Characteristics of Inner-City Children With Asthma:The National Cooperative Inner-City Asthma Study

Meyer Kattan, MD,1 Herman Mitchell, PhD,2 Peyton Eggleston, MD,3

Peter Gergen, MD, MPH,4 Ellen Crain, MD, PhD,5 Susan Redline, MD,6

Kevin Weiss, MD,7 Richard Evans III, MD, MPH,8 Richard Kaslow, MD,9

Carolyn Kercsmar, MD,6 Fred Leickly, MD,10 Floyd Malveaux, MD, PhD,11

and H. James Wedner, MD12

Summary. Asthma morbidity has increased dramatically in the past decade, especially amongpoor and minority children in the inner cities. The National Cooperative Inner-City Asthma Study(NCICAS) is a multicenter study designed to determine factors that contribute to asthma mor-bidity in children in the inner cities. A total of 1,528 children with asthma, ages 4 to 9 years old,were enrolled in a broad-based epidemiologic investigation of factors which were thought to berelated to asthma morbidity. Baseline assessment included morbidity, allergy evaluation, adher-ence and access to care, home visits, and pulmonary function. Interval assessments wereconducted at 3, 6, and 9 months after the baseline evaluations.

Over the one-year period, 83% of the children had no hospitalizations and 3.6% had two ormore. The children averaged 3 to 3.5 days of wheeze for each of the four two-week recallperiods. The pattern of skin test sensitivity differed from other populations in that positive reac-tions to cockroach were higher (35%) and positive reactions to house dust mite were lower(31%). Caretakers reported smoking in 39% of households of children with asthma, and cotinine/creatinine ratios exceeded 30 ng/mg in 48% of the sample. High exposure (>40 ppb) to nitrogendioxide was found in 24% of homes. Although the majority of children had insurance coverage,53% of study participants found it difficult to get follow-up asthma care. The data demonstratethat symptoms are frequent but do not result in hospitalization in the majority of children. Thesedata indicate a number of areas which are potential contributors to the asthma morbidity in thispopulation, such as environmental factors, lack of access to care, and adherence to treatment.Interventions to reduce asthma morbidity are more likely to be successful if they address themany different asthma risks found in the inner cities. Pediatr. Pulmonol. 1997;24:253–262.© 1997 Wiley-Liss, Inc.

Key words: asthma; children; inner-city.

1Mount Sinai School of Medicine, New York, New York.

2New England Research Institutes, Watertown, Massachusetts.

3Johns Hopkins School of Medicine, Baltimore, Maryland.

4National Institute of Allergy and Infectious Disease (National Insti-tutes of Health, Bethesda, Maryland).

5Albert Einstein College of Medicine, Bronx, New York.

6Case Western Reserve University, Cleveland, Ohio.

7National Institute of Allergy and Infectious Disease (National Insti-tutes of Health, Bethesda, Maryland) (now at Center for Health Ser-vices Research, Rush Primary Care Institute, Chicago, Illinois).

8Children’s Memorial Hospital, Chicago, Illinois.

9University of Alabama, Birmingham, Alabama.

10Henry Ford Hospital, Detroit, Michigan.

11Howard University College of Medicine, Washington, DC.

12Washington University School of Medicine, St. Louis, Missouri.

Contract grant sponsor: National Institute of Allergy and InfectiousDisease (National Institutes of Health, Bethesda, MD); Contract grantnumbers: UO1 A1-30751, A1-30752, A1-30756, A1-30772, A1-30773-01, A1-30777, A1-30779, A1-30780, and N01 A1-15105.

Correspondence to: Ernestine Smartt, RN, DAIT, NIAID, NIH, SolarBuilding, Room 4A42, 6003 Executive Boulevard, Bethesda, MD20892-7640.

Received 23 April 1996; accepted 30 June 1997.

Pediatric Pulmonology 24:253–262 (1997)

© 1997 Wiley-Liss, Inc.

INTRODUCTION

The morbidity and mortality associated with asthmahave increased in recent years. During the decade from1982 to 1991, mortality due to asthma has increased 42%among 5–34-year-olds.1 Weitzman et al., found a 39%increase in the prevalence of active asthma among chil-dren under 18 years of age.2 This increase is even moredramatic among the very young, under age 5.3 Poor andminority urban areas contribute disproportionately to thisincreasing asthma morbidity and mortality.4–6 The ma-jority of these studies of increased asthma morbidityhave focused on hospitalization or emergency depart-ment (ED) visits taken from large national utilizationdatabases.3,4,7–9 A number of other studies have mea-sured symptom incidence in samples of asthmatics, usu-ally at a single point in time.10 Those studies have fo-cused on factors which may be related to this increasedmorbidity and have typically focused on the relationshipof one, or possibly two factors and their impact on hos-pitalizations and ED use. The first phase of the NationalCooperative Inner-City Asthma Study (NCICAS) wasspecifically designed to capture extensive utilization andsymptom data on a large sample of inner-city children,with repeated measures of utilizations and symptomsover an extended period of time. This report provides adetailed description of utilization and asthma symptomsover a one-year period, the environmental milieu, andaccess to care in a geographically diverse inner-citypopulation of children with mild to severe asthma. Thepsychosocial characteristics of this cohort are publishedelsewhere.11 Subsequent analyses will allow us to relatespecific factors to morbidity among inner-city asthmaticsand provide the basis for an intervention.

