INTRODUCTION

In recent years, the vocabulary of empowerment has entered the discourse on HIV/AIDS prevention. “Empowering” interventions have been directed to AIDS educators, as well as to persons with AIDS, young gay men, prostitutes, and numerous other populations at risk. In published literature and conference presentations, empowering strategies are variously operationalized as AIDS education, partner negotiation training, community organizing, case-management, outreach, self-help groups, consciousness raising, organizational networking, leadership training, and individual and group-based problem solving.
Although relatively new to behavioral scientists working in the primary prevention of HIV, empowerment principles have a long history in public health and community psychology. The popularity and seeming durability of empowerment have not, however, produced broad consensus on its meaning, measurement, or implementation. Since empowerment rhetoric can serve many different agendas (Strawn, 1994), the need is urgent to examine empowerment, its core assumptions, and its implications for how we respond to the challenges of HIV/AIDS. 

BACKGROUND

The intellectual history of empowerment is long and branching. As old as democracy, with citizen participation as its cornerstone, as new as the current administration’s “empowerment zones”, empowerment goals and strategies figured prominently in the organizing of newly freed African-Americans during the Reconstruction period in the American South, the American labor movement of the 1930s and 1940s (Minkler, 1990), the American civil rights and women's liberation movements of the 1960s and 1970s, in public policy (the Great Society), and in the scientific debate on issues as diverse as marital decision making (e.g., Frieze and McHugh, 1992), the prevention of coronary heart disease (Bracht et al., 1994), and the mobilization of persons with disabilities (Fawcett et al., 1994). Empowerment principles have also entered the arena of HIV/AIDS prevention, primarily through the disciplines of public health and community psychology.
The roots of empowerment in public health extend back to the formation of the discipline in the mid-1800s. Since early efforts to improve sanitation in urban areas (Institute of Medicine, 1988; Winett et al., 1989), public health practitioners have recognized that environmental and community factors influence health. In addition, community involvement in the diagnosis and solution of health problems is a long-standing tenet of public health: practitioners were encouraged to listen to the concerns and problems identified by community residents (i.e. “starting where the people are”) (Minkler, 1990) and to “[promote] health and efficiency through organized community effort” (Winslow, 1920). Community organizing approaches were especially prominent in public health in the 1960s and 1970s, influenced by the work of grass-roots political activists (Alinsky,1972) and educators (Freire, 1983).
Despite early recognition of the embeddedness of health behavior in culture, geography, economic and political circumstances, health interventions continued to focus on individual characteristics, such as motivation and skill, to change premeditated and voluntary behaviors, such as getting vaccinated and complying with a treatment regimen (Green and Kreuter, 1991). By the mid-1980s, however, health promotion was being reformulated, with empowerment as its ideological center and citizen participation as its primary strategy. Within this framework, public health practitioners sought to promote health through the “empowerment of individuals to take control of their own health, of communities to decide how best to control their quality of life, and of the marginalized to define their own needs” (Stevenson and Burke, 1992).
Somewhat independently, the concept of empowerment germinated in a second discipline, community psychology. In the early 1970s, this discipline gained visibility when Iscoe (1974), Cottrell (1976), and others encouraged health professionals to focus on characteristics of communities, such as community competence, as well as on characteristics of persons, such as individual motivation. The role of the health professional was to facilitate the community's efforts to solve its own problems. Since Rappaport’s early formulations, the concept of empowerment has been widely discussed in the community psychology literature (see, for example, Rappaport, 1987; Serrano-Garcia and Bond, 1994). 

