Posted by Larry Hoover on May 29, 2005, at 9:37:41
In reply to Re: Walsh's papers » Jakeman, posted by Larry Hoover on May 27, 2005, at 21:24:21
> http://www.nationalautismassociation.org/library/Walsh%20Behavior%20paper6-04.pdf
>
> Physiol Behav. 2004 Oct 15;82(5):835-9.
>
> Reduced violent behavior following biochemical therapy.
>
> Walsh WJ, Glab LB, Haakenson ML.
>
> Pfeiffer Treatment Center, 4575 Weaver Parkway, Warrenville, IL 60555, United States. bill.walsh@hriptc.org
>
> Reduced violent behavior following biochemical therapy. We conducted an outcome study to measure the effectiveness of biochemical therapy for 207 consecutive patients presenting with a diagnosed behavior disorder. The treatment protocols were based on clinical evaluation and our past experience in the treatment of 8000 patients with behavior disorders at the Pfeiffer Treatment Center (PTC) over a 10-year period. Each test subject was screened for chemical imbalances previously found in high incidence in this population, including metal-metabolism disorders, methylation abnormalities, disordered pyrrole chemistry, heavy-metal overload, glucose dyscontrol, and malabsorption. The clinical procedure included a medical history, assay of 90 biochemical factors, and a physical examination. Standardized treatment protocols were applied for each imbalance that was identified. The frequencies of physical assaults and destructive episodes were determined using a standardized behavior scale before and after treatment, with follow-up ranging from 4 to 8 months. RESULTS: Seventy-six percent of the test subjects achieved compliance during the treatment period. The remaining 24% were reported to have discontinued the therapy. A reduced frequency of assaults was reported by 92% of the compliant assaultive patients, with 58% achieving elimination of the behavior. A total of 88% of compliant destructive patients exhibited a reduced frequency of destructive incidents and 53% achieved elimination of the behavior. Statistical significance was found for reduced frequency of assaults (t=7.74, p<0.001) and destructive incidents (t= 8.77, p<0.001). The results of this outcome study strongly suggest that individualized biochemical therapy may be efficacious in achieving behavioral improvements in this patient population.I am *so* not impressed.
Of 207 subjects, he diagnoses 495 different chemical imbalances. Some of them are most likely to be duplications (e.g. the pyrrole disorder would likely lead to zinc/copper disturbances), whereas others are clearly conceptually distinct. He should have also shown the distributions in a different way, to show the multiple diagnoses per subject.
The subjects' behaviours were assessed using verbal recall from patients' families. No observational information of any sort. No mention of confounds. (The intervention itself is a confound, being an open study.)
It is unclear if subjects were required to attend counselling, but it was assumed that subjects being counselled continued to do so. Those subjects receiving counselling should have been analyzed separately (a clear and obvious confound).
The data analysis section was woefully improper and inadequate.
The test Walsh applied was simply a time-course comparison, a paired t-test of earlier family observations against later ones. Without some sort of factor analysis, we cannot even begin to assess what interventions may have been at work on the putative changes in behaviour (third-party recall is not a good measure of behaviour).
To what, exactly, are we to attribute the changed behaviours? Walsh wants us to think it was his nutritional intervention, but where is the evidence for that?
The subjects were tested on metabolic parameters at intake. Where are the post-intervention results? Why aren't the data analyzed by intervention? Why aren't the final data analyzed by factorial analytical techniques (e.g. MANOVA)? How can Walsh explain the subjects who had no changes, or who became worse, despite compliance? Did subjects with multiple diagnoses fare differently than did those with simpler problems? With the general focus on nutrition (being involved with Walsh and Pfeiffer in any way), did that have an influence on dietary intake over the study period? Are we to assume the same pre-study insufficient diet continued throughout the study period, save for the nutritional interventions provided by Walsh?
The non-compliant subjects must have had behavioural observations made. Where are those data? Why weren't comparisons made between compliant and non-compliant groups?
In the discussion, Walsh invokes age as a variable, yet that was not once mentioned in the method or results. Along with age, Walsh then invokes issues of substance abuse, which he correlates to age. Failure to control for these variables in his statistical analysis invalidates his results. As do failing to assess the influence of counselling, and collapsing all his nutrient interventions into one group without showing that collapsing them is reasonable.
The nutrient interventions are non-equivalent. Each has a different rationale (valid or not). The effects may lie more with some interventions that with others. The effects may be completely independent of intervention.
He goes on to say, "The high incidence of biochemical imbalances in the behavior-disordered population and the major behavioral improvements following the correction of these imbalances..." without giving *any* evidence that any imbalance was affected in any way. He cannot use a third-party recall of behavioural frequency as a surrogate marker for chemical (im)balance. Walsh employs a non sequitur.
The study is total garbage, from a scientific perspective. I can't imagine how it got published.
It is useful, in one respect. It does suggest that a more rigorous test of the hypotheses is in order. (nota bene: Plural. There are multiple hypotheses in effect in this paper.) However, this test, this research report, is inadequate in all respects.
Aside about the other paper. He identified zinc/copper ratio as a potential marker for aggression. The subjects in this study had a 95% confidence interval that included virtually the entire normal range of values for zinc/copper. In any case, the high number of zinc disturbances seen here might well be a more controllable variable to consider in behavioural disturbance. Zinc deficiency is virtually a given in America today. Zinc interventions are really simple to manage and response measurements (blood tests) are readily available.
Lar
poster:Larry Hoover
thread:503534
URL: http://www.dr-bob.org/babble/alter/20050510/msgs/504633.html