METHODS

In the NCICAS project, 1,528 children with asthma,ages 4 through 9, and their caretakers were given anextensive baseline assessment including demographiccharacteristics, health history, medicine use, access tocare, adherence to treatment, environmental factors, psy-chological assessments, as well as symptoms and healthcare utilization. The full design and methodology

of NCICAS is detailed in accompanying articles pub-lished in this journal issue.11,12 Children were recruitedfrom EDs and clinics in inner-city areas based on phy-sician-diagnosed asthma and presence of asthma symp-toms for more than 3 days in the previous 12 months, orcough, wheezing, or shortness of breath that lasted morethan 6 weeks in the previous 12 months. This recruitmentprocedure yielded a sample in which 9.5% of the childrenhad not been diagnosed with asthma but were recruitedbased on asthma symptoms.

Skin-prick allergy tests, pulmonary function, and urinecollection for cotinine were done. Home visits were con-ducted to gather dust samples for roach, dust mite, andcat allergens in a large subsample of the families. Duringthis home visit, the environment was assessed and thecollection of nitrogen dioxide (NO2) specimens was ini-tiated. Following this baseline assessment, telephone as-sessments of symptoms and utilization were conducted at3-, 6-, and 9-month intervals. Recall over the prior 3months for utilization (hospitalizations, ED visits, andunscheduled physician visits) was assesed at baseline andat each 3-month follow-up telephone call, providing afull year of utilization information for the sample. Ateach of these four measurement periods, a 2-week recallof symptoms was obtained (loss of sleep, slow-down ofplay, and wheeze).

Home visits were conducted on 663 (43.4%) of studysubjects between November and July. Using a checklist,information was recorded both from direct inspection(such as general repair, carpeting, mattress covers, evi-dence of infestation, and of smoking in the home) orfrom questions to the family (such as laundry arrange-ments, pets, or infestations).

Skin Tests

Skin testing was performed by the prick puncturemethod on the volar surface of the forearm using a Mul-titestert device.13 Allergen extracts (D. pteronyssinus, D.farinae, cat pelt, dog pelt, rat pelt, mouse pelt, Germanand American cockroach mixture,Alternaria tenuis,Penicillium, mixed grasses, orchard grass, white oak,maple, and giant and short ragweed) were obtained inglycerosaline (Greer Laboratories, Lenoir, NC). All ex-tracts were 1:20 weight/volume except mites, which were10,000 Allergy Units (AU) per ml. Resulting whealswere outlined in ink 15 minutes after puncture and theoutline transferred to tape. A test was considered valid ifthe negative control wheal was at least 1 mm smaller thanthe positive histamine wheal. A test for any given antigenwas considered positive if its value exceeded the nega-tive wheal by at least 2 mm.

Smoking

Urine samples were collected from children at the timeof the baseline interview, then transferred to capped plas-

Abbreviations

AU Allergy UnitsBTPS Body temperature and barometric pressure, saturatedCCR Cotinine/creatine ratioED Emergency DepartmentETS Environmental tobacco smokeFEV1 Forced expiratory volume in one secondFVC Forced vital capacityNCICAS National Cooperative Inner-City Asthma StudyNO2 Nitrogen dioxide

254 Kattan et al.

tic tubes, frozen, and shipped in dry ice to a centrallaboratory at Henry Ford Hospital. Samples were ana-lyzed for cotinine content with an ELISA method usinga polyclonal antibody (Cotitraqt; Serex, Englewood,NJ).14 Urinary creatinine was assayed by the clinicalchemistry laboratory at Henry Ford Hospital and resultswere reported as continine/creatinine ratio (ng/mg).

Nitrogen Dioxide

Questions potentially relevant to NO2 exposure wereasked during the baseline interview and confirmed dur-ing the home visit. At the time of the home visit a Palmestube was taped to a wall in the child’s sleeping area. Thefamily was instructed to seal and mail them back to theasthma study center after 7 days. The tube was shipped toa central laboratory for analysis. The samplers have asensitivity of 600 ppb-hr, so that over 7 days a concen-tration above 3.6 ppb could be detected and samplesdiffering by this same amount could be distinguished.12

Peak Flow

Participants and their caretakers were instructed on theuse of a portable Mini-Wright peak flow meter during thebaseline visit and re-instructed by telephone quarterlyprior to interval assessments. Subjects were instructed touse the peak flow meter twice daily prior to use of aninhaled bronchodilator (in the morning and evening) fora 2-week period, recording the largest of three maximalexpiratory maneuvers on a peak flow/symptom diary thatwas returned in person or by mail. They also were askedto record information regarding concurrent symptomsand illnesses on this diary. Participants were told that thepurpose of peak flow monitoring was to help doctorsbetter understand how to treat asthma. They were notgiven specific instructions on interpreting peak flow val-ues; rather, they were instructed to follow their physi-cian’s usual guidelines for managing their asthma.

