Objective. To report on the development of the Questionnaire for Identifying Children with Chronic Conditions (QuICCC). This new instrument identifies children and adolescents who have chronic health conditions based on the noncategorical conceptual framework outlined in our earlier work. It uses the consequences of conditions as a method for identifying children with chronic health conditions and is completely independent of diagnosis.
Method. Through a combination of techniques, we developed and piloted items and created 39 brief question sequences that were designed to be administered to a parent or guardian of children <18 years of age. The prototype was field tested extensively and refined using data from local hospital-based samples representing 318 households and 666 children. The instrument was then administered to two large representative samples (local: 657 households, 1275 children; national: 712 households, 1388 children) to establish validity and reliability.
Results. Content, convergent, construct, and criterion validity each have been demonstrated. The QuICCC has good test-retest reliability. Parents find the questions easy to answer. It took 7 to 8 minutes on average to obtain information about all the children in a family. The QuICCC successfully identified children with a wide range of different conditions that are usually considered chronic, and excluded those with acute illnesses and those with conditions but no current consequences.
Conclusions. The Questionnaire for Identifying Children with Chronic Conditions is a practical instrument that can be used for epidemiological purposes. It offers considerable flexibility and has many potential applications in health care delivery research.
Condition lists are commonly used to identify children and adolescents with chronic health conditions and disabilities. The adequacy of this method has been the subject of much controversy and has led to questions about the accuracy of epidemiological data based on these lists.1-3 Estimates of the number of youngsters less than 18 years of age who are affected by chronic disorders range from 5% to more than 30% of United States children.4 The wide disparity in estimates may result both from inconsistencies in the definition of the population and from the use of condition checklists of varying length and content to determine whether children have any condition named on the list.
Evidence is also mounting that because disease-specific checklists do not capture the full range of conditions, determination of program eligibility based on diagnostic criteria is inequitable. This has generated legislative and judicial pressure for use of alternative approaches that are not based on diagnostic criteria.5-7One alternative, based on the notion of dimensions of illness, has been called the generic or noncategorical approach.6,8-11 The noncategorical or generic view, originally promulgated by Pless and Pinkerton,8 posits that the many consequences of childhood health problems are independent of the specific disease and that children with diverse medical problems have great similarities in life experiences and in the preventive and rehabilitative aspects of their lives.8,9,11 This approach assumes that although specific diagnosis is critical for biomedical intervention, other aspects or dimensions of chronic conditions are central for purposes other than biomedical diagnosis and treatment.
A noncategorical approach has been recommended for collection of epidemiological data and in the planning of program and policy.10,12 However, two problems must be addressed before a noncategorical method can be implemented. First, the field needs a clear definition of the target population that does not depend on diagnoses.1,2 Second, a noncategorical definition must be operationalized with sufficient precision so that the resulting tool is valid and reliable. We have previously proposed a noncategorical definition that relies on consequences of disorders.1
We report on the development of a practical instrument that can be used for epidemiologic purposes to identify children younger than 18 years old who meet the noncategorical definition that we outlined in our earlier work.1 It should be noted that throughout the discussion, we use the term “chronic health condition” to refer to any ongoing physical, behavioral, or cognitive disorder, including chronic illnesses, impairments, and disabilities. The term “children” is used to encompass children and adolescents from birth to 18 years. We summarize the key elements of the definition developed by our group; describe the new instrument, the Questionnaire for Identifying Children with Chronic Conditions (QuICCC); explain the methods used in its development; and present results of studies undertaken to test and refine it. Finally, we discuss its limitations and future potential.
KEY ELEMENTS OF THE DEFINITION AND ITS IMPLICATIONS
The framework we proposed for identifying children and adolescents who have chronic health conditions depends on the consequences of disorders and is independent of diagnosis.1 This framework uses three definitional concepts. All three elements must coexist for a child to be classified as having a chronic health condition: (1) disorder on a biological, psychological, or cognitive basis; (2) duration of at least 12 months; and (3) consequence(s) of the disorder. The definition further specifies three types of consequences: (a) functional limitations, (b) reliance on compensatory mechanisms or assistance, and (c) service use or need beyond that which is considered routine (Table 1). To operationalize the definition, each of the dimensions had to be translated into discrete and meaningful questions.
