The National Health Plan Collaborative: Overview of Its Origins, Accomplishments, and Lessons Learned
IV. Accomplishments on Targeted Activities
Most NHPC members perceived Phase II as more successful than Phase I, not only in terms of firms' commitment to and institutionalization of disparities work, but also in concrete progress made on targeted disparities activities. Here we discuss initial expectations and ultimate outcomes in the two areas that consumed much of the attention during Phase II: collecting race/ethnicity (and, to some extent, language) data and standardizing data, and improving access to and quality of language services.
Data Collection and Standardization
As initially conceived, Phase II was designed to support firms' efforts to capture primary data on race/ethnicity (with a growing emphasis on language data during this phase). This phase included developing guidance consistent with best practices and responsive to what NHPC participants anticipate will be a desire among users to collect these data with some uniformity across firms. (The intent is to report at least a minimum amount of consistent information). Early in the process however, firms were reluctant to endorse a uniform approach. Support organizations concluded that organizational structures were sufficiently different that it made more sense to focus less on promoting a specific standardized way of collecting race/ethnicity/language data than on encouraging firms to collect such data using whatever means and strategies were appropriate for their environment.
Data Collection Methods
Phase II was marked by firms' increased commitment to data collection. By the end of this phase, all of them were collecting direct data in some form but still varied substantially in the extent to which they were doing so. One firm's representative characterized the evolution of the NHPC's thinking on primary data collection as moving from a "prove it to me" attitude a few years ago to a strong conviction at this point of its value. Moreover, firms uniformly view primary data collection as the gold standard toward which they need to work. At the same time, they recognize that secondary sources (such as purchaser data on employees and Medicaid data on beneficiaries) provide another useful avenue for race/ethnicity/language collection. In addition, indirect or proxy methods (such as geocoding and surname analysis) likely will remain important, as discussed in more detail below.
Activities surrounding direct data collection were important in Phase II. At its start, NHPC firms shared information on how to collect primary data on race/ethnicity/language, including member surveys about race/ethnicity/language, questions asked during clinical encounters, and assurances given to members about how the data would be used. Firms also shared information on member responses and reactions to these inquiries. As one firm's representative stated, "We set the stage that health plans can ask for this information and not be totally lambasted in the marketplace."
Differences in firms' organizational structures and product mixes influenced the strategy to which they gravitated in addressing issues of race/ethnicity/language data. Firms like Molina, for example, serve a largely Medicaid population which provided them with unique access to data from States (the purchaser)—something that the other predominantly commercial-based firms in the NHPC did not have.ix The NHPC spent some time exploring the feasibility of obtaining similar data from employer accounts but found this avenue less fruitful. Some firms encountered employer resistance to sharing this information because of legal concerns, administrative costs, and other issues, and only one firm successfully obtained data through this source during Phase II.x
The other external source of race/ethnicity/language data is from providers (e.g., physician groups and hospitals) who collect these data during clinical encounters. Firms' use of this source is still relatively limited, as providers often are reluctant to share data with health plans for legal and other reasons. However, some firms with strong organizational ties to provider groups were able to capitalize on this capacity to obtain data for at least some of their patients.
Given most firms' limited ability to obtain member race/ethnicity/language data from employers and providers, most NHPC firms concluded that if they wanted such data, they would have to collect it themselves. Firms in the NHPC studied Aetna's experience with using member portals to collect these data, as well as the emerging experience of other NHPC participants with diverse strategies for data collection, including through member surveys and during member educational outreach calls.
As the NHPC firms became more sophisticated in collecting data, the importance of a mixed-methods approach to data collection emerged-that is, firms began using multiple sources of input data on race and ethnicity, with the goal of aggregating data from these sources across members (Table 1).xi Firms increased the number of sources from which they collected direct data, based both on growing awareness of alternative sources and the limitations of any single approach. NHPC firms now collect these data through various channels, including health risk assessments, Web portals, member surveys, disease management programs, clinical encounters, and enrollment. Individual firm strategies vary to match both their operational configuration and views of what approaches will be acceptable in the marketplace; a key emphasis is voluntary reporting by the relevant subgroups.
