Evaluating Two Mysteries: Camden Coalition Findings

Two fundamental mysteries characterize the health care sector’s embrace of the concept of population health [1]. Mystery A is why health care systems cannot control the extraordinary concentration of cost among the sickest 1–10 percent of patients. Mystery B is how to understand the relationship between those extraordinary costs of high-tech health care and the reality of neighborhoods near the hospitals, which hold the greatest disparities, most underserved groups, and thus a high proportion of the costs. The answer to both questions leads away from hospitals onto the streets and into the homes of community members with lower incomes and chronic illnesses. One of the most iconic population health programs working to understand why both mysteries persist, the Camden Coalition of Camden, New Jersey, was recently evaluated by Finkelstein, Taubman, and Doyle in the New England Journal of Medicine (NEJM) using a randomized controlled trial (RCT) methodology [2], prompting this perspective. In Memphis, Tennessee from 2008 to 2013, the population health work conducted by the authors wrapped highly complicated educational programming and referrals around all partners and produced years of encouraging data, now adapted into other social and institutional settings. Our similar, subsequent work statewide in North Carolina, the North Carolina Way, now engages over 500 congregations and hundreds of other community partners [3]. We argue that traditional educational programming, computer referrals, data analytics, and RCTs have not been constructed to accurately measure the eff ectiveness of these community-based, “messy” approaches to improving population health. That is why the RCT fi ndings of the Camden Coalition are so disappointing—the data cannot and should not be evaluated by methods that can’t accurately convey its impact. The Launch of Hot Spots and the Camden Coalition