For each 2-week period, the mean absolute peak flowand the mean percentage of predicted, peak flow, ad-justed for height, gender, race, and age, were calculatedfor all records containing at least 14 (out of a total maxi-mum of 28) entries.15 Summary measures (over all fourassessments) were calculated for those with valid records(i.e., $14 entries) for at least two surveys (n4 1,007).

Pulmonary Function

Spirometry (pre-and post-bronchodilator) was as-sessed at the baseline visit in children age$6 years usinga 12 L dry-sealed spirometer (Pulmo-Screen IIE/VRS;S&M Instrument, Doylestown, PA). Subjects were notasked to refrain from medication use prior to testing.Prior to use of an inhaled bronchodilator, at least five and

no more than eight attempts were made to obtain at leastthree acceptable forced expiratory maneuvers. In subjectswho had not used an inhaled bronchodilator within 2hours of testing, acute bronchodilation was achieved withinhalation of a nebulized 0.083% albuterol solution. Tenminutes following this, spirometry was repeated, with nomore than five attempts to obtain three acceptable forcedexpiratory maneuvers. All maneuvers were done in thestanding position, with the nose pinched closed manu-ally.

All studies were reviewed by a single observer foracceptability and quality of the FEV1 (flow) and FVC(volume). FEV1 was considered ‘‘acceptable’’ when theback-extrapolated volume was <5% of the FVC and thefirst 1 sec of the maneuver was free from artifact orcoughing, and ‘‘optimal’’ when the flow-volume curveshowed a sharp, rapid rise to peak expiratory flow (<80msec from start of test). The FVC was considered ac-ceptable when expiratory time was >3 secs and a plateau(no visible change in volume) was demonstrated duringthe last 1 sec of exhalation. The FVC was consideredoptimal when a plateau was demonstrated and the ma-neuver exceeded 4 sec. All data were adjusted for bodytemperature and barometric pressure (BTPS). Predictedvalues, specific for race, gender, and height, were calcu-lated based on the equations by Dockery et al., using themaximal acceptable values.16

RESULTS

Demographics of the NCICAS population

A total of 1,528 caretaker and child pairs completedthe baseline interview. The 3-, 6-, and 9-month assess-ments were completed by 91%, 92%, and 95% of theparticipants. Demographic information is shown in Table1. The average age of the children was 6.2 years and 62%of the children were boys. At the time of the baselineinterview, 79.9% were in school, ranging from preschoolto grade 4. According to the caretaker’s report, 73% ofthe children were African American, 20% Latino, and7% were white/other. The vast majority (96%) of thecaretakers were women. Approximately 23% reportedthat they were currently married, 53.8% were single, andthe remainder were divorced, separated, or widowed.Sixty-six percent of the caretakers and/or mothers and70.1% of the biological fathers had at least a high schooleducation.

Table 2 shows the household income. Sixty-one per-cent reported a household income below $15,000, whichis the federal poverty level for a family of four. Slightlymore than half (52.5%) of the children lived in a homewhere at least one adult had a regular paying job.

Among the recruited children, 24.7% were reported asbeing in a neonatal intensive care unit; 10.3% were on a

Inner-City Children With Asthma 255

respirator at birth, while 17.9% had low birth weight(<5.5 lbs.) More than half (57.4%) had a family historyof asthma.

Utilization and Asthma Symptoms

Table 3 shows health system utilization, asthma symp-toms and impact of asthma on the caretaker. Utilization(hospitalization and unscheduled visits to the ED, clinic,or physician’s office) was based on a 3-month recall. Allsymptom measures (wheeze, loss of sleep, reduced playactivities, etc.) were based on a 2-week recall with theexception of ‘‘percent school days missed,’’ which wasbased on the last 3 months. All of these measures, in-cluding utilization measures, were collected at baselineand at each 3-month interval assessment. For these chil-dren, they averaged approximately 3 to 31⁄2 days ofwheeze for each of the four 2-week recalls. The caretaker

indicated that the children had approximately 2 days perweek of reduced play activity due to their asthma, andjust less than 2 days per week of sleep disruption due toasthma. Since children had varying numbers of days ofschool during the prior 3 months, the results presentedhere are based on the percent of available school daysmissed due to asthma. Given that a 3-month periodwould permit approximately 65 days of school, thesechildren are missing approximately 4 days (6%) to 61⁄2days (10%) due to asthma.

Table 4 presents the distribution of unscheduled visitsdue to asthma in a year and the number of hospitaliza-tions due to asthma in a year, respectively. In this popu-lation, 34.3% had no unscheduled visits and 82.9% hadno hospitalizations.

Access to Care

Data on the usual place of follow-up asthma care andinsurance status are noted in Table 5. For 84.8%, thechildren received follow-up asthma care from the sameplace where they received primary care, but for only39.5% was this the same place that they also receivedcare for acute symptoms. A hospital-based pediatricclinic was the most common site for follow-up care(38.4%), followed closely by a health center (36.1%).However, only 67.4% saw the same provider for asthmafollow-up and primary care.

Medications

As part of the baseline interview, caretakers reportedthe asthma medications prescribed for their child duringthe previous 3 months. In Table 6 the results are dis-played separately for subjects recruited in the ED duringan acute attack, for those recruited during a non-acutevisit, and for the population as a whole.

