Theory and practice: applying the ecological model to formative research for a WIC training program in New York State (2024)

,

Article history

Abstract

Introduction

Using formative research methods to conduct needs assessment provides data for strategy development and identification of objectives (Helitzer-Allen and Kendall, 1992). They are used for development of behavior change programs both in the US and in developing countries (Flora and Farquhar, 1988; Lefebvre and Flora, 1988; Schechter et al., 1990; Spoth, 1990, 1991; Woods et al., 1991; Morris, 1993; Lloyd et al., 1994; O'Donnell et al., 1994; Mathews et al., 1995). Formative research presents information on target audience beliefs, values, attitudes, knowledge and behaviors related to the health problem of interest, and seeks to answer questions about the context that influences, and is influenced by, these individual factors.

In designing and implementing formative research, it is useful to apply a conceptual framework to help describe contextual influences on behavior and assess optimal intervention entry points (Clark and McLeroy, 1995). The Precede–Proceed model conceptualizes the reciprocal relationship between behavior and environment into three groups of influential factors (Green and Kreuter, 1999). Predisposing factors are those that make a health-conducive lifestyle change possible, such as information or availability of products necessary for the new behavior. Enabling factors are those skills needed to implement the new behavior; and reinforcing factors are supporting values and social norms that help individuals maintain the new behavior. The ecological model (McLeroy et al., 1988) adds further detail by systematically categorizing these factors into five levels of influence: (1) the individual level, including beliefs, values, education level, skills and other individual factors; (2) the interpersonal level, including interpersonal relationships between individuals; (3) the organizational level, which covers the way relevant institutions are organized and managed; (4) the community level, including the communities that individuals operate in (e.g. professional networks, associations, neighborhoods), community attitudes and the relationship among different institutions within communities; and (5) the policy level, which refers to policies and regulations affecting intervention participants and the institutions in which they function.

The ecological framework has gained increased recognition in the field of health promotion (McLeroy et al., 1988; Stokols, 1992; Green et al., 1996; Richard et al., 1996; Stokols, 1996; Stokols et al., 1996a) and has been applied to investigations of many different health issues (Kumpfer and Turner, 1990; Earp et al., 1995; Breslow, 1996; Stokols et al., 1996b; Wandersman et al., 1996). Some researchers have used social ecology to guide program development. Goodman et al. (Goodman et al., 1996), for example, applied the ecological approach to their evaluation of how well a community-based intervention to prevent alcohol, tobacco, drug abuse and related risky behavior intervened at multiple levels, and how appropriate each set of strategies was for the community's stage of readiness. In developing, implementing and evaluating the community-based North Carolina Breast Cancer Screening Program, Viadro et al. (Viadro et al., 1997) used a variety of instruments to collect process data about organizational, program and individual service deliverer characteristics, as well as about external environmental influences. These data were used both for formative (advising program development and refinement) and for summative (assessing program effect) purposes.

There is a paucity of literature, however, that critically examines the systematic application of the ecological model to formative research and issues inherent in such a process. This article discusses the application of the ecological model to formative research in a practical setting of developing a training program. The example is drawn from formative research designed to help develop session content for a training program devoted to plotting and interpreting growth data, nutrition education and counseling for growth monitoring for providers involved in the Child Growth Monitoring Project (CGMP) of the New York State Special Supplemental Nutrition Program for Women, Infants and Children (WIC). The objectives of the formative research were to obtain information on provider and client perceptions of growth monitoring and counseling, counseling practices of WIC providers, and the organization of the WIC visit around growth monitoring. Main research questions are presented in Table I. We chose the ecological model to guide us in developing this research because it offered a concrete framework to account for the behavior–environment interaction in a practical setting. Because we knew from the start that the intervention would focus on training, we focused our efforts on collecting data at those ecological levels that we considered potentially amenable to change through a training program—individual, interpersonal and organizational.

Methods

Data collection took place during April 1994, with full-day visits to 13 WIC agency sites participating in the CGMP. Sites were selected by state-level WIC staff familiar with the CGMP and selection was based on individual constraints of WIC agencies during the research period, such as availability of providers involved in growth monitoring and schedule of growth monitoring activities. Data collection was scheduled to coincide with times of intense activity at WIC sites when clients were certified or recertified to receive WIC benefits (every 6 months).

