NovaCare Health System · Healthcare

Clinical Workforce Retention & Wellbeing: Wellbeing Is Capacity

Maturity
L4
Domain
Grow & Keep
Analytics
predictive
Burnout risk index
~40%
Sponsor
Chief People Officer
Confidence
High

The situation

Which clinicians and units are at rising risk of burnout and attrition — early enough to intervene — and which levers (workload, scheduling, fatigue, support, leadership) actually move the risk for which population?

Grow & KeepL4Sponsor · James Whitfield

The recommendation on the table

Stand up leading-indicator burnout and attrition risk — explainable and support-oriented

Intervention before exit, prioritised by capacity-at-risk, and a break in the burnout-turnover loop.

Decision ownerChief People Officer · James Whitfield
MaturityL4
Priorityhigh

Trade-offWellbeing data demands strict privacy and ethical governance, and trust must be earned to keep signals honest.

The evidence

NovaCare's burnout–turnover–agency loop is self-reinforcing: clinician burnout (~40%) drives attrition (nurse ~19%, first-year ~30%), attrition worsens ratios and pushes premium spend, and worse ratios deepen burnout — and today it is seen only in the rear-view mirror. Healthcare's signature project and most distinctive contribution to the portfolio, HC-03 is the first to treat wellbeing and burnout as first-class workforce intelligence: a measured, leading indicator of lost capacity and patient-safety risk. It found ED burnout double the system average (fatigue-driven), experienced critical-care nurses at rising flight risk (few in number, large in capacity), and two units running with near-zero resilience buffer.

Clinical Workforce Retention & Wellbeing

Wellbeing, burnout, fatigue and retention as a leading capacity indicator.

Burnout risk index· clinicians at elevated, driver-attributed burnout risk
~40%+25vs target
Critical
Predicted attrition risk· modelled clinical attrition, capacity-weighted
~19%+12vs target
On watch
Fatigue exposure· clinicians exceeding safe fatigue thresholds
~22%+10vs target
Critical
First-year clinical turnover· attrition among first-year clinicians
~31%+20vs target
Critical
Illustrative preview
Burnout risk: unit × driver
WorkloadFatigueScheduleSupport
ED
3
3
2
2
Med-surg
3
2
2
2
ICU
2
2
1
1
Oncology
2
1
1
2
Ambulatory
1
1
1
1

Key takeawayED burnout risk is double the system average, driven primarily by fatigue exposure.

Clinicians by burnout-risk tier
100%TOTAL
  • Low38%38%
  • Moderate22%22%
  • High40%40%

Key takeaway~40% of clinicians sit at elevated burnout risk — a leading indicator of capacity loss.

Burnout driver attribution (share of high-risk)
012.52537.550Fatigue / overt…Chronic workloadSchedule instab…Support / auton…

Key takeawayFatigue and workload dominate — schedulable, relievable drivers, not pay.

Interactive view is best explored on desktop.

Key findings

ED burnout risk runs at double the system average, driven primarily by fatigue exposure, and experienced critical-care nurses show rising flight risk — few in number but large in capacity-at-risk. Wellbeing forecasts the losses that drive the agency loop.

What we can’t claim

Two units run chronically at the safe-staffing floor with near-zero buffer — so capacity pressure and burnout compound into one loop. The uncomfortable truth is that out-bidding on pay treats the symptom: the drivers are fatigue, workload and instability, which are schedulable, and a workforce kept permanently at the floor will keep burning out no matter the retention spend.

Recommendations

Stand up leading-indicator burnout and attrition risk — explainable and support-oriented

high priority

Intervention before exit, prioritised by capacity-at-risk, and a break in the burnout-turnover loop.

Trade-off

Wellbeing data demands strict privacy and ethical governance, and trust must be earned to keep signals honest.

Make rostering fatigue-aware and build buffer on the lowest-resilience units first

high priority

Lower fatigue-driven burnout and error risk, and units that can absorb shocks without tipping.

Trade-off

Buffer costs capacity up front, and fatigue-aware rules constrain scheduling flexibility.

Analytical framework

How we reached this

Predictive, glass-box wellbeing intelligence — predict and explain burnout, fatigue and attrition risk early enough to intervene, prioritised by capacity-at-risk.

ConfidenceMedium

Methods applied

Burnout/wellbeing driver modellingSurvival/time-to-attrition analysisFatigue exposure analysisResilience composite scoringIntervention-lever analysis

Statistical techniques

Logistic regressionSurvival analysisDriver/feature attributionSegmentationCorrelation

Algorithms

Interpretable logistic/survival models for attritionExplainable, driver-attributed burnout risk scoring

Data sources

Rosters/time-and-attendance (fatigue, overtime)Workload/acuityValidated wellbeing instruments (consented)Attrition historyEngagement signalsLeadership/span data

Outputs generated

Driver-attributed burnout riskCapacity-weighted attrition riskFatigue exposureUnit resilience indexIntervention-lever recommendations

Why this confidence

Wellbeing and burnout rest on softer, partly self-reported and proxy signals than HC-02's demand data; outputs are ranges, framed as risk to be relieved, with explicit ethical and measurement caveats. Deliberately glass-box.

The reasoning

Business context

Healthcare's signature theme, sponsored by the CPO with clinical leadership because retention here is a capacity-and-safety problem wearing an HR label. It owns burnout, wellbeing, fatigue, retention and resilience; HC-02 consumes its wellbeing signal but does not own it.

Expected value

Avoided turnover cost (recruitment, onboarding, lost productivity across thousands of exits), reduced premium backfill, and the safety/quality value of a less fatigued, more stable workforce. The cheapest retention is relieving the drivers, not out-bidding on pay.

Workforce landscape

ED burnout risk is double the system average, driven by fatigue exposure; experienced critical-care nurses show rising flight risk — small in number, large in capacity-at-risk; first-year turnover (~31%) concentrates in two high-acuity units; two units run with near-zero resilience buffer.

The analytics journey

Level 4, predictive, and deliberately glass-box given the sensitivity. Predicts burnout and attrition risk with driver attribution and explicit ethical and measurement caveats; outputs are framed as risk to relieve at the unit/driver level, never verdicts on individuals.

Under the hood

Interpretable logistic and survival models predict attrition (time-to-attrition, capacity-weighted by scarcity); explainable, driver-attributed scoring surfaces why burnout risk is rising; fatigue thresholds (consecutive shifts, overtime, rest gaps) feed scheduling; a resilience composite connects wellbeing to capacity. No opaque ML on people's wellbeing — every score is explainable and informs support, not punishment.

Confidence & evidence

Why you can rely on this

82%
Analysis confidenceHigh

The inconvenient truth

Two units run chronically at the safe-staffing floor with near-zero buffer — so capacity pressure and burnout compound into one loop. The uncomfortable truth is that out-bidding on pay treats the symptom: the drivers are fatigue, workload and instability, which are schedulable, and a workforce kept permanently at the floor will keep burning out no matter the retention spend.

Method

Confidence is a deterministic read of KPI strength, target and benchmark coverage across this project — shown on an illustrative reference dataset, computed the same way it would be on live data.

Take this further

Where this project connects