These notes, representing a summary of recently published papers, begin by referring to the debate over the use of secretin, with the tentative conclusion that there may well be benefits for some or many children, but not all children, and the extent of the benefits will vary across cases.

There follows a description of what is meant by 'Pervasive Developmental Disorder - Not Otherwise Specified' and its possible relationship with autism, along with a reference to the view that such a concept is of no practical usefulness (and, perhaps, of some potential damage).

The need to match a spectrum of need with a spectrum of provision is outlined, with implications for some question about the real or current meaning of 'inclusion' for children with autism; and the final section concerns the desirability of ongoing practitioner research into interventions 'in the field^' and of shared information.

M.J.Connor February 2000

The treatment of autism by intravenous infusions of secretin (a pancreatic hormone) has received much publicity over the recent past, with many anecdotal accounts of its effectiveness in reducing or ameliorating the observable symptoms of autism. However, there is little in the way of 'hard' research data by which to evaluate the claims of the benefits of secretin.

One study that has been published is that of Sandler et al (1999) who introduced their research with a reference to the media attention directed towards secretin-related effects which has 'escalated^' into reports that secretin may provide a cure.

In this study, a comparison was made of the effects of a single intravenous infusion of secretin with the effects of a single infusion of a saline solution (placebo) among a sample of 56 children diagnosed with autism or pervasive developmental disorder.

A baseline measure of each child's communication skills and autistic symptoms was established, against which to compare follow-up assessments 1 day, 2 days, 1 week, 2 weeks, and 4 weeks after the infusion.

It was reported that, when compared with the placebo condition, secretin produced no significant improvement in the majority (16) of outcome measures. Some improvement was observed in certain outcome measures (5 subtest scores in the Autism Behaviour Checklist and the speech subtest score in the Clinical Global Impression Scale) among both the secretin and placebo sub-samples of children. The extent of this improvement was similar in the two groups.

Among subjects for whom a full set of results was available, it was found that 9 of 27 (33%) in the secretin group responded to treatment, compared to 7 of 25 (28%) in the placebo group (a difference which is not statistically significant). No side effects were observed.

Nevertheless, it was reported that a majority of parents from both subgroups remained interested in secretin treatment for their children. However, while acknowledging that longer term treatment or a higher dosage of secretin might produce different (more positive) results, the authors concluded that parents should take a cautious approach and await the outcomes of further studies which were taking place.

In the same journal, the editorial section (Volkmar F.) offered a warning against pursuing unproven treatments, while a commentary from the director of the National Institute of Child Health and Human Development in Maryland has held that, pending further research evidence, these current findings contra-indicate the use of secretin to treat autism.

However, Bernard Rimland of the Autism Research Institute responded rapidly to this study and its conclusions. Rimland (1999) argues that there has been too rapid a dismissal of the possible efficacy of secretin after this one preliminary study involving single doses of the hormone; and he also takes issue with Volkmar's estimate that autism affects around 1 in 2000 children, claiming that the real incidence is at least 1 in 500.

He emphasises the limitations of the study, noting particularly the short time scale involved, with the concern that significant changes are unlikely to be brought about in a brain-based disorder within days or weeks. The point, acknowledged by Sandier et al, in respect of the possibility that multiple doses may achieve positive results is underlined, along with a further question whether the outcome might have been different if natural rather than synthetic secretin had been used.

Finally, a major criticism concerns the lack of measuring instruments sensitive enough to register changes due to treatment.

Rimland goes on to describe how several research studies concerning the use of secretin are nearing completion, but refers also to the considerable number of cases in which very positive and 'very hard to explain away' responses have been described in children. This series of single case histories submitted largely by parents, admittedly anecdotal, include common references to normalisation of bowel functioning, establishment of regular sleeping patterns, and to enhanced responsiveness or receptivity to communication. Many of the reports were submitted by parents said to be 'outraged' by the conclusions of the Sandler et al study which were repudiated by their own experiences.

Further, Rimland refers to direct communication with the Autism Research Institute on the part of Sandier himself who 'is impressed with the potential of secretin and wants to do a multi-dose trial', for which the Institute is willing to offer support.

