From Formulation to Practice: Testing PBS Hypotheses and Proving Impact in UK Care Settings

February 5, 2026

In many services, a behavioural formulation is written once, stored, and slowly becomes disconnected from day-to-day support. The result is familiar: staff do “what works for them”, incidents repeat, restrictive practice remains high, and families lose confidence. A better approach is to treat formulation as a testable set of hypotheses that must be implemented, checked and refined through governance. This article shows how to move from functional assessment and formulation into day-to-day delivery while staying aligned with PBS principles and values.

Start with a hypothesis you can actually test

A useful hypothesis is specific enough that different staff can apply it consistently. It should link:

Setting events (sleep loss, pain, medication changes, staffing change, trauma reminders)
Triggers (demands, transitions, sensory overload, uncertainty, waiting)
Behaviour (observable description)
Maintaining consequences (escape, access, attention, sensory regulation, control)

Then define the “replacement pathway”: what the person will do instead, and what staff must do to make that possible (communication supports, environmental changes, predictable routines, de-escalation and reinforcement).

Make the plan operational: “if this, then we do this”

For day-to-day delivery, staff need clear instructions that reduce decision fatigue. Strong plans include:

Early indicators of distress and the first response (before behaviour escalates)
De-escalation steps that are safe, lawful and consistent
What not to do (common escalators, inconsistent consequences, unnecessary demands)
Post-incident support (recovery time, repair, reflection, documentation)
Review triggers (repeated incidents, PRN use, restraint, safeguarding concerns)

This is also where restrictive practices must be explicitly addressed: what restrictions exist, why they exist, how they are monitored, and what the reduction plan is.

Fidelity: are staff delivering the plan as written?

Services often measure outcomes (incidents, restraint) but not whether staff are implementing the intervention reliably. “Fidelity” checks do not need to be complex:

Short observation by a competent lead using a checklist (5–10 minutes)
Spot-check of documentation: did staff use the agreed language and steps?
Reflective supervision: can staff explain the hypothesis and early indicators?

Fidelity is a safeguarding and quality issue. If the plan is not delivered consistently, you cannot interpret outcome data accurately, and families may experience the service as unpredictable or unsafe.

Operational example 1: Testing a demand-related hypothesis without increasing restrictions

Context: A person frequently throws items and shouts when asked to engage in daily living tasks. Staff begin avoiding tasks or escalating quickly to “firm boundaries”, and incidents increase.

Support approach: The formulation hypothesis is that behaviour is maintained by escape from demands, amplified by uncertainty and lack of control. The plan tests two changes: (1) increasing predictability and choice, (2) reinforcing participation in small steps.

Day-to-day delivery detail: Staff use a consistent “choice-first” approach: present two acceptable options, use a visual sequence, and break tasks into 2–3 minute steps with planned pauses. Staff reinforce cooperation immediately (specific praise, access to a preferred activity) and avoid bargaining during escalation. Early indicators (pacing, clenched fists) trigger a pause and a reset rather than pushing through. Staff record whether the person completed the first step and how much support was required.

How effectiveness is evidenced: Within four weeks, the service tracks reduced incident frequency and improved task engagement (percentage of days with successful first-step completion). Fidelity checks show whether staff used choice-first and paused at early indicators. Where fidelity is low, refresher coaching is provided and documented.

Operational example 2: Reducing restraint by improving post-incident learning and debrief quality

Context: In a registered setting, restraint episodes occur during peaks of agitation. Debriefs are inconsistent and focus on “what staff should have done”, without identifying patterns or changes to the plan.

Support approach: The service introduces a standard 24–48 hour post-incident review: what changed before the incident, what early indicators were missed, what de-escalation was attempted, what restrictions were used, and what should change in the environment or routine.

Day-to-day delivery detail: After each incident, the shift lead completes a short template and schedules a mini-debrief with involved staff (10–15 minutes) plus a check-in with the person where appropriate. Learning actions are assigned (for example, adjust transition timing; introduce a break card; change staff positioning during support). The manager reviews restraint data weekly, not just monthly, until the trend stabilises.

How effectiveness is evidenced: The service measures restraint frequency and duration, plus “near misses” where de-escalation prevented restraint. Governance minutes record the learning actions, completion dates, and plan version changes. Family updates are recorded when relevant, supporting transparency.

Operational example 3: Health-triggered behaviour and commissioner-ready outcomes

Context: A person’s self-injury spikes intermittently. Staff implement multiple strategies, but outcomes are inconsistent and the service struggles to explain progress to external professionals.

Support approach: The hypothesis is that behaviour increases with pain and sleep disruption, and is maintained by sensory relief. The service links PBS work with health monitoring and sets clear outcome measures.

Day-to-day delivery detail: Staff implement a weekly “health and wellbeing check”: sleep pattern, bowel pattern where needed, PRN use, and indicators of pain. The plan includes sensory alternatives (safe sensory input) and a clear escalation route for suspected pain (same-day GP request criteria). Staff keep a simple dashboard: incidents, sleep quality rating, PRN use, and engagement in preferred activities.

How effectiveness is evidenced: Over six weeks, the dashboard shows the relationship between poor sleep/pain indicators and incident spikes. After health interventions and routine changes, the service evidences sustained reduction in self-injury and improved engagement. This becomes commissioner-ready reporting because it shows data, actions taken, and outcomes achieved.

Commissioner expectation: outcomes that demonstrate both safety and quality of life

Commissioner expectation: Commissioners generally want evidence that PBS work reduces avoidable escalation and restrictive practice while improving day-to-day outcomes. Reporting is strongest when it includes:

Incident and restrictive practice trends (with narrative explaining what changed)
Quality of life indicators (community access, participation, routine stability)
Implementation evidence (training, coaching, fidelity checks)
Clear review points and governance oversight

Regulator expectation: CQC-ready assurance that practice is consistent and lawful

Regulator / Inspector expectation (CQC): CQC typically looks for consistency, learning, and least restrictive practice. Services should be able to show:

Staff can explain the person’s distress and the agreed plan in plain terms
Restrictive interventions are monitored, reviewed, and reduced with a clear rationale
Post-incident learning leads to plan updates and staff coaching
Governance routes exist for high-risk decisions (including safeguarding and MCA-related decisions where relevant)

Governance checkpoints that stop drift

To prevent formulation drift, build routine checkpoints:

2-week implementation check: Is the plan being used on every shift?
6-week outcomes review: Are incidents and restrictions reducing? Is quality of life improving?
Quarterly assurance review: Trends, restrictive practice oversight, training compliance, and supervision quality

When these checkpoints are in place, formulation becomes a living system: hypotheses are tested, practice improves, and outcomes can be demonstrated confidently.