Interagency Autism Coordinating Committee logo


Main content area.

Summary of RFI Responses: Diagnosis

Diagnostic Screening and Measurement:

  1. Look at the validity of the DMIC for diagnosis of developmental and learning disorders, as it strives to parse the symptoms of ASDs in a manner that provides finer descriptions of individuals in the broad spectrum, and through this lead to more specific, individualized understanding and therefore intervention for these persons.


  2. Re-evaluation tools need to be developed and standardized to measure progress of autism symptoms over time.


  3. Population attribute data gathering studies should store their data in RDF format so that inferencing and relationship visualization software can be utilized. Using RDF (graph) structured data will require pre-study database design planning in order to properly acquire, represent, and store this information.


  4. The job of identifying the key ASD environmental factor(s) should have major involvement by computer scientists with extensive expertise in probabilistic multivariate data mining. This involvement should begin at the earliest stages of research planning (pre-proposal) and continue throughout the study to ensure that the necessary data are collected, that the pedigree is maximized, and that the most advanced analytical techniques are used.


  5. Although the M-Chat and other screening instruments are available, there are not good data yet on the predictive validity of these instruments. No work has examined the diagnostic accuracy/screening efficiency of these measures. Furthermore, no one has adequately characterized the psychometric properties of these measures, which would be extremely useful for better understanding their utility and how to improve them.


  6. Effective early screening tools are needed! The MCHAT is unacceptably inaccurate and requires too much professional interpretation of positives for use in a busy pediatric practice. Effective diagnostic tools are needed that can be used in the presence of intellectual disability and/or comorbid psychiatric disorders (this may require biologic tests). Systems to screen, diagnose, and treat ALL children need to be developed. Low SES children are diagnosed later if at all. Spanish interpretation of tests and validation of diagnostic constructs cross culturally are needed.


  7. It is essential that a system of scales and evaluations be developed that have universal acceptance, that fully capture the extraordinary range of capacities and limitations characteristic of autistic people and that can be fitted into a regulatory framework. That is a formidable task but essential to the safety and well being of autistics who are still struggling to get the services they desperately need


  8. Should the autism diagnostic observation schedule (ADOS) be our only standard, and should all those who diagnose children with autism only use this one tool?


  9. It would be extremely helpful to look at clinician reliability when applying the DSM-IV criteria, particularly for PDD-NOS. Additionally, there should be some research examining the meaningful limits of a "spectrum" disorder at the mild end. Making sense of the increased identification of ASD requires sorting out whether clinicians are "overidentifying," in part by identifying children at the mild end of the spectrum. I have heard of clinic staff telling parents that PDD and ADHD are part of the same spectrum. At what point do ASD symptoms become so mild or subtle that it is no longer meaningful to consider the child to have an "autism spectrum disorder?"


  10. There is a need for instruments better able to differentiate variants within the autism spectrum (the ADI-R, for instance, tends to over-include people with Asperger syndrome as "autistic.")


  11. Existing instruments are designed to identify autism spectrum conditions but not especially to track them. There remains no widely accepted and validated instrument that can be used repetitively to evaluate response to behavioral therapies or other putative treatments.


  12. What is the best way to assess cognitive function in a person with ASD?


  13. It is essential that we determine the best means for determining IQ in children with ASD and Asperger. Which instruments are truly valid? IQ is a major factor in diagnosis. We must know how to measure it.


  14. Develop an instrument to include the quality of an autistic child's emotional well-being.


  15. What are the best ways to assess pain in people with communication and sensory disorders?


  16. Phenotypic measures ought to include not only diagnostic and psychometric information but also behavioral indices of perceptual function, attentional focus and shifting, executive function, and social cognition in the same individual subjects, rather than in separate studies with subject pools in which measures cannot be correlated. Behavioral tasks must have ecological validity and must whenever possible be presented in an event-driven format rather than a timed format, one that can accommodate variation in subjects' pace of responding. Points of convergence or "final common pathways" may occur at all levels of analysis, from genetic networks to abnormal learned behaviors and cognitive strategies.


  17. There needs to be a way to clinically screen/diagnose for different types of autism, such as the Neruodevelopmental Disorders of Relating and Communicating described in the DMIC.