Overall, 17% of the children were not using anyasthma medications. Nearly twice as many ED recruitswere prescribed two or more medications compared tonon-acute recruits (82% vs. 48%); this ratio was reversedfor one medication. Among children taking asthma medi-

TABLE 1—Demographics

N %

Race of childAfrican American 1111 73.5Hispanic 295 19.5Other 106 7.0Gender of child

Male 954 62.4Female 574 37.6

Age of child4 332 21.75 296 19.46 291 19.07 202 13.28 215 14.19 192 12.6

Gender of caretakerMale 56 3.7Female 1471 96.3

Marital status of caretakerMarried 353 23.3Not married 1165 76.7

Mother completed high school/GEDYes 930 66.7No 465 33.3

Father completed high schoolYes 949 70.1No 404 29.9

Household members with regular jobNone 718 47.5One or more 794 52.5

Number of people in household2 68 4.53 245 16.14 362 23.85 349 23.06 225 14.87 104 6.98 72 4.79 46 3.0

10+ 47 3.1

TABLE 2—Household Income

Income % (n4 1,528)

Under $5,000 17.0$5,000–9,999 21.7$10,000–14,999 16.3$15,000–19,999 8.8$20.000–29,999 14.1$30,000–49,999 10.3Over $50,000 1.6Don’t know 9.3Refused 0.7Missing 0.7

256 Kattan et al.

cations, 54% of acute ED recruits and 72% of non-acutesubjects were taking oral beta-agonists. Overall, 28% ofchildren taking medications were receiving inhaled ste-roids or cromolyn. Beta-agonists were the only medica-tion used by 42% of children taking any asthma medi-

cation. Of the 952 children attending school, 29% statedthat the school did not allow the administration of medi-cations. Nineteen percent of caretakers reported that theywere giving some medications less often than prescribed.

Home Visits

Only 17% of the families lived in detached houses.Most families lived in apartments (57%), duplexes(10%), or row houses (14%). Forty-four percent of thehomes were estimated to be more than 50 years old andwere frequently in poor repair with leaky roofs (12%),broken windows (13%), broken plaster (28%), and peel-ing paint (32%). Wall-to-wall carpet was common (41%of TV/living rooms, 32% of bedrooms), but only 38%had a functioning vacuum cleaner. Signs of roach infes-tation were seen in 66% and of mice in 29%.

Skin Tests

A total of 1,286 children had complete skin test resultswith valid positive and negative control test results. Ofthis number, 989 (77%) had at least one positive skin test,and 606 (47%) had three or more positive skin tests. Thefrequency of positive responses to various allergen ex-tracts is shown in Figure 1.

Smoking

At the baseline assessment, 887 families (59%) in-cluded at least one smoker and 588 caretakers (39%)

TABLE 3—Utilization, Symptoms, and Impact on Caretaker

Baseline(1,528)

Interval Assessments

3 Month(1,392)

6 Month(1,411)

9 Month(1,446)

UtilizationUnscheduled Visits1 Mean 0.70 0.39 0.46 0.48

S.D. 0.95 0.78 0.87 0.85Hospitalizations2 Mean 0.11 0.06 0.06 0.05

S.D. 0.37 0.29 0.28 0.25Symptoms

Wheezing1 Mean 2.93 3.21 3.66 3.57S.D. 3.64 3.72 4.02 3.96

Slow Play1 Mean 2.39 1.99 2.02 1.84S.D. 3.54 3.22 3.19 2.92

Sleep Loss1 Mean 1.96 1.67 1.82 1.75S.D. 3.24 2.92 2.99 3.02

Percent School Days Mean 10% 7% 6% 6%Missed2 S.D. 13% 12% 11% 8%

Impact on CaretakerCaretaker Sleep Loss1 Mean 2.52 2.07 2.33 2.10

S.D. 3.98 3.49 3.65 3.55Caretaker Changed Mean 3.74 2.76 2.53 2.81Plans2 S.D. 8.35 7.24 7.04 8.11

1Days per two-week recall.2Days per three-month recall.

TABLE 4—Distribution of Unscheduled Visitsand Hospitalizations

Numberper year

Unscheduled visits Hospitalizations

% (n 4 1,488) % (n4 1,364)

0 34.3 82.91 19.4 10.62 13.0 3.63 11.6 1.84 8.9 0.45 4.7 0.46 2.1 0.17 1.9 0.18 1.6 09 1.1 0.1

10+ 1.4 0Total 100% 100%

TABLE 5—Measures of Access to Health Care

Have used follow-up care (n4 1,297, 84.9% of total) N %

Have a usual place for follow-upasthma care for their child

1,246 96.1

Find it difficult to obtain follow-up care 692 53.4Insurance (n4 1,508, 98.7% of total)

Have insurance coverage for their child 1393 92.4Medicaid 1101 73.1Health maintenance organization 162 10.7Private insurance 130 8.6

Inner-City Children With Asthma 257

reported that they smoked. Someone was smoking in thehome during 10% of the home visits. Of 1,446 urinesamples, 48% collected from the asthmatic children atthe time of the baseline interviews had a cotinine/creatineratio (CCR) above 30 ng/mg, which is consistent withsignificant tobacco smoke exposure within the last 24hours.

