The Diagnosis of Autism: An Expert Interview With Catherine Lord, PhD

July 11, 2005

Editor's Note:
The early signs of autism can be hard to detect, but it is critical to diagnose autism spectrum disorders as soon as possible since prompt treatment improves the prognosis. To find out about the diagnosis of this disorder, Elizabeth Saenger, PhD, Medscape Psychiatry & Mental Health, interviewed Catherine Lord, PhD, Director, University of Michigan Autism and Communication Disorders Center, Ann Arbor.

Medscape: What are the earliest signs of autism spectrum disorder (ASD) and how can pediatricians and primary care providers (PCPs) identify them?

Dr. Lord: The earliest signs of ASD are usually not actions that kids do that seem abnormal, but rather the absence of normal behavior. That would include a child who doesn't respond to his name or only responds to his name if someone goes right up to him and makes faces. That would also include a history of having a few words and losing them, or a child who really seems to have a marked social regression some time in the second year of life. For example, a child who could play patty-cake and peekaboo and wave goodbye and then loses those skills. Or a child who has a few words that are used meaningfully, like "daddy" or "bottle" or "duck," and then gradually loses those words and doesn't develop other words. That seems to be something that happens in about a quarter of kids with autism. It doesn't happen in all children with autism, but very rarely happens in other disorders.

Medscape: So it's a bit like recognizing the dog that doesn't bark in the night.

Dr. Lord: Right. Many parents of kids with autism are concerned that their child might be deaf because he or she doesn't respond to something such as a parent walking into the room. On the other hand, many parents express that concern, but also say, "You know, I know he isn't deaf, because he can hear me turn on the television from 3 rooms away." Or, "He can hear the sound of the microwave from upstairs, if I'm warming something up that he wants to eat." Those are some of the most common early observations.

Another early example is children who learn to point in order to request something, but don't point in order to show people things or direct someone's attention. That's a tricky early warning sign, because it doesn't mean much when a very young child does not point at all. It may mean that the child is a little bit language-delayed. But, if you have a child who is really good at pointing when they want something, but doesn't ever point out things that are simply of interest, then that's more indicative of autism.

Many parents become concerned because their child isn't talking, and that probably is the most common reason parents go to a doctor and ask for help. The difficulty with this example is that there are lots of reasons why a child may not say words, including mild language delays, which are quite common. First, the physician has to be careful not to say to the parents "don't worry." And, the doctor has to ask enough about what else is going on -- in other words, gather information about other behaviors and treat the language issue as the entry window into what might or might not be happening.

Medscape: Would the pediatrician be able to form a diagnosis at that point or would that require referral to a specialist?

Dr. Lord: I think it would be very hard for a pediatrician in an office setting to make a diagnosis of autism in part because of what we just talked about -- the idea that what you're looking for is the absence of ordinary skills. There are a lot of reasons why a child in a doctor's office may not want to look at people he or she doesn't know, may not respond to what people are saying, may not play very well. It is important, if there are concerns or suspicions, for the general pediatrician to refer the child on.

The first steps, really, are to make sure that the child can hear and to send him or her to a speech pathologist who is experienced working with very young kids. Then, follow up is critical. Have the family come back in a month or so and go through questions about social behavior once again. If there are still any concerns, at that point, the child can be referred to a developmental disorders clinic. It is important to warn parents that this may not be autism, but whatever it is, it is best to figure it out sooner rather than later.

Medscape: What kind of screening tests are there?

Dr. Lord: Well, there are a number of screening tests, none of which work very well. There are some very clever screening tests, such as the Checklist for Autism in Toddlers or CHAT and the Early Screening for Autistic Traits or ESAT, developed in Holland and designed to try to identify children with autism around age 15 to 18 months. The problem with these screening tests is the opposite of what we expected. Experts were concerned that the tests might overidentify far too many children as having autism when those kids didn't have autism. But, in fact, what's turned out is that they miss many children. So, even though the screening tests are quite attractive and make theoretical sense, clinicians need to be very careful using them, because you may feel like you've done something, but the chances are that, even with a very expert person conducting the tests, at least 50% of the cases may be missed.

