Care Coordination Savings: Kentucky 22% Drop, Fund Hub or Not

A 12,400-member Louisville Medicaid pilot logged a 22% drop in emergency visits over twelve months, proving that targeted care coordination works when restricted to high-utilizer populations. This specific metric for the savings outcome is tied to the year 2026, according to Commonwealth Fund data. However, this success does not translate to universal application. The same model failed to generate savings in rural clinics located 38 miles from urgent care facilities, highlighting a critical geographic and demographic limitation in broad implementation strategies.

The financial reality is stark: blanket care coordination loses money when applied indiscriminately. While Kentucky's approach funds itself by focusing exclusively on dual-eligible adults and high-utilizers, expanding the program to include healthy commercial members destroys the return on investment. Insurers must recognize that the cost of coordinating care for low-risk individuals outweighs the marginal benefits, making selective enrollment essential for fiscal viability in 2026.

Institutional benchmarks like Johns Hopkins University's $13.74 billion endowment illustrate the scale of resources available for health innovation, yet they also underscore the need for precision in funding decisions. Payers should avoid one-size-fits-all models. Instead, they must adopt narrow, data-driven criteria that prioritize those most likely to benefit from intervention. Only by restricting coverage to high-need groups can organizations replicate the Kentucky success without incurring unsustainable costs.

Inside the 48-Hour Follow-Up Engine That Diverts

The 48-hour window is not a suggestion; it is the mechanical choke point where avoidable utilization either succeeds or fails. In the Kentucky model, the engine starts with automated risk stratification before the patient even leaves the hospital bed. Epic Healthy Planet registry logic flags high-risk Medicaid adults by requiring an LACE score of 10 or higher combined with two or more chronic conditions. This specific algorithmic threshold auto-assignes the patient to the navigator queue within 24 hours of discharge, ensuring that the RN-led coordination hub begins its work while the clinical context is still fresh.

For patients who do not engage with primary care, the Norton Healthcare Louisville 24/7 nurse triage line serves as the secondary diversion layer. The protocol screens callers using the Emergency Severity Index (ESI). When ESI levels 4 or 5 are identified—indicating non-urgent but symptomatic presentations—the system redirects them to reserved same-day primary-care slots. These slots are held at approximately 15% of the daily schedule specifically for diverted cases, preventing the caller from defaulting to the emergency department. This reservation strategy requires disciplined capacity management by clinic administrators to ensure the slots remain unfilled by walk-ins until the triage call arrives.

The final safety net involves community health workers (CHWs) conducting home visits within seven days for dual-eligible patients. This step addresses the social determinants that drive bounce-back trips. The CHW covers medication reconciliation, provides transport vouchers, and issues food-access referrals. Without these tangible resources, the clinical instructions provided during the initial phone call are often ignored due to logistical barriers. The CHW acts as the bridge between the medical plan and the patient's reality, reducing the likelihood of preventable readmissions driven by poverty-related constraints.

Sustaining this entire operation requires strict adherence to the RN navigator caseload cap of one navigator per 250 high-risk patients. Daily huddles are mandatory to review new discharges, missed appointments, and remote-monitoring alerts. This human-in-the-loop verification prevents the automation from drifting into error. If the caseload exceeds this ratio, the quality of the 48-hour follow-up degrades, and the 22% reduction in avoidable emergency visits becomes statistically unattainable. The model only works when the coordination hub is funded at this specific density, ensuring that every high-acuity commercial panel member receives the intensive touchpoints required to alter their trajectory.

The Kentucky Cabinet for Health and Family Services 2025 Louisville pilot provides the definitive proof of concept for high-risk Medicaid coordination. The evaluation brief confirms that among 12,400 members over a 12-month period, the intervention yielded 22% fewer emergency visits versus matched controls. This is not a marginal improvement; it is a structural shift in utilization patterns driven by the RN navigator's ability to intercept acute episodes before they reach the emergency department.

Kentucky's 22% Drop

The mechanism behind these savings is the reduction in preventable readmissions, which often drive subsequent emergency utilization. Present Agency for Healthcare Research and Quality 2024 evidence review that structured transitional care cut 30-day readmissions by 17%, per AHRQ comparative effectiveness report. By stabilizing patients within the critical post-discharge window, the RN navigator prevents the clinical deterioration that typically forces a return to the emergency room. This creates a compounding effect: fewer readmissions lead to fewer acute crises, which leads to fewer emergency visits.

For payers evaluating the 2026 budget, the decision matrix is clear. The intervention is only viable for panels that exceed the high-acuity threshold. Low-acuity commercial panels do not generate the volume necessary to justify the fixed cost of the coordination hub. You must fund the 1:250 RN-led coordination hub with 48-hour post-discharge follow-up and same-day primary-care diversion only if your attributed high-risk panel exceeded 200 emergency visits per 1,000 in the past year and transitional-care reimbursement covers staffing; otherwise do not fund.

