i autistic 我自闭

An autism therapy can have weak or incomplete evidence without being fraudulent, and a practitioner can sincerely believe in an intervention that later turns out to be ineffective. The purpose of this article is to help families decide what deserves further investigation, what may be reasonable to try cautiously, and what deserves a much higher level of evidence before committing substantial money, time or hope.

Families often have to decide before the science is complete

Scientific research takes time. Parents and autistic people may not have the luxury of waiting years for every useful intervention to be conclusively validated. A child who has difficulty communicating, sleeping, eating, learning or coping with everyday life needs support now.

This creates a genuine decision-making problem. “There is not enough evidence yet” does not necessarily mean “this cannot work”. “Nobody has proved it does not work” is also not evidence that it probably does.

A practical principle is: The greater the cost, risk, commitment and claimed benefit, the stronger the evidence you should demand.

Trying something inexpensive, reversible and relatively low-risk is a very different decision from spending tens of thousands of dollars, travelling overseas, withdrawing a child from school, abandoning established support, taking significant medical risks or committing hundreds of hours of family labour.

Small, controlled experiments can sometimes be reasonable

When evidence is incomplete but an intervention seems reasonably safe, one option is to treat it as a limited experiment rather than making a leap of faith.

• Keep the downside small. Prefer interventions with low financial cost, low medical risk and limited disruption to family life.
• Prefer reversible decisions. It should be easy to stop if there is no meaningful benefit.
• Define the target beforehand. Decide what you are actually trying to improve, such as sleep duration, constipation, communication attempts or ability to complete a particular task.
• Measure before and after. Otherwise normal fluctuations can easily look like improvement.
• Set a reasonable trial period. Do not continue indefinitely because improvement might supposedly appear eventually.
• Set stopping rules. Stop if there are significant adverse effects, unacceptable distress, escalating costs or no worthwhile improvement.
• Avoid changing many things simultaneously where practical. If five treatments begin together, it becomes difficult to know which one helped or harmed.
• Consider opportunity cost. Even a physically harmless intervention can cause harm if it consumes money, attention or time that would otherwise go towards something more useful.
• Do not replace necessary medical care merely because an experimental alternative sounds promising.

Ordinary nutritional supplementation illustrates the importance of proportion. A basic vitamin or mineral supplement at an appropriate dose may involve far less cost and commitment than an intensive programme requiring months of therapy or overseas travel.

[Note that this does not mean supplements always work or are safe. Excessive doses can cause harm and supplements can interact with medications. Correcting an identified nutritional deficiency is also a different proposition from taking very high doses in an attempt to treat autism itself.]

The same reasoning can apply outside nutrition. A low-cost intervention that is safe, reversible and objectively measurable may justify a lower threshold for experimentation than an expensive or invasive intervention making extraordinary claims.

Good intentions and good evidence are different things

Not every practitioner offering an inadequately proven treatment is deliberately deceiving people. A practitioner may genuinely believe what they have observed. Parents may genuinely report dramatic improvements. Researchers may pursue an interesting but ultimately incorrect hypothesis. Traditional practices may have developed from generations of sincere observations. All of these things can be true while the treatment itself still does not produce the claimed effect.

Humans are not especially good at determining causation from personal experience. Autism also develops over time, abilities can change unevenly, symptoms and behaviour fluctuate, and families frequently make several changes simultaneously. This makes sincere misinterpretation quite possible. Therefore, evaluate the claim and evidence separately from your opinion of the person’s character.

Plausibility matters, but it is not proof

Scientific plausibility is useful when deciding where to spend limited time and money. A treatment consistent with established biology and supported by evidence deserves a different starting level of confidence from one requiring several well-established principles of chemistry, physics or physiology to be wrong. However, plausibility alone does not prove that a treatment works. Many plausible medical ideas have failed when properly tested.

Homeopathy provides a useful example at the other extreme. Homeopathy traditionally proposes that substances associated with particular symptoms can have therapeutic effects when repeatedly diluted, with some preparations diluted to the point where none of the original substance would ordinarily be expected to remain. Major scientific reviews have found little reliable evidence that homeopathy effectively treats specific health conditions, and important parts of its proposed mechanism are unable to be reconciled with established chemistry and physics.

Even if, purely for the sake of argument, we assume that some extremely weak treatment-specific signal exists, another problem remains: can that signal reliably rise above the noise? The apparent improvement observed after any treatment can include:

• natural development;
• normal fluctuations in symptoms or behaviour;
• regression to the mean after treatment is started during an unusually difficult period;
• placebo and other contextual effects;
• changes in expectations or caregiver behaviour;
• other treatments or environmental changes occurring simultaneously;
• measurement error and subjective interpretation; and
• an unknown mechanism unrelated to the theory being promoted.

If a person appears to improve after homeopathy, it is not necessary to deny that the improvement occurred. The scientifically important question is whether homeopathy caused it, and if so, whether the proposed homeopathic mechanism is actually the explanation. The same reasoning should be applied fairly to conventional, complementary and alternative treatments.

