A referral management implementation can reach production in 90 days when the organization sequences it correctly: measure the baseline and stand up governance in weeks 1–2, integrate and configure in weeks 3–6, run a controlled pilot in weeks 7–10, and scale with training and reporting in weeks 11–13. Implementations that fail almost always skipped the first two weeks.
The failure mode is consistent. Software goes live, adoption is partial, nobody can prove whether performance changed because there is no defensible baseline, and the initiative loses executive support in the second quarter.
Before Week 1: Three Prerequisites
An executive sponsor with budget authority. Referral work crosses operations, IT, contracting, and clinical leadership. A sponsor who can resolve cross-functional disputes in days rather than weeks is required.
A named physician champion. Adoption is a clinical behavior change. It needs a clinician who will present it to peers as a professional improvement rather than an administrative mandate.
Agreement on the problem. Leakage, access, administrative cost, and closed-loop compliance are related but call for different first moves. Pick the primary objective and write it down.
Weeks 1–2: Baseline and Governance
This is the phase organizations most want to skip and most regret skipping.
Establish the baseline. Pull twelve months of referral orders and, critically, claims or HIE outcome data. Calculate leakage on completed encounters, median and 90th-percentile time to appointment, completion rate, closed-loop rate, and out-of-network spend — segmented by specialty and site. Freeze the definitions in a written data dictionary. Without this, no post-implementation claim of improvement is credible. The metric set is detailed in the 12 referral KPIs every healthcare executive should track.
Stand up governance. Charter the referral management committee, publish decision rights, and name the four owner roles. See referral governance for CINs and health systems.
Run the adequacy gap analysis. Demand by subspecialty and geography against effective supply. This determines whether the first-quarter constraint is routing or contracting.
Map the current workflow. Sit with referral coordinators for a full day. The documented process and the actual process differ, and configuration must fit the actual one.
Deliverables: baseline metrics report, data dictionary, committee charter with decision rights, adequacy gap list, current-state workflow map.
Weeks 3–6: Integration and Configuration
Establish EHR integration. Referral management that lives outside the EHR does not get used; the ordering clinician will not leave the chart. Integration should place matching and ordering inside the existing workflow. ReferralPoint maintains direct marketplace integrations with athena, eCW, NextGen, Veradigm, and Epic, with broader EHR coverage through Lightbeam Health Solutions — see integration and security.
Connect data feeds. Claims, HIE event feeds, eligibility, and the specialist master record. Outcome measurement depends on these, not on the EHR alone.
Build the specialist master. Consolidate registry, contract, claims, and attestation sources into a single record per clinician per location, with network participation keyed to plan and product.
Configure ranking criteria. Apply the weights the committee approved. Do not let this be an IT default; the criteria are a governance artifact.
Configure prior authorization rules. Payer-specific requirements, required documentation, and submission paths for the highest-volume specialties first.
Design patient outreach. Channels, languages, message content, cadence, escalation to a human. Test message copy with actual patients before launch.
Deliverables: working EHR integration in a test environment, connected data feeds, validated specialist master, configured ranking and PA rules, approved outreach design.
Weeks 7–10: Controlled Pilot
Scope it deliberately. Two to four specialties with high volume and clear leakage, at two to three sites, with clinicians who volunteered. Do not pilot with the most resistant group to prove a point, and do not pilot with the easiest group and then generalize.
Run the parallel path. Keep the legacy process available. Forcing exclusive use before the workflow is proven generates workarounds that are hard to unwind.
Instrument everything. Adoption rate, recommendation acceptance rate, override reasons, time to appointment, PA turnaround, and coordinator time per referral.
Hold weekly feedback sessions. Coordinators and pilot clinicians, thirty minutes, with a visible fix list. Speed of response during the pilot sets the tone for the rollout.
Fix data problems immediately. Every pilot surfaces directory defects — wrong numbers, stale panel status, missing subspecialty detail. Corrections need a one-business-day turnaround, or clinicians conclude the data is unreliable and stop trusting the recommendations. See provider directory accuracy.
Gate the rollout. Define exit criteria in advance: adoption above a set threshold among pilot users, measurable movement in time to appointment, override rate explainable, no unresolved critical defects.
Deliverables: pilot performance report against baseline, resolved defect log, refined configuration, documented workflow, go/no-go decision.
