect," or "asthma not likely." It also included a description of 4 medication regimens that physicians were asked to use as a basis for judging the severity of asthma as either severe persistent, moderate persistent, mild persistent, or mild intermittent.
During the summer of 1997, we mailed surveys to all known board-certified physicians (n = 111) practicing in each of the 5 pilot-test communities. We included a letter of introduction from the American Lung Association of Maine and a stamped return envelope. We sent reminder letters 1 month after the initial mailing. Just over half (53%; n = 59) of the physicians returned a completed survey.
We used the 1997 NHLBI guidelines11 to calculate the total number of correct diagnostic and severity assessments made by physician respondents. We computed the proportion of physician respondents who correctly answered at least 2 of the 4 diagnostic judgments and at least 2 of the 4 severity judgments.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Findings From Surveillance Databases
We first examined key indicators from individual databases included in the system. Findings are shown in Table 1
. Across the 5 HSAs studied, the prevalence of physiciandiagnosed asthma was 5.9% (± 1.84%) among 18- to 64-year-old adults and 9% (± 2.6%) among 3- to 17-year-old children. An additional 8% (± 2.1%) of adults and 10% (± 2.7%) of children reported symptoms consistent with asthma in the year before the survey, despite never having been diagnosed with the disease.
|
The medication-based algorithm that we developed to assess the severity of physician-diagnosed asthma successfully classified only 14 of 37 (38%) eligible adults and 16 of 39 (41%) eligible children as having severe persistent, moderate persistent, mild persistent, or mild intermittent asthma. Among classified cases, the overwhelming majority were mild intermittent (71% of adults and 88% of children). Across all 5 HSAs, the prevalence of mild intermittent asthma was 1.5% among adults and 3% among children. We did not identify any cases of severe persistent asthma and observed only 2 cases of either moderate or mild persistent asthma among adults and children.
The rate of asthma-related hospital admissions among 0- to 64-year-olds in the 5-community study area was 1.25 per 1000 population. The rate of asthma-related hospital ED/outpatient visits among 0- to 64-year-olds in the study area was 11.5 per 1000 population.
Finally, just under half (48%) of all physician survey respondents were correct in their diagnostic judgments of at least 2 of the 4 case studies. Incorrect responses were more likely to take the form of overestimating asthma (e.g., judging suspect asthma as likely asthma) than of underestimating it. Nearly all physicians (85%) were correct in their severity assessments of at least 2 of the 4 case studies presented.
Evaluation of the Surveillance System
Systems of disease surveillance should provide a richer understanding of disease patterns in a community than what can be obtained from any single source of data. These systems should also yield useful data for local prevention and/or treatment needs. With those 2 goals in mind, we evaluated our system by integrating data on asthma prevalence, hospital utilization, and provider practices in each of our pilot-test communities and then assessing their utility for elucidating local disease burden and health service needs.
The findings shown in Table 2. illustrate the value of the system in discerning meaningful patterns of risk, disease burden, and service utilization within communities. For example, relative to that in the urban/high-SES HSAs, the prevalence of physician-diagnosed asthma among adult men in the rural/low-SES HSAs was low (4.0% vs 1.0%; P < .10), whereas the prevalence of undiagnosed asthma symptoms was high (8.0% vs 17.7%; P < .05). At the same time, rates of asthma hospitalizations and of ED/hospital outpatient visits were elevated among males in the rural/low-SES HSAs relative to rates among males in the urban/high-SES HSAs (1.32 vs 0.67, P < .10; and 14.5 vs 6.7; P < .05).
|
Asthma admissions and ED/outpatient visits among females aged 0 to 64 years in the rural/low-SES HSAs were also high relative to those in the urban/high-SES HSAs (2.11 vs 1.29, P < .10; and 24.5 vs 10.4, P < .05), despite the fact that neither diagnosed asthma nor undiagnosed asthma symptoms were elevated among adult females or children in the rural/low-SES HSAs. Hospital-based asthma care was high in the rural/low-SES HSAs, even though nearly 60% of provider respondents from those communities correctly diagnosed 2 or more of the 4 case studies presented, and 86% correctly assessed the severity of at least 2 of the 4 asthma cases. Rather than indicating poor quality of care, these findings suggest that inadequate access to primary asthma care in the rural/low-SES HSAs or failure to comply with prescribed treatment regimens might contribute to more undiagnosed asthma and/or more intensive or emergency medical care for respiratory problems in those communities.