Reference is then made to one further contact from a physician whose 4 year old son received 6 infusions, 3 with secretin and 3 with saline solution, from a neurologist, and who described the immediate ease with which it was possible to identify which infusion involved secretin (because of his sleeping through the night, the relief from chronic diarrhea, and enhanced attention).

A study at the University of Maryland (Horvath et all 1999) is quoted by Rimland and as finding striking differences in many gastrointestinal functions, including reaction to secretin, among children diagnosed with autism compared to normally developing control children. Among the 36 children with autism, reflux esophagitis was found in 25 cases, chronic gastritis in 15, and chronic duodenitis in 24. More than half of the children also exhibited decreased activity of2 enzymes needed for the digestion of carbohydrates.

The authors describe the administration of secretin to autistic subjects and controls, and the significantly stronger reaction among the autistic children which was hypothesised to reflect a regulation of secretin receptors in the pancreas in contrast to the lack of previous and normal secretin stimulation.

Their conclusion refers to a demonstration of consistent physiological abnormalities associated with autism which are not known to occur in any other specific gastrointestinal disorder. 'The correlation of these findings with a clinical symptom, and the response to secretin, provide further support for a true physiological abnormality'.

Meanwhile, in a previous paper (Rimland 1998), reference is made to the origins of the interest in the use of secretin, and Rimland describes the case of Parker Beck whose parent had become convinced of the significance of digestive disorders or enzyme deficiencies through her observations of Parker's food preferences as well as his chronic diarrhea.

Following the infusion of secretin, there was said to be an immediate response not simply in pancreatic functioning but in a range of developmental areas, including language usage, eye contact, attention, toilet training, and responsiveness to requests.

After an interval of around 5 weeks, the positive effects appeared to wear off and secretin infusions were repeated 9 months and 12 months after the original treatment. The observed and claimed benefits were reinforced by before and after videotapes of Parker's behaviour along with before and after lab test data.

It is noted by Rimland that not all subjects respond positively to the secretin treatment, and an estimate, based upon existing reports of the use of these infusions, has it that up to 60% of autistic or autistic-like children will show some benefits which, in certain cases, may be considerable.

In addition to reports of behavioural and cognitive improvements, such as the establishment of eye contact, increased use of language, and rise in measured IQ, the reports collated by Rimland frequently refer also to significant and unexpected changes in physiological functioning. For example, reports include a shift in rubella or mumps antibody levels from very high to normal levels; significant improvement in blood flow; a drop in blood ammonia levels; and a normalisation of bowel function in numerous children previously showing chronic diarrhea or constipation.

The impact of secretin in the reduction of the autistic and physiological/gastrointestinal symptoms is hypothesised to be linked to secretin^'s role in stimulating the production and utilisation of the neurotransmitter substance serotonin, and in regulating the kind of gastrointestinal malfunctions shown to be common among children with autism according to research into food intolerance and immune system deficits.

While it would appear that this kind of treatment has potential benefits, there are still many uncertainties, particularly in determining which children (or adults) are the most suitable candidates for the use of secretin, the appropriate level and frequency of dosage, and the most effective means of administering secretin. As with so many reportedly effective treatments, caution is required lest this intervention is perceived as the all-encompassing answer when, in fact, benefits will not be observed in all subjects ; the level of benefit, where it is observed, will vary; and some cases will show no improvement.

The big issue in dealing with children diagnosed with autism is still that of determining criteria or indicators by which to match cases with particular treatment methods. The need for caution is underlined when one notes the comment of Rimland (1998) that the use of secretin appears to be the most promising intervention for the treatment of autism, following hard on his argument that early behavioural intervention was the most effective intervention, with dietary treatment ( vitamins and supplements ) close behind in its impact!

The concept of 'pervasive developmental disorder' is quite commonly discussed but it may not be easy to achieve a clear definition, with the result that one is left wondering whether such a description applied to a child is a kind of 'holding' diagnosis indicative of significant difficulties, but to be replaced in due course by a more specific diagnostic category.

