Ranking Community Health Status to Stimulate Discussion of Local Public Health Issues: The Wisconsin County Health Rankings

Abstract

United Health Foundation’s America’s Health Rankings, which ranks the states from “least healthy” to “healthiest,” receives wide press coverage and promotes discussion of public health issues. The University of Wisconsin Population Health Institute used the United Health Foundation’s model to develop the Wisconsin County Health Rankings (“Health Rankings”) from existing county-level data. The institute first released the rankings in 2004. A survey of the Wisconsin county health officers indicated that they intend to use the rankings for needs assessment, program planning, and discussion with county health boards. The institute implemented many of the health officers’ suggestions for improvement of the rankings in subsequent editions. The methods employed to create the rankings should be applicable in other states.

RANKINGS CONTRASTING cities, states, or nations for any number of characteristics—economic, education, health, and others—are continuously being released. Such rankings often receive substantial attention. In particular, United Health Foundation’s annual America’s Health Rankings,1 which ranks states from “least healthy” to “healthiest,” receives wide press coverage and promotes discussion of public health issues. The University of Wisconsin Population Health Institute in Madison released its first annual Wisconsin County Health Rankings (hereafter “Health Rankings”) report in January 2004, with the main purpose of encouraging discussion about important population health issues among Wisconsin public health and other policy communities. Our hope is that, by encouraging such discussion and raising awareness of variation in populations’ health within Wisconsin, the Health Rankings will yield an enhanced appreciation of the variety of factors that affect populations’ health and that are amenable to influence by public and private sector programs and policies.

The conceptual framework underpinning the Health Rankings is based on a model of population health improvement with health outcomes produced by a set of health determinants. Public and private sector programs and policies can enhance or limit these determinants.2 We describe briefly the design of the Health Rankings (more-detailed methods can be found elsewhere3), the method of annual release, and an evaluation of the Health Rankings’ usefulness to local public health authorities.

PROGRAM DESCRIPTION

Design of the Health Rankings

The Population Health Institute, in consultation with Wisconsin public health policymakers and other population health stakeholders, developed the first edition of the Health Rankings over an 18-month period. In creating the Health Rankings, we chose to focus on 2 categories of health measures: health determinants and outcomes (Figure 1 ▶). Outcomes measure the current state of health in a county, whereas determinants are predictors of future health outcomes. We used 2 broad subcategories to represent health outcomes: death and health status while alive. We assessed each with a single measure—years of potential life lost for death and self-reported general health status for health status.

FIGURE 1—

Structure of the Wisconsin County Health Rankings.

Note. Counties receive rankings for a summary of health outcomes and health determinants and subcategories of outcomes and determinants. We ranked counties for each population health category or subcategory shown in a box. Health determinants in bold were not in the 2003 Health Rankings but were added in 2004. Measures in italics were in the 2003 Health Rankings but not subsequent Health Rankings. Specific data sources for health determinants and outcomes are detailed in Vila et al.4

We used the Wisconsin State Health Priorities to select health determinant measures.5 We divided 18 health determinant measures into 4 subcategories: access to health care, behaviors, socioeconomic factors, and the physical environment. Each of these categories and subcategories comprises 1 or more underlying measures. Criteria for selecting the measures included being a direct or proxy measure of important aspects of public health, available publicly and updated periodically at the county level, consistently collected across counties, and of sufficient quantity to have moderately stable county-level estimates. We then calculated selected health measures for each county, averaging 1 to 7 years of data depending on sample sizes (more years were averaged for measures with sparser data). Rankings for the health categories and subcategories ranked weighted averages of standardized scores (z scores) for the health measures that made up each health subcategory. Table 1 ▶ is an excerpt of a table from the 2003 Health Rankings showing the counties with the 5 highest and 5 lowest overall health outcomes ranks—an average of mortality and health status. Detailed methods for weighting and scoring the health measures and derived ranks can be found elsewhere.3^,6^,7

TABLE 1—

Wisconsin Counties With the 5 Highest and 5 Lowest Overall Health Outcomes Ranks: Wisconsin County Health Rankings, 2003

		Mortalityb
County	Overall Health Outcomes Ranka	Rank	Years of Potential Life Lost	Health Status Rank (%)c
Ozaukee	1	3	8331	1 (5.1)
Waukesha	2	1	8132	2 (6.0)
Eau Claire	3	2	8301	3 (6.6)
Outagamie	4	13	9272	4 (6.8)
Marathon	5	9	9060	5 (7.3)
Wisconsin average	NA	NA	10 683	(9.6)
Iron	68	65	12 562	65 (11.0)
Ashland	69	55	11 863	71 (11.8)
Waushara	70	71	14 648	68 (11.3)
Milwaukee	71	69	13 653	72 (12.7)
Menominee	72	72	23 500	48 (10.1)

Source. Adapted from Peppard et al.6

Note. NA=not applicable.