The surveillance system also provided useful risk factor information that health care planners can use to target educational programs, clinical assessments, and treatment services. For example, the findings in Table 1
show that, across all study communities, adult men were more than twice as likely as adult women to experience respiratory symptoms consistent with asthma (11.7% vs 4.5%; P < .05), despite the fact that they were 70% less likely to have ever been formally diagnosed with or treated for asthma (2.7% vs 9.1%; P < .05) and had rates of asthma hospitalizations and ED/outpatient visits that were approximately half those of females.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Overall, the databases included in the system provided useful information about community-specific patterns of asthma-related risks, disease burden, health care utilization, and physicians’ diagnostic and treatment practices. In turn, that information was valuable in pointing to community needs for asthma-related education, prevention initiatives, and medical services.
The community survey provided otherwise unobtainable data on asthma symptoms among adults and children who had never been diagnosed with asthma. Because many individuals suffer from asthma without receiving a formal physician diagnosis,13–15 an exclusive focus on diagnosed disease precludes an understanding of the full burden of disease in the community. In our study, sole reliance on physician-diagnosed asthma would have led to the mistaken conclusion that risk for asthma was concentrated among women or that the burden of asthma was not elevated in the rural/low-SES communities.
The utility of the survey for assessing asthma severity was limited, however. The medication-based algorithm failed to capture a significant proportion of the treatment regimens reported in the survey. The failure of the algorithm might have been due to a lag between the publication of the NHLBI treatment guidelines and physician adoption of those guidelines, or to patient misunderstanding of or noncompliance with prescribed treatment regimens. Regardless of the cause of failure, more extensive development and validation work is necessary. In addition, we suggest that researchers explore other sources of data, such as pharmaceutical records, that can be used in a statewide asthma surveillance system to assess severity.
The survey was also limited by its cost. Implementing an ongoing statewide survey requires substantial resources, particularly if the sample size is large enough to generate reliable prevalence estimates at the community level. Most lung associations or state health departments would likely be unable to afford the cost of such a survey on a regular basis. Indeed, because of its cost, a statewide survey capable of generating community-level estimates of asthma prevalence and severity has not been adopted by any state agency in Maine. We suggest that it may be more cost-effective for states to use the 2-item asthma prevalence module developed by the Centers for Disease Control and Prevention as part of their Behavioral Risk Factor Surveillance System, supplemented with several key questions about symptoms of undiagnosed asthma. Including Maine, 18 states used the 2-item module in 1999. As of the 2000 survey, the 2 asthma prevalence questions had been incorporated into the core instrument. Surveillance system data on the prevalence of diagnosed asthma and symptoms of undiagnosed asthma could be reweighted to produce reliable estimates for regions within the state (e.g., HSAs, counties), especially if multiple years of data were combined. Periodic oversampling at the county or health district level—another economically viable alternative—would also produce valid substate estimates for local planning initiatives. Such an effort might be conducted as the need warrants and as support is available for additional funding.
Hospital admissions and ED databases provide valuable information about the accessibility and use of asthma-related health care in a community. In this study, high rates of inpatient and ED/outpatient asthma care, combined with low rates of physician-diagnosed asthma (despite apparent provider adherence to NHLBI guidelines) pointed to possible problems in accessing primary/outpatient asthma care in the rural/low-SES communities. Hospitalization and ED data are also readily available at a fairly low cost in many states, and community-level utilization rates are easily calculated from those databases. States that do not have access to hospital inpatient or ED data should work toward establishing usable databases and/or identifying special population databases (e.g., Medicaid) that can be used for asthma surveillance.
The physician survey yielded useful information about provider practices of diagnosing and treating asthma—information that helped us interpret community-level patterns of asthma prevalence and health care utilization. Because of the low response rate, however, it is possible that the high accuracy rates reflect a selection bias toward more knowledgeable and experienced physicians. Although we are unable to rule out this possibility, 64% of physicians reported that fewer than 10% of their patients were being treated for asthma. Therefore, it is probably unlikely that the sample was disproportionately composed of asthma experts.
We recommend that a physician survey be part of an ongoing statewide asthma surveillance effort that we believe would need to be updated only every several years. To ensure an adequate response rate, such a survey should be conducted in the context of an intensive and far-reaching physician awareness effort.
It might also be useful to include data on asthma mortality in a statewide system of asthma surveillance. We did not include mortality data in our prototype system because too few asthma deaths occurred in our pilot-test communities. Asthma mortality information would be informative at the state level, however, and like hospitalization data, it is inexpensive and easy to analyze.