According to the American Psychiatric Association (1994), PDDNOS is characterised principally by deficits in social reciprocity, and also by impairments in social communication and/or a restricted range of interests and behaviours.

Sicotte and Stemberger (1999) describe the commonly observed speculation that children with PDDNOS are more prevalent in the population than children with autism, but they acknowledge the limited research evidence. In particular, there is very little information about what unique features, if any, differentiate this condition from other disorders. Instead, PDDNOS seems to reflect a range of disabilities which happen to share the central feature of delayed or disordered social reciprocity, albeit to a milder degree than that associated with autism. A review of studies carried out by Sicotte and Stemberger indicates that PDDNOS can share symptoms with other pervasive disorders such as autism, Asperger syndrome, and Rett syndrome, with developmental disorders such as semantic-pragmatic disorder, and with conditions whose onset is in early childhood such as reactive attachment disorder or attention deficit disorder.

Nevertheless, it has been argued, including by Towbin (1997), that a diagnosis of PDDNOS can be helpful in clarifying a child's profile of strengths and weaknesses. For example, within the spectrum of pervasive developmental disorders, PDDNOS is a collection of symptoms relatively close to the lower end of the range of normal functioning and which can highlight the social disability even if behavioural, cognitive, and communicative features do not justify a diagnosis of autism. Such a diagnosis might also be useful when describing individuals who show a late onset of symptoms, or in those cases where impairment in social reciprocity is significant among an array of other symptoms which are not specifically related to autism, such as 'multiple developmental impairment'.

Sicotte and Stemberger (opp cit) focus upon the concept of a spectrum of pervasive disorders, and argue that it is necessary to clarify the position held by PDDNOS by examining whether children so-diagnosed exhibit a deficit in the critical area of social impairment that is strongly related to autism, such that it may be concluded whether or not autism and PDDNOS exist on the same spectrum.

Accordingly, the authors tested the hypothesis that children with PDDNOS would demonstrate a pattern of thinking similar to that of autistic children, but different from that of children diagnosed with other developmental disorders. For this study, children with language impairment were chosen as the comparison group because of the sharing of some social and communicative deficits leading to the possibility of confounding diagnostic categories.

The pattern of thinking chosen to explore autistic-PDDNOS similarities or differences was that of a deficit in 'Theory of Mind' - a lack of ability to attribute mental states to other people (or themselves), or to recognise that other people may have different beliefs or knowledge, thus limiting the probability of understanding and predicting actions, and underlying problems with eye contact or shared attention and with meaningful communication.

Two tasks were selected. The first involved 'false belief' whereby a clearly labelled sweet tin actually holds pencils, and children who are aware of this are asked what some uninformed child would say is in the tin. (The autistic child is less likely than controls to say that the other child would refer to sweets).

The second task involved 'brain function' where children are asked questions about the location and function of the heart and the brain ('Where is it? What does it do ?' ) with a view to determining whether or not there is any reference to the brain's function in thinking or feeling.

The results when comparing 14 pairs of subjects matched for mental age and gender, and diagnosed either as PDDNOS or language impaired, confirmed the hypothesis that the two diagnostic groupings could be differentiated in their theory of mind capacities. The PDDNOS children were significantly impaired in theory of mind in that only a minority could predict what other children might think, whereas nearly all the language impaired children could do so accurately.

In other words, PDDNOS is linked with the theory of mind deficits which are strongly related to autism; and the social and communicative difficulties observed in this group may be explicable in terms of the weakness in recognising the point of view or different knowledge/experience of other persons.

The brain function task did not significantly differentiate between the two diagnostic groups although the responses were in the predicted direction.... viz., that the PDDNOS children were less likely to attribute 'mental' as well as 'mechanical' functions to the brain.

(It was further noted by the authors that the small size of the sample may have contributed to the lack of significance in the findings here; and it was subsequently learned that the children had recently had a science lesson in which the function of organs in the body was taught, such that correct responses may have been attributable to rote memory rather than to a theory of mind!)

