There is not enough money and there are not enough people. What is left is how you run the schedule.
The Arithmetic Does Not Work, and It Is Not Going To
Start with the two numbers that govern this decade. The direct care workforce in the United States is about 5.4 million people. Between 2024 and 2034 it must fill roughly 9.7 million openings, counting both new positions and replacements for people who leave (PHI, citing Bureau of Labor Statistics projections). No sector in the American economy is projected to add more new jobs. None is worse positioned to fill them.
Demand is locked in. The population aged 65 and older grows from 57.8 million in 2022 toward 88.8 million by 2060, and the 85-and-older population nearly triples. The first baby boomers turned 80 this year. Meanwhile, the ratio of working-age adults to each person aged 85 and older collapses from 31-to-1 down to 12-to-1. Every operator is planning to hire through a period in which the pool of potential hires per resident shrinks by roughly 60%.
You are not managing a shortage that ends. You are managing a permanent structural condition, and the operators who understand that are already competing on something other than wage rate.
Turnover Is the Genuine Cost Line, and Nobody Prices It Honestly
Post-acute turnover is not high by the standards of healthcare. It is high by the standards of any industry that has ever been measured. In the CMS Payroll-Based Journal data, nursing home staff-count turnover rose between 2021 and 2023 from 63.1% to 87.7% for RNs, from 42.4% to 64.6% for LPNs, and from 49.1% to 67.1% for certified nursing assistants. Weighted by hours, the landmark Health Affairs analysis put mean nursing turnover at 128%. Home care sits in the same territory, near 79% in 2023. Hospital RN turnover in 2025 was 17.6%.
Read that comparison again because it is the whole argument. A post-acute operator replaces its clinical workforce three to five times as fast as the hospital down the road competing for the same licensed people. And the churn concentrates at the front end: 15% of all nursing care hours in nursing homes are delivered by staff hired within the previous 90 days (JAMA Internal Medicine, 2023), and in home care nearly four of every five caregivers leave within their first 100 days.
The First 90 Days Are Where the Money Leaks
Ask most operators where the turnover problem lives and they will describe a recruiting market. Ask the data and it describes something far narrower, and far more fixable. In home care, 57% of all caregiver turnover occurs within the first 90 days (Activated Insights). In acute hospitals, 22.7% of newly hired RNs leave within their first year, and those first-year departures account for 29% of all RN separations (NSI, 2026). Skilled nursing has no published 30, 60, or 90-day rate at all, because the only official measure, the CMS Payroll-Based Journal metric, is annual and is documented to undercount short-tenure churn. What skilled nursing does have is the finding that 15% of all nursing care hours are delivered by staff hired within the previous 90 days.
Now put a dollar on it. LeadingAge prices the replacement of a single direct-care worker at roughly $4,500 once direct and indirect costs are both counted, and notes that a facility averaging 40 terminations a year spends at least $180,000 simply replacing people. That figure excludes the overtime premium paid to cover the gap, the agency hour bought when overtime runs out, and the survey and star-rating consequences of a building staffed largely by people who arrived last month. For a 12-facility operator, the same arithmetic reaches roughly $2.2 million a year in replacement cost alone.
Exhibit 1. Where the Workforce Actually Leaks
Source: Activated Insights (home care); NSI National Health Care Retention Report 2026 (acute); Shen, McGarry and Gandhi, JAMA Internal Medicine 2023 (skilled nursing); LeadingAge Workforce Cost Calculator.
The takeaway is straightforward: The first 90 days are not just a hiring challenge. They are something operators can actively manage. When someone owns the process, tracks the results, and invests in improving it, retention improves. When nobody does, the costs still show up.
Most operators book this as a recruiting cost. It is not. It is a quality cost, a compliance cost, a census cost, and eventually a valuation cost. It simply arrives on four different lines of the P&L, none of them labeled turnover, which is exactly why it never gets managed as one number.
Wages Are Necessary. Wages Are Not the Answer
Here is the uncomfortable finding. Nursing homes delivered the largest wage increases anywhere in healthcare after the COVID-19 pandemic, roughly 3.4% to 3.6% in 2025, and 90% of providers still report difficulty recruiting (AHCA/NCAL). The academic literature explains why. Higher wages measurably reduce turnover among CNAs but show no comparable effect for RNs or LPNs, and a frequently cited estimate holds that an additional dollar per hour buys roughly 2.1 additional months of aide tenure. That is a real return. It is nowhere near the size of the problem.
If your workforce strategy is a wage strategy, you are in an auction you cannot win. Hospitals, health systems, retail, warehousing, and gig platforms bid for the same labor, and most of them have better margins than you do. 45% of nursing homes operated at a loss in 2024. You cannot buy your way out of this shortage. You need a different currency.
The Agency Trap, and How Operators Walk Into It
When the schedule breaks, operators buy agency hours. The premium is well documented: Agency RNs cost $64.19 per hour against $41.99 for employed staff, and CNAs $33.73 against $19.62 (Innovation in Aging, 2023). Between 2018 and 2022, agency use went from about a quarter of nursing homes and 3% of hours to roughly half of nursing homes and 11% of direct care hours. Usage has since fallen about 44% from its late-2022 peak, which is genuine progress, and it remains structurally embedded in the cost base.
The trap is that agency does not merely cost more. It makes the underlying problem worse. Facilities in the top quartile of agency use showed 7.7 percentage points higher permanent registered nurse turnover and 1.9 points higher aide turnover than their peers. Your own staff watch a traveler earn substantially more for the same shift, with none of the weekend rotation and none of the mandatory overtime. The agency hour you bought to cover a gap is manufacturing the next three gaps.
