Chapter V. Intervening to Reduce Disparities
Evaluation of a Learning Collaborative's Process and Effectiveness to Reduce Health Care Disparities Among Minority Populations
Launching pilot interventions intended to reduce racial and ethnic disparities was one of the four main areas of Collaborative focused on (bold type in Figure V.1). Because most of Phase I was devoted to geocoding and data collection, most firms began to develop pilot interventions only during the last six months of the Collaborative. This chapter discusses the role of the pilot interventions in the Collaborative and how expectations for interventions changed over time; the interventions themselves; challenges facing the firms as they implemented the interventions, and the role and significance of the pilot interventions to the firms.
A. Summary of Findings
During most of the Collaborative, the firms' efforts to implement interventions to reduce disparities took a back seat to data collection. As firms gained insight into disparities, they began to think more concretely about what they, as entities sponsoring health plans with different business practices, could do to reduce disparities. By the end of the Collaborative, seven of the nine firms had either completed or were in the process of completing pilot interventions, and two were in the process of developing them. Consistent with the agreed-upon Collaborative focus, diabetes in racial/ethnic minorities was the primary target of all pilot interventions, and four of the firms geared their interventions toward Hispanics and the others focused on other subgroups. Most pilots were small, though size varied by firm, and the interventions themselves varied markedly from one firm to the next. While it was too early at the time of this writing for most firms to know the outcomes of their interventions, most perceived them as creating a framework for future expansion and learning and planned to pursue related interventions even though Phase I was ending.
The firm's progress in developing interventions was challenging for a variety of reasons. First, firms were not sure where to begin, citing uncertainty about how to best intervene. Second, lack of data was a constraint for many firms because the ability to develop an appropriate intervention means having an understanding of the race/ethnicity of particular members and an ability to geographically target those members. Third, the scale and complexity of the firms themselves made it difficult to implement effective interventions because of the need to coordinate activities in the face of the split between corporate and regional responsibilities and between the various departments and other functional areas in the firm. Fourth, logistical issues, such as recruiting physicians to participate in provider-based interventions, were a challenge.
The Collaborative led firms to view their work on disparities as a part of their quality improvement effort rather than as an add-on or separate activity. This link created leverage to address disparities within firms. Still, firms were constrained by the tight fiscal environment in which they operated and by competition for resources. The ability to build a business case for addressing disparities was viewed by firms as important to obtaining the resources needed to address the quality improvement agenda and disparities in health care.
B. Expectations for Interventions Changed Over Time
From the start, expectations for the Collaborative were high. With technical assistance from CHCS and IHI, sponsors planned that firms would design, implement, and complete targeted interventions for their minority populations by the end of the Collaborative's two-year operational period. Sponsors hoped to be able to point to improvement in health outcomes, such as better diabetes indicators, to demonstrate the Collaborative's success.
As the Collaborative got off the ground, however, it became clear that the original expectations were not realistic. One sponsor noted, "I think we were just incredibly ambitious and off target for a lot of it." The majority of firms did not have critical information on their members—such as race, ethnicity, and language preference—and without it, firms would find it difficult to develop targeted interventions. Not surprisingly, the firms spent most of the first part of the Collaborative working with RAND on data collection activities, including geocoding and surname analysis (discussion in Chapter III).
Although the Collaborative discussed interventions earlier in Phase I, it wasn't until the last six months that most firms shifted their focus to developing interventions specifically related to their Collaborative efforts. According to an individual from one support organization, "I thought they would be further along on the interventions than they were. A lot of this is because of turnover or because of the time needed for data collection." Another support organization echoed the frustration, "It's disappointing to me that we couldn't get more folks to the intervention stage, but that was just unrealistic given time and measurement issues." In addition, by the time firms were ready to develop their pilot interventions, they had not yet built strong relationships with staff from CHCS and IHI. Probably for these and such other reasons as the firms' interest in keeping their internal processes confidential, they developed interventions with staff from their own organizations, often leveraging their own programs and/or activities rather than looking to support organizations for assistance.
Over the course of the Collaborative, the definition of an acceptable "intervention" broadened from a narrow pilot program to a wide range of firm activities, including cultural competency training, data collection, and long-term goals. The sponsors' definitions of "success" expanded as well. Over a year into the Collaborative, one sponsor noted, "Whether or not they moved the needle is less important than did they do something, did it work, and why or why not."
