Collecting useful clinical data without turning the client into their own surveillance system.
More data is not necessarily better data.
Your task is not simply to get the client to track. It is to design the smallest useful observation process
that can improve formulation or change the next treatment decision.
Target
What should be monitored?
Measure
What dimension matters?
Adapt
What makes it feasible?
Interpret
What can it tell us?
Stop
When has it done its job?
Approx. 25–35 minutes
Question
↓
Observe
↓
Notice pattern
↓
Generate hypothesis
↓
Test
↓
Revise or stop
Case orientation
Meet Maya
Maya is 27. She works rotating shifts at an emergency veterinary hospital while completing a part-time graduate degree.
She describes low mood, concentration problems, irritability, disrupted sleep, missed social plans,
procrastination, and long stretches spent awake in bed scrolling.
Maya
“I'm exhausted all the time.”
You ask how often she spends hours in bed while awake.
“Constantly.”
“Yesterday? Three hours? Maybe six.”
“Everything kind of blurs together.”
“Everything kind of blurs together.”
Decision 1 · Choose the target
What would you monitor first?
Clinical Support
Monitor the problem you need information about.
Clinical question
↓
What must we observe?
↓
What treatment decision could change?
If you cannot say what decision the data might help you make, the client may not need to collect it.
Decision 2 · What dimension matters?
Maya
“Do you just want yes or no?”
What is most useful?
Clinical Support
Match the measurement to the question.
Frequency
● ● ●
How often?
Duration
How long?
Context
After work
After school
After conflict
One five-minute episode and one four-hour episode count equally if you measure only frequency.
Decision 3 · Incomplete monitoring
One week later, Maya brings two entries.
Maya
“Sorry.”
“I kept forgetting.”
“By Thursday I figured I had already messed it up.”
How do you respond?
Monitoring redesign
Reduce the task to what Maya actually needs.
Explore the fields below. This interaction is not scored.
Collaboratively selected version
Start____
End____
What happened right before?________________________
Minimum viable monitoring.
Decision 4 · Measurement reactivity
Maya
“Every time I got in bed I thought, ‘Now I have to log this.’”
“So apparently annoyance is therapeutic.”
How do you interpret this?
Clinical Support
Measurement is not always neutral.
Behavior
↓
Notice
↓
Record
↓
Behavior changes
Ask what monitoring changed: awareness, delay, motivation, shame, effort, or deliberate choice.
Two weeks of monitoring
A pattern begins to emerge.
MON
After shift
TUE
Graduate seminar
WED
Reading assignment
THU
Family conflict
FRI
Socially demanding shift
SAT
Weekend
SUN
Weekend
Maya
“I thought I was just lazy all the time.”
Decision 5 · Pattern interpretation
What do you do with the pattern?
Clinical Support
Patterns generate hypotheses, not verdicts.
Physical fatigue after shifts
Social depletion
Avoidance after emotionally charged interactions
Transition difficulty between work and home
Move from “this always happens” toward “this seems more likely under these conditions.”
Decision 6 · Add variables?
Hypothesis 1
Physical exhaustion
Hypothesis 2
Escape from demand
What should you add?
Targeted addition
Start / end____ / ____
What happened right before?________________
Physical fatigue0–10
If I did not get into bed, what would I have been doing?________________
Add variables because you have a question—not because more information might be interesting.
Decision 7 · Subjective scales
Maya
“What's a seven?”
“Yesterday I put six, but today I think yesterday was an eight.”
What do you do?
Clinical Support
Measurement precision should not exceed the construct.
0 Not physically tired5 Noticeably tired but functioning10 Struggling to stay awake or continue
A 0–10 rating is useful for relative comparison and shared language. It is not a laboratory measurement.
Decision 8 · Monitoring becomes a project
Date
Duration
Fatigue
Context
Daily avg.
Weekly trend
Notes
Mon
142
6
After shift
74.3
↑
Reconstructed
Tue
31
5
School
69.8
↓
Complete
Wed
—
—
—
ERROR
?
Missing
Maya
“I made it better.”
“I got annoyed because Monday's numbers don't add up exactly.”
What is the clinical risk?
Clinical Support
Monitor the monitoring.
Observation
Reassurance
Control
Self-criticism
Self-monitoring is itself a behavior. It can have triggers, rules, consequences and functions.
Decision 9 · Missing data
M T W T F S S
● ● ? ● ? ● ●
Maya
“If I leave gaps, the week won't be accurate.”
What do you emphasize?
Clinical Support
Missing is not the same thing as none.
Day
Recorded bed time
Mon
2 hr 40 min
Tue
Missing
Wed
35 min
Thu
Missing
Missingness may reflect forgetting, burden, privacy, schedule, avoidance, or random interruption. Ask rather than assume.
Decision 10 · Memory versus record
Maya's memory
“The whole week was horrible.”
Recorded bed time
MON
2:40
TUE
0:15
WED
0
THU
0:40
FRI
3:05
How would you use the discrepancy?
Clinical Support
Data can refine memory without invalidating experience.
Lived impression
“The week felt globally bad.”
Recorded pattern
Long bed episodes clustered on two days.
The useful question is not “Which one is true?” but “What does the discrepancy teach us?”
Decision 11 · What is the bed doing?
“Had to open school portal.”
“Needed to call mom.”
“Had three unanswered messages.”
“Had to decide whether to go to friend's birthday.”
Maya
“So I'm avoiding things.”