Table I shows which methods tapped into which research questions at which ecological level. (For clarity, Table I presents all methods used in the research, although only examples of the resulting findings are discussed here. These are indicated.) Unstructured field guides were developed for each type of data collection method. All samples were purposively selected, according to appropriateness of informant (providers involved in growth monitoring, or designated caretaker of a WIC-enrolled infant or child), availability and willingness to be interviewed or observed. Informed consent was obtained from all informants. The study was approved by the Committee on Human Research of the Johns Hopkins University School of Hygiene and Public Health.

The data were analyzed as text using a coding scheme iteratively developed from the text. One principal coder (the first author) coded all data and all authors reviewed these initial coding results. Discrepancies in coding interpretation were resolved through discussion until consensus was reached. Domain analysis was conducted on the coded data (Spradley, 1979), allowing for themes and a theory of the problem (Helitzer and Wallerstein, 1999) to emerge out of, rather than be imposed on, the data.

Key findings and corresponding training program activity

Following are examples of key findings and how these affected development of the training program. (More detailed results are available from the first author.) The presentation by ecological level underscores the fact that data were not collected at all levels.

Individual: provider perspectives towards growth monitoring

In the literature, growth monitoring is defined as the entire process of tracking growth over time and counseling on observed patterns (Ruel, 1995). In interviews for the formative research, providers reported positive attitudes about the usefulness of growth monitoring in their work. Their definition of `growth monitoring', however, was limited to an assessment of anthropometric status at one point in time and did not include counseling about anthropometric results. With the information from the research, trainers attempted to expand providers' conception of growth monitoring by emphasizing the importance of using anthropometric data to track a child's growth over time, and using the observed growth pattern to discuss health and nutrition issues with the child's caretaker.

Interpersonal: provider–client interaction

Research has shown that clients who are involved in decision-making regarding follow-up and treatment feel in greater control over their health (Krouse and Roberts, 1989), and are more likely to have positive health outcomes (Greenfield et al., 1988). It is important to involve the client in discussion by getting her perspective from the beginning and building on this to negotiate realistic follow-up plans that the client is most likely to implement.

Many of the interactions observed between provider and client were not based on a negotiated approach to decision making. Counseling for certification tended to follow a pre-determined pattern of topics and was heavily driven by the forms that providers were mandated to review and complete. In the interactions observed, the child's growth pattern was briefly discussed toward the beginning of the visit. Few providers showed the growth chart to the client while discussing its contents. Although WIC providers appeared to be genuinely concerned about the clients they counseled and clients appreciated their efforts, there was little, if any, exploration of the client's perspective or back and forth discussion about what to do for the child.

During the training program, providers were encouraged to show the growth chart to clients as an entry point into the discussion on growth. The multi-step counseling process that providers were trained in was structured to work within the constraints of large caseload and limited time, and to build provider skills in finding out the client's perspective on her child's health and negotiating a follow-up plan.

Organizational: patterns of growth monitoring activities

To be most effective, anthropometry and counseling on growth and nutritional status need to be closely connected (Ruel, 1995). The research results indicated, however, that the sequencing of anthropometry and counseling on growth and nutrition within the WIC sites visited did not support a strong link between these two components of growth monitoring. For example, at three sites visited, counseling on growth and nutrition took place 1–2 months after anthropometric assessment. At the other 10 sites, although all clients received counseling on growth and nutrition on the same day as the child's anthropometric assessment, such counseling was kept brief at four of these sites. In-depth counseling took place 1–2 months later and, at some sites, was offered only to clients identified as high risk.

The link between anthropometry and counseling about child health and nutritional status was the overriding theme of the training program. Subsequent to the formative research, provider trainees were encouraged to consider growth monitoring as a framework for counseling WIC clients on their child's health. In order to promote a stronger anthropometry–counseling link among both staff and clients, the trainers advocated a team approach (nurse/nutritionists and paraprofessionals) to growth monitoring in those sites where staff of different professional grades conducted anthropometry and counseled about its results.

Discussion

Findings from the research showed gaps in the counseling skills of WIC providers, as well as organizational issues that did not promote the application of comprehensive growth monitoring as advocated in the literature. The CGMP training program targeted those issues that were amenable to change through a training intervention. The experiences of the training program, however, as well as post-training evaluation of the data by ecological level, pointed to important issues that could not be addressed by a training program, but that the trainers should have known about before the training program was designed. These related especially to issues at the interpersonal, community and policy levels, respectively, information that was not collected during the formative research.