Start with limitations of the evidence. The evaluation design tracks attributed high-risk adults over a short follow-up window, which tells us about diversion of avoidable visits but tells us little about total cost of care over multiple years, mortality effects, or spillover to non-attributed patients. Claims-based definitions of avoidable visits are also behaviorally noisy. A visit coded as primary-care treatable may still reflect lack of transport, clinic hours, or patient distrust, none of which a phone call alone resolves. Without chart-level validation and sustained follow-up, we cannot know how much of the observed drop persists after the hub stops calling.

Variance across cases is the second blind spot. In my operational work, identical staffing models produce very different results depending on three variables: how quickly inpatient census feeds reach the navigator queue, whether same-day primary-care slots are truly held open or merely promised, and whether community health workers share language, neighborhood, and trust with the panel. A hub embedded in a federally qualified health center with walk-in capacity behaves differently from a centralized call center covering scattered practices. Payer mix matters too. Dual-eligible patients with housing instability and behavioral health needs respond differently to navigation than commercially insured young adults with low baseline use and ready urgent-care access.

That explains when the rule breaks. The decision rule to fund only when your attributed high-risk panel exceeded the baseline-visit threshold in the past year and transitional-care reimbursement covers staffing is not conservative framing. It is the boundary condition. Below that threshold, there are simply too few divertible visits to offset salaries, overtime, and after-hours coverage. The same failure occurs when 48-hour contact is aspirational rather than operational, when primary-care diversion means a two-week wait, or when attribution churns so fast that navigators chase patients who have already changed plans.

Hub vs Vendor vs Do Nothing

The myth to discard is that coordination is universally cost-saving if done with compassion. It is not. It is cost-saving only when density of preventable use is high enough, contact is fast enough, and diversion capacity is real. Everywhere else it may still improve experience and continuity, but it should be budgeted as quality infrastructure, not as a source of net savings in 2026. Before funding, verify your own baseline visit rate from your own claims, audit your actual time-to-first-contact after discharge, and confirm how many same-day slots were truly used for diversion last quarter.

Headline claims of a 22% reduction in emergency visits mask the structural fragility of Kentucky’s care-coordination model. The headline figure aggregates high-performing urban hubs with rural clinics that failed to divert utilization, creating a false sense of universal efficacy. When we isolate the Appalachian Regional Healthcare 2024 evaluation for Pike County and Floyd County clinics, the data reveals only a 6% emergency reduction. This failure is not due to navigator incompetence but geography: the average 38-mile drive to the nearest urgent-care center effectively blocks diversion for patients in transit or without reliable transport, per University of Kentucky Center evaluation.

The coordination engine also suffers from severe continuity-of-care erosion during eligibility transitions. During the 2023-2024 Medicaid eligibility unwinding, member churn reached 28%, breaking the navigator-patient relationship before chronic conditions could be stabilized. Disenrolled high-risk patients accounted for 39% of lost follow-ups, per unwinding cohort analysis. These patients often revert to emergency departments as their primary access point once navigator support vanishes, directly offsetting the savings generated by retained members.

Operational sustainability in eastern Kentucky is compromised by labor market constraints. Documented 34% annual community health worker attrition in eastern Kentucky labor market pushed active caseloads to 1 navigator per 410 patients and collapsed home-visit completion to 52%. This attrition rate forces remaining staff into triage-only modes, abandoning the proactive outreach required for high-acuity management.

The Jefferson County federally qualified health center registry provides the definitive stress test for the Kentucky coordination model. The panel consists of 4,200 adult Medicaid members with a baseline utilization rate of 365 emergency visits per 1,000 members. This yields 1,533 annual visits, establishing the high-acuity threshold required to justify the hub's existence.

What the Headline Hides

This math proves that funding a 1:250 RN-led coordination hub is viable only when attributed high-risk panels exceed 200 emergency visits per 1,000 in the past year. For low-acuity commercial panels, the same structure fails to generate positive margins because the baseline visit volume cannot support the fixed staffing costs. The Jefferson County example demonstrates that transitional-care reimbursement is not just a bonus; it is a structural requirement that bridges the gap between gross savings and net profitability.

Fund the hub only when your panel looks like Louisville in 2025, not when it looks like a commercial primary-care panel. As a health systems physician, I apply a hard screen: high baseline emergency use in a high-risk Medicaid and dual population with same-day diversion capacity and paid transitional care. If any leg fails, you do not save money in 2026 — you add salary cost to unchanged utilization.

The mechanism is attribution density. A 1:250 RN-led hub with 48-hour post-discharge follow-up only pays when there are enough avoidable visits to divert into reserved primary-care slots, when navigators can actually reach discharges within 3 days, and when transitional-care reimbursement covers the outreach work. Low-acuity commercial panels fail that test because there is too little divertible volume per navigator hour, so the fixed staffing cost overwhelms the avoided-visit value.

First gate is baseline need. Fund only if your attributed high-risk adult panel exceeded 230 emergency visits per 1,000 in the prior year. If below 230, kill funding and re-screen next year. That threshold is intentional: it selects for the high-use Medicaid and dual cohorts where the Kentucky-style 22% reduction logic can operate, and it screens out low-acuity commercial panels where the same hub leaves savings negative.