Medscape: That seems like an enormous amount.

Dr. Lord: It is. And it's very interesting. The screening tests were created with a lot of care in mind, thinking about what could be done in 5 to 10 minutes that might help make the diagnosis of autism. But, the reality is that, in very young kids, autism is clearly not cut-or-dry. In most cases, you wouldn't say, "Boy, this child is really odd." Instead, the child may have a number of behaviors that look pretty normal, but then also not do other things that are very important. It takes time and persistence to be sure that this isn't just a child who's very shy or very uncomfortable, but that the appropriate level of social interaction and social reciprocity really are not there. That's something that's hard to do quickly.

Medscape: How early can ASD be identified?

Dr. Lord: There are children who, in the second year of life, very clearly have symptoms in all the areas of autism. In other words, there's a certain proportion of kids who can be identified between 12 and 18 months. But for many kids who turn out to have autism, it is not clear at those early ages. In our center, we have really focused on that first half of the second year and have seen maybe 30 or 40 kids in that age range. I think we've gotten better and better at being able to say which of the children we think will have autism by age 2 years. But, it's not straightforward.

Medscape: Is it important to catch them at 12 or 18 months or is it okay if they're not diagnosed until age 2 years?

Dr. Lord: The importance is that parents may want to do things differently with and for their child if they know that he or she is at risk for autism. We don't know how much difference knowing as early as the toddler years makes in terms of the progression of the condition. However, one of the current beliefs about autism is that many of the behaviors and difficulties that you might see in, say, a 10 year old with autism, are a product not just of their condition, but also of the social isolation that autism creates. And so, one of the premises is that if we can catch autism and related disorders early and bring the child into social interaction, not let him or her just look at corners of walls and play with string, but to really engage each child as much as possible, we may be able to decrease quite markedly the social deficits that children with autism have when they're older.

I think the goal is not to do anything particularly unusual with the children. Treatment of these very young children is not that different from education that you would do at a sort of toddler preschool. But, "therapy" per se must be persistent as well as organized around what that child is able to do. Because many children with autism do have other kinds of delays (for example, receptive and expressive language delays and fine or gross motor delays), a parent of a child with autism wants to find someone who can help identify where the child is in different developmental areas as well as some interesting, fun, and socially engaging activities that can be done with the child to minimize social deficit.

The other issue that often comes up is that some, but not all, children with autism can be very difficult babies. On the one hand, there is a notion that many children with autism are the best babies around because they don't complain; and that may be true for some. But, there are also children with autism who are very picky eaters, who can't be comforted easily, who just don't settle down, who may not sleep very well, and so on. In that circumstance, we want to help parents as soon as possible so that they are able to get through the ordinary demands of living -- to try to figure out ways to help their child to sleep better and eat and also, for example, not be afraid of specific things. A number of kids with autism, as toddlers, develop fears of things (for example, doorways or men with beards), which are really fears that interfere with them doing other things. And so, if someone can help the parent early on, the child may not build up a repertoire of odd or difficult behaviors.

Medscape: It seems that it would also help parents adjust, because it would be difficult to be the parent of a child with autism. Providing support to that parent in some way, whether it's a support group with other parents or counseling of some kind, might be helpful.

Dr. Lord: You're absolutely right. One of the hardest things for many parents of young children with autism is to deal with what other people say to them if their child has a temper tantrum or if they are unable to get their child to understand that you can't take food out of other people's grocery baskets or those kinds of things. One of the services we can provide is reassurance to parents that this is not something they caused; this is not their fault and they really do have to behave differently with their child from how other people behave with their children. We can also try to give parents ways of thinking and responding when they're besieged with advice from well-meaning people who can be very, very hard on them.

Medscape: How did you become interested in autism?

Dr. Lord: When I was an undergraduate, I took a class from Ivar Lovass, known as the "father" of applied behavior analysis. It was just in the days when kids with autism were being taken out of institutions with the idea that behavioral technology could teach anybody anything -- in other words, these asocial children would be taught to be social human beings. It worked in some ways and, in other ways, it did not. But at the time it gave me a chance to get to know 2 children with autism who were very different from each other. And, from that, I became more interested in what causes this condition, where does autism come from, what is really going on?