HIV prevention
From the beginning of the epidemic in the early 1980s, the response to HIV was shaped by a bourgeoning literature on cognitive behavioral models of behavior change, including the Stage of Change model (Prochaska et al., 1992), Social Learning Theory (Bandura, 1986), and the Theory of Reasoned Action (Fishbein and Ajzen, 1975). Persons practicing behaviors that put them at high risk for acquiring HIV were recruited into individual counseling and group-based interventions providing behaviorally specific information and skill training. These strategies were sometimes embedded in more comprehensive interventions including peer support, tailored services, and targeted media. The ability to implement theory-based, individual-level interventions in controlled settings and to evaluate them with relative ease made these approaches attractive to researchers and HIV prevention providers who needed to know “what works” to reduce the transmission of HIV (O'Reilly and Piot, 1996).
The effects of interventions focusing on individual-level change have not been trivial. Cognitive behavioral and skill-building interventions have produced small-to-moderate effect sizes in various populations, including runaway youth, men who have sex with men, and high-risk urban women (Choi and Coates, 1994; Kalichman et al., 1996; NIMH/NIH, 1996; NIH Consensus Meeting, 1997).
Despite these successes, there is growing interest in alternative approaches, sparked in part by the success of recent HIV interventions directed to entire communities or populations (Kelly et al., 1992; CDC, 1996; Kegeles et al., 1996). This interest also reflects a number of converging trends and issues in HIV/AIDS prevention. These include the practical difficulties in engaging and retaining some populations in government-supported research, the demand for more cost-effective and culturally sensitive intervention approaches, as well as increasing debate about the long-term sustainability of community-wide behavior change (e.g., Altman, 1995), the role of community participation in planning research and programs, and the special needs of women.
In addition, there is growing concern about the limitations of models that focus solely on proximal, individual-level determinants of behavior (such as attitudes or beliefs), that are shaped primarily by the priorities of funders and researchers external to the communities being studied, and that do not explicitly address the capacity of communities to define and address their own health concerns. According to some writers (e.g., Freudenberg and Zimmerman, 1995), public health has already embraced the concept of comprehensive or holistic approaches that link HIV/AIDS prevention to the “broader mission of community development and social justice”. In his recent plenary address to the XI International Conference on AIDS, Parker (1996) echoed this view, noting a gradual shift from “more individualistic and information-driven notions of AIDS education to more multi-dimensional models of community empowerment and community mobilization”. To fully understand what this shift entails, it is necessary to examine the meanings attached to the concept of empowerment. 

The empowerment approach
Empowerment has been variously defined as a value orientation, as a set of processes, and, more rarely, as specific outcomes. Rappaport, whose early work on empowerment framed the mission of a new discipline (community psychology), defined empowerment as “the process by which people, organizations and communities gain mastery over their lives” (Rappaport, 1984).
Wallerstein (1992), writing for a public health audience, defined empowerment as “a social-action process that promotes the participation of people, organizations, and communities towards the goals of increased individual and community control, political efficacy, improved quality of community life, and social justice”. Empowerment principles are also integral to participatory research in health promotion (Green et al., 1995).
Others (e.g., Hawe, 1994) have defined empowerment as specific outcomes, including increased political efficacy and social participation at the individual level, increased opportunities for shared leadership and decision-making at the organizational level, and increased competence to collectively identify and solve problems at the community level. Implicit in most formulations of empowerment is the assumption that empowerment can refer to different content areas (e.g., sexual negotiation, economic self-sufficiency) at different levels of analysis and practice (e.g., individuals, couples, groups, organizations). Individual-level, or psychological, empowerment has been widely discussed in the literature (e.g., Zimmerman, 1990, 1992, 1995). Our focus here, however, is on community empowerment as a framework for developing interventions to effect enduring and community-wide change in HIV transmission rates.

COMMUNITY EMPOWERMENT INTERVENTION

Definition and core assumptions
Our reading of the literature suggests the following working definition of a community empowerment intervention focusing on health: 

A community empowerment intervention seeks to effect community-wide change in health-related behaviors by organizing communities to define their health problems, to identify the determinants of those problems, and to engage in effective individual and collective action to change those determinants.

This definition is predicated on three assumptions: first, that health problems have multiple determinants, often including those which lie beyond individual volition and skill; second, that communities must participate in both the definition and the solution of health problems; and third, that the success of an intervention depends on the capacity of the community to engage in effective action.

