Prior authorisation is often treated as a clinical review problem. In practice, much of the delay occurs before a nurse or medical director evaluates medical necessity.
Requests arrive through portals, fax and phone. Eligibility information is checked in another system. Documentation is reviewed for completeness, then returned to the provider when something is missing. Straightforward cases wait behind complex cases, while scheduling teams may not receive the approval notification promptly.
Value Stream Mapping (VSM) exposes this entire pathway. It separates value-adding review from waiting, rework, batching and avoidable hand-offs. The objective is not to remove appropriate clinical oversight. It is to ensure that routine, complete and policy-aligned requests do not enter a clinical review queue unnecessarily.
Worked example: The figures in the tables below represent an illustrative composite payer operation designed to show how the method works. External benchmarks and case-study results are identified separately.
Define the Prior Authorisation Value Stream
The value stream should begin when a provider submits a request and end when the procedure is approved and available for scheduling.
The process boundary includes:
- Provider request submission through portal, fax or phone.
- Payer intake and case creation.
- Member eligibility and benefit verification.
- Clinical documentation completeness check.
- Nurse review triage.
- Medical director review for complex cases.
- Peer-to-peer discussion when required.
- Determination letter issuance.
- Approval notification and procedure scheduling.
A conventional process map shows sequence. A VSM adds the operational facts that explain delay:
- Cycle time: Active work time at each step.
- Wait time: Time the case sits before the next action.
- First-pass completeness: Percentage of requests that require no additional information.
- Information flow: Portal, fax, phone, EHR, work queue and notification mechanisms.
- Decision logic: The rules that determine whether a case is auto-approved, nurse-reviewed or escalated.
Healthcare VSM guidance recommends distinguishing value-added work from required non-value-added work and pure waste. For example, clinical assessment may add value, while a regulatory documentation step may be necessary even if it does not directly advance the patient toward treatment. Re-entering the same information into multiple systems, however, is avoidable overprocessing.
Current-State Map: Where the Clinical Review Queue Becomes the Bottleneck
Assume a health insurer processes 18,000 prior authorisation requests per month for elective procedures. Its current-state performance is as follows:
| Metric | Current-state result |
|---|---|
| Average determination turnaround | 5.8 calendar days |
| Requests pending additional information | 34% |
| First-pass completeness rate | 66% |
| Straight-through auto-approval rate | 18% |
| Avoidable fax or phone rework | 22% of requests |
| Average cost per transaction | $8.40 |
| Initial appeal rate | 9% of determinations |
| Approval-to-scheduling lag | 4.6 days |
| Nurse review queue | 2.1 days average wait |
| Medical director queue | 1.4 days average wait |
The process appears to have a clinical bottleneck, but the map reveals several upstream causes:
- Faxed requests are manually indexed and may lack a consistent member identifier.
- Phone requests require staff to collect information that an electronic form could make mandatory.
- Eligibility verification is repeated because intake and clinical review teams use different systems.
- Missing documentation is identified after the request has already entered the nurse queue.
- Routine cases wait alongside complex cases requiring medical director judgement.
- Approval letters are issued, but scheduling teams are notified through a separate manual process.
The largest delay is not necessarily the longest task. It is often the time between tasks. A request may require only 32 minutes of active work, yet remain open for nearly six days because it waits in queues or cycles back to the provider.
Virginia Mason Institute describes a prior authorisation improvement in which medication form completion fell from 30 minutes to 2 minutes, while payer response time fell from five days to less than one day. The team achieved this by mapping the real process, standardising information and activating an underused electronic medical record tool, not by asking staff simply to work faster.
Analyse the Data Before Redesigning the Process
The Analyse Phase of DMAIC should convert the map into evidence. Do not begin with a technology solution. First establish which inputs drive delay.
Useful analyses include:
- Pareto analysis of missing information by procedure type and provider.
- Stratification by submission channel: portal, fax and phone.
- Box plots showing turnaround variation by service category.
- ANOVA to test whether mean turnaround differs significantly among submission channels or provider groups.
- Control charts to distinguish common-cause variation from special-cause events.
- Root-cause analysis for rework, queue entry and appeal drivers.
- Time observation sheets for intake, eligibility and documentation review.
For example, an analysis of 2,000 monthly requests may show:
| Cause of delay or rework | Share of affected requests |
|---|---|
| Missing clinical notes or imaging | 31% |
| Incorrect or incomplete member details | 24% |
| Non-standard procedure coding | 18% |
| Fax unreadable or sent to wrong queue | 15% |
| Provider status calls | 12% |
The data may also show that portal submissions have a 2.4-day average turnaround, compared with 7.1 days for fax and 8.3 days for phone-originated cases. That difference is a strong signal that the process is designed around information quality, not merely staffing capacity.