In any event, the authors held that the pattern of results supported the hypothesis (Ozonoff et al 1991) that theory of mind capacities exist on a continuum; and in this study, the false belief task is the more difficult and, thus, more likely to highlight any theory of mind weakness.

In summary, there is evidence that PDDNOS children can be contrasted with language impaired children in their development of a theory of mind; and PDDNOS, unlike other disorders involving cognitive/social disorders, might be understood as existing on the same continuum as autism. Whether or not PDDNOS can be distinguished from autism on the basis of a less severe theory of mind deficit is not clear from these results, although they are consistent with the common view ( see, for example, Baron-Cohen 1989) that theory of mind is a broad construct involving lower and higher order abilities.

However, and while interested in the above study and its conclusions, the present writer (MJC) would return to the starting point of this section and ask how meaningful is a diagnosis of PDDNOS and why one cannot simply maintain the construct of 'Autistic Spectrum Disorder' with the existing implications for a range of severity or number of symptoms among children, albeit sharing categories of symptoms.

(It is noted that a similar query has already been posed by Rimland (1993) who expressed in no uncertain terms his view that a child may be autistic, or have a condition like autism, or many characteristics of autism, but that he does not have 'Pervasive Developmental Disorder' because there is no such thing. It is a label concocted to disguise uncertainty over diagnosis.

Rimland goes on to express the concern lest there may be an actual handicap conferred by such a diagnosis in that individuals who have had the PDD label affixed to them may be excluded from programmes and services designated for individuals with autism. Rimland supports the concept of ASD, recognising the many children who may have autistic-like symptoms but who would not qualify for a 'full' diagnosis of autism. His recommendation is that 'the pervasive developmental disorder designation along with the cumbersome bureaucratic baggage .. .. such as PDDNOS.... should be relegated to the Archives of Failed Attempts, where it will have plenty of company' !)

The concept of 'inclusion' for children with autistic spectrum disorders has been much discussed, especially following the 1992 endorsement in America of the principle of full inclusion of students with disabilities on the part of the National Association of State Boards of Education.

The concerns which followed did not relate to the principle, but to the practice of inclusion if it did not take account of the range of needs of students and the requirement for a whole range of educational goals, curricular levels and types, and teaching and learning strategies.

Rimland (1993) makes a distinction between inclusion and over-inclusion, and between the desirability of inclusion and the practicability of such an approach. He asks whether it makes sense to include a child in a regular classroom if that child functions far below the normal child intellectually, socially, and academically. It is held that it is not reasonable for inclusion to be seen as the only avenue, when parents need a range of options, including access to specialist programmes and to the experience and expertise of the staff involved for children with very special needs.

Lieberman (1992) expressed concern that the issue of inclusion has become a matter of morality, i.e. that any provision that is seen as separate is evil. Lieberman ponders whether a higher immorality than separateness is located within lack of progress or achievement or skills, and splintered learning of meaningless academic trivia.

One might stress again that it is not the principle of inclusion which is questioned, rather the current pressure for placing children with special educational needs into settings where the appropriate provisions, curricula, demands and expectations, criteria for measuring progress, and support are not yet available or adequate. As Lieberman put it ' Full inclusion is not the right thing to do. It is one right thing to do sometimes'.

The survey by Eaves and Ho (1997), however, has demonstrated that children with autism are being placed according to their needs rather than according to some all-encompassing ideology.

Their study of a large sample of children and young people diagnosed with autism showed that 35% were educated in specialist settings, 38% in mainstream classes with SNA support, and 16% in mainstream classes without SNA support. Age, IQ, and severity of autistic symptoms were all variables in determining placement decisions; with the younger, more able, and higher functioning children more likely to be placed in mainstream classes.

It was found that for mildly autistic and high functioning students, full inclusion can be quite effective, but problems commonly concerned independent working, completing tasks, and maintaining attention, along with frequently reported high levels of hyperactivity.