The person-in-environment
Empowerment recognizes that “virtually no behavior is under the complete and voluntary control of individuals” (Brown, 1991). In this view, talking to a sex partner about condoms, avoiding anal intercourse, and leaving a sexually abusive situation are not isolated, voluntary acts but part of “socially conditioned, culturally embedded, economically constrained patterns of living” (Green and Kreuter, 1991).
Extending rather than replacing psychosocial models of behavior change, empowerment widens the lens to include not only persons at risk but the social and structural context in which HIV transmission occurs (Coates and Greenblatt, 1990; Ewart, 1991; Aggleton et al., 1994; Institute of Medicine, 1994; Sweat and Denison, 1995; Gillies et al., 1996). For example, an intervention directed toward the female sex partners of injecting drug users may use a variety of strategies to motivate women to reduce their risk for sexually transmitted HIV, to foster the belief that other women in their situation are attempting to reduce their risk, and to increase women's skills in “negotiating” condom use with their partners. In recognition of the context in which women make and implement sexual decisions, intervention activities would also seek to change those determinants which lie beyond the control of individual women. These may include characteristics of: the cultural environment, such as gender roles which define women as subordinate to men; the physical environment, including access to appropriate services, such as battered women's shelters, and products, such as condoms and spermicides; the structural environment, such as opportunities for women to become economically self-sufficient; and the policy/legal environment, such as businesses providing paid leave for community service and child care. The primary audience for an empowerment intervention to reduce women's HIV/AIDS risk may be the women themselves, while secondary but no less important target audiences might include community leaders, employers, peers, providers, and others who control rewards and resources for women in a particular community.
The ecologic or comprehensive approach to prevention, in which combined strategies are directed to different segments of the community, is not new to public health. Best known are the generously funded cardiovascular trials (e.g., Farquhar et al., 1990; Luepker et al., 1994; COMMIT Research Group, 1995a,b), which have produced mixed, and some would say disappointingly modest, results (Susser, 1995).
The elements that uniquely distinguish community empowerment from other ecologic approaches to prevention are, first, the level of community participation in prevention planning and implementation, and second, the explicit focus on community capacity-building.

Community participation in research
In their review article on the strategies and principles of community involvement, Bracht and Gleason (1990) defined citizen participation as “the social process of taking part (voluntarily) in either formal or informal activities, programs and/or discussions to bring about a planned change or improvement in community life, services and/or resources” (p. 110).
Community participation has long been a basic principle in health education and research but has been inconsistently practiced. Community participation in prevention planning and intervention research is a growing trend in public health, a cornerstone of the major demonstration projects in chronic disease prevention such as the COMMIT project (Thompson et al., 1995), and recently embodied at the policy level in HIV Prevention Community Planning at the Centers for Disease Control and Prevention (Valdiserri et al., 1995).
From an empowerment perspective, community participation in prevention planning and research is expected to empower individuals as well as communities: at the individual level, by developing community- and health-relevant competencies, such as group decision making and knowledge of resources; at the community level, by creating new networks, resources, and opportunities. Practically, the expected rewards of expanded community participation include greater buy-in from community leaders, better penetration of communities with more acceptable and culturally relevant messages, and greater sustainability of the intervention activities and effects (e.g., Hatch et al., 1993; Bracht et al., 1994). Despite accumulating empirical support for some of these assumptions (e.g., Altman, 1995), it remains to be determined whether community participation in a health-related intervention produces higher risk reduction rates than a similar intervention without such participation, and whether the effects of a participatory approach are more resistant to decay than other approaches (Freudenberg and Zimmerman, 1995).
Moreover, Stevenson and Burke (1992) cautioned that community participation in the research enterprise is an important but limited form of empowerment. More important are “the kinds of power that enable communities to implement the results of that research; to use research to gain resources, overcome barriers to health, and effect social change”. As noted by Altman (1995), communities may feel a sense of ownership and aspire to full control of community interventions but may lack the capacity to do so. 