Start with the evidence

Understand exactly what an intervention claims to accomplish, then look for evidence addressing that particular outcome. Evidence that something improves sleep does not establish that it improves communication. Evidence that it reduces anxiety does not establish that it changes autism itself. Evidence that an ingredient produces a biological effect in a laboratory does not automatically establish that the finished treatment produces a meaningful benefit in autistic people.

When examining research, consider:

• study size;
• appropriate control or comparison groups;
• randomisation where appropriate;
• blinding where realistically possible;
• whether outcome measures are objective or suitably validated;
• length of follow-up;
• dropout rates;
• the actual size of the benefit rather than statistical significance alone;
• funding and financial conflicts of interest;
• publication quality;
• whether results have been independently replicated; and
• whether the conclusion remains convincing when unfavourable studies are considered as well.

Do not mechanically demand the same research design for every intervention. For example, parents obviously know whether they are providing a parent-mediated developmental intervention, so double-blinding may be impossible. In such cases, independent or blinded assessment of outcomes, appropriate comparison groups and other methods of reducing bias become more important.

Testimonials are useful information, but weak evidence of causation

A testimonial tells you what somebody believes happened to them. That can be worth reading. It does not establish why it happened or how often the same outcome occurs. Be especially cautious when almost all evidence for a treatment consists of dramatic success stories. Ask:

• How were these families selected?
• What happened to families who did not improve?
• How typical is the reported result?
• What else changed during the same period?
• How was improvement measured?
• Is there independent follow-up?
• Did the family receive discounts, free services, commissions, referral benefits or other assistance connected with promotion?

Read the bad reviews too

When investigating a treatment, practitioner or organisation, deliberately search for negative reviews, complaints and criticism rather than reading only material recommended by the provider.

Negative reviews are not automatically true. They can result from unreasonable expectations, misunderstandings, personal disputes, competitors or deliberate fabrication. Positive reviews can also be manipulated. Singapore consumer authorities have documented concerns involving fake reviews, incentivised ratings, undisclosed sponsorships and deletion of negative reviews, and there have been enforcement cases involving fabricated five-star reviews.

Instead of treating any individual review as proof, use reviews as leads for further investigation. A useful negative review contains details that can potentially be checked. Pay attention to:

• specific descriptions of what happened rather than insults or vague accusations;
• dates, amounts paid and descriptions of services received;
• screenshots, invoices, correspondence or other supporting material where legitimately published;
• similar complaints made independently by different people over time;
• problems involving refunds, unexpected charges, pressure selling or promises that differ from written terms;
• how the organisation responds to criticism;
• whether credible criticism appears to disappear from platforms controlled by the organisation; and
• whether a regulator, professional body or court record independently confirms any part of the complaint.

Search beyond the provider’s own website and social-media pages. Their own pages are naturally selected and moderated by them. At the same time, avoid repeating an anonymous review as established fact. “Several reviewers alleged X” and “X happened” are very different statements.

When families become part of the marketing system

Parent involvement in therapy is not itself suspicious. Some established interventions deliberately train parents and caregivers because much of a child’s development occurs during ordinary daily life. A different issue arises when families increasingly become part of the marketing, recruitment or fundraising machinery surrounding a treatment. Warning signs worth examining include:

• parents being strongly encouraged to promote the programme repeatedly on social media;
• families being given scripts, promotional material or suggested success stories to publish;
• existing families being used to recruit new families;
• discounts, benefits, referral rewards or other material connections that are not clearly disclosed;
• parents being encouraged to publicly fundraise substantial sums to send a child to an overseas centre;
• fundraising narratives that present a costly programme as the child’s exceptional or last realistic chance;
• social communities in which enthusiasm is strongly rewarded but difficult questions or disappointing outcomes are discouraged;
• a treatment’s public reputation depending heavily on testimonials generated by families currently participating in it;
• families being told that treatment failure reflects insufficient parental effort, commitment, positivity or fundraising rather than limitations of the intervention; and
• substantial unpaid family labour being presented as evidence of treatment effectiveness rather than being counted as part of the treatment’s true cost.

None of these proves that an organisation is dishonest. Parents may enthusiastically promote something because they sincerely believe it helped their child. Fundraising can also be entirely voluntary and legitimate.

The relevant questions are whether these relationships are transparent, whether families are under pressure, whether financial or other incentives are disclosed, and whether independent evidence exists outside the community promoting the intervention.

Beware of founder-centred evidence

Some therapies become closely associated with a persuasive founder, compelling personal story or movement-like community. There is nothing inherently wrong with being charismatic, enthusiastic or skilled at explaining an idea. Those qualities may help someone build an organisation and communicate effectively. However, a compelling founder is not evidence that a treatment works. Look more closely when:

• the founder’s personality and personal story receive more attention than independent outcome data;
• conference speeches, inspirational videos and media appearances substitute for published research;
• criticism of the therapy is reframed primarily as criticism of the founder or community;
• the organisation relies heavily on emotional transformation stories;
• families are encouraged to identify strongly with a movement surrounding the therapy; or
• claims depend heavily on what the founder says they have personally witnessed.