The definition represents a new conceptual approach to identifying children who have chronic health conditions because it reflects the health consequences experienced by children. According to this definition, the mere presence of a disorder (or diagnostic label) that has lasted or is expected to last for at least 12 months would not qualify a child as having a chronic health condition unless the child is currently experiencing consequences. However, it identifies those who have significant functional impairment or reliance on compensatory modalities even without a diagnostic label. Hence, the definition and the measure described below are designed to be independent of the label of the condition.
DESCRIPTION OF THE MEASURE
Because our primary purpose was to create an instrument to collect epidemiological data about children, the QuICCC was designed to be administered as a household survey to a parent or guardian using an interview format. We chose the parent or guardian, rather than a health care provider or child directly, to be compatible with traditional epidemiological survey practices and to be able to cover a wide spectrum of age groups and disabilities using a single consistent respondent. Moreover, we believe that parents are more knowledgeable about their children's day-to-day functioning and current health consequences than health care providers and there is support in the literature that parents can be accurate reporters for their children.13-15 The items were designed to be easily understood by a lay person and are not technical in nature. They are intended to be answered using knowledge based on living with a child. The QuICCC was designed to collect data about all the children in the household. After considerable discussion, we limited the instrument to children younger than 18 to be compatible with the National Health Interview Survey as recommended by the National Advisory Committee of the project. Table 2 shows the domains of the consequences included in the QuICCC and their operational representation.
The QuICCC consists of 39 question sequences. The first part of each question sequence asks about a specific consequence of having a chronic health condition: 15 in the functional limitation domain, 12 in the compensatory dependency domain, and 12 in the service use/need domain. If the respondent reports that a child in the household experiences the consequence, the interviewer moves to the second level of the question, which asks whether the consequence is the result of a medical, behavioral, or other health condition. If the second part of the question elicits an affirmative response, then the interviewer proceeds to the final part of the question, which inquires about the duration (or expected duration) criterion of 1 year or more. To meet the definition and be identified as having a chronic health condition a child must qualify in each component of at least one question sequence. A sample item is shown in Table 3.
The copyrighted instrument is available in English and Spanish and includes a detailed manual on its administration and scoring. We have also provided a version that can be used to obtain data about a single child. This version has not been separately validated, because their contents are identical except for the substitution of the phrase “Does [child's name] … ” for the phrase “Do any of the children … .”
Item Development and Refinement
Items were developed using existing literature and questionnaires, parent interviews, and expert review. We reviewed existing literature and national and regional health surveys to identify items that measure domains represented in the conceptual framework. Item sources included the Child Health Supplement and core of the National Health Interview Survey (NHIS) and the National Health and Nutrition Examination Survey III, both conducted by the National Center for Health Statistics; the 1989 National Medical Expenditure Survey conducted by the National Center for Health Services Research and Health Care Technology Assessment; and the Ontario Child Health Survey conducted by Statistics Canada. Some items were modified from these sources and original questions were also written for some domains when appropriate items could not be found.
Second, we interviewed parents of children who had chronic health conditions throughout the initial piloting and field testing to obtain information about the consequences their children were experiencing. To ascertain whether there were children who might meet the definition but were missed by the formal question sequence, we also asked respondents whether there was additional health information about their children that they thought would be important for a medical provider to know. This provided supplemental data for evaluating the efficacy of the instrument and feedback on item content and face validity throughout the development of the QuICCC.
Third, we used two types of expertise. We built on the extensive clinical and research experience of the authors in interviewing parents of children who have chronic health conditions about the nature of the consequences that their children experience. Additionally, a large interdisciplinary panel of experts, the National Advisory Committee of the National Child Health Assessment Planning Project, under a grant from the Maternal and Child Health Bureau (MCJ-367007), guided each step of the QuICCC's development. The National Advisory Committee was selected to represent a wide range of interests, perspectives, and technical expertise. Members included health researchers and methodologists, clinicians and health care delivery experts, educators, policy makers, potential users of the data, advocates, and parents and other family members of children who have chronic health conditions. National Advisory Committee members reviewed the instrument to assure that all relevant conceptual domains in the definition were represented. They also reviewed the preliminary data and provided further confirmation that the QuICCC reflected the conceptual model.
Instrument Development and Piloting
We evaluated and field tested versions of the QuICCC with three local hospital-based samples: (1) parents of children enrolled in a special program for children with chronic health conditions; (2) an outpatient population; and (3) an inpatient population. The hospital serves a wide range of children from healthy children being seen for routine check-ups to those with acute illnesses and those being seen for evaluation or continuing care for specific health problems. In each instance, we obtained information about all children in the household.