Community capacity-building
A community empowerment approach to primary prevention seeks to produce specific health outcomes, as well as to create or expand the community’s “capacity” (Mayer, 1996) or “competence” (Eng and Parker, 1994). From this perspective, a community empowerment intervention seeks to increase the community's capacity to be self-determining in a specific domain: in the case of HIV/AIDS, the ability to integrate effective HIV interventions into existing community structures; to adapt interventions to changing conditions; and to apply relevant skills and resources to other community health concerns.
Capacity is assessed in terms of access to and effective use of resources: financial and material resources, including money, goods, and services; technical resources, individual skills and organizational capacities; and social resources, such as community leaders, strong community based institutions, coalitions of community organizations, and high levels of civic engagement, or citizen participation. Citizen participation in community groups and institutions produces “social capital” (Putnam, 1996) Ð that is, the interpersonal trust, norms, and networks that can influence health behavior directly, through normative pressure to change individual-level characteristics (such as frequency of condom use), and indirectly, through collective action to change community-level characteristics (such as restrictive gender roles and employment practices).
Community health interventions tend to focus on the first two kinds of resources, for example, by improving services and by providing health-related skill building opportunities to at-risk individuals. A community empowerment intervention also focuses on the expansion of social resources, primarily through community development and mobilization, or “community organization”.
Bracht and Kingsbury (1990) described community organization as “a planned process to use [community] social structures and any available resources to accomplish community goals, decided primarily by community representatives and consistent with local values” (p. 67). However, communities may vary in their capacity or readiness for collective action. As noted by Altman (1994), the rapid and widespread mobilization of urban, predominantly White and gay-identified men in the United States around HIV/AIDS would probably not have occurred without the organizational base (gay clubs, bars, businesses) and community identification (Gay Pride) that has emerged over recent decades and was precipitated in part by the Stonewall riots a decade earlier in New York. In a community without identified leaders, mature community-based organizations, and a history of successful problem-solving, mobilization around a health issue may not be possible without a prior investment in community development: creating new networks, strengthening existing networks and community institutions, creating opportunities for civic engagement, building consensus on the relevance of a public health concern to community well-being, as well as motivating and training community members to lead and participate effectively. Development may also entail the provision of economic (services) and material resources (shelters, condoms, meeting space, and supplies).
Although it is not yet clear what baseline level of community capacity is needed to organize a community around a specific issue, there is growing appreciation of the importance of selecting intervention strategies that are compatible with the community's stage of readiness. Indeed, “the application of an effective strategy, but at an inappropriate stage, can actually delay or disable a project” (Goodman et al., 1996).
Various process models for organizing communities are described in the literature (e.g., Bracht and Kingsbury, 1990; Green and Kreuter, 1991; Fawcett et al., 1995). Differing in focus and specific strategies, these models have common elements, including varying levels of community participation in assessment, intervention planning, implementation, evaluation, and dissemination of findings; a focus on “community as the center of gravity for health promotion” (Green and Kreuter, 1991); and the commitment both to positive health outcomes and empowerment outcomes.

EMPOWERMENT AND HIV PREVENTION FOR WOMEN