Nitrogen Dioxide

At the home visit, 89% of homes used a gas stove and16% used the stove or a space heater to supplement homeheating. Only 13% of kitchens had an outside vent fanand 24% of kitchens either had no outside window or awindow that was painted shut. Study personnel placedPalmes tubes and 525 were returned. Nitrogen dioxideconcentrations were as high as 480 ppb and 24% of thechildren were exposed in their sleeping area to meanlevels higher than 40 ppb during the 1-week collectionperiod.

Peak Flow

A total of 3,616 peak flow diaries were received, but166 were excluded because they contained fewer than 14out of the required 28 readings. The remaining diariesrepresented 59%, 61%, 60%, and 59% of the patientswho were contacted at baseline, 3, 6, and 9 months.

Table 7 provides the absolute peak flows and theheight, gender, and race-adjusted percentage of predictedpeak flows for each 2-week monitoring period. Althoughthe mean values were within the normal range (84–89%of predicted), there was a wide coefficient of variation of25–26%.

Pulmonary Function

Of the 900 participants age 6–9 years, 827 (92%) hadspirometry performed. Of these, 683 subjects (83%) hadat least one acceptable FEV1 and 476 (56%) had at leastone acceptable FVC value obtained during a pre-bronchodilator maneuver. Bronchodilator inhalation wasperformed on 691 subjects, and resulted in at least oneacceptable FEV1 in 582 subjects (84%) and at least oneacceptable FVC in 451 subjects (65%).

Mean (± SD) percent predicted FEV1 and FVC were96.6% ± 16.9% (range 31.2 to 142.4%) and 99.5% ± 15.3(range 32.1 to 137.5), respectively, and the FEV1/FVCwas 0.83 ± 0.09. Only 9.7% of subjects with acceptableFEV1 values had FEV1 levels <75% of the predictedvalue, adjusted for height, gender, and race. Following abronchodilator, FEV1 increased by a mean of 10.2% ±13% (p < 0.0001) and FVC increased by 4.7% ± 9.4% (p< 0.001). Of the 528 subjects with acceptable FEV1 mea-surements before and after bronchodilator administra-tion, 182 (34.5%) had a$12% increase in post-bronchodilator FEV1.

DISCUSSION

This report describes the baseline characteristics andthe environment of the subjects participating in the inner-city study of children with asthma and recruited by theNCICAS project. Because of the variable nature ofasthma, a full year’s morbidity experience provided abetter characterization of the individual child withasthma than previous short-term studies. This study isunique in that asthma symptoms and health care utiliza-tion data were collected in a prospective manner by theuse of four contacts—the baseline interview and threetelephone contacts at 3, 6, and 9 months post-baseline.

TABLE 6—Asthma Medications

Medications reported to havebeen prescribed during theprevious three months

Recruited in EDduring acute attack

Recruited duringnonacute visits Total

Number of medicationsN 363 1,101 1,464None 2% 22% 17%One 16% 30% 27%Two or more 82% 48% 56%Total 100% 100% 100%

Type of Medication*N 355 855 1,210Oral beta-agonists 54% 72% 66%Inhaled beta-agonists 66% 49% 54%Oral xanthines 18% 20% 19%Oral steroids 59% 19% 31%Inhaled steroids or cromolyn 37% 24% 28%Only beta-agonists 20% 51% 42%

*Percent is based on 1,210 (355 acute and 855 not acute) patients prescribed at least one medication.Acute non-ED recruits (n4 20) are not included and 44 participants did not respond.

258 Kattan et al.

The objective of this prospective method of data collec-tion was to maximize the accuracy of the data collectedby minimizing the period of recall.

The primary focus of the NCICAS recruitment was tosample children with asthma living in the inner-city.However, NCICAS was not designed to be a population-based study of inner-city childhood asthma. The recruit-ment process was based in medical facilities and not incommunity facilities and it emphasized the recruitmentof moderate to severe asthmatics. As a consequence, themorbidity experience of the NCICAS population as re-flected by asthma hospitalizations was higher than thattypically reported in the literature. The direct impact ofrecruiting children during acute attacks at the ED can be

seen by their higher baseline morbidity, shown in Table2. Using the 1988 Child Health Supplement from theNational Health Interview Survey, Halfon, et al. found10.6% of poor children with asthma were hospitalized inthe last year, as compared to 17.1% in the NCICASsample.17 Lozano et al. found an average of 0.085 hos-pitalizations per person-years in asthmatic children ages3 to 17 years receiving Medicaid in Washington State, ascompared to 0.28 hospitalizations in the NCICAS popu-lation.18 In contrast, a number of population characteris-tics were as expected in an inner-city population. TheNCICAS sample largely consisted of minorities living ator below the poverty level. The prevalence of low birthweight in our population was twice the national average,

Fig. 1. Frequencies of responses to allergen extracts. Prick skin tests on 1,286 children.