Is that the conclusion?
Clinical Support
“Avoidance” is still a hypothesis.
Maya
“Nobody needs anything from me.”
“Nothing has to be decided.”
“I can just disappear for a little bit.”
Broad label
Avoidance
More functional description
Immediately available low-demand environment
Decision 12 · Monitoring → intervention
Maya
“So I guess I need to stop getting in bed.”
What is the next intervention?
Targeted behavioral experiment
Change the transition—not the moral status of rest.
Shift ends
↓
Commute
↓
Shower
↓
Food
↓
20 minutes low-demand music
Deliberate choice
Schoolwork
Social contact
Rest
Bed
Bed remains an available option.
Decision 13 · Early results
Tuesday
Transition routine
Bed: 35 min
Friday
No routine
Bed: 3 hr 05 min
Maya
“So the routine works.”
How confidently can you say that?
Clinical Support
Enough to keep testing ≠ enough to claim certainty.
“This is promising.”
“This pattern is worth testing.”
“We have preliminary evidence.”
Decision 14 · When to stop
After six weeks, Maya's total awake-in-bed time is lower overall.
Maya
“Can I stop tracking now?”
“Every time I lie down I hear a little therapist clipboard in my head.”
What do you do?
Monitoring exit check
Monitoring is a tool, not a lifestyle.
Do we understand the pattern?YES
Do we have a working hypothesis?YES
Has treatment changed because of it?YES
Does the next decision require daily data?NO
Daily monitoring can end.
Context complication
The graph does not make the schedule less demanding.
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Morning
Vet shift
Vet shift
Vet shift
Evening
Commute
Seminar
Commute
Seminar
Late commute
Maya
“We can make a very elegant graph of my exhaustion.”
“It doesn't make the schedule less insane.”
Decision 15 · Behavioral leverage and structural load
How do you use the data now?
Clinical Support
Monitoring can prevent overpsychologizing.
Behaviorally responsive
Transition difficulty
Task avoidance
Decision avoidance
Material / system load
Rotating shifts
Long commute
Late seminars
Schedule instability
Interacting — not mutually exclusive.
Decision 16 · More sophisticated data?
Maya
“What if I track sleep, heart rate, steps, and stress with my watch?”
“Then we'd have real data.”
What do you explore?
Clinical Support
Data sophistication should follow clinical utility.
What construct does it measure?
How accurately?
What clinical question does it answer?
What decision could change?
What burden might it create?
Self-Monitoring Learning Record
“Exhausted all the time”
↓
Time awake in bed
↓
Duration + context
↓
Pattern emerges
↓
Competing hypotheses
↓
Targeted additional data
↓
Refined formulation
↓
Behavioral experiment
↓
Reassess
↓
Stop daily monitoring
Adjacent CBT interventions
Self-monitoring
What actually happens across time and context?
Chain analysis
How did this particular episode unfold?
Thought record
What interpretation occurred, and how well does it fit the evidence?
Behavioral experiment
What deliberate change could test the prediction?
These interventions can inform one another without becoming a rigid sequence.
Final integration · Ten minutes remain
Maya
“The biggest thing I learned is that ‘all the time’ was hiding different things.”
“Sometimes I really am wrecked.”
“Sometimes I don't want to decide anything.”
“Sometimes bed is the only place nobody wants something from me.”
What would you do next?
Pattern synthesis
Your clinical attention tended to foreground…
This is not a score. It describes what your decisions tended to foreground.
Completed formulation
Too broad
Maya spends too much time in bed.
Still too broad
Maya uses bed to avoid things.
Working formulation
Under genuine schedule-related fatigue and high interpersonal/cognitive demand,
bed sometimes functions as necessary recovery and sometimes as an immediately available low-demand environment
that postpones decisions and tasks.
Self-monitoring helps distinguish these episodes well enough to target behavior
without reframing structural exhaustion as pathology.
Self-Monitoring Decision Rule
What information would actually help us decide what to do next?
If there is no question, do not assign monitoring.
Symptom, behavior, thought, activity, urge or context?
Frequency, duration, intensity, timing, sequence or context?
Paper, phone note, app, voice memo, calendar—or no monitoring at all.
How often, how many fields, how much time, and how much realistic capacity?
Awareness, avoidance, shame, compulsion, motivation or deliberate choice.
Do not assume. Ask.
Do not assign data you will not meaningfully examine.
Ask what is worth testing next—not what the data prove.
Define an exit condition.
Debrief
Reflect on the intervention process.
What clinical question did the first target answer?
Why did duration matter more than frequency?
When was it useful to add another variable?
What did incomplete monitoring teach you about design?
When did perfectionism enter the monitoring process?
What did the discrepancy between memory and record teach you?
Why did patterns support hypotheses rather than proof?
When did monitoring become sufficient to justify a behavioral experiment?
When did daily tracking stop adding clinical value?
How did the data distinguish behaviorally modifiable patterns from structural workload?
When might additional technology improve treatment—and when might it simply create more data?
Hold the question
If a client becomes excellent at measuring their distress
but no more able to live outside it,
what exactly has treatment improved?
No response is required.
Continue with The CBT Dive
See CBT use observation as a bridge to formulation—not as surveillance.
The transcripts extend these teaching modules into longer clinical conversations,
including uncertainty, pacing, contextual information, and the clinical judgment that disappears when interventions are reduced to worksheets.
Transcripts, Notes and Reflections from The CBT Dive Video Podcast