Meanwhile, even if a primary goal of inclusion is to allow the children with autism to make friends among their non-disabled peers, the authors cited reports from teaching staff that this was not occurring at a very high level. They found that only 17% of the autistic children had average social relationships with peers, and questioned whether even these represented real friendships in the light of parental reports in which only a very small minority (5%) described their children as having 'peer friends'.

The overall conclusion supported the view that a range of educational opportunities must be available to meet the range of abilities and needs of children and young people with autism

Logically following from the above section, one might summarise the work of Jordan (1999) in examining the various issues in the evaluation of interventions with, and management of, children with autism.

Jordan begins her paper with a reference to a need for all practitioners to work towards valid accounts of their own input to share with others and by which to develop principles of good practice. It is hoped that one might demystify research and remove the notion that research is the sole province of certain professionals; all practitioners are involved in research when they try some new approach with a child and observe outcomes. What is required is a means of making observations systematic, and not simply anecdotal, in order to increase the reliability of the information gained.

Salient questions include:

Will a given intervention work for me in my particular situation?

Is the intervention successful?

Is one intervention better than another in my particular circumstances?

Is an intervention being implemented in the way the originators intended?

Which children will do best on intervention x as opposed to intervention y?

Jordan then goes on to quote Weiss (1998) who sets Out S major purposes behind evaluation.

Ongoing monitoring - Where the effectiveness of an approach has already been demonstrated elsewhere, one may wish to determine whether similar benefits can be achieved in the case of a particular child, and, if not, to change or abandon this approach. There may be no readily available standardised measures for progress gained, so that criterion-based assessments are necessary.

The example is given of examining whether a personal visual timetable will help a child with autism to complete a series of tasks independently. It is not necessary to prove the efficacy of the methodology per se since there is ample existing evidence for the usefulness of visual cues ; the question is whether benefit is achievable in this case and in that school setting.

Evaluation will be continual, and will be based upon an accurate baseline measure of the child^'s ability to stay on task and the number/type of tasks that can be completed without direct supervision. Meanwhile, there will need to be an analysis of the circumstances in the classroom, including any distractors.

Given all this, one may then be in a position to assess the degree of similarity between this case and other cases, thus to determine the probability of success of approach with these other cases. Obviously it will not be possible to determine whether some other approach would have been as successful (or even whether as good an outcome might have arisen without any specific intervention ); what matters is that one can determine the value of some pre-validated intervention in a particular setting.

Determining (absolute) effectiveness - The issue here is how to interpret the value of an intervention when one has only pre- and post- intervention measures, and no control group. Even in the case of apparently considerable gains, one does not know if another intervention would have been as effective or more effective.

One way of coping with this situation is to use the children as their own controls. The procedure is to carry out a baseline assessment, maintain the status quo for a period equal to that of the proposed intervention, assess again, implement the intervention, and assess once more.

This will provide the opportunity to compare gains across similar time periods of action and no-action, bearing in mind the importance of 'disguising' which is the intervention period in order to avoid Hawthorn effects, or effects linked to self esteem (stemming from participating in a 'special' programme ) rather than to the intervention itself Equally, the child's perceptions of the intervention may bias the results in a negative direction.

Alternatively, one might compare variable levels of exposure to an intervention, where, for example, children might be treated according to the principles of applied behavioural analysis (Lovaas) but with different subgroups receiving different numbers of hours of treatment per week.

The example is given of a residential school whose staff are anxious to demonstrate the efficacy of its provisions. By assessing the children initially on a number of measures, then re-assessing at different points in time ( after 1 year, 2 years, etc.), one may indicate the success of the provisions. If the school is successful, one would expect a steady progression in intervention gains across time.

The results of such evaluations cannot be taken as unequivocal demonstrations of success given the number of intervening variables. Such variables include the precise choice of measures to be used at baseline and follow up; practice effects ; the fact that the staff who complete the assessments are the same staff who are seeking to demonstrate effectiveness ; small sample sizes; age effects ; and the access of students to experiences other than those directly provided via the intervention in question.

