For about two years I could not have told you whether our therapy was working. I had progress reports that said it was. I had a child who was falling apart every evening. I did not have a way to hold those two things in the same hand, so I mostly assumed the reports were right and I was the problem.
I was wrong about that, and it took me far too long to work out why. This is what I have learned since, including the part I got wrong, and the specific questions I wish someone had handed me at the start.
The gap that nobody was measuring
The clinic’s data always looked neat. Our evenings did not. My child held it together under bright lights and a strict routine for hours, then came apart at home, which is supposed to be the safe place. Sleep got worse. Meals got tense. Nobody at the clinic was measuring any of that, because none of it happened in the room where the data was collected.
That gap is not a quirk of one clinic. It shows up in the research too. The largest review of early autism intervention studies to date, Project AIM (Sandbank and colleagues, published in BMJ in 2023), pooled 252 studies covering 13,304 children. Two findings from it changed how I read a progress report.
First, the authors state plainly that adverse events were “poorly monitored, but possibly common.” Harm was largely not being looked for, so it was largely not being found. Second, when the analysis excluded outcomes rated by caregivers and teachers and studies at high risk of detection bias, most of the reported benefits did not survive. Of the intervention types examined, only naturalistic developmental behavioral interventions held a significant effect on autism characteristics, and it was a modest one.
I am not reading that as “therapy does not work.” I am reading it as: the measurement is weaker than the confidence with which it gets presented to parents. When a clinician tells you the data shows progress, it is fair to ask what the data covers, and what it does not.
Signs I learned to take seriously
These are the things I now think are worth paying attention to. Not one of them proves anything on its own. Several of them together is a pattern.
Your child is fine in the room and not fine afterward
The technical version of this is that gains are not generalizing, and the clinical version is that the child is spending everything they have to get through the session. If your child is compliant for three hours and then cannot speak, eat, or sleep, that is information, and it belongs in the progress note.
Skills are being performed, not used
My child could label emotions on a card and could not tell me they were overwhelmed. Mastery on a target and understanding are not the same thing, and only one of them is easy to measure.
Nobody can tell you what would count as done
I asked, more than once, what the exit criteria were. I never got a straight answer. A plan with no stopping condition is not a plan, it is a subscription.
The recommendation never changes
When the answer to every question is more hours of the same thing, the recommendation is not responding to your child. It is responding to something else.
Your child is telling you, in whatever way they have
I can’t think when the room is busy. Help me get calm first.
My child said a version of that to me before I was ready to hear it as data. It was the most accurate assessment anyone gave me that year.
The thing I got wrong
I assumed that because my child looked calm and cooperative in sessions, the approach was working and the problem was our home routine. What I was actually watching was effort, not ease.
The research literature on autistic adults has a name for the long-term version of this. In a 2020 study in Autism in Adulthood, Raymaker and colleagues defined autistic burnout as a syndrome resulting from chronic life stress and a mismatch between expectations and abilities without adequate support, marked by long-term exhaustion, loss of skills the person previously had, and reduced tolerance for sensory input. The participants in that study identified masking, meaning suppressing autistic traits to appear less autistic, as a primary contributor. Recovery was associated with acceptance from others, reduced demands, and being allowed to stop masking.
That study is about adults, and I want to be careful not to over-claim it. My child was five. But the description of holding it together in one setting and losing skills in another was the closest thing I had found to what I was watching, and it reframed compliance from an outcome into a cost.
Questions worth asking at the next meeting
I stopped going in with feelings and started going in with questions. These are the ones that got me real answers.
- What are the exit criteria for this goal, and what happens when we hit them?
- How are you measuring what happens at home, and who collects that?
- Are you tracking anything that would tell us this is not working, or is every measure a progress measure?
- What would you change if I told you sleep and eating have gotten worse since we started?
- My child said this to me. How does the plan respond to that?
- If we reduced hours, what specifically would you expect to lose?
The last one mattered most. It is the question that forces a specific answer instead of a general warning.
What actually changed for us
I said no to adding more ABA hours and asked for full occupational therapy and speech instead. The clinic pushed back. Insurance stalled. I felt shaky about it for months.
The environment changed first, and it changed fastest: less glare, less background chatter, and the same familiar people rather than a rotating staff list. That is not a therapeutic breakthrough, it is just a room that stopped costing my child everything to be in. Progress after that looked less like a chart going up and more like my kid having something left at four in the afternoon.
I want to be honest that this is one family and it is not evidence. What we did is not a protocol and I would not hand it to you as one.
What I am not saying
I am not telling you to stop your child’s therapy. I am not qualified to, and a blog post cannot see your child. The American Academy of Pediatrics clinical report on autism is worth reading precisely because it lays out how varied the picture is, and how much depends on the individual child and family.
Plenty of families have had good experiences with providers who listened and adjusted. The variable that mattered in our case was not the acronym on the door. It was whether anyone was willing to change the plan when the plan was not working.
What I am saying is that you are allowed to ask for evidence that this is working outside the room, you are allowed to say no to more hours, and the fact that your child looks fine during the session does not settle the question.
If any of this sounds like your evenings, the next step is not a decision. It is one question at the next meeting. Pick one from the list above.
