Health Economics Metrics

Topic 37 of 76 · Health system operational economics

Workforce Retention

Workforce retention economics quantify what staff turnover costs a health system — recruitment, onboarding, vacancy cover — and therefore what software that reduces administrative burnout is worth. Burnout from repetitive administrative data tasks is a primary driver of staff turnover and sickness absence in the NHS.

Why it matters

When a clinician leaves, the trust pays three times: to recruit a replacement (advertising, agency fees, interviews), to onboard them (months of reduced productivity, supervision), and to cover the vacancy meanwhile — typically with agency or locum staff at 2–3× substantive Agenda for Change rates (see avoidable outsourcing costs and hard cash-releasing savings). Because turnover costs are real cash, retention improvements are among the few workforce benefits a finance director can bank. Administrative friction is consistently among the top cited drivers of clinical burnout, which makes it a software-addressable cost.

The math

Cost per leaver = recruitment cost + onboarding/productivity-ramp cost
                + vacancy cover premium × vacancy duration

Annual turnover cost = headcount × turnover rate × cost per leaver

Value of software  = headcount × Δturnover rate × cost per leaver
                   + sickness-absence reduction × cover cost/day

The causal chain has two estimated links — software → burnout/friction, and burnout → turnover — so evidence both (staff surveys pre/post; published burnout-attrition associations) and keep the claimed Δ modest.

Worked example

A trust employs 1,200 nurses; turnover 11%/year. Cost per leaver:

Recruitment ≈ £4,500;  onboarding/ramp ≈ £6,000
Vacancy cover: 4 months × 0.6 WTE covered by agency premium ≈ £8,000
Total ≈ £18,500 per leaver
Baseline turnover cost = 1,200 × 0.11 × 18,500 ≈ £2.44M/year

Documentation-burden software (auto-populated assessments, single sign-on, dictation) plausibly moves turnover 1 percentage point:

Value = 1,200 × 0.01 × 18,500 = £222,000/year cash-relevant

A 1-point claim backed by staff-survey friction scores is credible; a 4-point claim is not. Run the tornado on Δturnover: it dominates everything else in the model.

Software engineering connection

Engineering retention math is identical and worse-documented: replacing a senior engineer costs 6–12 months of loaded salary (recruiting, ramp, lost context), so a 200-person org at 15% attrition burns millions annually on churn. Developer-experience investment (SPACE and DevEx) is the direct analogue of documentation-burden relief for nurses — and should be justified the same way: measured friction scores, a modest claimed effect on attrition, cost per leaver from your own finance data. The health-economics discipline to copy is costing the leaver honestly rather than arguing about whether people "really" leave over tooling.

Pitfalls

  • Attributing all turnover movement to your intervention — labor markets move turnover far more than software does; use control groups or at least sector trend adjustment.
  • Double counting: retention savings and agency-spend savings overlap (vacancy cover is agency spend); reconcile the lines.
  • Ignoring the lag: burnout-driven attrition responds to friction changes over 1–2 years, not the next quarter.

Sources