Nursing home scheduling software for long-term care facilities

• Long-term care facilities must submit direct-care staffing information to CMS through the Payroll-Based Journal, including agency and contract staff.Source: CMS PBJ staffing data submission
• PBJ submissions are due by 11:59 PM Eastern on the 45th calendar day after the last day of each federal fiscal quarter.Source: CMS PBJ staffing data submission
• CMS reports a full 12-hour shift paid to work as 11.5 hours after deducting the required 30-minute meal break.Source: CMS PBJ Policy Manual, version 2.7
• CMS nursing job title code 5 is for a registered nurse director of nursing, code 7 is for a direct-care registered nurse, code 9 is for an LPN or LVN, and code 10 is for a certified nurse aide.Source: CMS PBJ Policy Manual, version 2.7
• The surviving federal baseline requires a long-term care facility to use the services of a registered nurse for at least 8 consecutive hours a day, 7 days a week, except when waived, and to designate an RN as a full-time director of nursing.Source: Federal Register, repeal of minimum staffing standards (90 FR 55687)

For a 60 to 180 bed facility, a full schedule can still be wrong. One overnight shift may be assigned to the wrong reporting date. A nurse can be technically available while already committed to another unit. A replacement can fill the hole while creating overtime or violating a facility policy.

That is why nursing-home scheduling software should be judged as an operating control, not as a nicer calendar. It should understand units, roles, credentials, shift boundaries, paid hours, and the rules your facility has chosen to enforce. If it only fills open boxes, you are buying another place to hide errors.

Start with the hours your building must cover. One 24/7 post needs 168 coverage hours each week. A four-crew 12-hour rotation divides that into 42 hours per crew on average. That is arithmetic, not a feature choice. A fixed 3 by 12 model keeps a person at 36 hours and needs a separate relief plan.

Use this calculation when comparing patterns. The overtime result depends on the employee's workweek, state requirements, and facility policy.

Run the same employee through a 40-hour workweek and an 8/80 calculation. In the derived 2-2-3 example, the favorable 40-hour boundary produces 8 overtime hours per 14 days. Under 8/80, the same pattern produces 28 because each 12-hour shift has 4 hours over 8. The 8/80 option is not automatically cheaper for 12s. Confirm the setup with payroll counsel.

Ask a vendor to show how it turns a schedule into reportable data. CMS measures paid-to-work hours by calendar day, not the label on the schedule. Midnight is the cutoff. In the CMS example, an 11 PM to 7 AM shift places 1 hour on one date and 7 hours on the next. A full 12-hour shift paid to work is reported as 11.5 hours after the 30-minute meal-break deduction, whether the break was taken or not.

The software should keep the source record intact while exposing the fields needed for reconciliation. At minimum, the record needs:

Do not accept a PBJ-ready claim because a vendor can export a CSV. Ask whether the export uses paid-to-work hours, splits night shifts by date, and traces a reported line back to its assignment.

A published schedule that looks full can still be short in the places that matter. The three numbers a long-term care scheduler needs on screen are the same three numbers that show up later in reporting, so ask each vendor to display them before publication rather than in a month-end report.

HPRD against the day's real census

Hours per resident day is direct-care nursing hours for a calendar day divided by that day's census. It is not a shift ratio and it should not be computed from a monthly average census. A product that shows HPRD only as a monthly figure has hidden exactly the day you needed to look at, because the schedule can be unchanged while a census swing moves the result by half an hour per resident.

Ask the vendor which census input the number uses, where that input comes from, and whether the figure updates when you edit an assignment. If the answer is a spreadsheet export, you are buying a calendar and a reconciliation job.

Coverage by shift, not only by day

A day can reach an acceptable total while the night is thin. Coverage targets belong at the unit and shift level, with the role attached, because that is the level at which the work either happens or does not. The useful screen is a role-by-shift grid showing target, scheduled, and gap — not a roster with names in boxes.

The CNA, LPN/LVN, and RN mix

Calculate each role first, then the combined total. Any product that only shows one blended number is asking you to manage a symptom.