The potential relevance of empowerment to the primary prevention of HIV is particularly striking in the case of women at risk for HIV. Amaro (1995) writes that “condom use is a sexual behavior that is clearly under the control of men and is embedded in a socially sanctioned inequality between men and women ...[a] fact not captured in existing models of sexual risk behavior”. Models appropriate for urban, gay-identified men may have limited applicability to women whose choice and action are constrained by poverty, gender roles, and cultural norms (Ickovics and Rodin, 1992).
Qualitative and quantitative studies of women in the United States suggest that gender inequality constrains women's risk reduction in many ways. Poverty, particularly among women of color, limits women's opportunity to choose both their nonsexual relations (Belle, 1983) and their sexual relations with men (Merzel, 1991; Ickovics and Rodin, 1992). Threat of partner violence may reduce some women's willingness to challenge or refuse sexual relations or to define the terms of those relations (Frieze and McHugh, 1992; Heise, 1993). Gender- based inequality has also been linked to restricted access to appropriate information, services, and support (Gupta and Weiss, 1993), as well as low participation in community organizations or other social change institutions (Ickovics and Rodin, 1992).
Culturally based and reinforced by popular media, gender roles may reinforce women's dependence on their partners for sexual satisfaction, social acceptance, and self-esteem. Numerous studies suggest that restrictive gender roles are particularly relevant to the health-protective behavior of HIV-vulnerable women (Worth, 1989; Holland et al., 1992; Wyatt, 1994); African-American women (Fullilove et al., 1990; Gasch et al., 1991; Wingood and DiClemente, 1992; Pivnick, 1993; Shervington, 1993); and Latinas (Romero et al., 1993). There is growing concern that interventions designed only to change health-specific attitudes and competencies can benefit highly motivated program participants but do little to address cultural, structural, or other conditions that enhance women's vulnerability (e.g., Wingood and DiClemente, 1996).
In the absence of well-described theoretical and practice models for fostering empowerment, the term has been used somewhat promiscuously to cover an array of HIV/AIDS interventions directed toward women, including support and skill building in problem-solving, sexual assertiveness, condom use, or, more rarely, life skills (e.g. Rhodes et al., 1992; Levine et al., 1993; Prototypes, unpublished manual; Romero et al., 1993). Case management, outreach, and the provision of services have also been described as empowering for women (Gutierrez, 1990; Nyamathi and Lewis, 1991).
However, recent work by Robertson and Minkler (1994), Zimmerman (1995), as well as Israel et al. (1994) makes clear that empowerment focuses not just on health-specific risk beliefs and behaviors in individual women, but on beliefs and practices that are linked theoretically to interpersonal, organizational, and community change. At an individual level, these would include increased awareness of the link between individual circumstances and community characteristics; improved leadership, problem-solving and communication skills; and increased participation in community organizations (such as neighborhood coalitions) and activities (such as voting) which can exert influence on community characteristics through collective action.
Additional intervention activities may be directed to the organizational level (e.g., creating expanded opportunities for women to participate in the groups and organizations that influence their lives) and to the community level (e.g., increasing collaboration among women-focused community groups and institutions). Although a discussion of empowering organizations is beyond the scope of this paper, preliminary studies suggest four relevant organizational characteristics, including shared beliefs and values compatible with an empowerment perspective, opportunities to participate in decision making and leadership, supportive resources (such as communication skills training), and effective leadership or mentoring (Papineau, 1994). Key to community empowerment is the assumption that the women themselves, rather than social planners or external change agents, drive the social change effort. Not all community characteristics can be changed, at least in the short run. Community mobilization will not change the sex ratio (where women have too few eligible partners) but can influence the cultural norms that define women without men as inferior (e.g., by forming coalitions to change media depictions of women).

