|Home | About | Journals | Submit | Contact Us | Français|
The National Children’s Study will address, among other illnesses, the environmental causes of both incident asthma and exacerbations of asthma in children. Seven of the Centers for Children’s Environmental Health and Disease Prevention Research (Children’s Centers), funded by the National Institute of Environmental Health Sciences and the U.S. Environmental Protection Agency, conducted studies relating to asthma. The design of these studies was diverse and included cohorts, longitudinal studies of older children, and intervention trials involving asthmatic children. In addition to the general lessons provided regarding the conduct of clinical studies in both urban and rural populations, these studies provide important lessons regarding the successful conduct of community research addressing asthma. They demonstrate that it is necessary and feasible to conduct repeated evaluation of environmental exposures in the home to address environmental exposures relevant to asthma. The time and staff required were usually underestimated by the investigators, but through resourceful efforts, the studies were completed with a remarkably high completion rate. The definition of asthma and assessment of disease severity proved to be complex and required a combination of questionnaires, pulmonary function tests, and biologic samples for markers of immune response and disease activity. The definition of asthma was particularly problematic in younger children, who may exhibit typical asthma symptoms sporadically with respiratory infections without developing chronic asthma. Medications confounded the definition of asthma disease activity, and must be repeatedly and systematically estimated. Despite these many challenges, the Children’s Centers successfully conducted long-term studies of asthma.
In this article we outline information from the Centers for Children’s Environmental Health and Disease Prevention Research (Children’s Centers) that have conducted studies related to asthma. We do not include information on sampling strategies because the goal of the monograph is to inform the National Children’s Study. Rather, we focus on major issues related to the identification of asthma, asthma-related symptoms and end points, relevant exposures, biologic markers, and follow-up requirements.
Studies of asthmatic children and their homes were conducted at seven Children’s Centers. The protocols for these studies are summarized in Table 1, and further information is provided in the introduction to this mini-monograph (Kimmel et al. 2005). The birth cohort studies conducted at two Children’s Centers are described more completely by Eskenazi et al. (2005); they are included here because both Children’s Centers address respiratory cross-sectional data and outcomes. In addition, three Children’s Centers conducted cohort studies with older children with established asthma, and the study at Johns Hopkins University (JHU) included a nonasthmatic control group as a comparator in initial observations. Four Children’s Centers conducted intervention trials, with diverse designs ranging from a randomized primary prevention trial in school-age children to formal randomized controlled clinical trials.
The studies that recruited children with asthma had key similarities (Table 1). Children were recruited whose parents reported a history of doctor-diagnosed asthma. Other children were recruited as controls. There were few exclusion criteria in most studies except for other respiratory diseases. Because all participants in these studies were minors, the investigators had to obtain informed consent from a parent or other guardian. When recruiting children > 7 years of age, assent was obtained from the child in addition to the consent obtained from a parent or guardian. In the two Children’s Centers that conducted birth cohort studies, mothers consented to obtain information regarding their own and their child’s medical and exposure histories (Eskenazi et al. 2005). In certain cases, the study setting included prespecified areas of a city bounded by ZIP codes or prespecified counties. In other cases, the study setting was defined by participants seeking care at a specific clinic or medical center regardless of where they lived.
Although three studies used local schools (University of Iowa, JHU intervention study, University of Michigan), most studies identified the sample from a health care setting such as health plans, emergency departments, and physician offices. In schools, the protocol was approved both by an institutional review board (IRB) and by a school system IRB. Informed consent was mailed to the parents to be signed, and followed by a mailed questionnaire in the case of Iowa and Michigan; in the JHU intervention study, the mailed consent was followed by a visit from a recruiter. For other studies, informing and recruiting potential participants included both passive and active methods. The passive approaches included posting study information publicly and providing information when a potential participant’s parent or guardian called. These were generally less efficient in generating interested families to participate and generally were considered a supplement to active methods. For the studies of children, most active recruitment consisted of generating patient lists from computerized databases of children who met entry criteria. Study information was then mailed directly to participants, and study staff telephoned sometime later to determine the family’s interest, to confirm eligibility, and to initiate the informed consent process.