  18. Research is needed that provides state of the art assessment tools to determine levels of functioning for individuals with ASD at different stages in their lives.


  19. Develop autism assessment technology for high functioning children, aged four or higher. Such technology would be instrumental for treatment evaluation, general autism research, and diagnosis. While a highly instrumented display environment would make an excellent test bed for labor-intensive research, it also offers the opportunity to collect massive amounts of data on thousands of responses of thousands of normally developing individuals and individuals with high functioning autism. By using automated data analysis techniques designed for such massive amounts of data, it will be possible to categorize and quantitatively evaluate the subject pool. What will likely emerge is not a single measure of the degree of autism but rather a set of clusters of cognitive impairment. Ultimately, treatment and education programs can target specific cognitive deficits.


  20. Biological Measures:

  21. Develop a pre-natal test, similar to Down Syndrome and other disorders, that can inform potential parents either before they get pregnant or in the early stages of pregnancy as to their “risk factors” of being genetic carriers of autism. (This request is for my neurotypical son and for all the other brothers/sisters who may want to start their own families at some point in the future but worry about their gene pools and the risk factors.)


  22. Biomarkers that predict the development of autism, the likely severity of the disorder, the prognosis, and the response to specific treatments would have a tremendous impact on individuals with autism and their families.


  23. Develop biomarker testing for infants who have a family history of immune disorders and/or allergies to flag those children who have a higher susceptibility.


  24. Development of empirically validated biological markers of autism; imaging techniques for CNS inflammation such that drawing cerebrospinal fluid or doing a brain biopsy would not be needed; empirically validated biological markers of brain inflammation; non-endoscopic means to define gastrointestinal issues in ASD -- calprotectin, leukocyte esterase, eosinophil X, lactoferrin, etc.


  25. Empirical validation of the HPHPA marker for clostridia with DNA PCR of Clostridia; the prevalence of oxidative stress in autism – determine what are the best and easiest markers of oxidative stress.


  26. Include physiologically meaningful exposure measures and intermediary metabolism measures (e.g. measures from body compartments and with laboratory measures sensitive to chronic or persistent as well as acute exposures) that could inform identification of biomarkers and development of biomarker profiles that will aid in screening.


  27. Epidemiology/Surveillance:

  28. A critical question to answer is: Is there really an ASD epidemic? Our approach to research will be heavily influenced by the answer to this. If there is no epidemic, increasing numbers may be due to better case finding, evolving definitions and improved resources making autism a “desirable” educational or medical diagnosis. If rates of what we call “autism” today have been relatively stable, we can stop spending so much money and angst on searching for environmental causes and “cures” and spend more resources on helping people with ASD live happier, healthier and more productive lives.


  29. A national call should go out to all parents of children with ASDs to go to the Autism Speaks Web Site and complete an extensive survey. The survey would have questions that would document prenatal and postnatal data. Everything from prior drug and alcohol use, maternal and paternal, race of parents, median income, amalgam fillings, cravings and diet consumed, cookware used during gestation, taking of prenatal vitamins, city of gestation, problems/illness during gestation, what vaccinations received and when, vaccinations mother has had prior to pregnancy, formula or breast milk, if formula what type and where was water source, bottle heating method, etc.? A survey like this is long over due and may go far to get to the root of this epidemic.


  30. Follow-up assessments and research by every state on all individuals diagnosed with an ASD should be done at least every three years.


  31. Since there is a huge lack of quality data for other countries, one very important research target should involve acquiring the data necessary to identify populations with differing ASD rates followed by successively more focused, carefully-crafted surveys. Collaborate with WHO/UN. Is the Amish study useful, and does it suggest research approaches?


  32. If the NIH were to create a user-editable on-line database with unique password protected entry per individual, and the ability to create new symptom fields, and if enough people participated, the NIH might be able to tease out patterns, suggest specific people to genetically test, and perhaps isolate specific genes. This technique would not even have to be limited to autism spectrum disorders; it could be expanded to orphan genetic disorders and unusual pathologies.


  33. Do a zip-code-level analysis of number of births of children with autism by local/same level analysis of parents’ childhood zip code (where mother and father lived for 10 years).


  34. Is the prevalence of the disease changing? Are there geographical or population factors relevant to the prevalence of the disease? There may be important further clues to the disease available from this type of study.