To address content validity and comprehensibility of items, we conducted interviews with parents of children known to have chronic health conditions after we administered early versions of the QuICCC. This technique was designed to detect variability in question interpretation and whenever possible to incorporate the respondents' own language in the wording of the questions. As a result of this process, we developed question sequences in which complex questions were partitioned into elements administered sequentially. This led to the probing sequences that were easy for respondents to answer and saved time.
In terms of the need for medical care, we were concerned that questions that were based on utilization of services would be significantly biased toward older children and those in areas with more available services. To address the potential under-recognition in young children and in under served populations, we included two strategies. First, to identify young children who have functional limitations due to developmental delays that may not be recognized by parents or evaluated, we tested several types of questions commonly used to track developmental progress. Ultimately, we devised three questions based on norms in the literature to assess major deficiencies in development—the inability: to walk by the age of 2; to comprehend speech by the age of 2; and to speak so that people outside of the family can understand the child by the age of 3. Second, we developed and tested items about unmet health needs—services that were perceived as needed but were not obtainable. The final QuICCC retains one item in the service domain that asks directly about the need for services that were not received.
We tested the comprehensiveness and comprehensibility of the instrument. We continued to use qualitative techniques to determine how respondents interpreted each question and whether there were other consequences of their children's health status that they thought were important but were omitted. During this phase, we strategically recruited families including those with well children and acute illnesses as well as those enrolled in a special program for children who rely on medical technology or intensive health care services. The process of item revision and exploration was repeated through multiple versions of the QuICCC.
Over a 1-year period, a working prototype was developed that could be field tested. The preliminary QuICCC was then administered to parents of children in a hospital outpatient service and a hospital inpatient service of a public teaching hospital. We approached English-speaking parents or guardians who were waiting to see their child's doctor at a regularly scheduled appointment or in a walk-in clinic, or when they visited their child in the hospital. In the clinics, a random selection procedure was used; on the in-patient unit, all potential respondents were approached. Day and time varied with the interviewer's availability.
Respondents were asked to spend about 10 minutes answering questions about their children's health. Informed consent was obtained in accordance with the procedures of the Institutional Review Board of the Albert Einstein College of Medicine. Collectively, the hospital-based samples included 318 respondents who reported on 666 children.
The performance of the QuICCC was evaluated in an ongoing way in terms of several goals: how accurately it identified children with chronic conditions while simultaneously excluding children who were healthy or had acute illnesses; whether there were any particular conditions, impairments, or types of service that were missed by the QuICCC; and whether it included children with conditions that did not have consequences and who, therefore, did not meet our definition. In addition, we assessed whether the QuICCC performed consistently in different populations and how long it took to administer. The data from these assessments are reported in “Results.”
Community and Population-based Samples
Subsequently, the refined QuICCC was evaluated in two larger surveys conducted with a local community-based sample and a national sample. Both sets of data were collected by a large public opinion research firm (Schulman, Ronca & Bucuvalas, Inc, New York, NY) using telephone surveys of households with children that were conducted at SRBI's centralized telephone headquarters. Phone numbers were selected using a list-assisted random digit dialing (RDD) sampling procedure; a modification of the common Mitofsky-Waksberg method.16 The RDD technique is the best known sampling strategy for telephone surveys of the general population. It is estimated that RDD provides sampling coverage of more than 95% of the noninstitutionalized population of the United States.16 After the adult household population was geographically stratified by Census region, a sample of assigned telephone banks was randomly selected from an enumeration of the Working Residential Hundred Blocks within the active telephone exchanges. Every telephone number within the Hundred Block selected has an equal probability of being picked, whether it is listed or unlisted. The RDD sample of telephone numbers was then dialed by SRBI interviewers to determine which were currently working residential household telephone numbers. Nonworking numbers and nonresidential numbers were immediately replaced by other RDD numbers selected in the same fashion as the initial number. Ineligible households were also immediately replaced. The systematic dialing of those numbers to obtain a residential contact promotes an unbiased sample of telephone households. For the national survey, SRBI conducted the interviews in sample sweeps, so that they continued to “sweep” across the nation to insure the appropriate number of households contacted in each geographic region. More detailed information about the sampling methodology is available on request.