The analysis focuses on the primary intervention self-identified by the firms during our round three interviews. Seven of the nine firms had either completed or were in the process of completing their primary pilot interventions at that time; two firms were still in the development stage. We exclude from this analysis initiatives related to gaining broad support for firms pursuing disparities work or data collection activities related to the Collaborative, as these activities are the focus of other sections of this report. At this point, however, in some firms, these activities were at least as important as the interventions they piloted through the Collaborative.
C. Implementation Experience
Consistent with the agreed-upon focus of the Collaborative, all firms targeted their primary pilot intervention to members of racial/ethnic minorities with diabetes (Table V.1). The majority of the firms focused on subsets of their Hispanic membership; fewer targeted African Americans or other minority populations. Most firms started small before scaling up to the region or firm level.
The exact size of the populations targeted, however, varied significantly from firm to firm and over time. One firm, for example, targeted diabetic Hispanic members in one of their care centers; while it originally expected to reach out to 200+ patients who would be encouraged to pursue a three-drug lipid-lowering regimen, the firm discussed the fact that physicians had contacted many of these patients, some of whom already were following the regimen. As a result, the firm reached out only to the 52 patients who were not known to be on the regimen or to have been contacted. Another firm targeted 150 to 200 Spanish-speaking members in their diabetes disease management program. A third firm, lacking information on patients' race, reached out via a call campaign to all their known diabetic patients (around 25,000), letting them know that certain ethnic/minority groups in particular might be at risk for diabetes and asking them to respond if they wanted more information. About 5,000 members responded, giving the firm the opportunity to follow up.
Some firms used the results of their geocoding/surname analyses to identify their target population. One firm, for example, whose intervention had not yet become operational, planned to use geocoding results to identify "clusters" of Hispanic members with diabetics not receiving HbA1c testing and/or lipid testing. Other firms focused their interventions in sections of their service areas generally known to have a diverse racial/ethnic composition (e.g., Miami, Florida, which is known to have large Hispanic and African American populations). At least one firm combined these strategies. First, it implemented its pilot at a center composed of 90 percent Latino patients. After expanding the pilot to the region, however, the firm planned to use both surname analysis and GIS programs to identify the Latino population.
2. Intervention Strategies
Most large, national firms used a provider-based strategy as their primary intervention/activity for the Collaborative. In some cases, the intervention targeted providers employed by the firm. For example, one large national firm implemented a cultural competency program through which it aimed to train all nurse clinicians and physicians who interface with members as part of its diabetes disease management program. Another large national firm worked with a vendor to encourage its affiliated physicians to implement a Web-based patient registry that helped physicians to track patients with chronic diseases by showing past and needed future steps at upcoming appointments. A third national firm provided reminders to providers whose patients missed appointments or were overdue for preventive services.
Regional firms were more likely to use a variety of strategies and more often combined multiple interventions/activities "to try to move the needle" on diabetes interventions. For instance, once regional firm combined a provider- and a member-based intervention. First, it consulted with minority physicians about strategies for improving diabetes outcomes, thus identifying a lack of knowledge about diabetes among their patients as a barrier to controlling the disease (the provider-based intervention). Based on this information, the firm used an interactive voice recognition call campaign to reach its diabetic members and convey a culturally sensitive message explaining that their race/ethnicity could affect their risk for diabetes and urging members to respond if they were interested in additional information. The firm then sent information to members who requested it (the member-based intervention). Another regional firm partnered with Stop and Shop grocery stores to offer free interpreter services, eye exams, and other services to diabetic members (a community-based intervention) in combination with a dropped referral requirement (system redesign) and waived co-payments for diabetic retinal exams (member-based intervention,).
To date, only one of the firms (a large, national firm) involved in the Collaborative explicitly planned to judge the success of its intervention on improved outcomes of its targeted population, such as improved HbA1c test results (Table V.2). According to our round three interviews, the majority of the remaining firms, including all of the regional firms, defined (or planned to define) their success as improvements in process measures, such as HbA1c testing and LDL screening rates. One regional firm noted, "That [analyzing outcome data] would be so complicated it's not even on the radar screen right now."
While not explicitly assessing success in terms of outcome measures, one national firm reported that it modeled its pilot on studies that showed improved outcomes in their intervention populations. Thus, while not specifically examining outcome measures in its pilot program for the Collaborative, the firm assumed that improving the process measures would also improve outcomes.