Nevertheless, if it can be shown that children improve over time, and that children with the greatest experience of an intervention make the greatest gains, one can make a good case for the efficacy of the intervention. Further, one might gain additional insights, such as the greatest benefit being associated with the early application of an approach and a subsequent performance plateau.... in which case, one might wonder if the issue is more to do with the actual intervention, or with removal from a previously unstructured or unsettling situation.

Effectiveness comparison - Here, one investigates whether one intervention is better than another, or whether one intervention is better than another with a particular client or in a particular setting. The first question is a very difficult to answer because it will depend upon the definition of 'better' and upon the particular goals linked to the situation. For example, to a funding authority, a successful programme is one which can be delivered within a budget, while a parent might judge success in terms of the extent of normalisation of the child.

In respect of the second point, one would ideally use a proper experimental approach with randomised groups of subjects, and assessments completed by researchers who are independent of any of the interventions and also blind in respect of which subjects have had which intervention or no intervention at all.

Jordan refers also to the issue of ethics and points out that it is very unusual for there to be no evidence at all that one intervention is better than another, or that access to some intervention will bring about no more progress than a no-intervention situation. This being so, it will be hard to justify assigning children to what are recognised to be less favourable groups.

It is recognised that evidence does not emerge from a single study but from a series of studies which point in the same direction. Therefore, if one is to bring together a number of contributory studies, it is necessary to ensure standardisation of assessment measures and procedures. This in turn would involve co-operation between different research bodies or between different authorities.

Another approach is to compare the performances of a given group, firstly under one intervention programme and then under another. However, even here interpretation may not be easy because of the possibility, for example, that the second intervention appears more successful simply because the children are older or have more language, or have become somehow 'prepared¹ for intervention as a result of the initial experience.

An example given is that of comparing a TEACCH -based system against existing classroom organisation and methods in a specialist school, where the children in the different classes are assessed before and after a trial period and with regular probes in between. Criteria for success need to be agreed in advance and might include social interaction, behaviour, level and kind of communication, and academic progress; and the evaluation period needs also to be agreed in advance.

Even so, such a study cannot prove that one approach is the best, although it may provide evidence for the school to decide whether to continue or to abandon a particular approach. Success may depend not so much upon a method as upon the enthusiasm of the staff or upon its novelty. Failure may reflect limited staff training. In other words, results are not straightforward, but can still be useful as part of a constant and ongoing programme of evaluation.

Treatment integrity - An evaluation of this type examines how well the procedure is implemented compared to the approach as originally set out. For example, a programme may be subject to 'therapist drift' whereby those people working with the child in an intensive behavioral programme may shift towards a more natural form of interaction with that child so that structured elements of the programme are not pursued consistently or strictly in accordance with the guidelines.

One means of determining programme 'integrity' is to maintain unobtrusive video-recording of sessions.

The example is given of intervening between parents who are pressing for admission of their child to the Boston Higashi School, and LEA staff who are claiming that a local specialist school can provide equal benefits. A professional might compare observed progress in the local school against published evidence from the Boston school, and one possible outcome might relate to attributing differences to local culture rather than to strictness of adherence to a particular procedure. The conclusion might also include a recognition of the significance of a 24-hour curriculum, leading to a compromise whereby the local school might be funded to provide additional support beyond the normal school day.

Jordan makes the point that such an investigation would merit publication because it provides a protocol for dealing with similar questions arising elsewhere; and there is a further plea for co-operation in research among different Local Education Authorities in order to build up a bank of experience and of protocols for all established interventions.

Determining suitability for subgroups - Given the range of individual needs and their severity within children all diagnosed with autism, it would be unreasonable to expect children to benefit equally from a given intervention. What is needed is a means of determining how to match a child to an intervention.

For example, it is noted that, in the Young Autism Project (Lovaas 1987), there were marked discrepancies among the children in respect of the benefits observed as a result of the intensive behavioural treatment. Precursors to differential success were not identified, although subsequent analysis highlighted initial speech as a valid predictor of positive outcome, while pre-treatment IQ was held to be a relatively poor indicator

As a further example, Ventner et al (1992) found that language scores and IQ levels were the best predictors of eventual mainstream placement for children taught according to TEACCH principles.