For the local sample, eligible respondents were English- and Spanish-speaking adult caretakers of children from birth to 18 who resided in households with telephones. Data were confidential, and verbal consent was obtained to conduct the survey as approved by the Institutional Review Board. In this local sample, data were obtained on 1275 children in 657 households consisting predominantly of inner-city minority families. The second RDD was conducted on a national sample of 1388 children in 712 households, but this sample was restricted to English-speaking households. A few minor item revisions were made between the two RDD surveys.
In each RDD survey, more detailed data (eg, functional status of the child and the child's and mother's psychological adjustment) were collected on a subsample of 200 children in separate households who met the criteria of the QuICCC for having a chronic health condition. Comparison samples of an equal number of healthy children were drawn from families in which no child met the criteria of the QuICCC. Thus, in each survey there were data on the full range of QuICCC items on all children and additional detailed data on approximately 400 children, half of whom met the QuICCC's criteria.
The purpose of these two surveys was to test the performance of the QuICCC in two populations that differed in race, education, and income. The sample characteristics of the national and local inner-city samples are shown in Table 4. The RDD surveys illustrated both the feasibility of telephone administration of the QuICCC and the applicability of the QuICCC at a community and population-based level.
Content validity was further assessed in both RDD studies by adding a question that followed each positive response to a question sequence to ask the name of the condition responsible for the consequence reported. By knowing the type of condition that produced the consequence, we could assess whether each question was working properly. However, naming the condition is not a required part of the QuICCC's determination process for establishing the presence of a condition.
A pediatric clinician used the definition and the available information on the condition mentioned by the respondent along with the actual responses to other questions in the survey to determine if the child had been correctly categorized as meeting the definition or not. For the local RDD sample, if the data were insufficient or ambiguous and the clinician was unable to determine the appropriateness of the categorization, we used a follow-back phone call to obtain additional information from the respondent. Questionable cases for which additional information could not be obtained (N = 10) were omitted from analyses involving accuracy of the QuICCC. However, because condition naming is not required when using the QuICCC, these children were included in other substantive analyses of how the QuICCC performed.
There is no “gold standard” other than the definition itself against which to validate the QuICCC. Thus, at all stages of instrument development, we assessed how closely the QuICCC implemented the definition. As described above, we used qualitative methods with different groups of parents to confirm that the QuICCC had face validity. We then used data from the local and national RDD surveys to assess additional aspects of content validity and to assess the convergent, construct, and criterion validity of the QuICCC.
Content validation involves systematic and comprehensive investigation of items in a measure to determine whether the domains examined have been fully and accurately represented by the items.17,18 As reported earlier, content validity was established in several ways. Items were drawn from existing literature, the knowledge of active professionals in the field, and the experience of parents with children who have chronic health conditions to create the list of consequences. Multiple iterations of the QuICCC were pilot tested, and all pilot data were evaluated by a clinician, the members of the research team, and the National Advisory Committee. Items reflect a discrete set of concepts applicable to the broad range of childhood disorders. At this point, we are confident that the consequences of chronic conditions and disorders that we included in the QuICCC are sufficiently comprehensive to identify children for epidemiological purposes.
We used five criteria to further examine the content validity of the QuICCC. First, we assessed whether children who qualified had conditions that are traditionally considered to be chronic. Second, we evaluated whether the QuICCC identified children with a wide range of conditions that have very different consequences. Third, we determined whether children without a specific diagnosis could qualify. Fourth, we evaluated whether children who had acute illnesses but no evidence of chronic conditions were excluded. Finally, we determined whether children who had disease labels but no current consequences were excluded.
Children identified by the QuICCC had conditions that were considered to be chronic and represented a wide range of named disorders. Among children identified by the QUICCC in the two RDD surveys, we counted more 50 different conditions including asthma, gastrointestinal disorders, epilepsy, hydrocephalus, kidney disease, learning disabilities, developmental delay, mental retardation, and psychiatric conditions. Of children in the two RDD samples who met the QuICCC criteria, the parent did not know or was unable to provide the name of the condition beyond the description of its manifestations in 7% of the cases. This confirms the usefulness of the QuICCC for identification of children who have chronic conditions whose parents can only describe the consequences. The details of some of their answers to exploratory questions and the nature of their responses were compelling. For example, the mother of one young child reported that she had never been told the name of the underlying condition, but that her child had a growth deficiency, got regular medical care, was developmentally delayed, and received nursing care at home or school.