PARADIGM DRIFT IN HIV PREVENTION

Although there is little evidence in HIV prevention of a paradigm shift Ð that is, the widespread adoption of empowerment as a guide to theory, practice, and research Ð there are trends suggestive of paradigm drift, including the increasing interest in the social and environmental determinants of HIV transmission, as well as efforts to expand community participation in HIV/AIDS research, and to increase community capacity to develop and sustain HIV prevention programs.
One example of “drift” is a CDC-funded intervention developed for young, pregnant women from low-income neighborhoods in a midwestern city (Levine et al., 1993; Hobfoll et al., 1994). Women were recruited from obstetrics clinics and randomly assigned to three groups: a four-session HIV/AIDS prevention group, a four-session general health promotion group, and a control group.
The HIV prevention group concentrated on enhancing women's skills in negotiating condom use with their partners through role-playing, guided imagery, cognitive rehearsal, and other methods. The content of the group sessions addressed not only HIV prevention, but other health concerns, such as smoking during pregnancy. In addition, the sessions were designed to promote a sense of “communal mindedness” (Hobfoll et al., 1994), whereby women supported each other in their efforts to change their behaviors, and helped each other to overcome problems. Women participating in the HIV/AIDS prevention group did better than women in the other groups in their intentions and actual practice of safer sexual behaviors.
The authors note that the intervention was empowering because it increased women's perceived control over their lives. It is not clear, however, if their increased “communal mindedness” translated into increased community competence - that is, did the skills and commitment to mutual support in the small group experience prepare women for effective participation and collective action outside the group? It should also be noted that although women were encouraged to contribute to the content and topics of the group discussions, their participation in the design and implementation of the intervention was otherwise limited.
An example of broader community participation in an HIV prevention program is the Mpowerment Project, a community-level intervention for 18- to 29-year-old men who have sex with men. A “core group” of 12±15 young gay men in Eugene, OR, was formed to develop a peer intervention. Core group participants created the project name and logo, designed intervention materials, and planned and conducted program activities, including peer outreach, small group meetings, and a small publicity campaign. The goals of these activities were to create non-sexual venues where young gay men could meet other men, talk with each other about HIV prevention, and spread positive messages about being gay and adopting safer sex behaviors. Outcome data from a cohort of young men indicated a decrease in the proportion practicing unprotected anal intercourse with both non-main partners and boyfriends (Kegeles et al., 1996).
Consistent with an empowerment approach, young men from the community rather than outside experts had primary responsibility for designing and implementing key components of the intervention. This level of community participation is expected to promote both a sense of ownership of the current intervention and skills generalizable to other health concerns of young men who have sex with men.
Two recently completed multi-site HIV prevention intervention studies funded through the Centers for Disease Control and Prevention (CDC) also exemplify paradigm drift. The AIDS Community Demonstration Projects and the Prevention of HIV in Women and Infants Demonstration Projects (WIDP) enlisted populations at risk for HIV to deliver community-relevant HIV-prevention messages and condoms (or bleach kits for injecting drug users) to their peers, as well as provide social support for behavior change (O'Reilly and Higgins, 1991; CDC, 1996). The WIDP encouraged broader community participation by systematically identifying and recruiting multiple segments of the community (e.g. residents, businesses, and neighborhood organizations) to participate in the program (Person et al., 1996).
The interventions were effective in changing HIV risk behaviors (CDC, 1996), as expected. Another positive, and unexpected, outcome was the increased capacity of community members to address community problems. For example, women participating in one of the WIDP interventions created their own community-based organization to continue HIV prevention work after federal funding for the study ended; in another city, strong collaborative relationships between researchers and community members who participated in the ACDP study led to new intervention programs based on needs identified by the community (Person, personal communication, 1996; Guenther-Grey et al., 1996).
Each of the four projects described above have taken the first steps toward incorporating the con cept of empowerment into HIV prevention; however, two current CDC-sponsored projects suggest how the future of HIV prevention is “drifting” closer to empowerment as a guiding concept.
The Prevention Marketing Initiative (begun in five cities in 1993) is a demonstration project which seeks to test both an intervention method (combining social marketing principles and behavior science theory) and a process (development of the intervention for youth through extensive collaboration between local and national partners). Local planning committees (comprised of representatives from HIV and youth-services agencies) participate in all aspects of the project, including reviewing and interpreting data collected during formative research, and developing prevention marketing interventions. Youth committees also participate in the design and conduct of project activities. Through this project, community members are gaining social marketing and evaluation skills that are expected to strengthen these communities' capacity to implement successful HIV prevention programs in the future (Batelle, 1996).
Finally, the Urban Research Centers Project, funded by the Centers for Disease Control and Prevention in 1995, takes a comprehensive approach to health promotion, recognizing that community members may be concerned about, and affected by, a variety of health problems including (but not necessarily most important) HIV/AIDS. In three cities, health departments, universities, and community-based organizations have formed partnerships in order to systematically examine the quality of life and health status among inner-city urban populations. The goal is to better understand the relationships between socioeconomic or structural factors (poverty, low sex ratio, weak civic engagement) and endemic health problems, such as infectious diseases (tuberculosis, STDs, and HIV), teen pregnancy, domestic violence, and substance abuse. Each project site will implement new interventions or evaluate current interventions intended to promote individual behavior change, environmental change, and increased community capacity to address health concerns (CDC, 1994). Again, community representatives take an active role in all aspects of the design and implementation of this research, with the ultimate goal being the long-term maintenance of successful interventions with community resources (University of Michigan School of Public Health, 1997). 

CURRENT BARRIERS TO PARADIGM SHIFT

Despite its broad appeal, empowerment as a health goal is unlikely to have a dramatic impact on the way we do business in HIV/AIDS prevention until a number of issues have been resolved at both the theoretical and practical levels (e.g., Riger, 1993; Choi and Coates, 1994; Israel et al., 1994).