^aOverall health outcomes rank is a simple average of mortality and health status z-scores (the number of standard-deviation-units a county’s measure is from the mean of all 72 counties).

^bMortality as measured by age-adjusted years of potential life lost before age 85 years per 100 000 population.

^cThis is the prevalence (percentage) self-reporting poor or fair health.

Release of the Health Rankings

The Population Health Institute released its 2003 Health Rankings in early 2004. Local county health officers received a 13-page report by mail 2 weeks before the public release. Institute staff held 2 conference calls in which they described the Health Rankings to health officers and answered any questions. The Health Rankings were made public on the institute Web site and through a press release. Institute staff held an additional conference call with members of the press (primarily newspaper reporters).

DISCUSSION, EVALUATION, AND NEXT STEPS

The first annual Health Rankings (for 2003) took approximately one and a half years and $100 000 to develop. Roughly 85% of the cost was for personnel, 10% for publication and distribution (several hundred glossy reports and a Web posting), and the remainder for data purchase. Most data were publicly available, but the Wisconsin Department of Health and Family Services charged nominal fees for Wisconsin-specific survey data. Subsequent Health Rankings reports have cost less than half of the initial edition’s cost. We expect that public or private entities in other states would eventually be able to produce similar reports in 1 year for similar costs.

Several weeks after the 2003 Health Rankings’ release, we mailed surveys to each of Wisconsin’s 72 county public health officers. The surveys assessed officers’ awareness of the Health Rankings and their thoughts about the usefulness of the Health Rankings to their work. We sent a second mailing to officers who did not respond and called officers who did not respond to either mailing. The final response rate was 94%.

Table 2 ▶ presents the results of the survey. Most respondents (82%) reported that the Health Rankings were useful to their work. Forty-four percent reported local media coverage of the Health Rankings (data not shown), 29% suggested that the Health Rankings might be described with more clarity, and 16% recommended an expansion of the environmental component.

TABLE 2.

Wisconsin County Health Officers’ Most Frequent Responses to Survey Questions About the 2003 Wisconsin County Health Rankings

Question	Health officers responding, %
Rankings are useful to health officer’s work	82
Health officer plans to use Rankings in the community	69
Health officer will use the Rankingsa
To perform needs assessment and evaluation	50
In a presentation to the county health board	32
For program planning, resource targeting, or grant writing	26
In discussion with community partners and public health staff	24
Local news media contacts and coverage of the Rankings
Newspaper articles	40
Radio or television coverage	16
Health officer liked the Rankings fora
Easy-to-read, concise information source	24
Ability to compare with other communities and counties	24
Allowing tracking progress and change	13
Rankings could be improved ina
Clarity and explanation of Rankings	29
Environmental health component	16

Note. Out of 72 Wisconsin county health officers, 68 responded.

^aSummarized from responses to an open-ended question.

The survey demonstrates that our ranking efforts were useful to the majority of county health officers, many of whom planned to use the rankings in community communications. In response to the survey and additional informal feedback, the Population Health Institute revised the 2004 Health Rankings to improve communication clarity, increase accessibility to the data underlying the rankings (on the institute Web site), and expand the number of health measures used (Figure 1 ▶). We view this continuing process of evaluation and modification of the Health Rankings in response to future evaluations as an ongoing, key component of the Health Rankings product.

Media coverage and evidence from the survey indicate that our primary goal for releasing the Health Rankings—encouraging discussion of population health issues within Wisconsin communities—is, in part, being achieved. However, we have found that the Health Rankings were occasionally overinterpreted as though they were designed to be a statistically robust comparator of communities’ population health status. We discouraged such interpretation because of important limitations of the data underlying the Health Rankings, including survey data sparseness for small-population counties. Thus, with each subsequent release of the Health Rankings, we caution against the overextension of their use (e.g., for tracking changes in health within counties in response to local public health programs). Nonetheless, we strive to enhance our methods and communication strategies so that future editions of the Wisconsin County Health Rankings might provide a continually improving basis for discussion of the distribution and determinants of population health among Wisconsin communities.