Finally, we also recommend that statewide asthma surveillance systems include information about environmental risks for asthma. We evaluated the US Environmental Protection Agency’s Toxics Release Inventory for this purpose, but found that it did not contain data on all exposures from all releasing facilities. Future efforts should focus on identifying and/or developing more accurate and comprehensive sources of data on environmental risks for asthma. Those efforts should include a review of the state-based systems of occupational asthma surveillance developed as part of the Sentinel Event Notification System for Occupational Risks program of the National Institute for Occupational Safety and Health.
|
Acknowledgments |
|---|
This research was supported in part by a grant from the American Lung Association of Maine to the Public Health Research Institute.
The authors wish to acknowledge the following members of a technical advisory committee and of an oversight committee, who provided guidance to the project: Linda Niccolai, PhD, contributed substantially to the design of the project, and Willard Dyche, Shane Stoyer, and Jen Hilton provided valuable programming and administrative assistance during all phases of the project.
Human Participant Protection
No protocol approval was needed for this study.
|
Footnotes |
|---|
R. D. Deprez, L. C. Oliver, N. Anderson, E. Caldwell, and L. A. Baggott contributed to the conception of the project, the design of the asthma surveillance system, the design and implementation of the system evaluation, the interpretation of the findings, and the preparation, review, and approval of the article. N. L. Asdigian contributed to the design and implementation of the system evaluation, the interpretation of the findings, and the preparation and review of the article.
Accepted for publication January 29, 2002.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
1. Mannino DM, Homa DM, Pertowski CA, et al. Surveillance for asthma—United States, 1960–1995. MMWR CDC Surveill Summ. 1998;47(1):1–27.[Medline]
2. Centers for Disease Control and Prevention. Forecasted state-specific estimates of self-reported asthma prevalence—United States, 1998. MMWR Morb Mortal Wkly Rep. 1998;47:1022–1025.[Medline]
3. Weiss KB, Gergen PJ, Wagener DK. Breathing better or wheezing worse? The changing epidemiology of asthma morbidity and mortality. Annu Rev Public Health. 1993;14:491–513.[Medline]
4. American Lung Association. Prevalence based on revised National Health Interview Survey. Available at: http://www.lungusa.org/data/data_102000.html. Accessed September 10, 2002.
5. Healthy People 2010. Vol 2. 2nd ed. Washington, DC: US Department of Health and Human Services; 2000.
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.[Medline]
7. Centers for Disease Control and Prevention. Asthma surveillance programs in public health departments—United States. MMWR Morb Mortal Wkly Rep. 1996;45:802–804.[Medline]
8. Brown CM, Anderson HA, Etzel RA. Asthma. The states’ challenge. Public Health Rep. 1997;112:198–205.[Medline]
9. Venables KM, Farrer N, Sharp L, Graneek BJ, Newman Taylor AJ. Respiratory symptoms questionnaire for asthma epidemiology: validity and reproducibility. Thorax. 1993;48:214–219.
10. Toelle BG, Peat JK, Salome CM, Mellis CM, Woolcock AJ. Toward a definition of asthma for epidemiology. Am Rev Respir Dis. 1992;146:633–637.[Medline]
11. National Asthma Education and Prevention Program. NAEPP Expert Panel Report 2: Guidelines for the Diagnosis and Management of Asthma. Bethesda, Md: National Institutes of Health, National Heart, Lung, and Blood Institute; 1997.
12. International Classification of Diseases, 9th Revision, Clinical Modification. Hyattsville, Md: National Center for Health Statistics; 1980. DHHS publication PHS 80-1260.
13. Enright PL, McClelland RL, Newman AB, Gottlieb DJ, Lebowitz MD. Underdiagnosis and undertreatment of asthma in the elderly. Cardiovascular Health Study Research Group. Chest. 1999;116:603–613.
14. McLean DE, Bowen S, Rowe A, et al. Severity of asthma among homeless children in NYC. Presented at: American Public Health Association Annual Meeting; November 7–11, 1999; Chicago, Ill.
15. Findley SE, McLean DE, Mariko S. Potentially undiagnosed asthma among minorities. Poster D136. Presented at: 1999 American Lung Association/American Thoracic Society International Conference; April 23–28, 1999; San Diego, Calif.
This article has been cited by other articles:
|
|
 |
|
  M. J. Reeves, S. Lyon-Callo, M. D. Brown, K. Rosenman, E. Wasilevich, and S. G. Williams Using billing data to describe patterns in asthma-related emergency department visits in children. Pediatrics, April 1, 2006; 117(4 Pt 2): S106 - S117. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. P. Lanphear and P. J. Gergen Invited Commentary: Asthma Surveillance in US Children Am. J. Epidemiol., July 15, 2003; 158(2): 105 - 107. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |