Abstract

Understanding the nature and consequences of intellectual and developmental disabilities is challenging, especially when the condition is rare, affected individuals are geographically dispersed, and/or resource constraints limit large-scale studies involving direct assessment. Surveys provide an alternative methodology for gathering information but must be carefully designed and interpreted in light of obvious limitations. In this paper we discuss the potential of surveys in understanding a disabling condition; delineate characteristics of successful survey research; describe a survey of families of individuals with fragile X syndrome; and synthesize major findings. The survey has provided new information about the nature and consequences of fragile X syndrome in a cost-effective fashion, suggesting that survey methodology has a useful place in creating new knowledge about intellectual and developmental disabilities.

Understanding the nature and consequences of intellectual and developmental disabilities has posed a long-standing challenge, as typified by questions such as, What is the true incidence rate of autism? (Shattuck & Grosse, 2007). To what extent does the diagnosis of attention-deficit/hyperactivity disorder (ADHD) vary across specific disability types or as a function of variables such as race/ethnicity, gender, age, or socioeconomic status (Rowland, Lesesne, & Abramowitz, 2002)? Are different types of epileptic seizures systematically associated with specific syndromes (Cowen, 2002)? Are secondary conditions such as obesity or scoliosis a universal feature of developmental disorders such as spina bifida or do they occur in only a subset (Kirby, 2002; Simeonsson, McMillan, & Huntington, 2002)? What proportion of adults with Down syndrome are employed and what proportion live independently (Hodapp, 2007)?

Although survey research offers a methodology by which answers to questions such as these can be generated, for findings to be valid, surveys must be well-designed. In this paper we highlight the challenges inherent in understanding the nature and consequences of intellectual and developmental disabilities, discuss the benefits and limitations of survey research, and describe factors that need to be considered when designing and conducting surveys. Findings from a recent survey of families of children with fragile X syndrome are used to demonstrate the potential utility of survey research. we conclude this paper with some observations about the place of survey research within the broader array of research designs and instrumentation methodologies.

Challenges in Understanding a Disabling Condition

Researchers, clinicians, policymakers, and parents need answers to literally hundreds of important questions about the ultimate consequences of intellectual and developmental disabilities. Accurate data are needed to project service needs, define the potential burden of a condition for individuals and for society, and understand the extent to which features of a particular condition are universal for all affected individuals or only experienced by a subset. Valid answers to such questions, however, are elusive, for a variety of reasons. Perhaps the most obvious is the challenge of securing a large and representative sample of affected individuals. Most known causes of intellectual disability are relatively rare (e.g., Down syndrome, 1∶800 to 1,000; spina bifida, 1∶400, fragile X syndrome, 1∶4,000; Williams syndrome, 1∶7,500; Prader-Willi syndrome, 1∶10,000 to 15,000), and for some conditions the true incidence rate is unclear or changing (e.g., fetal alcohol syndrome may be as common as 1∶500 or as rare as 1∶5,000). Low prevalence, in combination with factors such as wide geographic dispersion, lack of comprehensive registries, changing criteria or incidence rates, and reluctance to report some conditions because of possible stigma (e.g., fetal alcohol syndrome, HIV/AIDS) make getting a large and representative sample difficult. Most researchers studying individuals with specific intellectual and developmental disabilities have used samples of 50 to 100 participants, often fewer. Low-income families and members of ethnic/racial minority groups typically are underrepresented, and whether the sample is representative of the full range of individuals with the condition that it purports to describe is virtually impossible to determine (Leonard & Wen, 2002). Changing perspectives on the true characterization of autism and the subsequent evolution of the “broad autism phenotype” exemplify these issues (Hurley, Losh, Parlier, Reznick, & Piven, 2007).

Instrumentation and study focus represent a second set of challenges. Most researchers address a particular aspect of development or function (e.g., neurocognition, memory, communication/language, reading, adaptive behavior, employment). Focused research provides important in-depth information but can preclude answering cross-cutting questions (e.g., in fragile X syndrome, is the rate of seizures for those who also have a diagnosis of autism different from those who do not have autism?). Further, when multiple investigators address the same construct, they often use different measures, making it difficult to compare findings. For example, Luo, Jefferson, Garcia, Ginsburg, and Topol (2009) described problems in defining the phenotype of individuals with coronary artery disease, due in large part to cross-study variability in measures used and differing definitions of affected versus unaffected status.

Basic science discoveries continue to expand the number of known genetic causes of intellectual disability. Inlow and Restifo (2004) reported that in 2004 there were nearly 300 known genetic causes of intellectual disability, with hundreds more likely to be discovered. Most of these will be rare, and characterizing the phenotypes associated with each will be difficult. Often the more severe cases seen in a clinic population are the focus of research projects, limiting understanding of the full spectrum of genotype–phenotype correlations. For example, some individuals with a genetic mutation may have very mild or even no symptoms, but in the absence of prospective population-based studies, these individuals may never be identified.

Bilder et al. (2009) argued that the cost of sequencing the human genome will be small relative to what is needed to describe the human “phenome.” To fully understand the ramifications of any one genetic mutation will require a transdisciplinary approach using multiple methods and measures with a large and representative sample. By representative, we include the traditional notions of sample adequacy (e.g., race/ethnicity, income, education, geography) as well as the notion of genetically representative (so that the full range of effects of an altered gene can be discerned). Of course, the latter will be virtually impossible without population-based genetic screening, but large and well-characterized samples can provide insights into possible variability that might not be detected with smaller samples.

Survey Research: Potential and Limitations

One approach to answering questions about the nature and consequences of a disabling condition would be through large, collaborative studies involving direct assessment of individuals over time, as exemplified by projects such as the National Down Syndrome Project, a multisite, population-based case-control study (Freeman et al., 2007). Although studies such as these are preferred in terms of thoroughness and accuracy of data, they are quite expensive and time-consuming.

Surveys offer a relatively fast and cost-effective alternative for collecting information from populations of interest compared with studies of comparable size involving direct assessments. Each year, hundreds of surveys are conducted by the government, private industry, and academia that are focused on the health and well-being of both the general population and individuals with specific conditions. These surveys offer a wealth of information on the characteristics of respondents, including their health status, current medical symptoms, and overall functioning. Although these data are somewhat limited by the fact that they are self-reported, they often provide the only available source of information on individual characteristics and functioning. This is particularly true of patients who have been understudied or emerging conditions where little is known about the variations in phenotype. Although not a replacement for large population studies, surveys do offer a way of collecting useful data in a standardized manner across a large sample.

In phenotypic research, surveys can offer useful information on symptoms, patient characteristics, and overall functioning or well-being. These data can be used to characterize both the overall population with a particular condition as well as subgroups who have clusters of similar characteristics in terms of symptoms or overall functioning. Of course, without direct observation and medical data (including genetic testing), relationships between these characteristics and various gene expressions cannot be definitively linked; however, data collected from surveys can still be valuable in identifying possible constellations of symptoms and patient characteristics that may represent different phenotypes. As suggested by Groth, Weiss, Pohlenz, and Leser (2008), it is often the case that clusters of symptoms that characterize a phenotype are identified before the genetic components of a disease or condition are fully understood. Examples from the recent literature of the use of surveys in phenotype research include studies on psoriasis (Ehrlich et al., 2006), pediatric respiratory symptoms (Spycher et al., 2008), substance abuse disorders (Bierut et al., 2002; Reich, 1996; Sullivan et al., 2008), Alpha-1 Antitrypsin Deficiency (Strange, Stoller, Sandhaus, Dickson, & Turino, 2006), fetal alcohol syndrome (Miller et al., 2006), and Aicardi syndrome (Glasmacher et al., 2007).

When considering use of a survey to characterize a population of interest, several key issues should be considered. First, it is critical to determine the purpose and specific goals of the survey (American Association for Public Opinion Research, 2009; Dillman, Smyth, & Christian, 2008). This will help focus the survey effort and narrow the scope of questions to those that are relevant. Also, this will help the research team determine what they can and cannot achieve with a given survey and can situate the survey in the larger context of ongoing research on the condition of interest. Anticipating the specific analyses to be conducted and the summary data desired will also greatly enhance survey development and ultimate utility.

Second, it is important to determine the audience of the survey (Tourangeau, Rips, & Rasinski, 2000). Will the survey go to known patients, their families, a general population sample, or some other group? The choice of the desired respondents should then guide the development of the survey, including the selection of scales or instruments. This will ensure that those selected for the survey will be able to respond appropriately to the questions they are asked.

Third, researchers will need to determine the mode of data collection. The most common modes of survey administration include mailed (paper and pencil), web-based, telephone, and in-person administration. All of these methods have different costs and benefits associated with them. For instance, although mailed surveys are the lowest in cost, they often result in relatively low response rates (Dillman et al., 2008). Web-based surveys often have faster response from participants and are of moderate cost, but they frequently have the lowest response rate in health or general population survey studies, and alternative methods would be needed for individuals who do not have access to the Internet (Dillman, 2006). Telephone and in-person surveys typically result in higher levels of participation but also require the highest level of effort and cost.

Beyond costs and participation, the choice of survey mode may also be driven by the information available on the target audience. This raises the fourth issue, namely, that a key consideration for any survey is that the research team is able to identify and contact the members of the target group. For small populations this may be done as a census (all members are included), whereas for larger groups it is more typical and cost effective to select a smaller representative sample (Dillman et al., 2008; Levy & Lemeshow, 1999). Another consideration when selecting a sample includes whether to use probability or nonprobability methods. In many cases, it may be difficult to determine the appropriate sample from which to select participants. This is especially the case with conditions where little is known or there are no systematic methods of identifying the population of interest. A good sampling statistician should be consulted to advise the research team on options and considerations for sample selection.

Fifth, researchers need to ensure that they use a high quality survey tool. The key to instrument development is that questions focus on collecting the necessary information with the specific population of interest. In some cases researchers may be able to use previously developed scales or survey items that have known psychometric properties (e.g., validity, reliability) or existing normative data, whereas in other cases new items must be developed. Regardless, once the survey instrument has been assembled, it is important to test the survey to determine whether there are any issues with instructions, format, or wording that could lead to error or variable interpretation of item intent on the part of respondents. Pilot testing can be achieved through field testing (i.e., administering the survey to a small number of members in the target group) or interview methods. The most common interview method is cognitive interviewing (or testing), which is used to evaluate the information presented, questions asked, and tasks required of participants (Beatty & Willis, 2007; Willis, 2004). The survey instrument should be reviewed by both subject-matter experts and survey methodologists to identify any technical errors in the way information is presented and questions are asked (Tourangeau et al., 2000). The ultimate goal is a survey that is easy to understand, provides adequate response options, and reduces any sources of respondent error (Biemer & Lyberg, 2003).

Sixth, when implementing the survey, it is important that researchers seek to maximize the level of participation within ethical limits. This is assured by careful monitoring of response rates. The use of reminders (e.g., letters, postcards, E-mail messages) and resending the survey (for mailed or web-based surveys) multiple times over the period of administration have been shown to enhance response rates (Dillman et al., 2008). Finally, when possible, researchers should consider the use of monetary (cash, gift card) and nonmonetary (some other desirable reward) incentives because they typically enhance the level of participation in surveys (Baumgartner & Rathbun, 1996; Church, 1993; Hawley, Cook, & Jensen-Doss, 2009).

The Case of Fragile X Syndrome

The evolving understanding of the nature of fragile X syndrome exemplifies the challenges inherent in fully understanding a genetic disorder. Fragile X syndrome results from the mutation of a single gene (FMR1) on the X chromosome caused by an abnormal expansion of nucleotide triplets in DNA. Most individuals have between 12 and 44 CGG (cytocine, guanine, guanine) trinucleotide repeats in the promoter region of their FMR1 sequence, leading to the normal production and regulation of FMRP, a protein known to be necessary for normal brain development. Typically, repeat length is stable across generations, but in some individuals it becomes unstable, expanding to larger lengths across subsequent generations. CGG repeat lengths of 45 to 55 are considered gray-zone alleles, with uncertain stability and no definitive known clinical effects. Individuals with CGG repeats between 55 and 200 are considered to be premutation carriers of FXS and, for the most part, have normal levels of FMRP, but are likely to pass on a gene with further expansion of repeats to offspring. Premutation carriers have an elevated risk of developing fragile X-associated tremor/ataxia syndrome or fragile X-related primary ovarian insufficiency (reviewed in Berry-Kravis et al., 2007), but this only occurs in a subset of the population.

Individuals with 200 or more CGG repeats are considered to have the full mutation fragile X syndrome. FMRP production is disrupted, leading to intellectual disability in most males and a large proportion of females. However, individuals with the full mutation can vary widely in the extent of intellectual impairment, which can range from mild to profound in males and from typical to severe in females (Loesch et al., 2002). Likewise, individuals with fragile X syndrome can have a number of secondary conditions, including anxiety, hyperactivity, self-injurious behavior (SIB), autism, and seizures (Bailey, Raspa, Olmsted, & Holiday, 2008).

Although fragile X syndrome is a single-gene disorder, it is now clear that FMR1 gene expansions can result in a wide range of effects on a continuum that includes both carriers and individuals with the full mutation. In fact, the National Fragile X Foundation (http://www.fragilex.org) now describes fragile X as a family of genetic conditions caused by changes in a single gene and uses the term fragile X associated disorders to encompass the range of possible effects. For the most part, however, the epidemiology of all of these features of fragile X syndrome and the co-associations among them has been based on small study samples, leading to fundamental questions about the true incidence and severity of both primary and secondary conditions. Practical data, such as independent living or functional skill attainment of adults with fragile X syndrome, have not been available due to the lack of a database or tracking mechanism.

The National Fragile X Survey

Method

The National Fragile X Survey was designed to provide answers to questions about the nature and consequences of fragile X syndrome, both for affected individuals and their families. We had two overarching goals: (a) locate and enroll a large, diverse group of families who have at least one child with fragile X syndrome; and (b) ask families questions about how fragile X affects their lives and their children.

Sample recruitment and enrollment

The first objective posed a major challenge because there is no national database of families affected by fragile X syndrome, and confidentiality guidelines precluded any individual or group from sharing names and contact information with the project team without parent permission. Therefore, to reach as many parents as possible, we partnered with three fragile X foundations (National Fragile X Foundation, FRAXA Research Foundation, Conquer Fragile X Foundation) and a number of researchers and clinicians. We worked with these research partners to compose a letter informing families of the study and inviting them to enroll. The letter was mailed in an envelope from the foundation or clinic to increase the likelihood that families would view it as coming from a trusted source. The letter, which was printed on specially designed stationary so that families would begin to connect the logo with the project, was signed by either the executive director of the foundation or the clinic director. Families were given the option of enrolling either online or by calling a toll-free number and speaking with a trained interviewer. We did not use a paper-and-pencil version because we wanted to avoid the costs associated with printing, distributing, and entering the data from a paper version.

One issue we needed to address in the early stages of enrollment was defining who was eligible to participate. In order to capture all types of families, we wanted those with biological, adopted, and foster children to participate. In addition, because fragile X syndrome is an inherited condition, we had to deal with the likelihood of multiple generations of families enrolling in the study. For example, the grandmother of a child with fragile X syndrome would enroll her daughter, who was a carrier, who in turn would enroll her son with fragile X syndrome. Because who participated and their responses was not shared with other survey participants, we did not ask whether extended family members completed a survey or attempt to link these files. In the end, we defined an appropriate survey respondent as any parent (limited to one spouse per nuclear family) of a child who was either a carrier of fragile X syndrome or had the full mutation, thus allowing multiple generations to respond independently.

The enrollment process took about 20 min per family. We used this process as an opportunity to gather demographic and contact information and asked survey respondents to provide birth date, gender, race/ethnicity, and genetic status (no fragile X, carrier, full mutation, not tested) of each family member. Parents rated each or their children (irrespective of genetic status) on characteristics such as thinking/reasoning/learning ability, mood, adaptability, and health. They also indicated whether each child had been diagnosed or treated for developmental delay or other co-occurring conditions, such as seizures or autism. This was done to reduce the length of time required to complete the full survey and to allow us to develop branching programs in the full survey based on enrollment responses.

A total of 1,250 families completed the enrollment portion of the study. Approximately 80% enrolled on line and 20% used the call center. The enrollment sample was diverse with respect to geography, age, and child gender. Families were generally well-distributed across four geographical regions: Northeast (23%), South (30%), Midwest (27%), and West (17%). A small percentage of respondents (3%) were from areas outside the United States. The enrolled families had a total of 2,964 children (1,845 males and 1,117 females). A total of 1,492 children had the full mutation fragile X syndrome, 276 had the premutation, and 565 had been tested and determined not to have fragile X syndrome. Another 600 children had not been tested for the syndrome. For the remaining 31 children, the respondents either did not know whether the child had been tested or did not report the results. The age at study enrollment of the children with the full mutation was varied, with 11% between the ages of birth and 4 years, 34% between 5 and 11 years, 24% between 12 and 18 years, 10% between 19 and 22 years, 10% between 23 and 30 years, and 11% who were 30 years or older.

Although this is the largest sample to date of families with a child who has fragile X syndrome, the sample was not truly representative. Many of the families who enrolled had higher incomes and more education than the typical American family. This may be due, in part, to our recruitment process. Because families had to be connected with a national organization, a clinic, or researcher, it is likely that we restricted the sample from the beginning. Many families from traditionally underrepresented groups (e.g., minority, Spanish-speaking, low-income) may not have access to researchers or clinics or may not be members of national organizations. In addition, financial limitations precluded translating the survey into languages other than English. When appropriate, we controlled for maternal education or income in publications resulting from the study, but the small number of families from ethnic minority groups has limited our ability to conduct informative analyses on the effects of race/ethnicity on factors such as age of diagnosis or access to services. Information on other children in the family who did not have fragile X syndrome provided comparison data that allowed us to match children with and without fragile X to test specific hypotheses.

Survey design

Our second goal in the survey was to gather as much information as possible about how fragile X syndrome affects individuals and their families. In the full survey, we decided that rather than focusing on one or two topics in depth, we would ask a few questions about a wide range of topics. Many of the items were developed in consultation with our research partners; and, in some cases, they came from previously developed scales. Both the topics and items were chosen because of their relevance for clinical and policy implications. We developed questions for families to determine how they discovered their child had fragile X syndrome, how much they knew about fragile X syndrome, how much and what types of support they had, in what areas they might have additional needs, their family history of fragile X-associated tremor/ataxia syndrome and fragile X-related primary ovarian insufficiency, the costs associated with raising a child with fragile X syndrome, and information about their insurance coverage. However, we also wanted to know about the children, both carriers and those with the full mutation. Families provided information on topics such as children's nutrition and physical activity, functional skills, sleep, seizures, self-injury, sensitivity to pain, medication use, education and schooling, services, and adult-specific topics such as employment and living arrangement.

