A COMPARISON OF 2 ANALYTIC METHODS FOR THE IDENTIFICATION OF NEIGHBORHOODS AS INTERVENTION AND CONTROL SITES FOR COMMUNITY-BASED PROGRAMS

Study Objective: Interest in community as the focus of public health i nterventions is growing. However, choosing intervention and comparison neighborhoods when designing community based programs poses a challen ge to program planners. Ideally intervention neighborhoods should be c hosen based upon risk profiles and demonstrated need for the program. Multiple sources of data that tap into neighborhood characteristics mi ght be used to facilitate the selection of intervention and comparison neighborhoods for program implementation and evaluation. Design: We p resent and compare selected characteristics of two analytic methods th at can be used to create perinatal risk profiles of neighborhoods with in cities. For our example, we used information several sources of rou tinely available data asn used census tract level low birthweight as o ur intervention or outcome variable. Main Results: At the neighborhood level, we found average household wealth of the census tract, proport ion of births to women with late or no prenatal care, proportion of te en births per census tract, per capital crime rates, proportion of hou sing violation, and number of community organizations as being importa nt factors identifying neighborhoods at risk for high rates of low bir thweight births. Advantages of both methods are discussed and risk pro files generated from either method can be used not only to identify hi gh risk areas of the city for adverse perinatal outcomes but also for the identification of intervention and comparison neighborhoods for im plementation of community based programs. (C) 1997 Elsevier Science Lt d. All rights reserved.