Five-Star Staffing Ratings: Connect PBJ Hours, Census, Case Mix and Turnover

Follow the calculation across two source systems

The CMS Five-Star Technical Users' Guide, pages 8–17, uses six staffing measures: adjusted total nursing hours, adjusted RN hours, adjusted weekend total nursing hours, total nurse turnover, RN turnover and administrator departures. The first three cover a quarter. The turnover measures describe annual experience but require six consecutive quarters of PBJ data.

A payroll reconciliation alone cannot reproduce the result. CMS divides included PBJ hours by resident days derived from MDS assessments, then adjusts staffing levels for case mix. Missing or incorrect MDS discharge information can change the census even when the hours were transmitted correctly. Compare the MDS census and resident-identity history with the operational census before blaming the payroll export.

Distinguish reported and adjusted hours

For the quarter, CMS aggregates eligible hours and resident census across days with at least one resident. Weekend measures use Saturday and Sunday hours and census. The guide specifies included job codes and excludes categories such as hospice staff and feeding assistants from these nurse-staffing calculations.

The adjustment formula is:

Adjusted hours = reported hours ÷ facility case-mix hours × national average case-mix hours.

CMS uses the facility's daily PDPM nursing case-mix distribution to construct case-mix hours. The national comparison inputs change by quarter. These are rating inputs, not a resident-specific staffing prescription or the Medicare payment calculation.

In a fictional reconciliation, 32,400 included nursing hours divided by 9,000 MDS-derived resident days gives 3.60 reported hours per resident day. If the applicable facility case-mix hours are 4.00 and the national average case-mix hours are 3.80, adjusted hours equal 3.60 ÷ 4.00 × 3.80 = 3.42. The dashboard's 3.60 is not necessarily wrong; it is a different measure. Label both values and their inputs.

Rebuild the points, then check exceptions

The following fictional packet starts with measure points already assigned using the applicable CMS cut points. It illustrates aggregation, not a shortcut for converting any raw staffing ratio into points.

Under Table 3 of the current guide, 290 points falls in the four-star range of 255–319, before considering exceptions. For example, four or more days with zero reported RN hours when residents were present produces a one-star staffing rating. Failure to submit staffing data by the deadline and specified audit failures have their own consequences. A favorable aggregate must not conceal those checks.

Now consider a separate, legitimate missing-measure case: the administrator measure is unavailable because there is no eligible administrator. If the other measures total 260 points out of a possible 350, the guide's rescaling produces 260 × 380 ÷ 350, rounded to 282. By contrast, an administrator measure receiving the minimum 10 points because required data were missing or erroneous is not simply removed and rescaled. With the same other points, the total is 270.

Explain turnover with employment spells

CMS's turnover denominator uses eligible work histories, including at least 120 hours in the specified 90-day eligibility period. A departure is inferred from at least 90 consecutive days with no reported work, with the gap beginning in the measurement period. Returning workers can generate another eligible employment spell. This differs from counting HR termination forms.

Preserve stable employee identifiers through payroll changes and agency submissions. An apparent mass departure on one date may be an identifier discontinuity. The guide contains exclusions and scoring consequences for implausible or incomplete data, so discovering that pattern does not authorize an analyst to quietly replace CMS's denominator with an HR headcount.

Finish with a cause, not just a recalculated star

Classify the difference: hours scope, MDS census, case-mix input, turnover eligibility, point threshold, missing-data treatment or scoring exception. Retain the quarter, six-quarter turnover window and source release for each conclusion.

The buyer's test is whether staffing, clinical census and historical worker identities can be reconciled without overwriting their source histories. The completed packet should explain why 3.60 became 3.42, how 290 points was assembled and whether an exception superseded it. It does not establish that staffing was sufficient for every resident or shift; that requires the separate facility assessment and professional staffing review.