The survey was designed to include branching programs that would easily move families only to relevant parts of the survey. For example, we would ask an introductory question about a given topic (e.g., Does your child currently experience any difficulty sleeping?). If the topic was applicable, respondents were asked several follow-up questions. If not, they were directed to the next topic. Some sections were automatically skipped, depending on age of the child or parent responses in the enrollment survey. For example, if parents had only young children, they were not asked questions about employment or independent living.

One issue related to the collection of child-level data was whether to ask about only one child per family or all affected children. We decided it would be too difficult to ask families to identify only one child (i.e., should it be the oldest child with fragile X syndrome or the most severely affected?), so we invited them to tell us about all their children with fragile X syndrome, both carriers and those with the full mutation. This meant that the survey had to loop families through several sets of questions for each affected child, resulting in a much longer survey for some families. Moreover, we needed to ask questions that were appropriate given the age of the child. For example, we did not want to ask families of a preschool-age child about elementary school activities.

We followed recommended practices with regard to survey methodology, limiting the use of jargon and keeping the language at an acceptable reading level. Response options were mutually exclusive. If appropriate, families were asked to “check all that apply” or to provide follow-up information if they selected the “other” response option. Prior to data collection, we conducted cognitive testing to make sure the items were written clearly, the response options were not too broad or too narrow, and the overall format of the survey was understandable.

The full survey was administered approximately 6 months after survey enrollment. A total of 1,075 families (86% of the enrollment sample) completed the full survey and another 51 partially completed the full survey.

Selected Findings From the National Fragile X Survey

The National Fragile X Survey has resulted in a number of papers addressing a variety of issues. The survey has led to new insights into the nature and consequences of fragile X syndrome, confirming some previous findings, elaborating on others, and, in some cases, questioning prior findings. Three papers in which researchers described findings related to seizures (Berry-Kravis Raspa, Loggin-Hester, Bishop, & Bailey, 2010), self-injury (Symons, Byiers, Raspa, Bishop, & Bailey, 2010), and obesity (Raspa et al., 2010) are included in this issue of AJIDD (November 2010). Brief summaries of other papers are provided here to exemplify the range of findings possible from such a survey.

Co-occurring conditions

In many clinical and research reports, investigators described a range of co-occurring conditions associated with fragile X syndrome, but these were typically based on small samples, and no researcher had attempted to link them together. We asked parents to indicate whether each of their children (including those who were carriers and those who did not carry an expanded FMR1 gene) had been diagnosed or treated for developmental delay or one of eight conditions that had previously been associated with fragile X syndrome: attention problems, hyperactivity, anxiety, aggression, SIB, autism, seizures, or depression (Bailey et al., 2008). We found that most individuals with the full mutation experienced multiple co-occurring conditions, an average of four for males and two for females. The most common associated conditions were the same for both genders, although, as expected, the relative frequency was less for females: attention problems (84% of males, 67% of females), anxiety (70% of males, 56% of females), and hyperactivity (66% of males, 30% of females), suggesting that these features are strongly associated with the fragile X syndrome phenotype. In males, aggression, SIB, and autism occurred in 38 to 46% of the sample, whereas seizures (18%) and depression (12%) only occurred in a small subset.

For the most part, the data on individual co-occurring conditions substantiated reports in the literature. For example, earlier researchers reported that seizures ranged between 16 and 19% of males with the full mutation; our finding of 18% suggests that this is probably the expected range because parents are likely to be good reporters of whether their child had experienced a seizure. The estimate of autism rates in males with the full mutation has varied widely in the literature, ranging from as low as 5% to as high as 60% (Loesch et al., 2007). Our finding of 46% suggests that males experience autism or autistic behavior at higher rates than reported in most studies, but this conclusion is severely limited in that we have no data to confirm whether these individuals met diagnostic criteria for autism. In contrast, our findings indicate that although self-injury is an important clinical problem in males (41%), it is likely not as high as the 58% previously reported in a much smaller sample (Symons et al., 2003). We had hoped to discern subtypes of fragile X syndrome in which certain co-occurring conditions clustered, but this was not possible due to the large number of unique patterns of co-occurrence.

A surprising finding from this study derived from an examination of premutation carriers. We matched carrier children with those who had tested as not having fragile X syndrome or carrier status and found that carrier males were significantly more likely to have been diagnosed or treated for attention problems (41%), developmental delay (33%), anxiety (33%), autism (19%), aggression (19%), and seizures (11%). Premutation females were significantly more likely to have been diagnosed or treated for anxiety (36%), depression (34%), attention problems (19%), and developmental delay (9%). Previous literature had suggested the possibility that carriers might have elevated risk for emotional or learning problems, but our data provide new insights into the likelihood of these conditions. Because we do not know whether we have a representative or accurately characterized group of carriers, our reported incidence figures should be viewed with caution, but the clearly elevated risk suggests that clinicians should pay attention to the possibility of behavioral and developmental challenges in this population.

Functional skills

Prior to the survey, virtually no data existed on the extent to which individuals with fragile X syndrome attained functional skills or on the patterns of skill attainment across the lifespan. To address this limitation, we asked parents to rate the extent to which each of their children with fragile X attained key functional skills in eating/feeding, dressing, toileting, communicating, reading, and self-care (Bailey, Raspa, Bishop, & Olmsted, 2009). Families rated each of 37 items (e.g., buttons clothing, cares for toileting needs without being reminded, uses complex sentences, recognizes letters) using four possible ability levels: does not perform this task, does this task but not well, does this task fairly well, or does this task very well.