To address the fourth and fifth criteria, we included the checklist of childhood health conditions in the 1988 NHIS Child Health Supplement (Section P5) in the local RDD survey. The QuICCC excluded children who had single or recurrent episodes of acute conditions or past conditions that had resolved. Examples of the former included sore throats, anemia, boils, chicken pox, ear infections, rashes, various respiratory infections and viruses, and urinary tract infections. Examples of the latter included repaired congenital cardiac condition or other surgical condition, as well as chronic conditions that carried no current consequences, eg, past history of asthma or epilepsy without any consequences in the past year while off medications.
In the absence of another instrument against which to measure convergent validity (ie, the extent of the empirical relationship between the QuICCC and another conceptually related measure), we used the data from the NHIS checklist of childhood health conditions. This diagnostic list approach has been used previously to identify a population of children who have chronic health conditions.4There is a certain irony in having to use diagnoses as the criteria against which to validate the instrument, because the deficiencies of this approach are well documented and have led to the need for alternative approaches. Nevertheless, despite the inherent inaccuracies and deficiencies, this remains one of the few commonly accepted standards.
We hypothesized that some conditions on the NHIS childhood health condition checklist would be almost invariably associated with consequences, and children who had these conditions usually should be identified by the QuICCC. Other conditions have only a probability of producing any consequence, and therefore a smaller number of children with these conditions would be identified by the QuICCC. Thus, we expected that there would be a large degree of overlap between the condition list and the QuICCC, but because the new definition is sufficiently distinct, we would expect some discrepancies in who would be identified by each of these two methods.
We found (Table 5) that 74% of children were classified the same way by both methods (groups 1 and 4) and 26% were classified differently (groups 2 and 3). Group 1 consisted of those identified by both the QuICCC and the diagnostic checklist as having a chronic condition; group 2 was identified only by the QuICCC; group 3 was identified only by the diagnostic checklist; and group 4 was not identified by either method and was conceptualized as representing children who did not have chronic conditions. On average, parents of children who were identified by both the QuICCC and the condition checklist endorsed more items on the QuICCC than those of children identified by the QuICCC alone.
Two findings are particularly noteworthy: one having to do with the range of conditions included in the checklist and the other with the number of children who have consequences. First, half of the conditions that the children identified by the QuICCC were not included in the checklist. Although these are not among the more prevalent conditions and in this sample account for only one-quarter of the children identified by the QuICCC, they represent an important omission and confirm the previous observation that many conditions are omitted from checklists. These children had conditions not typically included on diagnostic lists or were not labeled with diagnoses, but nonetheless experienced significant health-related consequences. Second, the QuICCC excluded children identified by the checklist who are not currently experiencing listed consequences of their conditions and, thus, from the point of view of our definition may be viewed as being incorrectly identified by the NHIS checklist (false positives). The QuICCC also excluded children who had conditions not generally considered to be chronic such as tonsillitis, urinary tract infections, pneumonia, and hay fever. These were the same types of conditions that a majority panel of pediatricians judged not to be chronic and were eliminated from the analyses of the Child Health Supplement to the 1988 NHIS.4
To assess construct validity, we tested the hypothesis that, as a group, children identified by the QuICCC would be more likely to have poor functional status than children who did not have a chronic health condition. We used the Functional Status-II(R) Measure (FS-II(R)),19,20 a measure designed to determine health status in children independent of condition or diagnoses. Items in the FS-II(R) ask the respondent to report on the child's ability to participate in a variety of age-appropriate activities and whether the child exhibits illness-related disruptions in behavior. The measure covers physical, psychological, cognitive, and social areas of functioning and is one of the few behaviorally based measures of child health status in wide use. It has a score range of 0 to 100 (maximal function). Although we hypothesized that there would be a relationship between the FS-II(R) and the QuICCC, the FS-II(R) measures dysfunction regardless of whether it is related to acute or chronic disorders. Further, it does not distinguish between children with normal functioning due to use of compensatory treatments (ie, a child who has well controlled diabetes or asthma) and those without need for compensatory mechanisms.
