Peer Benchmarking for Nurse Staffing

Compare workload and staffing before changing the model

Two hospitals can report the same nurse-to-patient ratio while their units face different workloads. Compare the staffing model, including nursing hours, skill mix, acuity, and coverage, with units doing sufficiently similar work. Then examine the difference alongside patient outcomes, workforce indicators, and labor costs. A ratio remains relevant to assignments and applicable requirements, but it cannot explain the full staffing need by itself.

Finance, clinical leaders, staff, and governing bodies need a shared account of what the benchmark measures. A reported 1:5 ratio on a medical-surgical unit leaves several questions unanswered: when it was measured, which staff counted, how admissions and discharges affected work, and whether the figure describes budgeted, scheduled, or actual staffing. Resolve those questions before treating the comparison as evidence for a change.

Use complementary measures of staffing and workload

Nursing hours per patient day (HPPD) adds a period-based measure to a point-in-time ratio. It does not fully capture workload or show whether staffing was adequate on every shift. Define productive versus paid hours and the treatment of education, float staff, sitters, and resource nurses. If a peer includes float coverage and your measure excludes it, reconcile that difference before drawing a conclusion about staffing. Retain shift and assignment information where it is needed to understand variation hidden by the average.

Add information on patient acuity and admissions, discharges, and transfers. These describe work that a census measure alone can miss. Define the throughput measure consistently and compare units with similar functions, including observation activity or care for patients awaiting another setting. A high-turnover unit may require a different staffing arrangement from a unit with the same average census and fewer transitions.

Review skill mix with the patient population, staff competencies, available support, and scope-of-practice requirements. State rules and collective bargaining obligations may also constrain the options. A benchmark can help identify a question about task allocation; it does not establish that one role can safely substitute for another. Clinical leaders should specify which proposed changes are feasible and what support they require.

Define the unit before choosing the peer group

An ICU label does not establish comparability. Match units by function and patient needs, distinguishing intensive care, stepdown, telemetry, general medical-surgical care, and observation as appropriate. Hospital-level averages can provide context, but unit-level decisions require more detail. Document meaningful differences that remain after peer selection so leaders know the limits of the comparison.

Budgeted staffing describes the plan, scheduled staffing describes the intended roster, and worked staffing describes what was delivered. Use the measure relevant to the question and align it with the peer definition. For a comparison of delivered staffing, worked hours are particularly useful. Internally, examine scheduled-to-worked differences, including absences, late changes, overtime, agency staff, and float coverage. Review their causes and effects without assuming that every variance reflects poor management.

Investigate a benchmark difference before changing the model

A unit above a peer range may have an avoidable cost, a different workload, or a measurement mismatch. Translating that position directly into a staffing reduction skips the clinical assessment. Examine the work, support roles, and patient needs first. Include likely effects on bed availability, emergency-department boarding, charge-nurse responsibilities, and other units so a local change does not simply move the burden.

Staffing changes take effect before some outcomes and retention measures can be assessed reliably. Establish immediate process and safety measures as well as longer-term follow-up. A short period without an adverse outcome is not sufficient evidence that a new model is safe, particularly when events are infrequent or the patient mix has changed.

Separate two decisions: whether the model differs meaningfully from comparable units, and whether a proposed alternative can meet local needs safely. A defined pilot may help answer the second question. Set its duration according to the change and the evidence needed, with regular clinical review, clear exception rules, and stop or escalation criteria. An initial 60- to 90-day period may support implementation assessment, but longer follow-up may be necessary for outcomes and retention.

Connect the comparison to outcomes and accountable decisions

The National Database of Nursing Quality Indicators (NDNQI) provides unit-level staffing and nursing-sensitive measures for comparison. AHRQ's review of nursing and patient safety also identifies acuity, patient transitions, skill mix, and available resources as relevant to staffing. Use such evidence alongside the rules applicable to the facility. California's ratio requirements, for example, are legal requirements in their covered settings, not an optional benchmark; state guidance also emphasizes staffing for patient acuity. Requirements from another jurisdiction do not replace local clinical assessment or applicable law.

Assign responsibility for definitions, approval of model changes, and exceptions. Keep the peer inputs, clinical review, assumptions, and reasons for each decision available for later examination. That record supports accountability but does not establish safety or compliance by itself. Systems with limited analytics capacity can begin with a narrower set of high-risk units while maintaining required staffing and oversight elsewhere.

Before using a benchmark to support a model change, document the staffing measure and the treatment of float, sitter, and charge coverage. Select comparable units and examine acuity and patient transitions. Choose patient-safety and workforce indicators relevant to the change, with enough coverage to detect foreseeable problems. Define thresholds for review and an immediate route for safety concerns, rather than wait for a scheduled reporting cycle.

The decision record should explain what was counted, why the comparison is useful, and how the proposed staffing model will be evaluated locally. Bring that account to the board alongside the percentile so leaders can assess the clinical reasoning and the plan for responding if the change does not work as intended.

Peer review can help identify differences in definitions that are easy to miss in a report. Connex convenes clinical executives in Trusted Peer Groups to compare unit-level definitions and staffing models. Ask how peers count hours and support roles, what changed after implementation, and which local conditions limited the result.

Sources

• AHRQ PSNet Nursing and Patient Safety
• Press Ganey NDNQI
• California Department of Public Health Staffing and patient acuity guidance