There were several barriers to recruitment of children to the studies. First among these was establishing the essential entry criterion for a diagnosis of asthma. This was a greater issue for those studies recruiting younger children (JHU longitudinal cohort study, Cincinnati) than in those focused on older children [JHU Intervention trial, University of Southern California (USC), Michigan, Iowa]. Older children were generally identified by a diagnosis being made by a physician and logged into the computerized database from which recruitment began. In some cases, children with typical asthma were entered into databases with related diagnoses such as recurrent bronchitis or reactive airway disease, which have slightly different codes in the International Classification of Diseases, 9th Revision [World Health Organization (WHO) 1978]. Study staff confirmed the diagnosis with screening questionnaires defining appropriate symptoms or with lung function tests. Younger children, who may have had asthmatic respiratory symptoms only during respiratory infections, were more difficult to qualify because they had less specific answers to respiratory questionnaires and could not perform lung function tests. In a birth cohort study, early wheeze may be recorded, with persistent wheeze or asthma being defined as an outcome later in life. However, entry criteria for the longitudinal cohort studies that recruited younger children used the same definition used with older children: doctor-diagnosed asthma. In a birth cohort study, cough or wheeze may be recorded as an outcome, with persistent wheeze or definable asthma used as an outcome later in life.
An important barrier was introduced by the need to preserve patient autonomy and privacy. IRBs have long required investigators to accommodate patients listed on health care databases who do not want to be contacted to participate in research studies. This has usually been accomplished by mailing an invitation letter containing a preaddressed and stamped postcard that families can return requesting that they not be called. In general, this has proven to be a minor barrier, excluding < 5% of identified families. The recently issued Health Insurance Portability and Accountability Act (HIPAA) regulations have introduced additional barriers by requiring that families give permission to use their health information even at preliminary stages of research (e.g., investigators having access to the health care database to obtain lists of children with a diagnosis of asthma). For example, HIPAA makes it illegal to create lists of persons with a specific diagnosis to whom to send an introductory letter and makes the approach used by most of these studies illegal. Various solutions have been developed, including having patients sign general waivers of HIPAA privacy rights as they enroll in a health care program or having their health care provider recommend them to the interviewer at the time of a health care encounter. Other approaches have included asking asthmatic patients or their parents who participate in health fairs, surveys, self-management programs, or other patient care activities to indicate in writing their willingness to participate in future research studies and creating a list of those who have so indicated. IRBs may approve a HIPAA waiver, but the investigators must justify why a waiver is necessary—that is, that the data being gathered are not sensitive, are not able to be linked with individual identifiers, or both are noninvasive and cannot be collected in another way. This requirement also involves clinic personnel, who are much less effective as recruiters that are trained, motivated study staff. Recruitment from school directories avoids these problems because health data are not used to identify potential subjects. One study that recruited from schools did so by conducting a mailed asthma screening survey of 10 school districts (87% response rate) to identify children with asthma (Chrischilles et al. 2004). Another reported a 78% response rate from 9,437 mailings (Lewis et al. 2004).
The two birth cohort studies (University of California at Berkeley, Columbia University) recruited pregnant mothers and thus faced the issue of participant identification to a smaller extent. At the time the studies were initiated, the HIPAA regulations had not been activated, so they did not face this issue. Were the studies to be conducted now, recruitment of identified pregnant mothers (with or without asthma) would still have to comply with the HIPAA requirements and, indeed, do so in the continued surveillance of the children.
A final barrier when dealing with families with lower socioeconomic status is obtaining a reliable means to contact persons on a list. Families move frequently (29% moved at least once during a 1-year observation) (Swartz et al. 2004), so the address and telephone numbers in a database may not be current. Additionally, as many as 25% of inner-city families may not have telephones (Wissow et al. 1988), and 52% may change their telephone number at least once during a year of follow-up (Swartz et al. 2004). This not only makes initial contact difficult but also interferes with follow-up.