Table 1. Methods used by NHPC firms for collecting race/ethnicity/language data
Disease management programs
Health risk assessments
Encounter data (e.g., physician office, hospital) Health plan direct outreach
Member Web portal
Member-initiated contacts (e.g., customer billing inquiries)
Centers for Medicare & Medicaid Services (Medicare beneficiaries)
State Medicaid agencies (Medicaid beneficiaries)
|Geocoding and surname analysis
Third-generation methods (using Bayesian statistics)
By the end of the formal NHPC, signs were emerging that firms' ability to make progress in race/ethnicity data collection on their own might have limits. Several NHPC firms that have been collecting direct race/ethnicity data for several years have reached a plateau of 25 to 35 percent of members for whom they can collect these data directly. For example, Aetna, which began collecting data in 2002, had direct race/ethnicity data for about 30 percent of its actively enrolled membership as of 2008. Similarly, HealthPartners has collected these data for about 32 percent of its membership. Firms are considering ways to boost the numbers of members with direct data beyond the 30 to 35 percent threshold, but this finding suggests that, unless a fundamental change occurs in the environment, indirect methods of estimating race/ethnicity may continue to play an important role in firms' disparities work in the future.
Standardization of Data Collection
Later in Phase II, the NHPC expanded its focus on the potential for standardization of data collection.xii While such data collection was a priority from the beginning of Phase II, standardization per se was not. About midway through this phase, NHPC members began coalescing around the idea of promoting standardization at the national level. The impetus for movement in this area was State efforts (e.g., California and Massachusetts) to encourage and specify standards for primary data collection of race/ethnicity/language. These State efforts generated concern within the industry about lack of national uniformity if each State pursues its own unique approach. In addition, UnitedHealth Group's leadership in this area and its push for health plans to be "out in front" on this issue industrywide helped to galvanize support for standardization within the NHPC.
Several meetings of the standardization workgroup in winter and early spring 2008 resulted in the NHPC seeking external help on this issue. As one firm's representative said, "There was a clear declaration on the necessity of settling the definition of R/E/L once and for all.a clear sense that we need health plan leadership as well as policy leadership to drive the process." Another NHPC organization concurred, saying, "The standardization [workgroup] crystallized the issue. They realized the problem was beyond what they could solve."
Ultimately, NHPC firms met with Carolyn Clancy, the Director of AHRQ, and others to discuss the possibility of an IOM study on race/ethnicity/language standardization. Since that time, AHRQ has commissioned the IOM to develop a study that will recommend appropriate racial and ethnic categories, and perhaps promote standardization in data collection. The results of this study are expected in mid-2009.
Federal legislation also may have an impact on data collection and perhaps its standardization. The Medicare Improvements for Patients and Providers Act of 2008 is likely to have important implications for data collection by health plans.8 The act requires a report within 4 years of enactment that includes recommendations for improving identification of health disparities among Medicare beneficiaries based on beneficiary data on race, ethnicity, and gender.
By the end of Phase II, NHPC firms, particularly national firms, continued to believe in the importance of standardization and remained concerned about the possible proliferation of varying State standards if national standards are not developed. While some firms characterized the NHPC's push for standardization in Phase II as a major accomplishment, others described progress on this front as slow and disappointing. Standardization could limit firms' flexibility in data collection and increase their costs, but NHPC firms generally seemed to perceive the benefits of a uniform approach as outweighing the costs. Moreover, a few large firms appeared reluctant to embrace primary data collection fully until an approach was codified, so standardization may begin to push these firms forward.
Access to Language Services
The second area of emphasis in Phase II was on work to enhance firms' ability to provide high-quality care to patients with limited English proficiency. The goal was to develop recommendations on how to select and evaluate vendors for language services, support provider adoption of such services, and encourage employers and members to use language services.