We found that the majority of adults with fragile X syndrome had mastered independent performance of many functional skills. Most were verbal, used the toilet, dressed independently, and took care of basic self-help skills, such as bathing or using a towel. Females were much more likely to have attained independent use of most functional skills than were males, but results sometimes varied by specific items. For example, almost 100% of both adult males and females with fragile X syndrome could use a spoon or fork, but only 60% of males ate at a normal pace compared with 90% of females. In terms of other high profile skills, the majority of males (60%) could engage in a conversation with others, but 10% were nonverbal, reflecting the considerable range of ability often seen in the fragile X syndrome population. Only 20% of males could read a book containing new words or concepts, but more than 60% could recognize some words by sight. In contrast, nearly 80% of females could read books containing new words or concepts, and almost 100% recognized letters. Overall, the most challenging skills for males were eating at a normal pace, tying shoes independently, wiping independently, brushing hair and teeth, using complex sentences, engaging in conversations, speaking clearly and at a typical rate, and reading. These were also typically the most challenging skills for females, but with significant differences in the level of attainment of all skills.

Although the data were cross-sectional, we were able to examine age-related differences in skill attainment for both males and females by testing whether there were significant differences in the percentage of skills attained when comparing one age group with the next. For males, we found significant increases in each subsequent age group across the lifespan and across all skill areas, with one exception—reading skills did not significantly increase after the 6 to 10-year age group. In contrast, females attained most functional skills by age 10, with no significant increases after age 10 except for dressing skills, which continued to show increases until age 20.

These data provide the first comprehensive look at functional skills in individuals with fragile X syndrome across the lifespan. They offer baseline descriptive data, giving parents and clinicians a beginning perspective on the probabilities of attaining specific skills. They show that individuals with fragile X syndrome continue to attain new skills throughout the lifespan, but some skills are more difficult to attain than others. A substantial proportion of the population of both males and females were able to attain the highest level of functioning within each of the reported skill areas, showing that none of these skills is impossible to achieve. This finding should provide hope to both parents and clinicians and, we hope, will stimulate research and curriculum development activities that would lead to even higher levels of skill attainment for all individuals with fragile X syndrome.

Adult life circumstances

An issue of particular importance to families regards the future life prospects for their son or daughter with fragile X syndrome. The survey included items that addressed functional concerns, such as employment and independent living, as well as more subjective concerns, such as happiness, friendships, and quality of life. The sample included 328 families who had at least one adult child age 22 years or more with the full mutation of fragile X syndrome. These families answered a range of questions regarding educational attainment, residential settings, employment circumstances, level of assistance needed in daily life, friends, and leisure activities (Hartley et al., in press).

We found that 70% of adult males and 51% of adult females still resided with their parents. About 10% of males and 44% of females lived independently. The majority of males (60%) and females (73%) were employed, although about two thirds of the employed males worked parttime. Forty-three percent of males and 81% of females were rated as needing none or minimal assistance in everyday life. Only 19% of males and 9% of females had no friends. Most males (84%) and females (96%) engaged in three or more leisure activities. Most males had a certificate of high school completion (56%) or a high school diploma or GED (40%), and 4% had a vocational/trade school certificate or community college degree. In contrast, 12% of women had a high school completion certificate, 56% had a high school diploma, 15% had a vocational/trade school certificate or community college degree, and 17% had a Bachelor's or graduate degree.

This study provides the first systematic insights into the life circumstances of adults with fragile X syndrome. We found that many males still lived with their parents and needed moderate to considerable assistance in everyday life. This was less true for females, but half were still living with parents and nearly 20% needed moderate to considerable assistance in everyday life. A composite score of overall independence found that 57% of males and 23% of females experienced a low or very low level of independence. However, the majority of males and females were employed, had friends, and engaged in a variety of leisure activities. These data suggest that the challenges of parenting a child with fragile X syndrome continue to exist in the adult years, pointing to the ongoing need for adult services and supports. The fact that 10% of males and 44% of females had high or very high levels of independence, however, demonstrates the potential for successful life circumstances and points to the need for further research to identify the nature of family and community supports necessary to maximize independent living.

Sleep

Sleep problems are frequently reported in individuals with intellectual and developmental disabilities, more so than the general population, so it would not be unusual to see this in fragile X as well. However, basic research with animal models of fragile X syndrome suggests that flies who do not produce FMRP and mice that are double knockouts for FMR1 and FRAGILE XR2 (a similar gene with functional overlap with FMR1 in higher animals), demonstrate substantially impaired ability to maintain a normal sleep–wake cycle, prompting questions as to whether sleep disturbance is related to the biological disruption caused by the absence of FMRP. There is a limited amount of research in which investigators have examined sleep in humans with fragile X syndrome and, with the exception of a recent study involving 90 children (Kronk, Dahl, & Noll, 2009), these researchers have described sleep disturbance using very small samples of 9 to 14 children. Kronk et al. found that 47% of parents reported clinically significant sleep problems in their child, but in the other studies results imply showed elevated sleep problems without answering questions regarding the true prevalence, nature, and correlates of sleep problems.

As part of the National Fragile X Survey, 978 families who had one or more children with the full mutation fragile X syndrome answered sleep questions on a total of 1,295 children with the full mutation (1,013 males and 282 females) (Kronk et al., 2010). One question was, Is your son or daughter currently experiencing any sleep difficulties? If so, parents answered nine additional questions regarding age of onset, nature, and frequency of sleep problems as well as the nature and effectiveness of various treatments.

We found that about one third (32%) of the sample experienced current sleep difficulties, with approximately equal rates seen in males (33%) and females (29%). Most of their parents (71% of males, 64% of females) reported that sleep problems began early and were evident at age 3 years or younger. Most (84%) had two or more sleep problems, the most common were problems falling asleep and frequent nighttime awakenings. For individuals who had sleep difficulties, they occurred frequently, with more than a third occurring 6 or 7 nights per week. Parents had tried a variety of environmental (e.g., changing sleep environment or routine) or medication treatments. About 60% of parents felt that medical interventions had been at least somewhat effective, and about 40% felt that behavioral or environmental treatments had been at least somewhat effective, but obviously none had been completely successful because all parents reported current sleep problems. We did not, however, ask whether individuals not currently reporting sleep problems had required sleep treatments in the past. This study suggests that there is a lower rate of sleep disturbance than that reported in other studies, but because our data only addressed current sleep problems, parents of children with successfully treated sleep problems were not represented. It appears unlikely that significant sleep disturbance is a fundamental characteristic of fragile X syndrome but, rather, one that occurs in a subset. When sleep problems do occur, however, they are likely to emerge early in life and persist across the lifespan. Sleep problems almost certainly are disruptive to individual and family life and, as such, represent an important target for further treatment studies.