What Keeps Caregivers, According to the Evidence
The research on post-acute retention is more consistent than most operators realize, and almost none of it points at pay. A study of nurse aide scheduling published in Manufacturing and Service Operations Management found that assigning aides to consistent co-worker teams reduced turnover by 24% at no additional cost and cut operating costs by as much as 7%. Facilities scoring high on leadership quality and staff empowerment were roughly six times as likely to reach the top tier of aide retention (JAMDA). And when home care workers were asked directly, they ranked control, community, and respect ahead of compensation (JAMA Network Open, 2025).
Exhibit 2. The Post-Acute Retention Hierarchy
Source: Ankura analysis of the published retention literature; see Sources.
Nearly every operator competes on the bottom tier. The differentiation lives in the three tiers above it, and each of them is a scheduling and workforce management capability before it is a culture initiative.
Notice what those findings have in common. Consistent teams, predictable hours, self-scheduling, shift-swap, a voice in the roster, transparent access to pay: Every one of them is a workforce management capability before it is a culture initiative. You cannot build consistent co-worker teams on a laminated grid, and you cannot detect that an aide is chronically under-scheduled, and therefore quietly on her way out, unless somebody is looking at hours by person rather than hours by building.
The Scoreboard Is Already Public
Since 2022, CMS has published nurse turnover and weekend staffing data on Care Compare and folded turnover into the Five-Star rating. Your retention performance is no longer an internal metric. It is a marketing asset for your competitors, visible to every discharge planner and every adult child comparing options. One-star facilities carry median turnover of 135.3% against 76.7% at five-star facilities, and each 10-percentage point rise in nursing turnover produces 0.241 additional citations per inspection against a mean of about six (JAMA Internal Medicine, 2023).
Turnover also drives rehospitalization and worse quality measures. In a referral market moving toward narrow networks and value-based arrangements, that chain runs directly from an unfilled shift to a lost referral to an empty bed. The supply picture is tightening at the same time: Roughly 30% of direct care workers are immigrants, rising to 33% in home care against 18% of the workforce overall, and at least one in five comes from a country affected by the current visa pause (KFF, 2026).
Technology Is the Enabler. Adoption Is the Intervention
Everything the evidence recommends requires a system capable of doing it. Consistent teaming requires scheduling logic that respects continuity rather than merely filling holes. Predictable hours require visibility into hours by person across every site. Self-scheduling and shift-swap require a mobile experience a CNA will use on a phone between rounds. Payroll-Based Journal integrity requires a timekeeping layer you can defend to CMS. None of this is achievable on spreadsheets, paper grids, and a payroll system chosen because it was cheap.
But we have watched enough of these programs to be blunt about the other half. Buying the platform changes nothing by itself. The failure mode in post-acute is adoption, not selection. A scheduling engine that a director of nursing overrides every Friday because she does not trust it is worth less than the paper it replaced. Self-scheduling nobody enabled because the regional team feared losing control is a licensing cost with no return. Operators who tell us they own good technology and are not getting value are almost always describing a change management gap, not a product gap. Adoption must be a funded workstream with an owner, a budget, and measured outcomes.
Where We Would Start
- Measure turnover as one number with a dollar attached, reported to the board at the same cadence as census and margin. Include recruiting, orientation, overtime backfill, agency premium, and the quality and survey consequences.
- Invest in the first 100 days. That is where the losses concentrate, and almost no post-acute operator manages it as a distinct stage with its own owner and metrics.
- Rebuild the schedule around continuity rather than coverage. Consistent co-worker teams are the highest-return, lowest-cost intervention in the published literature.
- Give the workforce control it can see, through self-scheduling, shift-swap, transparent open-shift access, and predictable hours. Control, community, and respect outrank pay in what caregivers say keeps them.
- Fund adoption and change management explicitly, at the building and regional level, and hold it to outcomes: agency hours, overtime hours, first-100-day retention, and CMS turnover as published.
An Honest Conversation First
Ankura does not sell workforce management software and is not a systems integrator, which means we have nothing to defend when we tell an operator the technology is not the binding constraint. Our work is labor strategy and adoption: designing the staffing and scheduling model the evidence supports, then doing the change management that gets it used by charge nurses, schedulers, and administrators who have every reason to be skeptical of the last three initiatives. If your turnover is above the sector median, if agency is still in your run rate, or if you already own a workforce management platform that has not moved a single labor metric, start with 30 minutes with one of our post-acute experts. No product, no deck, no obligation. When the conversation points toward a modern workforce management platform as the enabling layer, we will say so directly, and our post-acute team can take the technology discussion from there. What we will not tell you is that the software alone will fix it. It will not, and you have probably already paid to learn that once. Reach out to Christine Ishak.
Sources
PHI, Direct Care Workforce Key Facts and Policy Priorities (2025), citing U.S. Census Bureau and BLS projections; Gandhi, Yu and Grabowski, Health Affairs (2021); Pradhan et al., Healthcare (2025); Shen, McGarry and Gandhi, JAMA Internal Medicine (2023); Activated Insights home care benchmarking via HCAOA (2024); NSI National Health Care Retention Report (2026); AHCA/NCAL Long-Term Care Workforce Report (2026) and State of the Sector (2024); Bowblis and Grabowski et al., and Sharma and Xu, Innovation in Aging (2023, 2022); Mayo et al., Manufacturing and Service Operations Management (2023); Berridge et al., JAMDA (2020); Gusoff et al., JAMA Network Open (2025); CMS Care Compare staffing and turnover data; KFF direct care workforce analysis (2026).
© Copyright 2026. The views expressed herein are those of the author(s) and not necessarily the views of Ankura Consulting Group, LLC, its management, its subsidiaries, its affiliates, or its other professionals. Ankura is not a law firm and cannot provide legal advice.