TABLE 7—Peak Flows

Baseline 3-Month 6-Month 9-Month

Peak flow (1/min)Mean 198.5 201.5 204.0 206.1S.D. 65.4 67.7 67.7 69.9min 61.4 67.2 61.5 62.0max 439.3 413 450 445.3n 908 844 849 849

% Predicted peak flowMean 84.2% 86.0% 87.1% 88.7%S.D. 21.1 21.6 22.4 23.5min 20.5 20.7 23.0 22.9max 199.2 228.7 170.9 199.3n 838 776 766 772

Inner-City Children With Asthma 259

but was similar to levels reported previously for pre-dominantly poor African Americans and US-born His-panics.19,20

It is evident that, in this population, symptoms werefrequent, adding up to more than 2 months per year. Thesymptoms were severe enough to have an impact onschool attendance, sleep, and caretaker activities. Despitethe frequency of symptoms, hospitalization was an infre-quent event. Only 3.7% of the population had two ormore hospitalizations and 83% had none. Two percent ofthe population had six or more unscheduled visits andone-third had none. A small proportion of patients re-quired hospitalization or made multiple ED visits; otherstudies have shown that these events account for 43% ofthe costs of asthma.21

The presence or absence of a regular source of primaryor preventive care is considered particularly importantfor individuals with ongoing health conditions such asasthma and has traditionally been used to measure accessto care. It was surprising, given the morbidity in ourpopulation and the importance of access to good qualityhealth care for preventing morbidity, that nearly all re-spondents reported having a usual place for follow-upcare. However, over half of the respondents reported thatit was difficult to get care for an asthma attack as well asfor follow-up visits. This is particularly intriguing since,in our sample, over 92% of children were covered byinsurance. Nearly three-quarters were covered by Med-icaid, which has been shown to improve access to carefor children. Our data support the findings of St. Peter etal. and Halfon and Newacheck17,22 that poor childrenwith Medicaid are very likely to report a usual source ofroutine care, but are likely to receive that care in a neigh-borhood health center or hospital-based clinic rather thana physician’s office and are unlikely to have continuitybetween usual sources of routine (follow-up) and acutecare.

If the variable ‘‘having a usual place for follow-upasthma care’’ adequately measured access, we would nothave found that more than 50% of respondents reportedthat it was hard to get follow-up care, particularly sincenearly all children in our sample were covered by insur-ance. We will have to investigate specific barriers to careas well as different ways to measure access to care.Moreover, variations or limitations in access do not tellus much about the quality of care delivered. It may bethat attributes of the care and the relative lack of conti-nuity of care for these children may contribute more tothe relationship between health care delivery and mor-bidity among children with asthma than more traditionalmeasures of access.

These data suggest that inner-city children with asth-ma are more likely to receive two or more medicationsduring the previous 3 months if they have had an asthmaattack. While there is a possibility that some of the chil-

dren recruited during a non-acute visit had an acute ex-acerbation during the three months prior to the baselineinterview, these children were nearly twice as likely toreport being prescribed a single medicine and half aslikely to be prescribed two or more medicines.

Among the non-acute recruits, the reported prescrip-tion of inhaled anti-inflammatory medications was 24%,compared to 22% for US pediatricians in 1993.23 Thepopulation in this study had an overall level of morbidityand medications prescribed that indicate a symptomaticpopulation. If there were adherence to current treatmentguidelines, we would have expected greater use of anti-inflammatory medications for the degree of symptomsreported. Use of beta-agonists as the only medication wasmore widespread than inhaled anti-inflammatory medi-cations. Although this suggests sub-optimal therapy, fur-ther investigation is needed to assess whether the patternof medication use is due to poor access to quality medicalcare or poor adherence to a preventive medical regimen.

Although the frequency of positive skin tests seen inthis population (77% with one or more positive skintests) is similar to that seen in other studies of asthmaticchildren, the pattern of positive skin tests seen amonginner-city children in this study differs from previousreports. In studies of unselected or middle-class asth-matic children, 50–60% had positive skin tests to housedust mite, 30% or more had positive skin tests to cat ordog, and positive skin tests to cockroach was uncom-mon.24,25Previous reports regardingAlternaria sensitiv-ity in children have indicated that 4 to 13% had a positiveskin test, which is substantially lower than in this popu-lation.26,27 The pattern of observed skin test sensitivityclearly reflects the unique allergenic environment towhich inner-city children are exposed and may be animportant consideration in designing environmental in-terventions in this population.

This study demonstrates that exposure to environmen-tal tobacco smoke (ETS) is common among inner-citychildren with asthma. The percent of homes with at leastone smoker, the percent of asthmatics who smoke, andthe percent of children with a CCR greater than 30 ng/mgis similar to a previous study of inner-city asthmatics.28

The question arises as to whether these percentages areoverestimates because of the deliberate recruitment of asubstantial proportion of acute asthmatics from the ED.This is unlikely because Ehrlich et al. did not demon-strate differences in exposure between those with acuteand non-acute asthma.28 Since ETS is a risk factor forasthma and associated with increased bronchial reactiv-ity, the data suggest that reduction of this risk factorshould be an important goal for an intervention program.

