PDGM Analytics

PDGM, effective January 2020, pays each 30-day period based on a case-mix group built from admission source (community or institutional), timing (early for the first 30-day period, late for subsequent periods), one of 12 clinical groupings from the primary diagnosis, a functional impairment level (low, medium, or high, derived from OASIS items), and a comorbidity adjustment (none, low, or high) from secondary diagnoses. That is five dimensions, 432 combinations, recalibrated by CMS annually, most recently in the CY2026 final rule using CY2024 data. No manager can hold that in their head, so agencies need reporting that shows how their intake mix, coding, OASIS accuracy, and visit patterns translate into revenue and margin.

A focused PDGM dashboard covers a handful of measures:

• Case-mix weight trends, overall and by clinical grouping, referral source, and team
• LUPA rate and periods currently trending toward the threshold, which ranges from 2 to 6 visits depending on the HIPPS group
• Visit utilization per period by discipline against the plan of care
• Admission source and timing mix, since institutional and early periods pay differently than community and late
• Comorbidity adjustment capture, the share of periods with low or high adjustments
• Gross margin per period, combining payment with cost per visit

The value shows up when metrics change behavior. A falling case-mix weight in one clinical grouping usually points to coding or OASIS functional scoring drift, which is a QA education fix. A LUPA rate spike concentrated in one team is often a scheduling or missed-visit problem, catchable mid-period if the report runs daily rather than monthly. A margin gap between two referral sources with similar patients may reflect authorization friction or documentation quality at intake. The discipline is to route each metric to an owner: coding trends to the QA lead, LUPA-risk periods to schedulers the same day, and mix and margin views to leadership monthly.

Analytics should make the agency accurate, not aggressive. Chasing higher functional impairment levels through OASIS scoring, padding secondary diagnoses to reach a comorbidity adjustment, or adding clinically unnecessary visits to clear a LUPA threshold are the patterns that draw Targeted Probe and Educate reviews, UPIC audits, and False Claims Act exposure. The defensible use of the same data is symmetrical: find undercoded charts and unscored functional deficits that documentation supports, and find missed visits that left ordered care undelivered. A useful internal test is whether the correction would survive an auditor reading the chart, because eventually one will.