Theory and measurement
If empowerment theory is to provide a framework for designing interventions, the linkages must be fully described between changes in macro-structure (such as the formation of new coalitions to decrease discrimination against homosexuality in communities of color), changes in micro-structure (increased number of supportive peers in the social networks of previously isolated young gay men), and changes at the individual level (increased participation among young black men in community groups and institutions).
Moreover, current theoretical models do not specify the conditions in which empowerment approaches are most appropriate. Labonte (1994) warned that empowerment is not a remedy for all problems in all contexts. Yet no theorist makes clear when an intervention focusing both on cognitive behavioral change and increased community capacity is more likely to have an impact on community-wide HIV transmission than an intervention designed to increase access to condoms, build partner negotiation skills, and stimulate normative support for risk reduction.
In an e.ort to build the necessary empirical foundation for empowerment, investigators in community psychology, social epidemiology, and health promotion have been systematically examining the links between social participation, community identification, and perceived control or influence (Wandersman, 1981; Chavis and Wandersman, 1990; Florin and Wandersman, 1990; Zimmerman, 1992); operationalizing concepts such as community capacity (Speers, personal communication, 1996); developing protocols for assessing and increasing community competence (Eng and Parker, 1994); and attempting to design measurement models capable of assessing outcomes of multifaceted interventions tailored to community-defined goals and priorities (Steckler et al., 1995). In an era when scientific rigor is equated with controlled trials with random assignment, the last task seems especially formidable.
Despite recent advances, considerably more work remains to be done in decomposing and operationalizing such constructs as community capacity and, even more basic, “community”. Another issue to be explored is the determination of community readiness-for-change. Assuming that some irreducible minimum of social, financial, or technical resources are needed to mobilize a community around a specific issue, then how much is enough? In addition, since social change is likely to be slow, interim change measures will be needed, perhaps analogous to stage-based behavior change at the individual level.
Hawe (1994) cautioned us that if empowerment is to become more than a value stance, concepts must be translated into recognizable outcomes. Flora et al. (1995) suggest that empowerment approaches may have multiple outcomes at di.erent stages of the change process and at di.erent levels of intervention. Interim outcomes of a community intervention to reduce the transmission of HIV/AIDS might include increased community awareness of community-level influences on HIV transmission, attitudinal changes in community leaders, proposed policy changes in community institutions, such as school boards and civic groups, and increased leadership capacity in community segments most affected by HIV. Endpoint outcomes, on the other hand, might include aggregate change in condom use and perceived normative support for change among high-risk community subgroups, as well as increased community-based resources for HIV/AIDS prevention and other community problems.
Similarly, intermediate outcomes in a community empowerment intervention to reduce HIV infection rates in women might include: increased opportunities for women to participate in group-based learning and support activities, expanded gender sensitive health services, and training in communication and community organization skills for natural leaders in the women's community (e.g., Israel et al., 1994; Amaro, 1995). Longer-range objectives might include changing gender-role stereotypes through media and school-based education and by recruiting men into AIDS caregiving roles; decreasing female poverty through job training and the development of community-based child care services; and facilitating the mobilization of women around reproductive and sexual health needs, such as affordable female condoms (Fullilove et al., 1990; du Guerny and Sjoberg, 1993; Amaro, 1995).
New outcomes require new measures. The Community Indicators Project, now underway in CDC's Behavioral Intervention Research Branch, Division of HIV/AIDS Prevention, seeks to develop valid and reliable indicators of community characteristics, such as civic engagement and community capacity, that are theoretically linked to HIV/AIDS-specific outcomes, such as aggregated condom use. The results of this study are expected to provide a foundation for more accurately assessing the impact of community-level approaches to HIV/AIDS prevention and for systematically intervening in social or structural conditions that exert direct or mediated influence on individual health behaviors (Kraft, personal communication, 1997). 

Implementation
More challenging than the need for theory and measurement development are heavily contested issues concerning implementation, not least of which is discovering how to implement social change strategies within the constraints of categorical funding. Also problematic are: how to practically define “community”; how to insure the “parity, inclusiveness, and representativeness” of community participants (Gruber and Trickett, 1987; Raeburn and Beaglehole, 1989; Woelk, 1992; Gomez, 1994); health professionals' lack of experience in group process, consensual decision making, and other community-focused skills (Valdiserri and West, 1994); and conflicting priorities of funders and community stakeholders (Eisen, 1994; Fawcett et al., 1994). The challenge of the empowerment paradigm is to find a model for priority setting that balances community concerns with the priorities of researchers and the constraints of categorical funding.
Despite the market appeal of “expanded community participation” to expert-wary consumers and politicians, the barriers to implementing community participation are daunting. Not least of these is the challenge of balancing the needs for scientific rigor and fiscal responsibility with the urgent realities of multiply stressed communities (Eisen, 1994; Fawcett et al., 1994). Assuming that some level of community participation is both necessary and desirable, it remains to be determined which participatory models - advisory boards, councils, coalitions, or other forms of partnership - produce better health outcomes, more sustainable activities and effects, and more empowered communities.
In addition, after more than a decade of chronic disease prevention trials that sought to mobilize communities around health issues, such as smoking or cardiovascular disease, the impact of community participation and social change on specific health outcomes remains to be determined. As noted by Fawcett et al. (1993), such projects relied on different models of community participation and community mobilization and in general failed to adequately document the extent to which intervention effects and activities were sustainable and generalizable to other community problems.
Cowen (1985) urged restraint in embracing empowerment theory as a framework for large-scale interventions, noting that “social reform [remains] an ideal, surrounded by uncertainty about procedural steps and outcomes”. Until we know whom to empower, as well as how to implement, evaluate, and pay for structural change interventions, person centered strategies “can at least help to achieve a shorter-term objective” (p. 45). As previously noted, however, community empowerment approaches seek not to replace the traditional person-centered approaches (ie. psychological models of behavior change) but to widen the lens to include the person-in-environment. Continuing efforts to operationalize core constructs and to test specific empowerment strategies in small-scale studies will provide a foundation for the large-scale intervention studies that are needed to assess the utility of community empowerment as a framework for HIV prevention. 