External data provides useful context. KFF’s analysis of 2025 reported metrics found median response times of approximately 0.9 to 1 day for standard requests across Medicare Advantage, Medicaid managed care and ACA Marketplace insurers. It also found standard-request denial rates ranging from 12% to 18%, while 43% to 67% of appealed denials were overturned, depending on market segment. These figures reinforce the importance of measuring first-pass quality and appeal causes rather than relying on an overall average.
Future-State Map: Route by Complexity, Not Arrival Sequence
The future-state design should create a single source of truth for the case and a pull-based flow for clinical work.
A practical future-state pathway is:
- Structured electronic submission: Mandatory member, provider, procedure and clinical fields.
- Real-time eligibility verification: Confirm coverage before clinical work begins.
- Completeness gate: Reject or return incomplete cases before they enter a clinical queue.
- Rules-based routing: Separate standard, low-risk cases from complex or ambiguous cases.
- Straight-through processing: Auto-approve eligible requests when policy criteria and documentation are satisfied.
- Targeted nurse review: Review cases requiring clinical judgement.
- Medical director escalation: Reserve physician review for complex cases, exceptions and peer-to-peer discussions.
- Automated determination letter: Generate consistent communication with reason codes and next steps.
- Scheduling notification: Send the approval number and status directly to the scheduling work queue.
The target state for the illustrative payer could be:
| Metric | Current state | Future-state target |
|---|---|---|
| Average determination turnaround | 5.8 days | 2.0 days |
| Additional-information rate | 34% | 12% |
| First-pass completeness | 66% | 88% |
| Auto-approval rate | 18% | 45% |
| Avoidable fax or phone rework | 22% | 6% |
| Cost per transaction | $8.40 | $4.10 |
| Appeal rate | 9% | 5% |
| Approval-to-scheduling lag | 4.6 days | 1.2 days |
A higher auto-approval rate must be governed carefully. The rules engine should apply approved clinical and administrative criteria, maintain an audit trail and include exception handling. Autonomation, or Jidoka, is relevant here: the system should detect incomplete or inconsistent information and stop or redirect the case rather than allow an error to flow downstream.
Sequence the Kaizen Improvements
A disciplined improvement sequence prevents the organisation from automating a fragmented process.
Kaizen 1: Stabilise intake
- Create one standard request form for each high-volume procedure family.
- Make essential fields mandatory.
- Establish a single member and provider identification standard.
- Publish clear documentation requirements.
- Track first-pass completeness by provider and procedure.
Kaizen 2: Remove avoidable rework
- Replace fax status calls with a case-status dashboard.
- Scan and index legacy fax requests automatically where possible.
- Use standard reason codes for missing information.
- Send one consolidated information request instead of multiple messages.
- Assign ownership for provider follow-up.
Kaizen 3: Design the clinical review pull system
- Create separate queues for routine, expedited and complex cases.
- Set work-in-process limits for nurse and medical director queues.
- Route cases according to complexity, not arrival sequence alone.
- Use daily visual management to expose ageing cases.
- Apply Andon-style alerts for cases approaching service-level thresholds.
Kaizen 4: Enable controlled straight-through processing
- Pilot auto-approval on one high-volume, low-variation procedure.
- Validate approval accuracy through retrospective sampling.
- Monitor denials, appeals and adverse exceptions.
- Expand only when quality and compliance remain stable.
Kaizen 5: Connect approval to scheduling
- Trigger an electronic notification when approval is issued.
- Include the authorisation number, approved service and validity dates.
- Track “approved but not scheduled” cases daily.
- Measure procedure cancellation and rescheduling rates before and after implementation.
Measure What Matters
A strong VSM control plan should include both speed and quality. Focusing only on turnaround can encourage premature approvals or incomplete documentation. Balance the dashboard with:
- Determination turnaround by request type.
- First-pass completeness.
- Additional-information rate.
- Auto-approval rate and accuracy.
- Nurse and medical director queue age.
- Peer-to-peer frequency.
- Appeal rate and appeal overturn rate.
- Cost per transaction.
- Approval-to-scheduling lag.
- Elective procedure cancellation rate.
In the illustrative case, reducing scheduling lag from 4.6 to 1.2 days could release substantial appointment capacity. If 6,000 approved elective procedures are processed monthly, a 3.4-day reduction represents approximately 20,400 procedure-days of earlier scheduling opportunity across the pathway, not necessarily 20,400 additional procedures, but significantly less idle time between approval and booking.
The purpose of VSM is to make this opportunity visible, measurable and governable. It connects provider experience, payer efficiency, clinical quality and patient access in one operating model.
Prior authorisation will always require appropriate clinical judgement for some cases. The improvement opportunity is to ensure that the clinical review queue is reserved for cases that genuinely need it.
Build the capability to map, analyse and redesign complex healthcare processes by pursuing Lean Six Sigma training and professional certification with Lean 6 Sigma Hub, including practical Green Belt training.
Kaizen. Kai-Care. Kai-Done. ( Lean Six Sigma)