The question is begged over how one might predict outcomes in the case of children whose ages and measured ability levels cover a relatively wide range.

Knowledge of precursors to success would be helpful not only in choosing between interventions for given children, but also for identifying key skills necessary for entry to any given approach.

The example is given of seeking how best to improve social interaction skills among adult autistic individuals by exploring responsiveness to music as a means of facilitating interaction, involving a comparison of interactive therapy with and without musical support. Reference was made to the difficulty of defining social responsiveness, and to the problems in controlling for practice effects making for difficulties in generalising results beyond a fairly narrow setting.

Again, the implication is for accumulating findings from a range of settings and for gaining evidence from shared findings for more general effects.

Jordan seeks to summarise much of the foregoing by reiterating that research is not to be limited to some select few, but may be regarded as an attitude of mind -... a seeking after results which can be made public and challenged, with the probability that what might appear to be the 'truth' at one point in time will be subsequently found not to fit all cases, and perhaps even shown eventually not to be the truth at all.

In addition to commenting again about the difficulty of gaining control groups, of avoiding bias, and of maintaining an ethical approach, Jordan refers to the desirability of gaining informed consent from all participants, even if these are children.... failing which, in the case of subjects with marked learning difficulties, the consent of parents/carers should be sought.

Research may be directed towards the general interests of individuals with autism, but any given research project needs to demonstrate the potential benefits for the individuals who are involved directly in that project.

Further, it may be that a great deal of academic research may appear a little remote from the day to day experiences of children or young people with autism, and the implication is for practitioner-based research to seek to redress this situation by focusing upon natural settings and gaining data by means of which to improve practice and to provide a rational basis for decision-making.

M.J.Connor February 2000

American Psychiatric Association 1994 Diagnostic and Statistical Manual of Mental Disorders (DSMA). Washington, APA

Baron-Cohen 5. 1989 The autistic child's theory of mind. Journal of Child Psychology and Psychiatry 30 285-997

Eaves L. and Ho H. 1997 School placement and academic achievement in children with autistic spectrum disorders. Journal of Developmental and Physical Disabilities 9(4) 277-291

Jordan R 1999 Evaluating practice. Autism 3(4) 411-434

Lieberman L. 1992 Letters to the Editor. Education Week ( December 16)

Lovaas 0 1987 Behavioural treatment and normal intellectual and educational functioning in autistic children. Journal of Consulting and Clinical Psychology 55 3-9

Ozonoff S., Pennington B., and Rogers 5. 1991 Executive function deficits in high functioning autistic individuals. Journal of Child Psychology and Psychiatry 32 1081-1105

Rimland B. 1993 Plain talk about PDD and the diagnosis of autism. Autism Research Review International 7(2) p.3

Rimland B. 1998 Editor's notebook. Autism Research Review International 12(3) p.3 and 12(4) p.3

Rimland B. 1999 Secretin: positive, negative reports. Autism Research Review International 13(4) p.7

Sandier A., Sutton K., DeWees 3., Girardi M., Sheppard V., and Bodfish 1. 1999 Lack of benefit of a single dose of synthetic human secretin in the treatment of autism and pervasive developmental disorder. New England Journal of Medicine 341(24)1801-1806.

Sicotte C. and Stemberger R. 1999 Do children with PDDNOS have a theory of mind? Journal of Autism and Developmental Disorders 29(3) 225-233

Towbin K. 1997 Pervasive developmental disorder. In Cohen D. and Volkmar F. (Eds) Handbook of Autism and Pervasive Developmental Disorders. New York: Wiley

Ventner A., Lord C., and Schopler E. 1992 A follow up study of high functioning autistic children. Journal of Child Psychology and Psychiatry 33 489-507

Volkmar F. 1999 Lessons from secretin. New England Journal of Medicine 341(24) 1842-1844

Weiss B. 1998 Routine monitoring of the effectiveness of child psychotherapy. Journal of Child Psychology and Psychiatry 39 943-950