Both measures were administered to a subset of children in the two RDD studies, so that the children identified by the QuICCC could be compared to a sample of children from households in which no child was identified as having a chronic condition. As expected, mean FS-II(R) scores were significantly different between the two groups, with children identified by the QuICCC exhibiting more functional problems (lower FS-II(R) scores) than comparison children in both the local sample (mean FS-II(R) ± score SD: 96.5 ± 7.8 vs 99.5 ± 2.6; range 53.6 to 100 vs 71.4 to 100; F = 25.1, df = 1390, P < .001) and in the national sample (mean FS-II(R) score 94.3 ± 10.5 vs 98.8 ± 4; range 53.6 to 100 vs 71.4 to 100; F = 32.6, df = 1397P < .001). These scores for both the healthy and the disabled children are somewhat higher than those obtained in an institution-based sample.19,20 Not surprisingly, many index children had a high mean FS-II(R) score, a finding that has also been observed in children who have chronic conditions defined by diagnostic lists approaches19,20 (Perrin J, personal communication). However, the SD for identified children was more than two and one-half times that for children without chronic health conditions in both RDD samples, reflecting the far greater range of illness-related disturbance in functioning experienced by children who have chronic health conditions.
To further examine the construct validity of the QuICCC, we examined the proportion of children with significant dysfunction who were classified as having chronic conditions by the QuICCC. More than three-quarters of children whose functioning was two standard deviations below the mean on the FS-II(R) were identified by the QuICCC (78% and 76% in the local and national samples, respectively). Moreover, when 3 SD below the mean was used as the cut-off for the identification of children who have significant dysfunction, as recommended in the scoring instructions of the FS-II(R),19,20 the QuICCC identified 87.5% of children meeting this criterion. This distribution is significantly different from the expected probability (P < .05).
To test criterion validity, we assessed the ability of the QuICCC to identify children in our national RDD sample whose parents reported they were receiving Supplemental Security Income. Supplemental Security Income is currently received by children who meet Federal guidelines for disability and who are income eligible. The QuICCC identified 100% of children who were reported to be receiving this benefit.
The QuICCC is a composite of items that measure a range of constructs related to the consequences of having a chronic health condition and is not intended to be a unidimensional instrument. In addition, each domain contains items that are not expected to correlate (for example blindness and deafness or medication use and special equipment). Therefore, we did not attempt to design a summary measure, scale, or index of the items in the QuICCC. Thus, it is inappropriate to look at internal consistency reliability or to perform some of the customary psychometric tests that depend on the presence of unidimensionality.
However, test-retest reliability of the QuICCC over a 2-wk interval was examined in a random sample of 201 mothers of children interviewed in the national RDD survey. Test-retest reliability, measured as Cohen's Kappa, was 0.73 (P < .001) using the criterion of meeting or not meeting the definition on repeated administration (88% agreement).
Feasibility and Implementation
To determine efficiency we timed the administration of the QuICCC. It took 7 to 8 minutes on average to assess all the children in the household. Furthermore, there were virtually no missing data (<1%), indicating that respondents were able to answer questions with relative ease.
Because we sought to develop an instrument that could be implemented for epidemiological purposes, we compared the proportions of children identified by the QuICCC as having a chronic health condition across a range of different populations. Because we collected data on all the children in the household in each sample, we would not expect very large discrepancies in proportion across the samples, unless the QuICCC performed differently in demographically distinct samples. The percentage of children identified was 21% in the combined hospital-based sample (excluding the deliberate chronic health conditions sample); approximately 19% were identified in both the local and the national RDD surveys.
The noncategorical approach has been accepted as a valid perspective from which to examine the implications of having a chronic health condition.8-1021-23 However, despite the conceptual shift taking place at the programmatic level, the noncategorical approach is considerably less familiar in epidemiology, health services research, and health policy arenas. The lack of a valid, reliable, and practical instrument using the noncategorical framework appears to be a major barrier to widespread application.
We sought to operationalize a noncategorical definition of children who have chronic health conditions that does not require reference to a diagnostic list. The data presented above support the validity and reliability of the QuICCC for epidemiological and research purposes. Moreover, the data demonstrate the ability of the new instrument to identify children with a wide range of disorders who are not identified by a childhood condition checklist such as that used in the 1988 Childhood Supplement of the NHIS. The QuICCC takes only a few minutes to administer and has very little missing data indicating its ease for respondents.
Cultural, social, and political values are embedded in any definition and measure. Thus, it is not surprising that there are many value-laden decisions underlying the QuICCC. However, when judgments are explicit, they are far more accessible for discussion and potential modification than when they are hidden. We believe that one of the strengths of the QuICCC is that the wording of the questions in the QuICCC makes the judgments clear.
Adoption of this new approach to the identification of children who have chronic conditions would provide the opportunity to compare rates using the noncategorical conceptual paradigm in populations over time and across sites. It could be used to evaluate changing health care policies on child health. The QuICCC is offered as a model of operationalizing a noncategorical definition. We believe that important data could be obtained by its consistent use in the form we have described and tested.
