Medicare Home Health Research: Case Mix Project Findings for Feb.

Below is a portion of the information HCFA’s Office of Strategic Planning provided to home care groups Feb. 10, 1999:

I. Rationale for case mix research

• Problem: Under home health PPS, need to balance cost control incentive to treat individual patients appropriately (e.g., accept cases with high care needs)

Solution: Develop case mix adjuster for PPS rates

Approach: define patient groups with differing expected care needs, as measured by patient assessments

II. Data resources for research

• Previously poor data on patient characteristics and little data on resource utilization

Case mix project collects improved data from a representative sample of home health agencies: patient assessments (OASIS); resources used in care (time spent on visits)


• Demographics and patient history

Living arrangements

Supportive assistance

Health domains (e.g., sensory, respiratory)



Equipment management

Emergent care

IV. Case mix performance criteria

• Ability of case mix groups to explain variation in resource use (R2)examples elsewhere: Hospital PPS (DRGs): R2=26%-39%

SNF PPS: R2=56%

Reasonable for HH PPS: R2=20%-40%, in view of caregiver role and other factors

Clinical coherence

V. Components of CM system

• Clinical factors

Functional status factors

Utilization factors

VI. Selecting factors for the components from OASIS+

• Incentive effects

Good predictor

Reliability, subjectivity


Administrative considerations

Clinician feedback

VII. Clinical component

• Presence/absence of specific conditions:

pain (M0420)

ulcers (staged) (M0460, M0468)

wounds (M0488)

dyspnea (M0490)

incontinence (urinary or bowel) (M0530, M0540)

ostomy (M0250)

PEN nutrition (M0250)

selected diagnoses and impairments (M0230, M0390, M0610)

VIII. Functional status component

• 7 ADLs (all current):

grooming (M064)

dressing (upper, lower body) (M0650, M0660)

bathing (M0670)

toileting (M0680)

transferring (M0690)

locomotion (M0700)

IX. Utilization component

• 2 elements:

pre-admission location in preceding 14 days: hospital; nursing or rehab facility

amount of therapy (i.e., physical, occupational, speech therapy) at least 8 hours

X. Combining variables within components

• Score for each condition based on impact of condition on mean resource cost

Scores summed within a component and split into groups (e.g., high. moderate, low, minimal)

XI. Examples: scoring functional component

• Grooming:

If not independent in grooming, add 1 to score


If requires minimal assistance or uses assistive device, add 1 to score

If unable to transfer or bedfast, add 6 to score

XII. Combining components in the case mix model

• Tree structure defined by combination of levels on each component

Final case mix classification determined at end of the payment period (depending on amount of therapy used)

XIII. Results

• Relatively simple decision tree

80 groups

Explanatory power (R2) approximately 33%

8+ hours of therapy use a strong contributor

Source: Health Care Financing Administration