At the start of Phase II, the NHPC considered developing an operational pilot related to language services in one community (partially driven by Molina's suggestion for such joint work). However, several firms did not wish to commit to joint work, largely because it might have required them to coordinate competing internal and external demands. Moreover, some NHPC participants may not have been well positioned organizationally to grasp everything their firm was doing in the area of language access. This was particularly true for firms that use a central contract to support a variety of administrative services or that purchase language services through one or more vendors. NHPC participants instead opted for collecting and developing broad-based tools, such as standards for interpretation services, which each could use in their own way.
Firms' awareness and knowledge of language services increased substantially as a result of this Phase II activity. One firm's representative described the major accomplishment surrounding this component of Phase II as "consciousness raising," noting that, in the past 4 to 5 years, the firms had been focused more on collection of race/ethnicity data, while language access "wasn't on our radar." NHPC members generally agreed that language access was a productive area, in which firms were "creative and willing to share" during Phase II.
California's SB 853 played a major role in catalyzing NHPC firms' interest in language services. Although SB 853 was passed in 2003, it took some time for regulations to be developed. The regulations, developed by the Department of Managed Health Care (DMHC), mandated that health plans operating in California provide language assistance services at all points of care to members with limited-English proficiency. Services included translating vital documents in threshold languages and providing translation and interpreter services as needed. Plans are required to comply with the regulations beginning in January 2009. National plans were particularly attuned to this policy development because California represents a large market and other States might follow suit. Indeed, several States likely will be looking to California's experience to understand what works before considering their own possible implementation.9
Language access materials shared between NHPC members in Phase II included:
- Vendor selection criteria for translators and interpreters.
- Information on various approaches to interpretation and translation, including advantages and disadvantages.
- Guidelines for in-house translation reviews.
- Various translation checklists and worksheets, some of which were compiled based on information from NHPC firms.
UnitedHealth Group's Web conference presentation of its in-language provider directory for Asian Americans was considered by several firms as one of the more useful Phase II seminars.xiii (Information on the in-language provider directory is included in the NHPC toolkit, as is a discussion of promising practices in interpreter services based on Kaiser Permanente's qualified bilingual staff model and health care interpreter certificate program (go to www.rwjf.org/qualityequality/product.jsp?id=34010).
ix While health plans serving Medicare and Medicaid beneficiaries can obtain information on beneficiaries' race and ethnicity from Federal and State governments, only limited race/ethnicity distinctions are possible in Medicare data, and Medicaid data vary substantially by State in quality and completeness.
x For more information on the barriers to employers sharing these data, see Higgins PC, Au M, Taylor EF. Employer perspectives on health plan efforts to reduce disparities. Manuscript submitted for publication. Mathematica Policy Research; 2009.
xi For detailed information on which sources and strategies each NHPC firm was using at the end of Phase II, refer to the National Health Plan Collaborative Toolkit, Chapter 4, at www.rwjf.org/qualityequality/product.jsp?id=33988
xii For general information on recent Federal and State activities surrounding race/ethnicity data collection, refer to Higgins PC, Taylor EF. Emerging Federal and State activities on the collection of data to measure racial and ethnic disparities. Policy brief prepared for the Agency for Healthcare Research and Quality. Washington, DC: Mathematica Policy Research; 2009. Also refer to Sec. 185 of H.R. 6331.
xiii UnitedHealthCare won a "Recognizing Innovation in Multicultural Health Care Award" from NCQA in 2007 for its work on the in-language provider directory for Asian Americans. Since the final NHPC meeting, NCQA has announced its 2008 awards recognizing innovation in multicultural health care. Five of the seven awardees were NHPC firms (Aetna, Highmark, Molina, UnitedHealth Group, and WellPoint). For more information, go to www.ncqa.org/tabid/826/Default.aspx.