  35. Advantage has not generally been taken of the straightforward contrast offered by clinically unaffected sibs. These individuals share many of the risk factors for autism spectrum conditions, yet have been developmentally rescued from autism spectrum conditions - either because their accumulation of genetic and environmental risk factors is subthreshold or because they possess protective factors not held by their affected sibs. The effects of these protective and liability factors may very well be traceable on many levels, from genetic variation to biochemical profiles to physiology, anatomy, and behavior. Comparing sibs to individuals with autism and to unrelated individuals, using wide-ranging and integrative sets of assays, will help to identify these risk factors and protective factors.


  36. Conduct a prevalence study comparing rates of autism and ASDs among school age children age 8 to 18 to determine if the rate of autism is really rising and not just due to diagnostic changes.


  37. Conduct a prevalence study among adults, using current diagnostic criteria, to determine if the 1 in 150 rate is recent or 'has always been there and just not noticed'.


  38. Determine if the phenotype of ASD has changed over time, for example, the rate of mental retardation, the frequency of co-morbid conditions, or the proportion of classic autism, PDD-NOS, and Asperger's


  39. Phenotypes:

  40. I believe it is imperative to fund more research into the causes of severe autism specifically as opposed to higher functioning autism. (While the latter is also important, I feel that this is an entirely different group to study.) Specifically, I believe that the language development should be a major focus as this is a severe detriment to quality of life for those who suffer with severe autism.


  41. Accurately define the natural history of autism, redefine subtypes of autism through objective, quantitative methods such as cluster analysis without preconceived classification bias, and define risk groups of good/poor outcome. Antiquated, mathematically unsubstantiated labels such as Aspergers, High Functioning Autism, etc., should be thrown out.


  42. A better, more scientifically-driven and data-driven diagnostic classification of ASD is needed. This requires more specific phenotyping of “the autisms” and should lead to more accurate prognosis and treatment decisions.


  43. It would good to have an investigation into the intelligence of autistic individuals, noting the difference between the ability to express oneself intelligently and intelligence itself, which often have been confused in previous research.


  44. What are the boundaries of the ASD phenotype? What are the core features, present in all cases, including neurocognitive and neurobiological core features (possibly superior temporal sulcus abnormalities or other aspects of the mirror neuron system)? What are the associated features that occur frequently in autism cases but not in every case? What is the latent structure of each of these features/constructs? Understanding the latent structure will help us better understand whether that feature is unitary or multi-faceted (uni-factorial or multi-factorial) as well as whether that feature is likely to be strongly linked to single or multiple genetic causes (categorical latent structure versus continuous latent structure).


  45. Study the bases and criteria for the identification of subtypes. This may include a combination of genetic factors, laboratory physiological manifestations (e.g., elevated levels of hormones, antibodies, neurotransmitters or other matter), anatomical manifestations, and brain processing (e.g., function localization and pathways.) Considerable existing data from a large number of xstudies currently existing may prove useful. It is essential that researchers be given incentives to "pool" data as needed to allow sufficient sample sizes.


  46. Refinement of prognosis would better serve professionals, families and governmental agencies in planning activities across the lifespan. Study of longitudinal course of cohorts characterized by potential prognostic signs, including those with and without various treatment strategies and with varying degrees of language development. Data base for revision of prognosis, which currently appears overly pessimistic for those diagnosed by the modern, more liberal criteria.


  47. Gain a better understanding of the degree to which those with "early signs" progress to an actual ASD diagnosis, the reasons for progressing/not progressing, and whether progression varies by exposure history.


  48. Is ASD one spectrum or a variety of separate disorders?


  49. Track severity of symptoms and onset of co-morbid conditions with prior exposures.


  50. Subgroups of autism will enable parents to find programs that will truly address their child's individual needs without putting them at odds with administrators, programs, etc.


  51. Identify subgroups while considering concomitant medical problems.


  52. Identify and classify subtypes and which treatments work best based on the symptoms of the subtype.


  53. There is no construct for determining "mild/moderate” or “severe" in respect to autism spectrum disorders. Many families and providers alike use these words. However, since there is no standardization there cannot be consensus of what mild/moderate or severe actually describes.