Age of diagnosis

In earlier papers our research group described family experiences in getting a diagnosis of fragile X (Bailey, Skinner, Hatton, & Roberts, 2000; Bailey, Skinner, & Sparkman, 2003). In both of these studies an average age of diagnosis around 3 years was reported, often preceded by an extended series of appointments with various clinicians until a correct diagnosis was finally made. As a result, children often missed the opportunity to participate in early intervention programs, and many families had a second child with fragile X syndrome before their first child was diagnosed.

Since these papers were published, much work has been done by foundations and professional organizations to promote earlier identification, including a recommendation from the American Academy of Pediatrics that all children be screened for developmental problems at 9, 18, and 30 months of age (American Academy of Pediatrics, 2006). We wanted to know whether there had been any changes in the age of diagnosis in recent years (Bailey, Raspa, Bishop, & Holiday, 2009). Of the more than 1,000 families who participated in the survey, 249 had their first child diagnosed between 2001 and 2007 and did not know about fragile X syndrome in their family before the diagnosis. We found essentially no change in the age of diagnosis over this 7-year period; the average male with fragile X syndrome was diagnosed between 35 and 37 months of age, and approximately 25% of families had a second child with fragile X syndrome before the first child was diagnosed.

In this paper we have demonstrated that widespread advocacy efforts to promote earlier identification have not resulted in measurable changes in the age of diagnosis of fragile X, although we did detect a trend toward earlier diagnosis of a delay in development. The delay in getting the fragile X diagnosis apparently happens because pediatricians do not request genetic testing but instead refer families to other specialists (e.g., a neurologist or a developmental–behavioral pediatrician), who then order the test. In this paper we have highlighted the challenges inherent in promoting earlier diagnoses of genetic disorders in the absence of systematic genetic screening programs.

Summary

In this article, we demonstrate how a survey of parents can be used to enhance the understanding of the nature and consequences of intellectual and developmental disabilities. Drawing on a sample of more than 1,000 families and using a relatively brief survey designed to answer specific questions, we now have a much better picture of how fragile X syndrome affects both individuals with the syndrome and their families.

A number of key limitations of the National Fragile X Survey and of a survey approach in general must be noted. First, this survey was exclusively based on parent report and, thus, characteristics such as autism, anxiety, or attention problems were not documented through professional assessment or diagnosis. It is likely that variability exists across items with respect to parents' accuracy of reporting. For example, autism status is likely to have been assessed in a number of different ways, and some children who did not meet strict diagnostic criteria may be in classrooms for children with autism.

Likewise, anxiety or attention problems are prevalent, but whether they meet or exceed clinical thresholds is unknown. On the other hand, other data are likely to be quite accurate (e.g., whether the child has ever had a seizure and at what age the child was diagnosed with fragile X syndrome), can only be reported by parents (e.g., living conditions, friendships, recreational activities), or are perceptual in nature and must be self-reported (e.g., quality of life, family stress).

Second, although large, the survey sample is not representative of the general population of families of individuals with fragile X, especially with respect to ethnicity, income, and maternal education. This likely reflects the fact that these factors are associated with who decides to affiliate with a national organization or who can attend a specialty clinic, cohorts from which we drew many of our respondents. Although we were able to factor variables such as income into many of our analyses, this problem will continue to be a challenge whenever surveys are used, both in terms of locating potential survey respondents and in the characteristics of individuals more likely to participate in and complete a survey.

Third, we asked parents about the genetic status of each of their children, but this was not confirmed through laboratory reports. Given the complexity of fragile X syndrome and the confusion that families sometimes experience in differentiating carrier from affected status, especially for females, it is possible that some children are mischaracterized by their families. Consequently, we also do not know whether the sample is representative from a genetic perspective, capturing the full range of genotype–phenotype expression. For example, some children in the family were never tested for fragile X syndrome status, and they could be carriers or even have the full mutation. The survey (like most other fragile X syndrome studies) likely underrepresents more mildly affected individuals.

Finally, although the survey data provide a wealth of new information about fragile X syndrome, our decision to address a large number of topics meant that we had to limit the number of questions we could ask about any one topic. For example, we asked whether medications had been used to treat problems such as sleep disturbance, attention, or self-injury and the perceived effectiveness of medications, but we did not ask which particular medicines were tried, as this would have greatly complicated the survey and the accuracy of parent report would likely be low. An alternative approach would have been to either have a much longer survey (a decision that would almost certainly have lowered the response rate) or to focus questions on a more limited set of topics, which would have limited the ultimate utility of the first such survey.

Despite the aforementioned limitations, the National Fragile X Survey has provided much useful data clarifying the nature and consequences of fragile X syndrome. In addition to the phenotypic data described above and in the other papers published in this issue, the survey also provided a means to document national health issues surrounding fragile X syndrome, such as lack of change in the age of diagnosis, which points to the need for new strategies to promote earlier identification (Bailey, Raspa, Bishop, & Holiday, 2009). We are planning papers on early intervention and school services, impact on families, and economic costs associated with having a child with fragile X syndrome; more than 94% of the families who completed the survey have agreed to allow us to keep their names for re-contact about participating in future survey research.

Surveys provide a unique opportunity to gather information about individuals with intellectual and developmental disabilities and will likely gain in importance as more rare causes of disability are discovered. By partnering with relevant organizations, following well-established survey research methodologies, taking advantage of new and emerging technologies, using input from diverse sources to identify specific questions that need to be answered, and acknowledging the limitations of data collected, survey researchers can use relatively modest financial resources to gain considerable information of relevance to policy and practice.

Acknowledgments

Preparation of this article was supported in part by the Centers for Disease Control and Prevention (CDC) and the Association for Prevention Teaching and Research (APTR) Cooperative Agreement No. U50/CCU300860, Project TS-1380. The findings and conclusions in this publication are those of the authors and do not necessarily represent the views of the CDC or APTR. The authors express their appreciation to the research collaborators and organizations who supported the recruitment of study participants; the families who completed the survey; RTI staff who assisted in survey programming (Venkat Yetukuri), data management (Anne Kenyon and Kirstin Miller), and call center services (Ellen Fay); and Elizabeth Berry-Kravis, who reviewed and provided input on the manuscript.