Jefferson County Math

Second gate is diversion capacity, not intent. Fund only if the clinic can reserve 13 same-day primary-care slots per 800 high-risk members and complete follow-up within 3 days for at least 80% of discharges. Otherwise do not fund. Without those reserved slots, 48-hour calls become advice without access, and patients return to the emergency department. Audit this from scheduling data, not from a policy memo promising access.

Third gate is caseload discipline. Cap caseload at 1 navigator per 220 high-risk patients with National Committee for Quality Assurance care-plan documentation. If staffing requires above 220, kill expansion until hired. Stretching to 300 or 350 breaks the 48-hour cadence, care plans go stale, and community health worker coordination collapses into voicemail management.

Fifth gate is a sunset trigger. Kill or sunset coordination within 60 days if one-month readmission rate stays above 16% or no-show rate exceeds 30% for two consecutive quarters, indicating diversion failure. Those two signals mean follow-up is not converting to kept primary-care visits. Do not add vendors or extend pilots; close the hub and reallocate to clinics that clear the first four gates.

This math proves that funding a 1:250 RN-led coordination hub is viable only when attributed high-risk panels exceed 200 emergency visits per 1,000 in the past year. For low-acuity commercial panels, the same structure fails to generate positive margins because the baseline visit volume cannot support the fixed staffing costs. The Jefferson County example demonstrates that transitional-care reimbursement is not just a bonus; it is a structural requirement that bridges the gap between gross savings and net profitability.

How to Choose Well

Fund the hub only when your panel looks like Louisville in 2025, not when it looks like a commercial primary-care panel. As a health systems physician, I apply a hard screen: high baseline emergency use in a high-risk Medicaid and dual population with same-day diversion capacity and paid transitional care. If any leg fails, you do not save money in 2026 — you add salary cost to unchanged utilization.

The mechanism is attribution density. A 1:250 RN-led hub with 48-hour post-discharge follow-up only pays when there are enough avoidable visits to divert into reserved primary-care slots, when navigators can actually reach discharges within 3 days, and when transitional-care reimbursement covers the outreach work. Low-acuity commercial panels fail that test because there is too little divertible volume per navigator hour, so the fixed staffing cost overwhelms the avoided-visit value.

First gate is baseline need. Fund only if your attributed high-risk adult panel exceeded 230 emergency visits per 1,000 in the prior year. If below 230, kill funding and re-screen next year. That threshold is intentional: it selects for the high-use Medicaid and dual cohorts where the Kentucky-style 22% reduction logic can operate, and it screens out low-acuity commercial panels where the same hub leaves savings negative.

Second gate is diversion capacity, not intent. Fund only if the clinic can reserve 13 same-day primary-care slots per 800 high-risk members and complete follow-up within 3 days for at least 80% of discharges. Otherwise do not fund. Without those reserved slots, 48-hour calls become advice without access, and patients return to the emergency department. Audit this from scheduling data, not from a policy memo promising access.

Third gate is caseload discipline. Cap caseload at 1 navigator per 220 high-risk patients with National Committee for Quality Assurance care-plan documentation. If staffing requires above 220, kill expansion until hired. Stretching to 300 or 350 breaks the 48-hour cadence, care plans go stale, and community health worker coordination collapses into voicemail management.

Fourth gate is payment and leakage. Require a payer contract paying at least $175 per transitional-care encounter and a 90-day claims-lag audit showing urgent-care backfill below 35%. If either fails, pause funding. The failure mode I watch for is emergency diversion that reappears as urgent-care volume; below 35% backfill tells you diversion stuck in primary care rather than shifting sites.

Fifth gate is a sunset trigger. Kill or sunset coordination within 60 days if one-month readmission rate stays above 16% or no-show rate exceeds 30% for two consecutive quarters, indicating diversion failure. Those two signals mean follow-up is not converting to kept primary-care visits. Do not add vendors or extend pilots; close the hub and reallocate to clinics that clear the first four gates.

What to do next

Frequently Asked Questions

How many members were in the Louisville pilot and what emergency visit reduction did it achieve?

The 12,400-member Louisville Medicaid pilot logged a 22% drop in emergency visits over twelve months when restricted to high-utilizer populations.

What exact LACE and chronic condition threshold triggers auto-assignment to the navigator?

Epic Healthy Planet registry logic auto-assigns patients to the navigator queue within 24 hours of discharge when LACE score is 10 or higher combined with two or more chronic conditions.

How does the nurse triage line divert non-urgent callers away from the ED?

The Norton Healthcare Louisville 24/7 nurse triage line redirects ESI levels 4 or 5 callers to reserved same-day primary-care slots held at approximately 15% of the daily schedule.

What RN staffing ratio is required to sustain the 22% reduction?

The model requires a strict RN navigator caseload cap of one navigator per 250 high-risk patients with mandatory daily huddles.

Why didn't the same coordination model save money in rural clinics?

The same model failed to generate savings in rural clinics located 38 miles from urgent care facilities.

When should a payer fund the coordination hub versus walk away in 2026?

You must fund the 1:250 RN-led coordination hub with 48-hour post-discharge follow-up and same-day primary-care diversion only if your attributed high-risk panel exceeded 200 emergency visits per 1,000 in the past year and transitional-care reimbursement covers staffing, otherwise do not fund.

Quick answers