Findings and Status. The majority of firms are in the process of measuring and/or analyzing their results. One large national firm, for example, did not find statistically significant differences in screening rates between intervention and comparison groups, but they did find outcomes of the treatment group to be slightly better than the comparison group. Preliminary data for one regional firm, on the other hand, show that areas participating in the intervention had improved screening/testing rates over those in the non-participating regions. Several firms recognized that while most may not actually reduce disparities under the short timeframe of the National Health Plan Collaborative, firms developed the capacity to reduce disparities. "We've built a framework that we can now expand upon to identify and address disparities. It's foundation work that wouldn't have happened without the Collaborative. We have a way of identifying people for interventions to reduce disparities."
Future Work. While it is too early for most firms to determine the success of their interventions, most firms made plans to continue and/or expand their pilot activities/interventions to additional members, regions, and/or disease conditions. One national firm, for example, planned to expand its pilot intervention from the initially targeted 52 members to an entire region of the firm with approximately 475,000 patients, half of which are Latino. Similarly, one regional firm expanded its pilot activities from Spanish-speaking members with diabetes in three counties to all Spanish-speaking members enrolled in the health plan. Pilot interventions will continue into Phase II of the Collaborative (Chapter VIII).
D. Challenges Facing Firms
Firms faced a number of challenges in implementing pilot interventions to reduce disparities among their member populations. First, firms lacked an evidence-based on how to develop interventions. One firm noted, "Literature wasn't there on interventions." Another firm said that "where to begin, how to target disparities, and what interventions to implement" was a challenge. Also according to the firm's spokesperson, "I'm still wrestling with... what's the big lever to close the gap? This is not clear to me going forward.... Right now, I'm flying blind, and just selecting interventions and approaches on face validity."
Firms also cited a lack of data as an impediment to developing interventions. "Until we have concrete information on our members, it's awfully hard to create interventions that are directed in the right way." Similarly, obtaining resources to implement interventions proved challenging for many. One regional firm noted, "Getting resources for work we want to do is hard."
Firm structure was also cited as a challenge. Making any changes in large national firms is difficult. But the regional structure of national firms made it doubly difficult for them to focus on change at a macro-level because the regions have significant authority over care delivery and decisionmaking. Even at small firms, it is difficult to make changes when they put corporate-level decisions into play. For instance, it took one regional firm almost one year to implement a "no referral necessary" policy change for diabetic retinal exams.
In addition, firms cited logistical issues as challenges to implementation. "Bringing all of those resources together—making sure everyone was trained and making sure everyone knew what to do—was a Herculean effort." One firm, for example, had trouble recruiting physicians to participate in their provider-based intervention. Even after recruiting them, the firm was faced with the issue of familiarizing and training the physicians on the Web-based system used for the intervention. "It takes a lot of education to get the doctors used to the system." Another firm could not integrate data from its pilot intervention into the rest of its quality improvement data system.
E. Intervention Role and Importance at Firms
Many firms view their activities/interventions to reduce disparities as part of their quality improvement efforts, rather than as a separate activity (discussion in Chapter III). "Disparities are just another way to look at quality." This view is useful in terms of funding and otherwise sustaining disparities initiatives. One firm noted that the link between quality improvement and disparities work helped it in the way of funding because there was no budget specific to disparities projects. That is, "we definitely see disparities work as relating to quality improvement—it's the main argument we're using for resources." Also, another firm recognized this important link for the future sustainability of these activities. "Future sustainability means institutionalizing disparities work into quality improvement work, so it's just something that happens."
While most firms recognized the potential value of their activity/intervention if it was successful, the specific activity/interventions at this point are often one of many of the firms' quality improvement efforts and are not seen as a main priority. One firm noted, "Reducing disparities is ranked well below survival." It's possible that this viewpoint is the byproduct of competition from outside forces. For example, one firm is tied up with integrating its new IT platform, and another has been focused on a recent merger. Despite these "distractions," however, many firms claim that their pilots would persist even if the Collaborative ended. One large national firm noted, "In many cases, the pilots would continue." Regional firms agree, "The Collaborative started the original interest in these areas, but I think it would be self-sustaining at this point."
Many firms are trying to use the interventions to build a business case for disparities interventions. One large national firm claims that initiatives have already shown a return on investment, and several other firms are approaching their disparities work with this in mind. One firm is attempting to conduct an analysis to support a business case for disparities, "To the extent that we can say, because we now have more folks managing their diabetes, and we know what it costs to treat diabetes when it's not managed.... Purely from the financial standpoint, we can say that we can save money." Overall, firms said that developing the business case for disparities work is a priority as they move forward. This issue will be a major focus of Phase II activities (Chapter VIII).