Private Duty Nursing Disputes Are Won and Lost on One Word: Continuous
Two children each need eight nursing interventions in twenty-four hours. The first has all eight during waking hours. The second has four by day and four across the night. Same daily total — completely different answers. The first may need a twelve-hour day shift; the second may need around-the-clock coverage, because nobody can safely sleep between those night events.
Any model that adds a day's needs into one total has already erased the distinction the whole benefit turns on. Six lessons from building a PDN instrument around it.
1. The federal test is continuity, not acuity. Private duty nursing exists for children who need "more individual and continuous care than is available from a visiting nurse." The question is never just how sick — it's whether the skilled need recurs at intervals short enough that intermittent visits cannot serve it.
2. Windows, not totals. Assess the day window and the night window separately. In each: how often do nurse-level events occur, what's the real interval between them, and what happens in the gap if no one is there? Presence in one window and not the other is a common, correct answer — and a summed model can't produce it.
3. Time-to-harm is the gate people skip. An event every three hours doesn't require presence if a caregiver can be summoned in time. An airway that can close in ninety seconds does — whatever the frequency says. Frequency and time-to-harm are separate routes to the same conclusion, and honest tools test both.
4. Hours of ventilation are not events of ventilation. Continuous therapies — a vent, continuous feeds — occupy time, not moments. We learned this the hard way: counting a ventilator as an "event" made a ventilated child look 3.5% utilized. Measure continuous care in hours occupied, or the sickest children score the lowest.
5. The caregiver is context, not a deduction. Courts have been consistent: a state may consider caregiver capacity; it may not reduce skilled hours because a parent exists, or condition the benefit on an unpaid second caregiver. A parent asleep is not coverage for a ninety-second airway.
6. A second read is a record, not a recomputation. When a reviewer disagrees with the model, the disagreement should be recorded in their name next to the model's figure — never silently blended. The gap between the two numbers is information the family's appeal is entitled to.
We built this as an instrument: one child's day mapped event by event, the interval test and the time-to-harm gate run per window, jurisdiction comparisons, legal integrity flags cited to the cases, and a notice that shows its work. Browser-only — no child's data ever leaves the machine.
The thread through it all: families don't lose these cases because the child needs less care. They lose because the paperwork measured the wrong thing.
If you review PDN requests: does your criteria set ask about the interval between events — or only how many there were?