Medscape: And then what did you go on to do in terms of your research?

Dr. Lord: Our research is focused on a number of different aspects of autism. One of our greatest interests is in how things change over time -- what are the trajectories in autism? And how, for example, does social development relate to language development and how does that relate to fine motor skills?

Another aspect I've been involved in is trying to develop standardized instruments for diagnosis. If different researchers use the same techniques to reach diagnoses, then we can put our data together and work with each other. That would allow for larger samples, giving a better way to look at the huge individual differences among children with autism.

Medscape: Can you tell me a little bit more about your work in developing standardized instruments?

Dr. Lord: Well, there are 3 instruments that a number of researchers in the United Kingdom and the United States, including Michael Ritter, and I developed. One of them is called the Autism Diagnostic Interview-Revised or ADIR. It's an interview given to parents or other caregivers covering what the child is like now in terms of social reciprocity, communication, and repetitive behaviors and also what the child was like during preschool years.

We've also developed something called the Autism Diagnostic Observation Schedule or ADOS, which is actually 4 different modules. The interviewer selects which module to use based on the age of the child and how well he or she can talk. Each module consists of about 10 tasks that are designed to evoke different kinds of social behaviors, such as sharing and requesting, playing with someone and following, and enjoying attention. The test can be performed in about half an hour during an office visit as part of the diagnosis. The idea is that to make a diagnosis, information from parents, because they know the child better than any clinician or researcher, should be combined with the observations of the clinician.

The third instrument is the screening questionnaire, which is designed to be sent out to parents before they come to the clinic. It asks very similar questions to the parent interview, but gives them a chance to think and reflect about these issues before they get there.

Medscape: What do you think the particular challenges are in diagnosing autism for physicians?

Dr. Lord: As we have been discussing, early detection is particularly hard because the signs at this stage are very subtle. And, when they're not subtle, like bad temper tantrums, they are often not specific to autism. Related to the difficulty of reading early signs is the question of how to refer families on without scaring them. Ideally, we should develop a common strategy where we first focus on the communication deficit and warn families that that might include many things such as a language impairment, a cognitive delay, or autism, and then move on from there to more specific diagnoses. I think that's quite a difficult process, and physicians who see small children don't have a lot of time to talk to families in a calm, thoughtful manner. If autism is a specific concern of the doctor or the family, he or she should try to set aside a time to talk about that specifically. Otherwise, the clinician may be constantly caught off guard with complicated questions.

We also don't know what the active ingredient is in early treatment; therefore, it's hard to give parents advice about what they should or should not do, other than maintain social engagement. There are millions of treatments that are being marketed, including vitamins, chemicals and medicines, diets, and various kinds of odd person-to-person treatments. Parents often access these through the Internet. And it's hard to help parents sort out what is really worth trying to do even though there isn't good scientific data for it, vs what is really quackery and a waste of time and money and, sometimes, even dangerous.

Medscape: What do you think the current challenges are in diagnosing autism from the point of view of researchers in the field?

Dr. Lord: For researchers, one question is: Where is the boundary between autism spectrum disorder and milder disorders with overlapping symptoms, like ADHD, nonverbal learning disability, and speech and language problems? I would say that we're fairly good about making the diagnosis of kids who are classically autistic; but, as you move away from that specific disorder, it gets harder. It is also harder to diagnose children who do not meet criteria for mental retardation. For kids who are of average, close to average, or above average intelligence, it is difficult to sort out what is eccentricity vs what is a real social deficit.

Medscape: Is there anything else you would like to add?

Dr. Lord: Overall, I would like to encourage physicians to not eliminate autism as a possibility when talking to parents of particular children -- especially when you haven't seen that child in a positive social relationship. In other words, be very careful not to tell parents that this isn't autism. Those are words that parents will hang on to for many years. Again, it's harder to rule out autism in very young children than it is to identify it. So, when in doubt, it is really important for the physician to get help and refer the child on to a specialist.

Medscape: Well, thank you for sharing your views on the diagnosis of ASD.

Dr. Lord: Thank you for having me.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.