How the workload review works
Healthcare IT Workload Review describes recent workload patterns using the schedule and answers you provide. It is descriptive, not diagnostic.
Three evidence types, kept separate
The review keeps three kinds of information visibly separate and never merges them into a score.
What was scheduled during the review period — scheduled-event time, events outside your normal hours, open gaps, and back-to-back stretches. In this manual MVP these are your own estimates. Scheduled events show what was on the calendar; they do not confirm you attended, and calendar availability does not prove you had productive focus time.
What you tell us directly — on-call days, after-hours work, a high-intensity work phase, and how often complexity made work hard to finish. These are reported answers.
Your own words about drain, recovery, priority control, and support. Reflection is kept separate, is never scored, and is never combined with your schedule into a number.
Provenance labels
Every value carries a provenance label with visible text, so nothing relies on shape or colour alone.
Estimate — your own estimate of your schedule. In the manual MVP, most schedule metrics carry this label.
Reported — something you told us directly. Workload-context and reflection inputs carry this label.
Calendar — derived from connected calendar timing. This appears only on the future connected path, which is not enabled here.
Review period and normalisation
- You choose the previous 14 or 28 days.
- 14-day totals are normalised to a weekly rate by dividing by 2.
- 28-day totals are normalised to a weekly rate by dividing by 4.
- Weekly categories never compare a raw 28-day total directly with a weekly threshold.
Configured weekly capacity
Configured weekly capacity = selected workdays × (end time − start time − unpaid break)
- The unpaid break is visible and user-configurable on the schedule step.
- No 40-hour week is assumed.
- 08:00–17:00 with a 60-minute unpaid break across five workdays equals 40 configured hours.
Product-defined categories
These categories are product-defined. They are not norms, health thresholds, validated benchmarks, or recommended targets.
Scheduled-event load (share of your normal workweek)
- Light
- under 20%
- Moderate
- 20–39%
- Heavy
- 40–59%
- Very heavy
- 60% or more
Events outside your normal hours (per week)
- None
- 0
- Occasional
- greater than 0, fewer than 2
- Frequent
- 2 to fewer than 4
- Very frequent
- 4 or more
Workdays with no 90-minute open gap (per week)
- Generally available
- 0–1 days
- Sometimes limited
- 2 days
- Frequently limited
- 3 days
- Consistently limited
- 4 or more days
Back-to-back scheduled events (share of workdays)
- Never
- 0% of workdays
- Rarely
- up to 10% of workdays
- Sometimes
- 10–29% of workdays
- Often
- 30–59% of workdays
- Most workdays
- 60% or more of workdays
Work outside your normal schedule (per week)
- None
- 0 hours
- Limited
- less than 2 hours
- Elevated
- 2–5 hours
- High
- 6–10 hours
- Very high
- more than 10 hours
On-call exposure (share of the review period)
- None
- 0%
- Limited
- greater than 0% through 20%
- Regular
- greater than 20% through 40%
- Frequent
- greater than 40%
How a result is selected
- One headline is selected from deterministic rules.
- One cross-signal observation may be shown when signals combine meaningfully.
- The first eligible rule wins; only one cross-signal appears.
- Up to two non-duplicative observations may follow.
- No AI model selects or scores the result.
- No composite or burnout score exists anywhere.
Missing data
- Missing answers are omitted, never guessed.
- When a value is missing, its metric and any dependent findings are left out.
- The reflection section may be skipped; the result then describes workload patterns only.
- Zero scheduled-event time does not mean low total workload — unscheduled support, build work, and interruptions are not captured.
- Manual estimates may be imperfect, and are labelled as estimates.
Current and future sources
- User-entered estimates and reported answers only.
- No Outlook connection.
- No raw calendar data.
- No employer access.
- Not enabled in this MVP.
- Requires a separate Microsoft Graph feasibility and permissions review.
- Would process event timing metadata into aggregate values only.
- Would never read or store titles, bodies, attendees, locations, join links, or attachments.
- Field availability remains spike-gated and unverified.
See also Privacy and deletion.