As shown in Table 1, retention rates were generally > 80% in those studies that involved follow-up. A number of barriers had to be overcome to achieve these high rates, including frequent telephone number changes, residential address changes, and general reluctance in making follow-up appointments. The usual solutions to the lack of telephones or frequent changes of telephones and addresses are to maintain frequent contacts with participants (generally every 1–2 months), to ask families for secondary (and tertiary) alternative contact information, or to provide incentives to notify the study personnel in the event of a change in address or telephone number (Mitchell et al. 1997). Most study staff members were able to conduct home visits even when telephone contact strategies had proven unsuccessful. However, families who moved without providing new contact information interfered with this strategy as well as with follow-up home inspections that were part of several protocols. For some Children’s Centers, missed appointments were important barriers to retention. Reasons for missed appointments included inclement weather and illness among family members needed to help transport the child, but in many cases no explanation was provided. Some studies categorized repeated “no-shows” as having passively withdrawn from the study. This practical measure was necessary to preserve study resources. Children’s Centers that conducted the research in the homes of subjects (i.e., did not require participants to come to the study center) also faced appointment cancellations.
The Children’s Centers employed multiple strategies to maintain contact with participants and to encourage their continued participation. Commonly used incentives included cash, gift certificates, toys for the children, food stamps, and infant car seats. A wide variety of gifts were provided, including T-shirts, tote bags, hats, key chains, stuffed animals, games, and musical instruments. These incentives were provided either at follow-up visits or on successful completion of the study. Additional incentives included health-related devices such as peak flow meters, spacers for metered dose inhalers, and allergen-proof mattress and pillow covers. One Children’s Center provided up to $200 as reimbursement for electricity costs incurred during the study. The Columbia Children’s Center, faced with participants who moved to a new location (e.g., Florida), paid for air travel to do follow-up assessments, and in one case a research worker has flown to the Dominican Republic to obtain follow-up data. The Berkeley Children’s Center created a movable laboratory in a recreation vehicle and went on the road once a year to other areas in California to include participants who moved. Another common incentive was to provide reimbursement for travel expenses related to study participation.
In some cases, the incentives had no monetary value but focused on information provided to the family. In the JHU cohort study, participants learned about home environmental measurements, including allergens and pollutants; many of these families indicated that this was the principal incentive that attracted them to the study. All investigators felt that the most important factor in participant retention was the Children’s Centers’ staffs. Retention was highest when the staff members were able to create an empathetic bond with the participating families, who felt that the staff would try to help them not only with their child’s illness but also with other difficulties associated with their social and economic circumstances. Families frequently sought help with problems such as transportation, referrals, and problems of daily living and valued the “ear” provided by the staff regardless of whether the staff member could actually help with the issue.
As shown in Table 2, all the Children’s Centers employed extensive questionnaires, collecting demographic, social, medical, and environmental exposure data similar to those collected in other studies. These data were collected repeatedly during longitudinal follow-up, although generally in abbreviated questionnaires.
Most Children’s Centers conducted ambient air sampling at one or more sites (Table 3). In addition, all Children’s Centers conducted home visits and in many cases conducted visits repeatedly (see Table 4). At these visits, the participants’ homes were formally inspected with checklists and families provided additional information through questionnaires. In most cases, settled dust samples were collected and were assayed for indoor allergens; concentrations of endotoxin and pesticides were also measured. Most Children’s Centers conducted repeated measures of environmental exposures, including inspection, settled dust sampling, and air sampling; completion rates for these evaluations ranged from 73 to 91%. The data were generally used to describe exposure at the times indicated and was compared with asthma morbidity at these times. Because the data were collected as several time points, time series analyses were conducted.
The number of studies that collected blood samples from children was remarkable (Table 5). These samples were generally used for radioallergosorbent tests (RAST) for specific immunoglobulin E (IgE) antibody to supplement or replace skin tests. Prick skin tests and RAST have been shown to correlate well for sensitivity to inhalant allergens (Wood et al. 1999) and food allergens (Sampson and Albergo 1984). When blood is saved for RAST, the investigator has the advantage of being able to test for additional sensitivities that were not considered at the time the study was conducted. The specific sensitivities determined in this manner were essential both to determine that a child was atopic and to detect specific IgE to the indoor allergens measured in home visits. Atopy, defined as a genetic predisposition to produce long-lived IgE antibody to environmental allergens in association with a constellation of chronic diseases including food allergy, eczema, allergic rhinitis, and asthma, is the most important risk factor for both incident asthma and asthma severity. The combination of specific IgE to environmental allergens and the presence of high concentrations of these allergens in the child’s home is the strongest known risk factor for asthma severity and morbidity (Rosenstreich et al. 1997).