RN coverage is a requirement, not a preference

The federal baseline that survived the 2026 repeal of the numeric staffing standards still requires an RN for at least 8 consecutive hours a day, 7 days a week, subject to waiver, and a full-time RN director of nursing. A day without that coverage is not a thin day — it is an uncovered required shift, and PBJ is where it becomes visible to CMS. Configure it as a hard constraint that blocks publication, then check that the system can still tell you which specific days failed and why.

Agency reliance, measured rather than felt

Every building knows it uses agency. Few can say which role, which shift, and which trend. Ask the product to show agency hours as a share of direct-care hours by role and by shift over a quarter. A Tuesday-night CNA line that is 60 percent agency is a vacancy you have not posted, not a call-out problem you can schedule around. The same record also has to carry the agency or contract flag through to reporting, or the hours disappear from the staffing picture entirely.

Coverage rules differ by facility. A software demo should show how your rules behave, not how many color choices the calendar has.

• Unit demand: set targets by unit, shift, role, and census input. A building may need a different mix on memory care than on a short-stay wing.
• Hard constraints: block expired credentials, unavailable employees, overlapping assignments, and a shift that would exceed your configured overtime or rest policy.
• Call-out ranking: rank qualified available replacements using the rules you actually use. Include overtime status and keep a record of offers, declines, and acceptance.
• Publish review: show unresolved gaps and conflicts before publication. After publication, preserve who changed the assignment, when it changed, and why.

A warning is not the same as a control. If an expired credential appears as a yellow note while the user can still publish the shift, the system has made the scheduler responsible for remembering one more thing. A hard block is appropriate for a true eligibility rule. A visible warning is better for a preference that a supervisor may intentionally override.

Mandatory overtime, staffing ratios, staffing plans, and overtime pay are separate questions. A product that says it handles nursing compliance should tell you which one it means.

The research behind this guide found 17 confirmed state restrictions on mandatory overtime, but coverage varies sharply. Only 4 states clearly reached nursing homes in the facility-type review. Some laws cover hospitals only. Others protect a nurse who refuses extra hours rather than setting a simple hour cap.

Store the state policy, bargaining terms, on-call rules, rest requirements, and emergency workflow as separate inputs. In several reviewed laws, chronic short staffing is expressly not an emergency exception. Have counsel confirm the policy before turning legal language into an automatic block.

Do not evaluate a product with a clean Monday morning schedule. Use a scenario that forces the system to reveal its assumptions.

1. Create a 7 PM to 7 AM RN shift and assign it to a unit with another role due during the same window.
2. Check the PBJ view. The night shift should split at midnight, with the 11.5-hour treatment after the meal-break deduction.
3. Add an agency worker, an expired credential, and an employee who would cross your overtime threshold. Try to publish each assignment.
4. Call out the RN two hours before the shift. Confirm that the replacement list filters eligibility, then records the offers and final decision.
5. Edit the published assignment. Ask for the audit record and reconciliation export.

Make the vendor show the raw row, not a dashboard screenshot. You are testing whether the system can explain a result when someone audits it later.

Use the same five checks with every vendor during a 30-day pilot. Record the result while the scenario is still on screen so a polished follow-up deck cannot replace what the product actually did.

Keep this scorecard with the purchasing record. It gives operations, payroll, and compliance the same evidence when they compare the pilot.

• A vendor exports scheduled hours as PBJ hours without using paid-to-work time or the 30-minute meal-break deduction.
• An overnight shift stays on its start date instead of splitting at midnight.
• The product markets 8/80 as a shortcut for 12-hour overtime without showing the daily overtime math.
• The system applies a hospital rule to a nursing home without checking the facility type and state policy.
• An expired credential produces a warning but still allows the assignment to publish.
• A call-out workflow forces the person already on duty instead of documenting voluntary coverage efforts.
• A combined HPRD figure looks healthy while the RN row sits below target, and adding aide hours is treated as a fix.
• HPRD is shown only as a monthly average, which hides the high-census day you were trying to find.
• A vendor claims its software will raise your CMS staffing star rating. It cannot: that rating tracks the nursing hours you purchase. What scheduling can genuinely fix is an uncovered required shift and a PBJ data defect.