Weeks 11–13: Scale
Roll out in waves, not all at once. Group by site or specialty. Each wave should have a named local owner and a two-week stabilization window before the next begins.
Train by role, not by system. Coordinators, schedulers, PA staff, and ordering clinicians need different fifteen-minute sessions focused on what changes in their day — not a platform tour.
Recruit clinician-to-clinician advocacy. Pilot participants presenting to peers outperform any training deck.
Launch executive reporting. Same metrics, same definitions as the baseline, with delta reporting against it. Auto 360° VISIBILITY is designed around this comparison.
Establish the operating rhythm. Weekly working group, monthly committee, quarterly board review. The implementation ends; the operating model continues.
Deliverables: all waves live, role-based training complete, executive dashboard in production, standing governance cadence.
The 90-Day View
| Phase | Weeks | Primary output | Gate to proceed |
|---|---|---|---|
| Baseline and governance | 1–2 | Frozen baseline, chartered committee | Sponsor sign-off on definitions |
| Integration and configuration | 3–6 | Working integration, configured rules | Successful test-environment referral |
| Controlled pilot | 7–10 | Measured pilot results | Exit criteria met |
| Scale | 11–13 | Full rollout, live reporting | Each wave stabilized |
The Five Ways Implementations Fail
- No credible baseline. Improvement cannot be demonstrated, so funding lapses.
- Bolt-on workflow. A separate portal outside the EHR produces low adoption regardless of feature quality. This trade-off is examined in EHR referral modules versus dedicated referral platforms.
- Dirty specialist data at go-live. One bad recommendation costs more trust than ten good ones earn.
- Technology without governance. Configuration drifts from policy and nobody owns the metrics.
- Big-bang rollout. Every defect surfaces simultaneously across every site, and the support model collapses.
What Happens After Day 90
Ninety days delivers a functioning system and a first measurable delta. Sustained performance comes from the operating rhythm: quarterly adequacy analysis, quarterly directory accuracy audits, monthly override review, and continuous expansion into additional specialties, prior authorization payers, and post-acute workflows. Organizations preferring to outsource part of this operating load can staff it through Referral Staff Outsource.
Key Takeaways
- Sequence matters more than speed: baseline and governance in weeks 1–2, integration in 3–6, pilot in 7–10, scale in 11–13.
- A frozen baseline with a written data dictionary is the prerequisite for proving any improvement.
- Integration must be inside the EHR; separate portals fail on adoption regardless of features.
- Pilot with volunteers in high-volume specialties, keep a parallel path, and gate the rollout on written exit criteria.
- Fix directory defects within one business day during the pilot or clinicians will stop trusting recommendations.
- Roll out in waves with local owners and a stabilization window between them.
Frequently Asked Questions
Q: How long does a referral management implementation take? A: A focused implementation covering baseline measurement, EHR integration, configuration, a controlled pilot, and a waved rollout can complete in about 90 days. Timelines extend when EHR integration requires custom work, when claims data access is not already established, or when governance has to be created from scratch.
Q: What should happen before any software is configured? A: Establish a twelve-month baseline using claims-based outcome data with frozen metric definitions, charter the governance committee and name owners, run an adequacy gap analysis, and map the actual current workflow by observing referral coordinators for a full day.
Q: Why is a baseline so important? A: Without a pre-implementation baseline calculated on definitions that do not change, post-implementation results cannot be attributed to the initiative. This is the most common reason otherwise successful implementations lose executive support in the second or third quarter.
Q: Should we pilot before rolling out organization-wide? A: Yes. Pilot two to four high-volume specialties at two to three sites with volunteer clinicians, keep the legacy path available, instrument adoption and outcomes, and define written exit criteria before starting. A big-bang rollout surfaces every defect simultaneously and overwhelms the support model.
Q: What is the biggest risk to adoption? A: Two risks dominate. First, a workflow that lives outside the EHR, which clinicians will not adopt regardless of feature quality. Second, poor specialist data at go-live, because a single bad recommendation erodes more trust than many good ones build.
Q: What resources does the organization need to commit? A: An executive sponsor with budget authority, a physician champion, a referral operations lead, an analytics resource for baseline and ongoing reporting, an IT or informatics resource for integration, and coordinator time for workflow mapping and pilot feedback. Governance participation from contracting and compliance is periodic rather than full-time.
A 90-day plan only works if week one starts with measurement. Request an implementation scoping session.