In addition to blood samples, prick–puncture skin testing was usually done, and samples of meconium, urine, saliva, and hair were collected, usually to detect exposure to environmental tobacco smoke (ETS) and pesticides. What is most remarkable is that biologic samples were collected from children in most Children’s Centers on multiple occasions and that these collections did not have noticeable impact on participant retention rates (Table 1).
All of these studies included questions regarding asthma disease activity as shown in Table 6. Assessing disease morbidity in epidemiologic studies of children with asthma involves the creation of a composite of symptoms, health care, and medication use from questionnaires, together with pulmonary function tests in older children (Kattan et al. 1997). Although there is a general consensus about asthma-associated symptoms, there is less agreement with regard to specific questions that provide the most appropriate assessment of these symptoms. Major symptoms include cough, wheeze, chest tightness, and dyspnea, but these are quite variable day to day and depend on exposures to specific stimuli or “triggers” such as respiratory viruses, allergens, or irritants such as ETS and air pollutants. It is now generally accepted that symptoms should be assessed over 2-week intervals and that questions be framed in terms of the number of days during which these symptoms are experienced rather than frequency per day or symptom intensity. The two most widely used questionnaires were developed by the American Thoracic Society (ATS) (Ferris 1978) and the Children’s Health Survey for Asthma (CHSA) (Asmussen et al. 1999).
Several scales have been created to synthesize the symptoms into a description of disease activity. The most commonly accepted comes from the National Institute of Health (NIH)’s National Asthma Education and Prevention Program (NAEPP) (National Heart, Lung, and Blood Institute 1997). This scale, summarized in Table 7, was originally intended to categorize severity in untreated asthmatics in clinical settings but has been used in epidemiologic studies, as well (Diette et al. 2001). More recently, as the use of daily medication for asthma has become more common, the NAEPP scale has been felt to more appropriately describe disease control rather than severity.
Some of the Children’s Centers included questions from the CHSA (Asmussen et al. 1999) or from the pediatric version of the ATS questionnaire. The JHU intervention study used the ATS questions as they had been adapted for the Children’s Asthma Management Program study (Szefler et al. 2000). The difference between these questionnaires is that the CHSA uses a Likert scale to define symptom frequency in the preceding 2 weeks, whereas the other questionnaires ask how many days the child had experienced individual symptoms; the reliability of the CHSA is better documented than is the case with other questionnaires. Because of the diurnal variation in asthma disease activity, symptoms occurring at night and during the day are always considered separately. Symptoms occurring with exercise or on days that the child does not have an upper respiratory tract infection are also recorded separately.
Other aspects of asthma morbidity also recorded in the Children’s Centers’ investigations included interference with a child’s activity (exercise, play, school, sleep), interference with parents’ activities, and acute events. The latter were generally defined as those requiring systemic corticosteroid use, unscheduled physician visits, visits to emergency rooms, or hospitalization. Because these were uncommon events, the recall period was generally longer, that is, 2 months or longer, in almost all Children’s Centers.
In the early 1990s, an international group created five core questions regarding childhood asthma to facilitate an international comparison of varying prevalence rates or severity indices. This scale from the International Study of Asthma and Allergy in Childhood (ISAAC) (Asher et al. 1995) has widespread support among epidemiologists. The ISAAC core asthma questions were commonly included in the asthma-specific questionnaires used in the Children’s Centers’ studies.
Juniper et al. (1996, 2000) developed a series of questions to assess the effect of symptoms on a child’s quality of life. These provide a composite assessment of disease activity and tend to correlate better with daily lung function measures than do symptom questionnaires. In addition, they have the advantage of providing a single summary number that has proven to be sensitive to change across time, an extremely valuable property in longitudinal cohort studies and interventions.