References

References
American Academy of Pediatrics
2006
.
Identifying infants and young children with developmental disorders in the medical home: An algorithm for developmental surveillance and screening.
Pediatrics
118
:
405
420
.
American Association for Public Opinion Research
2009
.
Best practices in survey research
.
Bailey
,
D. B.
,
M.
Raspa
,
E.
Bishop
, and
D.
Holiday
.
2009
.
No change in the age of diagnosis of fragile X syndrome: Findings from a national survey.
Pediatrics
124
:
527
533
.
Bailey
,
D. B.
,
M.
Raspa
,
E.
Bishop
, and
M.
Olmsted
.
2009
.
The functional skills of individuals with fragile X syndrome: A lifespan, cross-sectional analysis.
American Journal on Intellectual and Developmental Disabilities
114
:
289
303
.
Bailey
,
D. B.
,
M.
Raspa
,
M.
Olmsted
, and
D. B.
Holiday
.
2008
.
Co-occurring conditions associated with FMR1 gene variations: Findings from a national parent survey.
American Journal of Medical Genetics Part A
146A
:
2060
2069
.
Bailey
,
D. B.
,
D.
Skinner
,
D.
Hatton
, and
J.
Roberts
.
2000
.
Family experiences and factors associated with the diagnosis of fragile X syndrome.
Journal of Developmental and Behavioral Pediatrics
21
:
315
321
.
Bailey
,
D. B.
,
D.
Skinner
, and
K.
Sparkman
.
2003
.
Discovering fragile X syndrome: Family experiences and perceptions.
Pediatrics
111
:
407
416
.
Baumgartner
,
R. M.
and
P. R.
Rathbun
.
1996
.
Prepaid monetary incentives and mail survey response rates
Paper presented at the 1996 Joint Statistical Meetings, Chicago.
Beatty
,
P. C.
and
G. B.
Willis
.
2007
.
Research synthesis: The practice of cognitive interviewing.
Public Opinion Quarterly
71
:
287
311
.
Berry-Kravis
,
E.
,
L.
Abrams
,
S.
Coffey
,
D.
Hall
,
C.
Greco
,
L.
Gane
, et al
.
2007
.
Fragile X-associated tremor/ataxia syndrome: Clinical features, genetics and testing guidelines.
Movement Disorders
22
:
2018
2030
.
Berry-Kravis
,
E.
,
M.
Raspa
,
L.
Loggin-Hester
,
E.
Bishop
, and
D. B.
Bailey
.
2010
.
Epilepsy in fragile X syndrome: Characteristics and comorbid diagnoses.
American Journal on Intellectual and Developmental Disabilities
115
:
461
472
.
Biemer
,
P. P.
and
L. E.
Lyberg
.
2003
.
Introduction to survey quality
.
Hoboken, NJ
Wiley
.
Bierut
,
L. J.
,
N. L.
Saccone
,
J. P.
Rice
,
A.
Goate
,
T.
Foroud
,
H.
Edenberg
, et al
.
2002
.
Defining alcohol-related phenotypes in humans: The Collaborative Study on the Genetics of Alcoholism.
Alcohol Research and Health
26
:
208
213
.
Bilder
,
R. M.
,
F. W.
Sabb
,
T. D.
Cannon
,
E. D.
London
,
J. D.
Jentsch
,
D. S.
Parker
, et al
.
2009
.
Phenomics: The systematic study of phenotypes on a genome-wide scale.
Neuroscience
164
:
36
42
.
Church
,
A. H.
1993
.
Estimating the effect of incentives on mail survey response rates: A meta-analysis.
Public Opinion Quarterly
57
:
62
79
.
Cowen
,
L. D.
2002
.
The epidemiology of epilepsies in children.
Mental Retardation and Developmental Disabilities Research Reviews
8
:
171
182
.
Dillman
,
D. A.
2006
.
Why choice of survey mode makes a difference.
Public Health Reports
121
:
11
13
.
Dillman
,
D. A.
,
J. D.
Smyth
, and
L. M.
Christian
.
2008
.
Internet, mail, and mixed-mode surveys: The tailored design method (3rd ed.)
.
Hoboken, NJ
Wiley
.
Ehrlich
,
A.
,
T.
Koch
,
B.
Amin
,
D. J.
Liewehr
,
S. M.
Steinberg
,
M. L.
Turner
, and
A.
Blauvelt
.
2006
.
Development and reliability of a standardized questionnaire to assess psoriasis phenotype.
Journal of the American Academy of Dermatology
54
:
987
991
.
Freeman
,
S. B.
,
E. G.
Allen
,
C. L.
Oxford-Wright
,
S.
Tinker
,
C.
Druschel
,
C.
Hobbs
, et al
.
2007
.
The National Down Syndrome Project: Design and implementation.
Public Health Reports
122
:
62
72
.
Glasmacher
,
M. A.
,
V. R.
Sutton
,
B.
Hopkins
,
T.
Eble
,
R. A.
Lewis
,
P. D.
Park
, and
I. B.
Van den Veyver
.
2007
.
Phenotype and management of Aicardi syndrome: New findings from a survey of 69 children.
Journal of Child Neurology
22
:
176
184
.
Groth
,
P.
,
B.
Weiss
,
H.
Pohlenz
, and
U.
Leser
.
2008
.
Mining phenotypes for gene function prediction.
BMC Bioinformatics
9
:
1
15
.
Hartley
,
S. L.
,
M. M.
Seltzer
,
M.
Raspa
,
M.
Olmsted
,
E.
Bishop
, and
D. B.
Bailey
.
in press
.
Adult life for men and women with fragile X syndrome.
American Journal on Intellectual and Developmental Disabilities
.
Hawley
,
K. M.
,
J. R.
Cook
, and
A.
Jensen-Doss
.
2009
.
Do noncontingent incentives increase survey response rates among mental health providers? A randomized trial comparison.
Administration and Policy in Mental Health and Mental Health Services Research
36
:
343
348
.
Hodapp
,
R. M.
2007
.
Families of persons with Down syndrome: New perspectives, findings, and research and service needs.
Mental Retardation and Developmental Disabilities Research Reviews