On average, peak flows were within normal limits butvaried greatly among participants. There was limited ad-herence to this aspect of the protocol, i.e., less than two-

260 Kattan et al.

thirds of participants returned peak flow diaries. Thislack of adherence is likely related to the demands madeby twice-daily monitoring and the limited clinical feed-back provided to participants regarding the personal useof peak flow data. In pilot studies, we demonstrated thatactual adherence to peak flow monitoring in a similarresearch setting was limited by large amounts of missingor fabricated data that increased markedly after twoweeks of monitoring.29 These data emphasize that re-search protocols that make daily demands on participantsmay have limited success unless clear personal benefitfrom participation can be demonstrated directly to par-ticipants.

Although results from spirometry and peak flow moni-toring showed a good level of agreement between abso-lute FEV1 and mean peak flow (r4 0.65,P < 0.0001),on average, the percent of predicted peak flows werelower than percent predicted FEV1. It is likely that spi-rometry and peak flow monitoring provide somewhatdifferent types of information on airway physiology.Other possibilities include different reference popula-tions used for calculating percent predictive values anddifferences in effort between supervised and unsuper-vised maneuvers.

Because of the young age and potential morbidity inthis study population, the criteria for acceptability of pul-monary function curves were modified from those rec-ommended by the ATS to allow analysis of FEV1 or FVCdata, even if maneuvers were not available that met bothstart and end of test criteria. Data were also analyzedeven if they were based on only one acceptable value(205 FEV1 values and 89 FVC values measured pre-bronchodilator); however, mean FEV1 and FVC fromchildren with only one acceptable value did not differsignificantly from data of subjects with two or more ac-ceptable maneuvers. Interestingly, mean levels of FEV1

and FVC were close to 100% of the predicted values, andonly 9.7% of subjects had values <75% of predicted.This suggests that in our high-risk population, those chil-dren who could adequately perform the forced expiratorymaneuvers, had ‘‘baseline’’ levels of spirometry thatwere largely within the normal range. For those recruitedduring an acute attack, baseline studies were performedat least 3 weeks following the episode and, thus, partici-pants may have been under better asthma control than atother times. The exclusion of children unable to performthe spirometric maneuvers because of poor lung functionmay result in an underestimation of the mean FEV1 val-ues. This possibility is less likely to be a major factorbecause younger age was related to the inability to per-form the maneuvers. These data nevertheless suggest thateven among a ‘‘high risk’’ inner-city population, as rep-resented in NCICAS, many children have normal lungfunction measured at a single point in time, apart fromduring an acute episode.

A significant bronchodilator response is often consid-ered a sine qua non for asthma diagnosis. Only one-thirdof our population of children demonstrated a 12% orgreater increase in FEV1. The relatively high levels oflung function may indicate that most participants weretested during times of relatively minimal bronchospasm.These data emphasize the potential limitations of using abronchodilator response as a diagnostic tool, and the in-fluences of medication use, intrinsic variability in re-sponses, and baseline lung function must be considered.

SUMMARY

The data presented in this report address baseline char-acteristics of a broad-based population of children withasthma. A large number of possible risk factors forasthma morbidity were clearly present in this population.Further analysis is required to assess the interactions ofasthma morbidity with the environment, psychosocialfactors, access to care, and adherence. In addition, therelationship with ethnicity and severity of asthma needsfurther analysis.

Data regarding morbidity of inner-city children withasthma have mainly addressed hospitalization and EDvisits. This study provides comprehensive data regardingsymptoms and limitation of daily activities, in addition toutilization of hospital services in geographically diverseinner-city populations with asthma.

In the past, the cost-effectiveness of asthma interven-tions have been evaluated by their ability to reduce the‘‘direct’’ expenditures for asthma such as hospitaliza-tions and ED visits. These interventions are aimed at thehigh users of such services—a relatively small propor-tion of asthmatics. A much larger portion of the asth-matic population can be targeted by interventions aimedat reducing asthma symptoms to lessen the burden ofasthma on the daily life of the family. Potentially greatercost savings are possible for this type of intervention asthe ‘‘indirect’’ costs of asthma associated with time lostfrom work by the parents, disruption of family life, andthe child’s school activities are almost twice as high asthe ‘‘direct’’ cost for children under 18 years of age.21

REFERENCES

1. Current Trends in Asthma—United States, 1982–1992. MMWR.1995; 43:952–955.

2. Weitzman M, Gortmaker SL, Sobol AM, Perrin JM. Recent trendsin the prevalence and severity of childhood asthma. JAMA. 1992;268:2673–2677.

3. Gergen PA, Weiss KB. Changing patterns of asthma hospitaliza-tion among children: 1979–1987. JAMA. 1990; 264:1688–1692.

4. Carr W, Zeitel L, Weiss KB. Variations in asthma hospitalizationsand deaths in New York City. Am J Public Health. 1992; 82:59–65.

5. Lang DM, Polansky M. Patterns of asthma mortality in Philadel-phia from 1969 to 1991. N Engl J Med. 1994; 331:1542–1546.

6. Targonski PV, Persky VW, Orris P, Addington W. Trends in

Inner-City Children With Asthma 261

asthma mortality among African Americans and whites in Chi-cago, 1968 through 1991. Am J Public Health. 1994; 84:1830–1833.