However, we also recognize that the explicit nature of the items in the QuICCC provides an opportunity for flexibility. Although the QuICCC is designed to reflect a single definition, it is likely that different users may need to be more or less inclusive in detecting children who have chronic health conditions. The QuICCC could be modified to change the level of inclusivity in two ways. First, individual items could be modified by adding probes or altering decision rules about what does or does not qualify as a consequence. For example, in the category of appliance or assistive device, it can be specified whether an orthodontic mouth guard or orthopedic support worn only during specific contact sports activities should be excluded. Similarly, in an item on special diet, it could be specified whether eligibility included diets for inborn errors of metabolism, to avoid allergic symptoms (mild or life threatening), or to treat obesity. Or if one wanted to include all children requiring corrective lenses, this could be done instead of our current choice, which is to include only to those for whom the corrective lenses do not achieve functional vision. A second way to change the level of inclusivity of the QuICCC would be to require that a child have more than one consequence or more than one type of consequence.
This flexibility remains an essential strength of the design of the QuICCC, allowing its adaptation for multiple purposes and assuring that whatever values are inherent in the choices are made publicly and explicitly. The resulting clarity is helpful to those who use modifications of the QuICCC and to those who interpret the data collected with it, provided there is consistency in implementation within any given application. Nevertheless, we should caution that if modifications are undertaken in the future, they would need to be validated and should be distinguished from the current form of the QuICCC so that differences are clear. Otherwise, there is substantial danger of confusing interpretations of data collected with different forms of the instrument.
As different applications of the QuICCC are considered, it is important to stress that at this stage the QuICCC is intended only as a method for the identification of children who meet the Stein et al. definition.1 It is not intended to measure the severity of their conditions relative to one another. In the future, it may be possible to develop an instrument based on the QuICCC that would be sufficiently accurate for use in the screening of individuals and the determination of service eligibility. Such an instrument would have to be sensitive to and specific for individual screening, able to quantify severity of consequences, and incorporate other criteria for eligibility besides the presence of a condition with consequences. Moving ahead with such an application is integrally intertwined with understanding the value judgments underlying criteria for service or program inclusion or the intent of the application.
In the interim, the QuICCC can be used to study the implications of different types of consequences of health impairments for children and their families. How do the life experiences of children with functional impairments differ from those with reliance on compensatory modalities? Do children with multiple consequences or with consequences in multiple domains differ from those with fewer consequences or consequences in only one area? To begin to examine some of these questions, we are currently using several data sets to assess the relative effects of differing health consequences identified by the QuICCC.
The QuICCC is being applied in epidemiological research on a national level. A version of the QuICCC has been incorporated into the Children's Section of the 1994 and 1995 National Health Interview Survey on Disability, Phase I. The NHIS Disability Supplement will provide one of the best sources of general population-based epidemiological data, because it has both the largest sample size and the most detailed descriptive information about children who have disabilities.
The QuICCC also is being tested for its applicability at the state level. The Omnibus Budget Reconciliation Act 1989 amendments mandate that state Children with Special Health Care Needs programs funded under Title V of the Social Security Act broaden service programs beyond diagnostic categories so that eligibility of children is based on a noncategorical framework.24 As a result, state MCH Services Block Grants to Children with Special Health Care Needs Programs now require Needs Assessments of the populations served by these programs as part of their minimum data reporting requirements. The deficit of state-level information on children with special health care needs is even more extensive than that at the national level.21 A Needs Assessment Tool based on the QuICCC is currently under development. The QuICCC is consistent with the direction promoted by national Maternal and Child Health Bureau objectives for the states, because it encourages a shift away from the current provider-based data collection systems, which only identify children already in the service system, to a population-based data collection system that could identify children in the general population.
SUMMARY AND CONCLUSION
In summary, the QuICCC was designed to operationalize the specific definition proposed by Stein et al.1 The instrument identifies children from birth to age 18 who have physical, psychological, or cognitive conditions that have lasted or are expected to last for at least 12 months and that currently cause consequences. The QuICCC is completely independent of diagnosis: respondents need not know the name of the child's condition. It is not intended to be used to describe the level of functioning (severity) for any individual child nor to discriminate among children functioning at different levels.