13
:
279
287
.
Hurley
,
R. S. E.
,
M.
Losh
,
M.
Parlier
,
J. S.
Reznick
, and
J.
Piven
.
2007
.
The Broad Autism Phenotype Questionnaire.
Journal of Autism and Developmental Disorders
37
:
1679
1690
.
Inlow
,
J. K.
and
L. L.
Restifo
.
2004
.
Molecular and comparative genetics of mental retardation.
Genetics
166
:
835
881
.
Kirby
,
R. S.
2002
.
Co-occurrence of developmental disabilities with birth defects.
Mental Retardation and Developmental Disabilities Research Reviews
8
:
182
187
.
Kronk
,
R.
,
E. E.
Bishop
,
M.
Raspa
,
J. O.
Bickel
,
D. A.
Mandel
, and
D. B.
Bailey
.
2010
.
Prevalence, nature, and correlates of sleep problems in fragile X syndrome.
SLEEP
33
:
679
687
.
Kronk
,
R.
,
R.
Dahl
, and
R.
Noll
.
2009
.
Caregiver reports of sleep problems on a convenience sample of children with fragile X syndrome.
American Journal of Intellectual and Developmental Disabilities
114
:
383
392
.
Leonard
,
H.
and
X.
Wen
.
2002
.
The epidemiology of mental retardation: Challenges and opportunities in the new millennium.
Mental Retardation and Developmental Disabilities Research Reviews
8
:
117
134
.
Levy
,
P. S.
and
S.
Lemeshow
.
1999
.
Sampling populations: Methods and applications (3rd ed.)
.
New York
Wiley
.
Loesch
,
D. Z.
,
Q. M.
Bui
,
C.
Dissanayake
,
S.
Clifford
,
E.
Gould
,
D.
Bulhak-Paterson
, et al
.
2007
.
Molecular and cognitive predictors of the continuum of autistic behaviors in fragile X.
Neuroscience Biobehavioral Reviews
31
:
315
326
.
Loesch
,
D. Z.
,
R. M.
Huggins
,
Q. M.
Bui
,
J. L.
Epstein
,
A. K.
Taylor
, and
R.
Hagerman
.
2002
.
Effects of the deficits of fragile X mental retardation protein on cognitive status of fragile X males and females assessed by robust pedigree analysis.
Journal of Developmental and Behavioral Pediatrics
23
:
416
423
.
Luo
,
A. K.
,
B. K.
Jefferson
,
M. J.
Garcia
,
G. S.
Ginsburg
, and
E. J.
Topol
.
2009
.
Challenges in the phenotypic characterization of patients in genetic studies of coronary artery disease.
Journal of Medical Genetics
44
:
161
165
.
Miller
,
L. C.
,
W.
Chan
,
A.
Litvinova
,
A.
Rubin
,
K.
Comfort
,
L.
Tirella
, et al
.
2006
.
Fetal alcohol spectrum disorders in children residing in Russian orphanages: A phenotypic survey.
Alcohol: Clinical and Experimental Research
30
:
531
538
.
Raspa
,
M.
,
D. B.
Bailey
,
E.
Bishop
,
D. B.
Holiday
, and
M. G.
Olmsted
.
2010
.
Obesity, food selectivity, and physical activity in individuals with fragile X syndrome.
American Journal on Intellectual and Developmental Disabilities
115
:
482
495
.
Reich
,
T.
1996
.
A genomic survey of alcohol dependence and related phenotypes: Results from the collaborative study on the genetics of alcoholism.
Alcoholism, Clinical and Experimental Research
20
:
133A
137A
.
Rowland
,
A. S.
,
C. A.
Lesesne
, and
A. J.
Abramowitz
.
2002
.
The epidemiology of attention-deficit/hyperactivity disorder (ADHD): A public health view.
Mental Retardation and Developmental Disabilities Research Reviews
8
:
162
170
.
Shattuck
,
P. T.
and
S. D.
Grosse
.
2007
.
Issues related to the diagnosis and treatment of autism spectrum disorders.
Mental Retardation and Developmental Disabilities Research Reviews
13
:
129
135
.
Simeonsson
,
R. J.
,
J. S.
McMillen
, and
G. S.
Huntington
.
2002
.
Secondary conditions in children with disabilities: Spina bifida as a case example.
Mental Retardation and Developmental Disabilities Research Reviews
8
:
198
205
.
Spycher
,
B. D.
,
M.
Silverman
,
A. M.
Brooke
,
C. E.
Minder
, and
C. E.
Keuhni
.
2008
.
Distinguishing phenotypes of childhood wheeze and cough using latent class analyses.
European Respiratory Journal
3
:
974
981
.
Strange
,
C.
,
J. K.
Stoller
,
R. A.
Sandhaus
,
R.
Dickson
, and
G.
Turino
.
2006
.
Results of a survey of patients with Alpha-1 Antitrypsin deficiency.
Respiration
73
:
185
190
.
Sullivan
,
P. F.
,
P.
Kuo
,
B. T.
Webb
,
M. C.
Neale
,
J.
Vittum
,
H.
Furberg
, et al
.
2008
.
Genomewide linkage survey of nicotine dependence phenotypes.
Drug and Alcohol Dependence
93
:
210
216
.
Symons
,
F. J.
,
R. D.
Clark
,
D. D.
Hatton
,
M.
Skinner
, and
D. B.
Bailey
.
2003
.
Self-injurious behavior in young boys with fragile X syndrome.
American Journal of Medical Genetics
118A
:
115
121
.
Symons
,
F. J.
,
B. J.
Byiers
,
M.
Raspa
,
E.
Bishop
, and
D. B.
Bailey
.
2010
.
Self-injurious behavior and fragile X syndrome: Findings from the National Fragile X Survey.
American Journal on Intellectual and Developmental Disabilities
115
:
473
481
.
Tourangeau
,
R.
,
L. J.
Rips
, and
K.
Rasinski
.
2000
.
The psychology of survey response
.
New York
Cambridge University Press
.
Willis
,
G. B.
2004
.
Cognitive interviewing: A tool for improving questionnaire design
.
Thousand Oaks, CA
Sage
.

Author notes

Editor-in-charge: Leonard Abbeduto