Back to playbooks

Standard Plays (minimum viable interventions)

These are the default moves to reduce admin drag safely. Pick one. Ship it. Measure it.

Rule of thumb: start with clarity, routing, and deleting dumb loops before you reach for heavy automation.

Play 1: Inbox triage rules (human-owned)

Goal: reduce clinician inbox time without missing safety signals.

• Define categories (urgent clinical / routine clinical / admin / FYI)
• Route admin to staff; route routine to protocol queues
• Use drafted replies for common items (clinician approves)
• Add clear escalation triggers

Measure: inbox minutes/day, response times, safety events.

Ship this: a routing table with escalation triggers.
Safety backstop: random weekly audit of urgent items and missed escalations.
Failure modes: urgent items buried, staff overload, patients unsure who owns the reply.

Play 2: Drafted patient replies (templates + personalization)

Goal: stop rewriting the same message 30 times.

• Build 10-20 templates for common questions
• Include “red flag” language (when to call/come in)
• Clinician edits and sends

Measure: time per message, patient satisfaction proxy, callback rate.

Ship this: 10-20 reviewed templates with red-flag language.
Safety backstop: clinician approval stays required for send.
Failure modes: wrong tone, missing nuance, callback volume shifts instead of falling.

Play 3: Note drafting from a structured visit outline

Goal: cut note time without degrading quality.

• Use a consistent structure (HPI, assessment, plan)
• Draft from problem list + prior notes + visit structure
• Require citations/links to source in chart for key claims
• Clinician reviews and signs

Measure: note closure time, after-hours time, documentation quality audits.

Ship this: one structured note format for one visit type.
Safety backstop: clinician signs and spot-audits source references.
Failure modes: hallucinated details, stale carry-forward text, hidden quality drop.

Play 4: Prior auth packet builder (assembly, not decisions)

Goal: reduce the hunt-and-gather burden.

• Standard packet per payer and condition
• Auto-assemble: history, meds tried, labs/imaging, guidelines
• Human reviews; clinician signs
• Track where packets still bounce so the form gets better over time

Measure: cycle time, approval rate, number of rework loops.

Ship this: one packet checklist for one payer and one condition.
Safety backstop: human review and sign-off before submission.
Failure modes: missing document, payer-specific mismatch, packet assembly shifts burden to the wrong role.

Play 5: Plain-language denial cleanup

Goal: stop wasting time on denials nobody can decode.

• Require a simple denial reason field people can actually understand
• Group denials into a small set of usable categories
• Build the next action into the workflow: resend / appeal / switch / done
• Track repeat denial reasons by payer and service line

Measure: time from denial to next action, repeat denial categories, avoidable rework.

Ship this: denial categories plus one plain-language next-step grid.
Safety backstop: appeal/escalation owner named for unclear denials.
Failure modes: denials categorized prettily but still unresolved, staff confusion about next move.

Play 6: Referral handoff owner

Goal: reduce referral friction and bounce-backs.

• Assign one owner for outbound referral handoffs
• Standard summary + question being asked
• Include relevant labs/imaging and timeline
• Use a checklist so nothing is missing
• Measure time to scheduled visit, not just time to fax sent

Measure: referral turnaround time, incomplete referral rate, time to scheduled visit.

Ship this: a handoff checklist plus one owner assignment.
Safety backstop: weekly audit of stalled referrals.
Failure modes: fax sent counted as done, patient still unscheduled, ownership ambiguity persists.

Play 7: Protocol-driven med refills (rules + flags)

Goal: move routine refills out of clinician brainspace.

• Define refill protocols
• Auto-flag exceptions (labs overdue, interactions, controlled meds)
• Human ownership stays clear

Measure: refill turnaround, exception rate, safety events.

Ship this: one refill protocol with exception flags.
Safety backstop: exceptions route to clinician review automatically.
Failure modes: overdue labs missed, controlled-med exceptions mishandled, refill delay shifts to staff queue.

Play 8: Fast-lane approvals for repeat low-value denials

Goal: stop repeating the same approval fight for the same low-risk cases.

• Identify services with high approval likelihood after routine rework
• Create a gold-card or fast-lane proposal for those categories
• Start with one payer or one service line if needed
• Track whether time saved comes without extra downstream problems

Measure: approval cycle time, repeat-appeal volume, clinician/admin time saved.

Ship this: one fast-lane proposal for one payer or one service line.
Safety backstop: track downstream denials or safety issues after approval shortcuts.
Failure modes: payer refuses the lane, hidden rework shifts downstream, low-value fight remains unchanged.

Play 9: Reduce duplicate work (the delete-a-step play)

Goal: remove a step that exists only because “we’ve always done it.”

• Identify duplicates (double entry, double review, duplicate portal upload)
• Propose one removal with a safety backstop
• Trial for 2-4 weeks; keep if safe

Measure: minutes saved, error rate.

Ship this: one deleted step plus a 2-4 week trial.
Safety backstop: rollback trigger if error rate rises.
Failure modes: hidden duplicate was actually a safety catch, work reappears elsewhere.

Play 10: Patient instructions that prevent repeat calls

Goal: reduce confusion loops for patients and staff.

• Rewrite top repeat-confusion instructions in plain language
• Put next steps, timing, and who-to-call in one place
• Check whether the message works with real patients, not just staff

Measure: repeat call volume, message turnaround time, patient confusion rate.

Ship this: one rewritten instruction bundle for one high-repeat workflow.
Safety backstop: test with real patients or staff before broad use.
Failure modes: instructions read clearly to staff but not to patients, confusion shifts to portal messages.

Play 11: Lab and diagnostic results routing rules

Goal: keep abnormal results visible while moving routine review through a cleaner path.

• Define what requires clinician eyes
• Define what can route through protocol or staff
• Name turnaround expectations
• Build escalation for abnormal or unclear results

Measure: result turnaround time, missed follow-up rate, patient callback rate.

Ship this: one routing rule table for one result category.
Safety backstop: abnormal results audit with rapid escalation.
Failure modes: abnormal values misrouted, staff uncertainty about escalation, delayed callbacks.

Play 12: One-click orderset plus patient instruction bundle

Goal: reduce repeated assembly work for common conditions.

• Bundle orders, education, and next-step instructions
• Keep clinician review and customization
• Start with one high-volume condition

Measure: order-entry time, repeat clarification messages, completion rate.

Ship this: one orderset plus one instruction bundle.
Safety backstop: clinician review before sign.
Failure modes: stale defaults, wrong patient instructions, hidden downstream work.

Play 13: Visit prep and pre-charting guardrails

Goal: move safe prep work upstream without shifting judgment away from the clinician.

• Define what can be prepped safely
• Define what must stay clinician-owned
• Standardize prep inputs and red flags

Measure: visit prep time, note closure time, missing-info rate.

Ship this: one pre-charting checklist for one visit type.
Safety backstop: clinician confirms key judgment fields.
Failure modes: staff drift into assessment, missing nuance, prep becomes duplicate work.

Back to playbooks