Two versions are available, one to obtain household-level data, the other to obtain information about a single child. Parents or other adults living in the household who are knowledgeable about the overall health and day-to-day functioning of the children can be respondents and have little difficulty providing the information. The QuICCC is in an interview format, appropriate for in-person or telephone administration, and takes only 7 to 8 minutes. Preliminary analyses indicate that the QuICCC is a valid and reliable method by which to identify children who have chronic health conditions as defined by Stein et al1 across a range of populations.
This work was supported by grants from the Maternal and Child Health Bureau (MCJ-367007) and the National Institute of Mental Health, Preventive Intervention Research Center for Child Health (P50-MH-838280).
We are grateful to the members of the National Advisory Committee of the National Child Health Assessment Planning Project for their dedication to the successful completion of this project and for their helpful expertise and advice: Paul Newacheck, DrPH (Chairman); LuAnn Aday, PhD, Victor Baldwin, PhD, Peter Budetti, MD, JD, Richard Curtis, Antoinette P. Eaton, MD, Juanita W. Fleming, RN, PhD, FAAN, Dale C. Garell, MD, Gerald S. Golden, MD, Carolyn D. Gray, Esq, Bernard Guyer, MD, MPH, Robert J. Haggerty, MD, Cheryl D. Hayes, PhD, Henry T. Ireys, PhD, Paul S. Jellinek, PhD, Corinne Kirchner, PhD, Mary Grace Kovar, DrPH, David A. Mrazek, MD, MRCPsych, Barbara Starfield, MD, MPH, Eleanor S. Szanton, PhD, Peter C. Van Dyck, MD, MPH, Deborah Klein Walker, EdD, Nora Wells, and Karl R. White, PhD.
We also want to express our sincere thanks to all the members of the staff of the Maternal and Child Health Bureau who assisted in this effort and to the large number of representatives of other Federal agencies who participated in this project and to Ellen J. Silver, PhD, for her valuable input in the preparation of this manuscript.
- Received December 28, 1995.
- Accepted April 29, 1996.
Reprint requests to (R.E.K.S.) Preventive Intervention Research Center, Department of Pediatrics, 1300 Morris Park Avenue, Bronx, NY 10561.
- QuICCC =
- Questionnaire for Identifying Children with Chronic Conditions •
- FS-II(R) =
- Functional Status-II(R) •
- NHIS =
- National Health Interview Survey •
- RDD =
- random digit dialing
- Perrin EC,
- Newacheck P,
- Pless I,
- Drotar D,
- Gortmaker SL,
- Leventhal J,
- Perrin JM,
- Stein REK,
- Walker DK,
- Weitzman M
- Newacheck PW,
- Hughes DC,
- Stoddard JJ,
- Halfon N
- ↵Sullivan v Zebley, 88–1377 (US Supreme Court, 20, Feb 1990)
- Ireys HT,
- Nelson RP
- Perrin JM,
- Stein REK
- ↵Pless IB, Pinkerton P. Chronic Childhood Disorder: Promoting Patterns of Adjustment. Chicago, IL: Year Book Medical Publishers; 1975
- ↵Hobbs N, Perrin JM, Ireys HT. Chronically Ill Children and Their Families: Problems, Prospects, and Proposals From the Vanderbilt Study. San Francisco, CA: Jossey-Bass; 1985
- ↵Stein REK. To be or not to be … noncategorical. J Dev Behav Pediatr. 16;17:36–37
- ↵Aday LA. Designing and Conducting Health Surveys. San Francisco, CA: Jossey-Bass; 1989:195–216
- ↵National Center for Health Statistics. Data Evaluation and Methods Research. Vital and Health Statistics, Series 2, No. 8). Washington, DC: Government Printing Office; 1965
- Casady R,
- Lepkowski J
- ↵Anastasi A. Psychological Testing. 5th ed. New York, NY: Macmillan; 1982
- ↵Aday LA. Designing and Conducting Health Surveys. San Francisco, CA: Jossey-Bass; 1989
- ↵Stein REK, Jessop DJ. Manual for the Functional Status II(R) Measure. New York, NY: Albert Einstein College of Medicine, PACTS PAPERS/AECOM; 1991
- ↵Haggerty RJ, Roghmann KH, Pless IB. Child Health and the Community. New York, NY: John Wiley & Sons; 1975
- Pless IB,
- Douglas JWB
- ↵Silverman MM, Koretz DS. Preventing mental health problems. In: Stein REK, ed. Caring for Children with Chronic Illness: Issues and Strategies. New York, NY: Springer; 1989
- Copyright © 1997 American Academy of Pediatrics