7. Evans R, Mullally DI, Wilson RW, Gergen PJ, Rosenberg HM,Graham JS, Cherarley F. National trends in the morbidity andmortality of asthma in the US: Prevalence, hospitalization anddeath from asthma over two decades: 1965–1984. Chest. 1987;91:65S–74S.

8. Wissow LS, Gittelsohn AM, Szklo M, Starfield B, Mussman M.Poverty, race and hospitalization for childhood asthma. Am JPublic Health. 1988; 78:777–782.

9. Weiss KB. Seasonal trends in US asthma hospitalizations andmortality. JAMA. 1990; 263:2323–2328.

10. Crain EF, Weiss KB, Bijor PE, Hush M, Westbrook L, Stein REK.An estimate of the prevalence of asthma and wheezing amonginner-city children. Pediatrics. 1994; 94:356–362.

11. Wade S, Weil C, Holden G, Mitchell H, Evans R, Kruszon-MoranD, Bauman L, Crain E, Eggleston P, Kattan M, Kercsmar C,Leickly F, Malveaux F, Wedner HJ. Psychosocial Characteristicsof Inner-city children with Asthma: A description of the NCICASpsychosocial protocol. Pediatr Pulmonol. 1997;24:263–276.

12. Mitchell H, Senturia Y, Gergen P, Baker D, Joseph C, McNiff K,Wedner HJ, Crain E, Eggleston P, Evans R, Kattan M, KercsmarC, Leickly F, Malveaux F, Smartt E, Weiss K. Design and meth-ods of the National Cooperative Inner-City Asthma Study. PediatrPulmonol. 1997;24:237–252.

13. Berkowitz RB, Tinuelman DG, Lutz C, Crummie D, Smith K.Evaluation of the multi-test evidence for immediate hypersensi-tivity skin testing. J Allergy Clin Immunol. 1992; 90:979–985.

14. Langone JJ, Cook G, Bjercke RHJ, Lifschitz MH. Monoclonalantibody ELISA for cotinine in saliva and urine in active andpassive smokers. J Immunol Methods. 1988; 144:73–78.

15. Hsu KHK, Jenkins DE, Hsi BP, Bourhofer E, Thompson V, HsuF, Jacob SC. Ventilatory functions of normal children and youngadults—Mexican-American, White, and Black. II. Wright peakflowmeters. J Pediatr. 1979; 95:192–196.

16. Dockery DW, Berkey CS, Ware JH, Speizer FE, Ferris BG. Dis-tribution of forced vital capacity and forced expiratory volume inone second in children 6 to 11 years of age. Am Rev Respir Dis.1983; 128:405–412.

17. Halfon N, Newacheck PW. Childhood asthma and poverty: Dif-ferential impacts and utilization of health services. Pediatrics.1993; 91:56–61.

18. Lozano P, Connell FA, Koepsell TD. Use of health services byAfrican-American children with asthma on Medicaid. JAMA.1995; 274:469–473.

19. Collins JW, David RJ. The differential effect of traditional riskfactors on infant birth weight among Blacks and Whites in Chi-cago. Am J Pub Health. 1990; 80:679–681.

20. Collins JW, Shay DK. Prevalence of low birth weight amongHispanic infants with US-born and foreign-born mothers: Theeffect of urban poverty. Am J Epidemiol. 1994; 139:184–192.

21. Weiss KB, Gergen PJ, Hodgson TA. An economic evaluation ofasthma in the United States. N Engl J Med. 1992; 326:862–866.

22. St. Peter RF, Newacheck PW, Halfon N. Access to care for poorchildren, separate and unequal. JAMA 1992; 267:2760–2764.

23. Szefler SJ, Bender BG, Jusko WS, Lanier BQ, Lemanske RF,Skoner DP, Stempel DA. Evolving role of theophylline for treat-ment of chronic childhood asthma. J Pediatr. 1995; 127:176–185.

24. Murray AB, Ferguson AC, Morrison BJ. The frequency and se-verity of cat allergy vs. dog allergy in atopic children. J AllergyClin Immunol. 1983; 72:145–149.

25. Willies-Jacobo LJ, Denson-Lino JM, Rosas A, O’Connor RD,Wilson NW. Socioeconomic status and allergy in children withasthma. J Allergy Clin Immunol. 1993; 92:630–634.

26. Krilis S, Baldo BA, Basten A. Analysis of allergen-specific IgEresponses in 341 allergic patients: Associations between allergensand between allergen groups and clinical diagnoses. Aus NZ JMed. 1985; 15:421–426.

27. Kalliel JN, Goldstein BM, Braman SS, Settipone GA. High fre-quency of atopic asthma in a pulmonary clinic population. Chest1989; 96:1336–1340.

28. Ehrlich R, Kattan M, Godbold J, Saltzberg D, Grimm K, Land-rigan P, Lillienfeld D. Childhood asthma and passive smoking:Urinary cotinine as a biomarker of exposure. Amer Rev Res Dis.1992; 145:594–599.

29. Redline S, Wright EC, Kattan M, Kercsmar C, Weiss K. Short-term compliance with peak flow monitoring: Results from a studyof inner-city children with asthma. Pediatr Pulmonol. 1996; 21:203–210.

262 Kattan et al.