You write a careful, function-based plan. Six weeks later, the data aren't there, the technician is asking to be reassigned, and the caregiver still isn't following through.
It is tempting to conclude the staff aren't doing their job. Before you do, ask a harder question: is this a fidelity problem, or was it a design problem from the start?
Here is a case that shows what I mean. It's completely made up, but it's built from the kind of situations I see all the time.
Meet Mateo
Mateo is seven and has severe self-injurious behavior: head hitting and hand biting, mostly during transitions and when he is denied preferred items. He lives in a two-bedroom apartment with two younger siblings and a single parent who works rotating shifts.
After two years in a 40-hour-a-week clinic program, his hours were reauthorized at 15 per week at home following an insurance review, and he started public school.
The plan on paper was clinically sound. It used a prompted functional communication response, a token economy, frequent reinforcement checks, and a padded rest area with documented blocking and safety protocols. If you only have six to eight hours to assess a child like this, you design for the highest level of support you think he needs. That's reasonable.
What happened by week six
There was no room for the rest area. In a two-bedroom apartment with several people, there was nowhere to put it.
Nobody had a free hand. The token economy needs a free hand to deliver tokens, and during an episode of self-injury, neither the technician nor the parent had one.
Caregiver training happened once. It was squeezed between the parent's shifts, and a second session never made it onto the calendar.
The technician asked to be reassigned, citing safety. Trial counts were too low to show up in the data, and the BCBA and clinical director began to wonder why the technician wasn't following the plan.
That is the moment I want you to notice: a frustrated BCBA looking at a technician and a caregiver and concluding that they aren't doing what they are supposed to do.
Run it through a fit check
When I run a plan through a contextual fit review, I look at three areas.
Environmental fit. Does the setting support the plan as written? The apartment had no space for the reset corner. Not in place.
Implementer fit. Does each person have the skill, training, and support to sustain it? The technician understood the plan well. The caregiver had one training session, no coaching in the moment, and no follow-up. Not in place.
Organizational fit. Are there systems keeping the plan alive over time? There was no supervision touchpoint after the two-week onboarding window closed. Not in place.
One point from the tool matters here. When you assess both a technician and a caregiver, the lowest score drives the decision. A technician who scores beautifully doesn't rescue a plan the caregiver can't carry out. You may need two versions of the plan, and ideally they build on each other, not diverge.
Redesigning without changing the clinical target
The clinical target for Mateo never changed. What changed was whether the plan could survive contact with his actual home.
Environmental: Replace the fixed rest area with a portable calm kit in a backpack, with headphones, a weighted lap band, and whatever helps him feel safer, usable anywhere in the apartment or out in the community.
Implementer: Replace the single long training block with 15-minute coaching sessions built into existing handoff times. Do some of the training with the caregiver and some with the child present.
Organizational: Add a standing 15-minute biweekly check-in as a required touchpoint. More would be better, but start with what you can sustain while still meeting your billing and every other requirement.
Feedback: Decide how the plan will be checked, who is acknowledged for following it, and what the organization does if it isn't.
Some of my most powerful parent trainings have been on hard days, when I switched a protocol modification session to family guidance and coached live while the parent and I worked together. They left having watched what I did, what they did, and how to do it themselves.
If staff behavior doesn't change, the system hasn't changed.
Why this matters for how you assess
Think about your most intense cases. Even with 20% supervision, six hours of a 30-hour week, there are still 24 hours when a newly trained technician is working without you. Those 24 hours are where the plan has to fit.
So run an environmental, implementer, and organizational check before you treat any plan as final, not after something has gone wrong. If you plan to land on the moon and you land on Mars, the plan to go to space was not the problem. The atmosphere was different.
If you remember one thing: when a plan isn't being followed, check the design before you question the people.
You can try the fit check yourself with the free B-FIT Clinical tool on our Clinical Tools page. It has one track for the technician and one for the caregiver, saves nothing, and runs entirely in your browser, so no client data is stored.
If your team is struggling with a plan that was never going to work, learn about Clinical and Agency Consulting. I offer a free 15-minute consultation if you want to talk it through.
Casey Gressmen, MBA, MEd, BCBA, LBA, is a licensed special education teacher and board certified behavior analyst and the founder of Phoenix Rising Behavior and Academic Consulting.
