NSG/577 Week 2: Developing a Performance Measure, sample paper

Hospitalized Within Sixty Days: Writing a Precise Definition of a Home Health Agency's Lead Outcome Measure and the Process Measures That Feed It

[Student Name]

University of Phoenix

NSG/577: Continuous Quality Monitoring and Outcomes Improvement

Week 2 Assignment

[Instructor Name]

[Date]

The agency, its data and all figures are a composite written for a model paper.

Last week the board accepted hospital admission as the lead quality focus, since an admission is hard on patients, visible to referral sources and shaped by nearly everything the agency does. This paper turns that focus into a measure. A measure that different people calculate differently cannot guide improvement, so the definition matters as much as the data.

The Lead Outcome Measure

Name: acute care hospitalization within 60 days of the start of care.

Numerator: patients in the denominator who had at least one unplanned inpatient hospital admission within 60 days after the start of home health care, identified from transfer assessments, discharge records and hospital notifications.

Denominator: all patients admitted to home health care during the measurement period who had a start-of-care assessment completed.

Exclusions: patients admitted for planned procedures documented in the plan of care, such as scheduled surgery or chemotherapy; patients who transferred to hospice without a hospital admission; and patients who died at home without a hospital admission within the window.

Time frame: reported quarterly, by the quarter in which care started, with results final 90 days after the quarter ends so all 60-day windows are complete.

Data sources: the agency's standardized assessments, visit records and hospital admission notifications from the regional health information exchange.

The 60-day window was chosen because it matches the period in which home health care is most intensive and in which most avoidable hospitalizations occur.

Why These Choices

Counting unplanned admissions only focuses the measure on hospitalizations home health might prevent. Using the start-of-care quarter rather than the quarter of hospitalization keeps each patient in a single cohort, so results are not distorted by patients whose windows cross quarters. Excluding hospice transfers avoids penalizing appropriate end-of-life decisions. Emergency department visits without admission are measured separately, since they have different causes and costs.

A Risk Tool to Guide the Process

Hospitalization risk is not the same for every patient. Using assessment, plan of care, medication and record data from 46,366 patients at an urban home health agency, Rosati and Huang (2007) built a model that classified patients into seven risk groups at the start of care and found that demographic, financial, clinical and health status factors together could predict hospitalization within 60 days. The agency will use a simpler version of this approach: a risk score at admission, based on factors the literature and our own data identify, such as recent hospitalizations, heart failure or chronic lung disease, more than nine medications and living alone.

Process Measure 1: Timely Initiation of Care

Numerator: patients whose first skilled visit occurred within 48 hours of referral or hospital discharge, or on the physician-ordered start date. Denominator: all admissions. Rationale: the first days after hospital discharge carry high risk, and a delayed first visit can miss early problems.

Process Measure 2: Medication Reconciliation With Follow-Up

Numerator: patients whose medications were reconciled at the start of care and whose discrepancies were communicated to the physician within 24 hours. Denominator: all admissions. Rationale: medication errors after discharge are a common cause of readmission.

Process Measure 3: Front-Loaded Visits for High-Risk Patients

Numerator: patients in the top two risk groups who received at least three skilled nursing visits in the first seven days. Denominator: patients in the top two risk groups. Rationale: concentrating visits early for high-risk patients is intended to catch deterioration before it requires hospitalization.

Linking Process and Outcome

The process measures are chosen because they plausibly affect hospitalization and are within the agency's control. Donabedian (1988) argued that process measures are valid indicators of quality only when there is evidence linking them to outcomes; the agency will test that link in its own data in Week 5 by comparing hospitalization rates for patients whose care met and did not meet each process measure.

Stratifying the Measure

The rate will be reported for the whole agency and separately for each clinical team, for each of the seven risk groups and for patients admitted from a hospital compared with those referred from the community. Patients referred directly from the community tend to have different needs and hospital risks than those discharged the day before, and mixing them could hide a problem in one group. Stratified results will only be shown when a group has at least 30 patients in the period, to avoid reading meaning into tiny numbers.

The Emergency Visit Companion Measure

Emergency department visits without an admission will be defined in the same way as the lead measure, with the same denominator and 60-day window and the same exclusions. Keeping the definitions parallel lets the two be read side by side. If the agency succeeds in keeping patients out of the hospital only by sending them to the emergency department instead, the companion measure will show it.

Collecting Measures Without Adding Burden

Clinicians already complete the standardized assessment and visit notes, so all three process measures can be drawn from existing fields. The one new element is a structured field in the start-of-care note recording the risk group, which the scheduling system will fill automatically from the assessment. No new forms are needed. A measure that requires nurses to fill in a separate form will be completed well for a month and then quietly abandoned.

Linking Measures to Accountability

Each measure will have a named owner. The quality director owns the lead measure, and the three clinical team managers own the process measures for their teams. Ownership means reviewing the monthly result, explaining any signal and leading the response; it does not mean being blamed for chance variation, which Week 4 will address with control charts. Assigning owners at the definition stage avoids the common pattern in which a measure is reported faithfully each month and acted on by no one.

Validating the Definitions

Before any result is reported, two quality staff will calculate the lead measure independently for one quarter from the same data, and a sample of 40 records will be reviewed by hand to confirm that hospitalizations were correctly identified and classified as planned or unplanned. The standardized assessment's items vary in reliability (O'Connor & Davitt, 2012), so the review will also check whether transfer assessments are completed consistently.

What the Measures Cannot Show

The lead measure counts hospitalizations but cannot say which were avoidable. A patient admitted for a new hip fracture and one admitted for heart failure that worsened over a week both count. Case reviews of a sample of hospitalizations will add that judgment, as described in later weeks.

Conclusion

The lead measure, unplanned hospitalization within 60 days of the start of care, is defined with a numerator, denominator, exclusions, cohort rules and data sources precise enough to reproduce. Three process measures, timely initiation, medication reconciliation with follow-up and front-loaded visits for high-risk patients, connect the outcome to actions the agency controls. Week 3 will set benchmarks and targets and examine how the measure could be misread or gamed.

References

Donabedian, A. (1988). The quality of care: How can it be assessed? JAMA, 260(12), 1743-1748. https://doi.org/10.1001/jama.1988.03410120089033

O'Connor, M., & Davitt, J. K. (2012). The Outcome and Assessment Information Set (OASIS): A review of validity and reliability. Home Health Care Services Quarterly, 31(4), 267-301. https://doi.org/10.1080/01621424.2012.703908

Rosati, R. J., & Huang, L. (2007). Development and testing of an analytic model to identify home healthcare patients at risk for a hospitalization within the first 60 days of care. Home Health Care Services Quarterly, 26(4), 21-36. https://doi.org/10.1300/J027v26n04_03