MMHA 6540 Week 9 Baseline and Target Framework Example

Baselines, Targets and a Fair Comparison: An Evaluation Framework for Prairie Gate's Falls Program

Student Name

Walden University

MMHA 6540: Population Health Management Applications

Instructor Name

Month Day, Year

Baselines, Targets and a Fair Comparison: An Evaluation Framework for Prairie Gate's Falls Program

Purpose

This framework sets out how Prairie Gate Health Plan will judge its falls prevention program. It defines outcome and process measures with baselines and targets, adds a measure members report themselves, and describes a comparison design. The main threat to a fair judgment is built into the program: participants are selected because they recently fell, and people selected at a high point tend to do better afterward even without help (Barnett et al., 2005). A framework that compares participants only with their own previous year would credit the program with improvement that would have happened anyway. All values are illustrative.

Measures

Table 1 lists the measures with baselines, targets and sources.

Table 1

Falls Program Measures (Illustrative Values)

Outcome Measures

The two outcome measures count fall-related emergency visits and admissions per 1,000 participants per year, using diagnosis and external cause codes in claims. The baselines come from the target segment's prior year, but because the segment was defined by having fallen, those baselines are inflated: they include the very falls that qualified members for the program. The targets are therefore stated against a comparison group rather than as a fixed reduction from baseline. The admission target is tied to the break-even point in the payer brief: at least 44 fewer admissions than expected among 600 participants. Because fall claims take roughly a quarter to settle, so the first full year of outcomes can be read only about fifteen months after launch.

Process Measures

The process rows mirror the blueprint's four steps one for one, and unlike claims they can be read within weeks. They answer a different question from the outcomes. If outcomes disappoint, these rows reveal whether the fault lay in the idea or in its execution. If a step is slipping, for example if modifications lag because of the repair bottleneck, the program can fix it long before claims would reveal any effect on falls. Exercise reduces falls mainly when people actually attend (Sherrington et al., 2019), so class attendance is the process measure most closely tied to the outcome.

Member-Reported Measure

Fear of falling is common after a fall and leads many older adults to restrict their activity, which can weaken them and increase future risk. A short questionnaire at enrollment and at six months, asked by the coordinator by phone, will show whether participants feel safer. This measure matters for two reasons: it captures a benefit members value that claims will never show, and it helps interpret the outcomes, since a participant who feels safer may move more and therefore have more opportunities to fall.

Comparison Design

The program will be rolled out by county in stages: the metro county first, as a pilot, then Ash and Corley. Members who meet the segment definition in counties not yet served will form the comparison group, and their fall-related visits and admissions will be tracked in the same way. The program's effect is estimated as the difference between the change among participants and the change among comparison members over the same months, the logic of difference-in-differences (Dimick & Ryan, 2014). Because both groups are selected by the same rule, both will regress toward lower rates, and the comparison removes that shared improvement. Staggering rollout in this way is a recognized design for programs an organization intends to extend to everyone (Hemming et al., 2015).

The comparison has limits. The rural counties differ from the metro county in age, housing and access, so their trends may differ for reasons unrelated to the program; the analysis should adjust for age and prior falls. And the comparison ends when the last county is served, so the design yields its strongest evidence in the first year.

Reporting

Process measures will be reported weekly to coordinators and monthly to the population health director. The member-reported measure will be summarized at six months. Outcomes will be reported fifteen months after launch, when claims for the first year are complete, to the executive committee, together with the process measures and the comparison results. The report will state plainly whether the break-even threshold was met and what the program achieved for members either way.

Data Readiness

Several pieces must be in place before the first member is enrolled. The analyst must confirm that fall-related emergency visits and admissions can be identified reliably in claims, using injury diagnosis codes together with external cause codes, and must test the definition on last year's data to see whether it matches the counts in the burden profile. The care management system needs fields for the date a member was identified, the date of first contact and the date of enrollment, so that the seven-day measure can be calculated automatically. The repair nonprofit's completion reports must include the date of the home visit they respond to, or the thirty-day measure cannot be computed. The area agency's attendance file must carry a member identifier the plan can match. And the coordinators need a script for the fear-of-falling questions, so that every participant is asked the same way at enrollment and at six months.

Each of these is a small task, but missing any one would leave a measure blank in the first report. The population health analyst will own a readiness checklist and report its status to the director two weeks before launch, and launch should wait until every item on it is complete.

References

Barnett, A. G., van der Pols, J. C., & Dobson, A. J. (2005). Regression to the mean: What it is and how to deal with it. International Journal of Epidemiology, 34(1), 215-220. https://doi.org/10.1093/ije/dyh299

Dimick, J. B., & Ryan, A. M. (2014). Methods for evaluating changes in health care policy: The difference-in-differences approach. JAMA, 312(22), 2401-2402. https://doi.org/10.1001/jama.2014.16153

Hemming, K., Haines, T. P., Chilton, P. J., Girling, A. J., & Lilford, R. J. (2015). The stepped wedge cluster randomised trial: Rationale, design, analysis, and reporting. BMJ, 350, Article h391. https://doi.org/10.1136/bmj.h391

Sherrington, C., Fairhall, N. J., Wallbank, G. K., Tiedemann, A., Michaleff, Z. A., Howard, K., Clemson, L., Hopewell, S., & Lamb, S. E. (2019). Exercise for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews, 2019(1), Article CD012424. https://doi.org/10.1002/14651858.CD012424.pub2