  54. Seek to define various subtypes of autism based on symptoms and severity of symptoms (including sensory disturbances, evidence of seizure disorders, gastrointestinal issues, etc.), so that comparisons of possible causes and treatments can be as useful and significant as possible.


  55. Children with very little eye contact should have their vision examined. Side glancing may indicate an inability to see normally. It may be a physical problem rather than a social one. They may also be a subtype as well.


  56. Seek out individuals who have lost their autism diagnosis and characterize them at multiple levels – genetics, metabolism, brain, behavior, neurophysiology. Develop prospective studies that may capture such individuals, facilitated by clinical treatment tracking using a clinical database and/or repository for detailed case studies


  57. Diagnosis:

  58. What is the frequency of misdiagnoses that both affirm and dismiss ASD?


  59. Better understand how autism "presents." Not all children with autism are socially withdrawn. Some love to deep-pressure touch; some shun it. The variables need to be better understood so that treatments can be better designed for different subclasses. A "one size fits all" approach doesn't work.


  60. What is the relationship between ADHD, sensory integration disorder, bipolar disorder, Asperger's, PDD-NOS, and autism?


  61. Examine alternative diagnostic classification systems that examine underlying sensory /motor/cognitive/communicative /social/emotional/ regulatory/ interpersonal domains, using a “front-end” approach. This includes an examination of the “traditional” sensory systems -- auditory processing, visual-spatial processing, tactile process and difficulties, but also the proprioceptive, vestibular, and kinesthetic systems. This approach does not simply examine “what” the deficits and symptom profiles are but considers the “why” and “how” underlying deficits are manifested in the behaviors that are observable.


  62. The DSM-IV provides a static view of autism spectrum disorders as if they were individual static entities. Children who are growing and developing can attain skills which suggested they have moved, for instance, from autism to Asperger's.


  63. Look for subtle perception issues at the earliest possible age. The standard tests for vision and hearing, for example, would miss all sorts of problems, and they could go unnoticed for years if they are not causing the person immediate. Then, do research on the effect of early intervention of the application of sensory therapies, before they have a chance to lead to other problems.


  64. It is my experience that many children with Asperger's fit the diagnostic criteria for autism due to the difficulty in communication such as conversational skills and social use of communication, including pragmatic language deficits.


  65. The prevalence of adults without diagnosis of ASD/Asperger/PDDNOS but in the BAP should be determined.


  66. Studies should be done of families of ASD/Asperger/PDDNOS children (parents, grandparents, relatives) to look for misdiagnoses.


  67. We need better guidelines for determining whether we wish to consider Asperger's a separate diagnosis from ASD. We may need cross disciplinary data to help us make these discriminations.


  68. The role of the doctor attending an autistic child/teen/adult in the proper detection of physical illness should be investigated, even in non-verbal autistic people of all ages. How are physical symptoms read by parents and doctors in non-verbal patients? How does an autistic person participate in the diagnosis, especially when it is possible to communicate his/her symptomatology in different ways? How is this information considered in the planning of adequate and proper and necessary testing and treatment? How do families decide for certain treatments in autistic children/teens/adults? How are autistic teens/adults consulted in the decision in verbal or non-verbal ways?


  69. Should IQ testing be a determining factor in diagnosis and/or characterization of autism spectrum disorders?


  70. Please refine the "ASD" and "PDD-NOS" definitions by breaking them down into sub-types.


  71. Consider including a Type A, B, C for Autistic Disorder and Asperger and PDD NOS to reflect autism as a “spectrum.” This would assist with diagnosis when diagnostic features are “atypical.”


  72. Work at the diagnostic level to distinguish Asperger’s Syndrome from Nonverbal Learning Disorder (NLD). Questions to be addressed include whether these are two separate disorders or if NLD should be included on the autism spectrum.


  73. There needs to be a better more accurate and universal evaluation process to determine whether the diagnosis of Autism or ASD truly fits. It seems to me that there are a number of children diagnosed with ASD who are probably mentally handicapped but that ASD is a more acceptable diagnosis to present to a family.


  74. Can we develop algorithms for investigating changes in an autistic child's behavior (such as increased aggression or self-injury) that are thought to be due to physical discomfort?