AGENDA

Technologic advances in containing the in vivo replication of the AIDS virus and in treating HIV-related infections and diseases do not moderate the urgent need to develop effective behavioral interventions for both HIV-infected persons and persons at risk. New models are needed for HIV prevention efforts that take into account social, cultural, structural, and political influences on behavior and that incorporate community development and mobilization strategies (Amaro, 1995).
Community empowerment has been proposed by some theorists (e.g., McKnight, 1985; Minkler, 1992; Schulz et al., 1995) as a new paradigm for addressing health problems and other community concerns. However, numerous issues must be dealt with before empowerment can fully inform our efforts to intervene in the deadly epidemic of HIV/AIDS. The concept of empowerment, its core assumptions, and illustrative examples of HIV interventions inspired by the concept have been discussed in this paper. However, a fully articulated model, specifically and systematically adapted to HIV prevention, has not yet been developed. Despite the expressed commitment to community participation in the development of culturally relevant, effective, and sustainable interventions, HIV prevention research and programs continue to focus on persons in virtual isolation from the political, economic, and cultural realities of their lives.
The gap between empowerment rhetoric and empowerment practice is unlikely to be bridged without further progress in operationalizing new concepts and testing hypothesized relationships between such variables as community participation and increased community capacity to effectively address health issues. As previously noted, the measurement of these concepts and their application to public health problems have only begun (Zimmerman, 1992; Eng and Parker, 1994; Israel et al., 1994; Schulz et al., 1995). As interventions based on empowerment theory are developed and implemented, it will be critical to determine whether community development and empowerment are linked empirically to positive health outcomes. Such an undertaking is likely to demand new forms of partnership and long-term involvement with communities.
Finally, implementing research and health promotion programs based on an empowerment model will require a reexamination of the role of the researcher or the health professional and the skills and competencies needed to fulfill that role. There is already a greater awareness in the professional community of the need to enlist the participation of communities in more than a superficial or perfunctory way in programs and research. To facilitate the meaningful participation of community members in the planning and implementation of HIV prevention programs, the role of the researcher or the health professional must shift from that of an expert who enters a community to deliver an intervention toward that of an advocate, collaborator, or mentor who assists community members in developing the resources, skills, and social networks to implement and maintain these programs (Kelly, 1995; Serrano-Garcia, 1984; Robertson and Minkler, 1994; Fahlberg et al., 1991; Swift and Levin, 1987; Biegel, 1984; Hatch et al., 1993).
In this new role, the researcher or health professional may be called upon to exercise skills in community organizing, group facilitation, conflict resolution, public speaking, or advocacy (Braithwaite et al., 1994; Valdiserri and West, 1994). Unfortunately, such skills are rarely taught in schools of public health or behavioral science (Eng et al., 1992; Steckler et al., 1995).
Adapting an empowerment perspective to HIV prevention and other public health concerns is a daunting challenge, because it requires us to consider major changes in the way we think about our public health mission, where we direct our interventions, how we work with communities, and how we develop, and provide funding for, research and programs. Yet this effort also represents an opportunity for public health professionals to invest in long-term partnerships with communities that have the potential to facilitate broad, sustained, positive changes in health conditions well into the 21st century.