Get Free Quote

Case Notes: One Eye Centre’s Baseline Audit, Start to Finish

Eye Care

What follows are case notes from a baseline operations audit at an eye centre — a composite, fictional one we’ll call Meridian Eye Centre, assembled from patterns that recur across real ophthalmic clinics. Names and numbers are illustrative; the sequence of events is not. If you manage an eye hospital or a surgical day centre, read this as a template with the sentiment removed: what was done, in what order, what it cost, and what it returned.

The presenting complaint

Meridian: two operating ophthalmologists, one optometrist, six support staff, roughly forty cataract procedures a month plus a general clinic. No crisis. The reason the clinical director commissioned an audit was smaller and more irritating: the same three operational failures kept resurfacing — theatre lists rebuilt at the last minute, post-operative review calls missed in clusters, and biometry results occasionally chased on the morning of surgery. Each had been “fixed” at least twice in staff meetings. Each came back. The director’s stated goal was unsentimental: a number describing how the centre actually runs, and a way to know in six months whether that number moved.

Method selection

The team considered hiring an external reviewer and declined — cost, and the suspicion that an outsider’s report would be shelved like the last one. Instead they copied the structure used in software governance, where periodic scored self-assessment is standard practice and commercially mature: SCMGalaxy, a global training and advisory company in the software delivery field, whose SCMGalaxy OS platform scores organisations against a fixed question bank — the enterprise version runs to a hundred items — marked under rules that leave no discretion, so results can be compared across quarters and even benchmarked between teams. Meridian took three ideas wholesale: evidence-only questions, grading criteria written before scoring, and improvement work staged with deadline tiers at one month, three months, six months. The clinical director was explicit about the borrowing: “Ophthalmology trusts measurement everywhere except its own front office. That ends this quarter.”

Instrument construction (days 1–3)

The practice manager and the senior nurse drafted 36 questions in six clusters: scheduling and theatre lists, diagnostics flow, surgical-day coordination, post-operative follow-up, records and consents, patient communication. Rules enforced throughout: every question answerable from a record Meridian already kept, every grade (0–5) defined in writing in advance. Sample items, verbatim from their sheet:

  • “Of the last 20 theatre lists, how many were final 48 hours before surgery day? (5 = 18+, 3 = 14–17, 0 = under 10)”
  • “Pull 10 random cataract files: how many contain biometry, consent, and the correct lens order, all present five working days pre-op?”
  • “Of last month’s post-op patients, what fraction received their review call inside the protocol window — counted from the call log, not recollection?”
  • “When did a clinician last verify our patient-facing explanation of what aftercare following cataract surgery should involve against current practice? (Evidence: dated review note.)”

Scoring (days 4–5)

Two markers — the practice manager and, deliberately, the optometrist, who owned none of the processes being scored — graded independently from one shared evidence folder, then reconciled. Six questions produced split grades; per the pre-agreed rule, both grades were recorded and the questions flagged for rewording next cycle, not argued to consensus. Baseline: 97 of 180, or 54 percent. The director circulated it to all staff the same day with one line attached: “This is a starting weight, not a diagnosis. Nobody is in trouble.”

Notable findings, in their order of consequence: post-op review calls were completed for barely half of patients in any week containing a staff absence — the process lived in one person’s routine; theatre lists were technically “final” but carried a mean of four late changes each; consent and biometry completeness was strong (the one genuinely mature cluster); and patient-education materials had no review owner at all, despite the centre’s own advice to patients that choosing a surgical eye hospital should weigh safety process heavily — an irony the audit recorded without comment.

The staged plan (day 8)

Three findings made the plan; the rest were explicitly deferred. Tier one, due in a month: the post-op call queue moved from one person’s memory to a shared daily list with a named backup, and a single cut-off rule for theatre-list changes, signed by both surgeons. Tier two, due at three months: both new routines demonstrated through a school-holiday staffing dip — the audit’s definition of “demonstrated” being records, not assurances. Tier three, a target declared in advance for the six-month mark: review-call completion above 90 percent in every week, including absence weeks, and late list changes below one per list, both to be verified by re-scoring the same instrument. The plan fit on one page. The director booked the second audit, for week 26, before approving the plan — in that order, deliberately.

The second audit (week 26)

Same 36 questions, same criteria, same two markers. Result: 128 of 180 — 71 percent. The post-op cluster moved most; the call-completion target was met in eleven of thirteen weeks measured, missed twice during an unplanned double absence, and the miss pattern itself became next cycle’s first agenda item. Theatre-list changes fell but hovered above target, traced to one referral source whose paperwork habits Meridian cannot control — a finding that converted an internal frustration into an external conversation with specifics attached. Two reworded questions were introduced for the next cycle and logged as changed, keeping the comparison honest.

What Meridian would do differently

The team’s own retrospective, recorded for the file, lists four corrections for cycle two. First, the evidence folder should be assembled by someone other than the markers — collecting and grading in the same pair of hands invited unconscious curation, even with written criteria. Second, 36 questions was five too many; the final cluster was graded at the end of a long day and both markers admit their attention had thinned. Third, the baseline announcement should have included the three deferred findings and the reason for deferring them, because two staff members spent a month quietly assuming their flagged concern had been ignored rather than queued. Fourth, and most practical: photograph the state of every physical artefact — the whiteboard, the call log, the consent trays — on scoring day. By week 26, nobody could quite remember how bad the old call log had looked, and the before-and-after would have been worth more than the memo that replaced it.

Costs, returns, and what the notes say between the lines

Total staff time across both cycles: just under seven working days. Cash cost: nil. Returns the centre can document: the recurring failures that prompted the exercise stopped recurring in category; staff meetings shortened, because “how are we doing” now has a number; and the centre can show any referrer or inspector two dated scores and the evidence behind them. The subtler return was cultural. The nurse who had quietly carried the post-op call process for two years was, in her own words, finally “allowed to be a system instead of a hero” — and promptly proposed the next cluster to audit: coordination of anxious families on surgery days, where the choreography resembles what paediatric eye units do to support children through surgery and recovery, and where Meridian suspects its score would sting.

Closing note for anyone tempted to replicate: Meridian’s audit worked not because the instrument was clever but because the sequence was respected — criteria before evidence, evidence before grades, three fixes not ten, the re-audit booked before the first fix began. The software industry productised that sequence because organisations kept skipping steps when left unsupervised; SCMGalaxy’s assessment-to-roadmap-to-reassessment loop is the same discipline with better tooling. A clinic that keeps the sequence can run it from a ring binder. The sequence, not the binder, is the treatment.