  75. Early Identification:

  76. Research into diagnosis has been quite successful. This work needs to be developed to a maturity that private practice pediatricians are able and willing to perform this screening. As with treatment, standards need to be developed so that nonspecialty hospitals and private and school psychologists are equipped to address the referrals generated by private practice pediatricians and others.


  77. What can be done to shorten the lengthy wait time to see a developmental pediatrician (12-18 months in my city) to get an early diagnosis and begin treatment? Most mothers I know commented that they knew something was different about their child (usually based on lack of eye contact) by 6 months. We are losing valuable time, and the general pediatricians and even neurologists seem unable to make a diagnosis if the child is not severely autistic (e.g., "He smiles--he can't be autistic." "An autistic child wouldn't look at you like that.").


  78. Focus on the best way to educate pediatricians on the early signs and symptoms of autism. The pediatrician is often the parents’ first line of communication, and they are not recognizing the early symptoms. Psychologists cannot handle all of the work themselves. We must insert better training in to medical schools.


  79. More age-appropriate examples of how autism looks in a child under two or three are needed. This is critical because autism can and should be identified under the age of three, but the diagnostic criteria are not easily applied to children under age three.


  80. Efforts at educating pediatricians, teachers, and other front-line personnel on risk factors and warning signs of autism spectrum conditions need to continue and to expand.


  81. Are there approaches to the physical exam or ways to make a physician's office visit more autism friendly?


  82. Public health and community outreach should be a significant focus for all centers of autism research, and applications for funding should be evaluated with this criterion in mind.


  83. Cultural Differences:

  84. Research is needed to determine disparities in age of diagnosis across class and ethnic lines. The scope should be expanded to include other countries. The research should include testing of validity and reliability of screening and assessment tools in culturally different communities.


  85. Diagnosis is a huge issue for communities of color. Many minority groups are under-identified in prevalence estimates, which may be largely based on not receiving timely diagnoses. For example, Latinos have been reported to have low prevalence of autism through various surveillance studies. However because prevalence studies typically rely on counting children who are already diagnosed, these rates may be underestimates. Often researchers rely on mainstream clinics to locate children who have been diagnosed, rather than also gathering data from clinics where different minorities may go. Furthermore, it may be necessary to conduct screenings and not just rely on data about children already diagnosed in order to catch children who are usually missed in the diagnosis process. One approach may be to look to one of the NIH funded epidemiological studies on Latinos and other groups specifically and develop a screening process for autism among these large population based samples. American Indians represent another population that is grossly underrepresented in the diagnosis process and will need specific culturally based approaches.


  86. What do parents first notice that is different about their child, and does what they notice varies by culture? Parents’ role is very important in leading to a diagnosis because the system relies on parents to report problems to the pediatricians and service providers. If certain child behaviors are not seen as problematic to the parent, they may not discuss them with providers.


  87. Does symptom expression of autism vary by culture? For example in cultures that emphasize greater social interaction, do children with ASD adopt social behaviors in a different way than other cultures, or are the problems exacerbated at least from the perspective of the parents?


  88. Miscellaneous Comments:

  89. Earlier diagnosis is so critical. We were seeing developmental delays in my daughter at 9 months and signs of autism started to poke through around her first birthday. I was told to “wait and see”. I was told that they won’t diagnose for children that young. Those responses simply aren’t good enough. Any sign of delay should be a red flag. These children are the future of our country.


  90. There are doctors in my community who see a child once in an office (instead of the home) and give an autism diagnosis if the child is timid or stacks some blocks. I have seen significant damage done to families by this kind of unprofessional behavior. One or two doctors may not sound like a big deal but, if even a single doctor is doing this, tens of kids can get misdiagnosed in a single community in a single year! There are no checks and balances and I don’t know what to do about wild card doctors, but families are being damaged by doctors with hasty diagnostic practices.


  91. I continue to regret listening to our pediatrician who at age 2 said, "He's not talking very much...try talking more to him at home and requiring him to talk more and we'll see how he is doing when he turns 3 years old." Obviously, by then it was too late for us to gain the benefit of that 2nd year for developing speech with him.


HHS Home | Contacting IACC | Accessibility | Privacy Policy | FOIA | Disclaimer | USA.gov | IACC Webmaster

U.S. Department of Health & Human Services • 200 Independence Avenue, S.W. • Washington, D.C. 20201