Charisma can make a good idea easier to understand. It can also make a weak idea easier to sell. Evaluate the evidence separately.

Investigate the practitioner and organisation

Verify important information independently where practical:

• What qualifications does the practitioner actually hold?
• Which institution awarded them?
• Is professional registration required for the service, and if so, is it current?
• Does the treatment fall within the person’s professional scope?
• Are claimed affiliations and endorsements genuine?
• Who owns the organisation?
• Who profits from the treatment, associated products, referrals or training programmes?
• Are fees and refund conditions clear before payment?
• What happens if treatment does not work?
• What adverse effects or unsuccessful cases does the organisation acknowledge?

The title “doctor” should also be interpreted carefully. It can describe a medical practitioner or somebody holding a doctorate in another field. What matters is the person’s actual qualification, registration and relevance to the particular claim being made.

Other red flags worth investigating

• Guaranteed, miraculous or exceptionally rapid outcomes.
• A single dramatic success story presented as proof.
• Claims much stronger than the research being cited.
• Credentials or affiliations that cannot readily be independently verified.
• Endorsements from impressive people whose expertise is unrelated to the treatment claim.
• Stock photographs or artificial imagery presented in a way that could be mistaken for actual patients or testimonial providers.
• Extensive discussion of benefits with little meaningful discussion of uncertainty, failure rates, limitations or adverse effects.
• Selective citation of favourable studies while stronger contradictory evidence is ignored.
• Marketing expenditure or social-media activity that appears disproportionate to the available independent evidence.
• Multi-level or recruitment-based financial incentives.
• Pressure to pay quickly because a special opportunity will supposedly disappear.
• Expensive prepaid packages before the person’s response to treatment is known.
• Discouraging independent professional advice.
• Framing reasonable scepticism as proof that the family is negative, closed-minded or insufficiently committed.
• Repeatedly moving the goalposts when measurable improvement fails to occur.
• Claims that treatment failure means the treatment was started too late or not pursued intensely enough, regardless of circumstances.
• Substantial overseas travel where the reason the treatment cannot be provided locally is unclear.
• Unclear regulatory oversight, complaint procedures or legal entity.
• “Natural” being used as if it means “safe.”

A red flag means look closer. One red flag does not establish fraud. Several independent red flags pointing in the same direction deserve progressively greater caution.

Questions to ask before making a major commitment

• What exact problem are we trying to improve?
• How will improvement be measured?
• What is the strongest independent evidence?
• How many people do not improve?
• What adverse effects have been reported?
• What is the total financial cost?
• How many hours of parent and child time are required?
• What established services might we give up to do this?
• Can we try a smaller or cheaper version first?
• Can we stop easily?
• What happens if there is no measurable benefit?
• Does anyone recommending the intervention receive a financial or other benefit?
• Why must this treatment be provided at this particular centre or country?
• Can we take the information away and obtain an independent opinion before paying?

Distinguish “unproven” from “fake”

These terms should not be used interchangeably.

• Limited evidence can mean that research exists but remains small or preliminary.
• Conflicting evidence means reasonably credible studies have produced different results.
• Experimental can describe an intervention that is still being investigated.
• Scientifically implausible means the proposed mechanism conflicts substantially with established knowledge, although empirical claims can still be tested.
• Unsupported means the available evidence does not adequately justify the claim being made.
• Ineffective should generally be reserved for situations where suitable evidence supports that conclusion for the particular claimed outcome.
• Fraudulent, deceptive or fake goes further because it can imply something about conduct or representation. Exercise particular care before applying such descriptions to identifiable people or organisations.

Be precise when publishing criticism

If you discover something concerning, state what can actually be established.

If you cannot find somebody in a professional registry, say that you were unable to verify the claimed registration in that registry as of the date you checked. Do not automatically conclude that the person invented the credential.

If a regulator opens an investigation, describe it as an investigation. If the regulator later makes findings, link to and accurately describe those findings.

If reviewers make allegations, attribute those allegations to the reviewers unless independently established.

If research is retracted, corrected or criticised, describe what the relevant journal, institution or authors actually said rather than speculating about motives.

This is both fairer and more useful. Readers can see the underlying facts and form their own conclusions.

Note: This page provides general educational information and a framework for evaluating treatment claims. It is not individual medical or legal advice and does not diagnose, recommend or rule out treatment for any particular person. Health decisions should take account of individual circumstances and, where appropriate, advice from suitably qualified professionals. References to warning signs are prompts for further investigation, not findings that any particular practitioner, organisation or intervention has engaged in fraud, deception, professional misconduct or unlawful activity.