IHP 604 Module 2 Measurement Paper Example

Measuring to Learn: A Diabetes Measurement Plan for Crestline Medical Group

[Student Name]

Southern New Hampshire University

IHP 604: Healthcare Quality and Improvement

Module Two Paper

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

Measuring to Learn: A Diabetes Measurement Plan for Crestline Medical Group

Crestline Medical Group learns how well it manages diabetes once a year, when a health plan sends a report showing what share of its patients had poor control. By then the data are months old, the patients are unnamed and no clinic can tell what changed. This paper designs a measurement plan that can guide improvement: what to measure, how to define each measure exactly and how often and in what form to look at the results.

Three Purposes of Measurement

Solberg et al. (1997) argued that performance measurement serves three different purposes, improvement, accountability and research, and that confusing them causes problems. Measurement for improvement needs to be fast, frequent and good enough to show whether changes are working; it can use small samples and simple definitions. Measurement for accountability, such as payer reports and public rankings, needs precise definitions, risk adjustment and audited data because it is used to judge and compare. Research demands the greatest rigor of all. They warned that holding improvement data to accountability standards slows learning, while using improvement data to judge people discourages honest reporting.

What This Means for Crestline

Crestline will keep reporting the payer measure for accountability exactly as specified, because contracts and a quality bonus depend on it and because its fixed definitions allow comparison with other groups in the region. Alongside it, the group will build an internal set of improvement measures, refreshed monthly from the electronic record registry and shared with clinic teams, not used for individual performance reviews. Keeping the two uses separate protects the improvement data from the defensive behavior that judgment tends to create.

Choosing a Family of Measures

Donabedian (1988) proposed judging quality through structure, meaning the staff, equipment and organization behind care; process, what is actually done for patients; and outcome, the resulting changes in health. He argued that the three are linked, so that adequate staffing and tools make correct care likelier, and correct care in turn makes better health likelier, and that measuring all three helps explain results. Crestline's plan uses a small family drawn from each category, plus a balancing measure to catch harm.

Outcome Measures

The primary outcome measure is the share of adults with diabetes whose most recent A1c is above 9%. A secondary outcome is the share with A1c below 8%, which shows movement toward good control rather than only escape from the worst range. Longer-term outcomes such as amputations and dialysis starts are too rare and slow to guide monthly improvement but will be reviewed yearly.

Process and Structure Measures

Process measures track the steps most likely to move A1c: an A1c drawn within the last six months, treatment intensified within 90 days of a result above 9%, a completed diabetes education referral and annual kidney screening with a urine albumin test. Structure measures describe capacity: the number of patients per diabetes care manager and whether each clinic reviews a registry list of patients above 9% at least monthly.

A Balancing Measure

Pushing A1c down too quickly can cause dangerous low blood sugar, especially in older adults taking insulin or sulfonylureas. The balancing measure counts, month by month, hospital or ED encounters coded for hypoglycemia among patients with diabetes. If improvement in A1c came with a rise in hypoglycemia events, the changes would need review.

Operational Definitions

Measures are only useful if everyone calculates them the same way. For the primary measure, the denominator is patients aged 18 to 75 with a diabetes diagnosis and at least two visits in the past two years. The numerator is those whose most recent A1c in the past twelve months is above 9.0% or who have no A1c recorded in that period. Counting missing tests as poor control, as the payer specification does, prevents a clinic from appearing to improve simply by testing fewer patients.

Table 1. Crestline Diabetes Measurement Family

Note. Composite definitions prepared for the medical group.

Stratifying to See Inequity

Averages can hide gaps. Every measure will be reported by clinic, by preferred language, by insurance type and by race and ethnicity where recorded. Crestline's clinic rates of poor control already range from 19% to 44%, and early registry work suggests Spanish-speaking patients have rates about 10 points higher than English speakers. Improvement that lifts the average while widening those gaps would not count as success.

Displaying Data Over Time

Berwick (1996) observed that every system is designed to achieve exactly the results it gets, and that improving results requires changing the system and learning from those changes. Learning in that way depends on seeing data over time. Crestline's monthly results will be shown as run charts for each clinic and for the group, so teams can see whether a change is followed by a shift in the pattern rather than comparing one annual figure with the last. Later modules will add control charts.

Data Quality and Burden

All measures come from existing registry fields, so clinicians do not enter extra data. The main data quality risk is A1c results done at outside labs that never reach the record, which would inflate the missing category. Care managers will reconcile outside results monthly, and the share of missing values will itself be tracked.

Conclusion

A measurement plan built for improvement looks different from one built for judgment. By separating purposes, choosing a small family of measures across structure, process and outcome, defining each exactly, stratifying for equity and showing results monthly over time, Crestline can turn a once-a-year verdict into a tool that clinic teams use to learn.

References

Berwick, D. M. (1996). A primer on leading the improvement of systems. BMJ, 312(7031), 619-622. https://doi.org/10.1136/bmj.312.7031.619

Donabedian, A. (1988). The quality of care: How can it be assessed? JAMA, 260(12), 1743-1748. https://doi.org/10.1001/jama.260.12.1743

Solberg, L. I., Mosser, G., & McDonald, S. (1997). The three faces of performance measurement: Improvement, accountability, and research. Joint Commission Journal on Quality Improvement, 23(3), 135-147. https://doi.org/10.1016/S1070-3241(16)30305-4