• Interpersonal level: working relationships between WIC staff at different levels. Tensions between staff of different professional grades became evident during the training program, as did the lack of consensus among trainees as to the appropriate roles for different professional grades of WIC staff involved in growth monitoring. This suggested less than full support for a team approach to growth monitoring that paired paraprofessionals with nurses or nutritionists.

• Community level: supporting networks among WIC staff. Lack of information about formal and informal professional networks of WIC staff meant that trainers could not suggest ways that such networks could help support providers in implementing their new skills.

• Policy level: support from regional and state-level WIC management for a growth monitoring approach that closely linked anthropometry and counseling. The assumption that such support existed (without confirmatory formative research data) meant that strategies for reinforcing new ways of implementing growth monitoring at the site were not made explicit during the training program.

It is often the case in behavior change program planning that the type of intervention is decided on before information has been collected about intervention target groups—information that might suggest other or additional types of interventions as necessary for reaching the behavior change goal. Within these realities, and in light of resource constraints, intervention planners and formative researchers may easily focus on collecting information about factors that can be directly addressed by the planned intervention, thereby neglecting other important barriers to behavior change.

In this study, training was selected as the intervention before the formative research was planned. We focused our research on assessing individual, interpersonal (provider–client) and organizational issues, issues that could be targeted by a training intervention. As a result, we had significant gaps in information needed to develop an effective comprehensive intervention.

Two types of information are necessary in designing a training program: (1) information to develop the technical content of training sessions, and (2) information to understand how this technical content can be integrated and maintained by trainees in their professional routines following training. We collected the first kind of information, realizing only during and after the training program the importance of the second. Including the management and policy-making level in needs assessment is especially important for developing a training program, as a discord between the goals of the training program and of the trainee organization can hamper the efficacy of the training as well as sustained implementation of new skills (Fielding and Llewelyn, 1987).

Considerations for other research

Given the lessons we learned from designing and implementing the CGMP training program, we argue that full application of the ecological framework is crucial. The advantage of the ecological model is that it provides a framework to collect both content-specific and contextual information in a systematic way. A matrix of the model (ecological level with corresponding main research questions by source of information) can most effectively serve as a kind of backward decision-making tool (Andreasen, 1985) throughout the research process. The matrix can be used in developing the research to assure that research questions are posed at all levels. It can be used during the research to monitor data collection at all ecological levels and it can be used when the research data are being evaluated to organize findings by ecological level.

Table II presents an example of how such a matrix could have been used for this study. The crosses point to which kind of method would provide data at which ecological level to help answer which key research question. For example, had we implemented the ecological model fully, we would have added the following research questions: `What are characteristics of the professional relationships between providers and how do these relationships affect the implementation of a team approach to growth monitoring?', `What are formal and informal networks of providers across sites, and how can these support comprehensive growth monitoring?', `What are state-level WIC regulations and policies related to growth monitoring and how do these affect the implementation of the CGMP program?'. The latter question would have necessitated additional in-depth interviews with state-level WIC staff.

We argue that data need to be collected in the formative process to complete at least one cell in each row of the matrix. This is not to say that, in the context of limited resources and already-decided intervention strategies, equal weight should be given to all rows (ecological levels) in the data collection process. Technical expertise of the area of interest is crucial to help identify which rows in the ecological matrix need most attention in the data collection stage. Other rows should not be overlooked, but should be filled at a less intense level.

To take full advantage of the ecological approach, one needs to use it not only as a model for interpreting behavior patterns, but also as a framework for data collection. This promises to guide researchers to collect comprehensive and programmatically useful data.

The work reported in this paper was supported by a subcontract from Irwin J. Shorr Productions, under a prime award from the State of New York. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of Shorr Productions or the State of New York. The authors would like to thank the WIC staff and clients who participated in this research and without whom the study would not have been possible. The authors would also like to thank Mr Irwin Shorr and the staff at the New York State WIC office in Albany for their support of the research and their comments on earlier drafts of this paper. At the time of the study, G. N.-A. was a doctoral candidate in the Department of International Health, The Johns Hopkins University School of Hygiene and Public Health.

References

© Oxford University Press

Issue Section:

Original Articles

Email alerts

Citing articles via

More from Oxford Academic