Pulmonary function tests are important measures of disease activity but correlate modestly with reported symptoms or other measures of morbidity. Daily measures usually are limited to peak expiratory flow rate (PEFR), although forced expiratory volume in 1 sec (FEV1) may also be measured by portable spirometers. Children, parents, and staff must be trained in the use of these devices; most preschool children given training can perform accurate PEFR measures. With proper coaching, 83% of children 3–6 years of age can complete technically acceptable and reproducible maneuvers (Eigen et al. 2001); the Children’s Center’s experience with technically acceptable data ranged from 52 to 99%.
Medication use both modifies symptoms and provides an independent measure of disease morbidity. Medications that are taken to reverse symptoms of obstruction [“relievers” is the term used in the NIH consensus guidelines (National Heart, Lung, and Blood Institute 1997)] are recorded as equivalent to symptoms during a day. Medications that are taken daily [“controllers” in the NIH consensus guidelines (National Heart, Lung, and Blood Institute 1997)] may modify symptoms, as may medications taken before exercise or other stimuli to prevent attacks. In this setting, it is appropriate to talk about disease “control” rather than “morbidity.” Validated scales have been published to describe disease control in adults (Nathan et al. 2004).
Diaries can potentially give more accurate records of disease activity, but retrieval and consistency are problematic. Diaries and periodic questionnaires generally correlate well (Gold et al. 1989), and the studies conducted in the Children’s Centers used repeated questionnaires to avoid the logistical problems of diary retrieval and data verification.
Five of the Children’s Centers conducted interventions to test the efficacy of environmental control measures on improving indoor environmental exposures and asthma-related health. Strategies employed are summarized in Table 8. Although all Children’s Centers emphasized an environmental education program for families, they varied in the breadth and intensity of other components of the intervention programs. Some Children’s Centers focused on strategies targeting a few key triggers, whereas others chose a more comprehensive approach. In addition, some Children’s Centers chose to supplement home environmental strategies with education for families on asthma management, or education targeting physicians treating the asthmatic children. The relative benefits and challenges of these various strategies remain the subject of intense investigation.
Although childhood asthma is an issue of national and international significance, it is important to remember that it occurs within a local context. Cultural, social, and linguistic factors vary by location, as do systems of health care and community resources. Sources of environmental exposure, housing stock, and population behavior patterns relative to the exposure may vary tremendously. In addition, historical relationships between academia and local communities may range from strained to quite cooperative. Each of these factors influences the way asthma studies and interventions can be practically implemented in any given location. Some of the variation seen in study design between Children’s Centers is a direct result of variation in local priorities, circumstances, and resource constraints. What may work in one setting may not be feasible or relevant in another. Each of the Children’s Centers described here used a community-based participatory research approach (Israel et al. 2005) in which community partners contributed significantly to the research process. The Children’s Centers universally report that the involvement by community members enhanced their ability to accomplish their research goals. Multicenter studies, such as the proposed National Children’s Study, have an additional challenge of balancing the need for protocol uniformity across sites with the very real need to adjust to local contextual issues.
The experience gained in the studies conducted by the National Institute of Environmental Health Sciences/U.S. Environmental Protection Agency Children’s Centers for Children’s Environmental Health and Disease Prevention provide important lessons for the National Children’s Study. The lessons that specifically relate to asthma can be summarized as follows:
This article is part of the mini-monograph “Lessons Learned from the National Institute of Environmental Health Sciences/U.S. Environmental Protection Agency Centers for Children’s Environmental Health and Disease Prevention Research for the National Children’s Study.”
We acknowledge the support of the National Children’s Study in the development of the articles in this mini-monograph. We also gratefully acknowledge the outstanding, dedicated staff and co-investigators at each of the Children’s Centers, the community members who helped shape the studies, and the many families and children who participated.
This work was supported in part by the National Institute of Environmental Health Sciences (ES009606, ES09589, ES009605, ES009581, ES09607, ES011261, ES009600) and the U.S. Environmental Protection Agency (R826724, R826710, R826709, R826708, R826711, R829389, R827027).