Cancer among Black Families: Diffusion as a Strategy of Prevention and Intervention

Article excerpt


The Department of Sociology and Anthropology, the Association of Black Sociologists (ABS), and the National Outreach Initiatives Branch (NOIB)joined a cooperative agreement to strengthen and enhance the National Black Leadership Initiatives on Cancer (NBLIC). The general mission of this project is to reduce the incidence of cancer among African Americans. The specific aims were threefold: 1) to enhance cancer prevention in target communities by enlisting a broad coalition of African American social scientists in identifying and making accessible behavioral models and resources to change individual cancer-related behaviors; 2) to promote the dissemination and adoption of cancer prevention behaviors through NBLIC coalitions that are linked with participants from the National Congress of Black Churches (NCBC), local colleges and universities, and the National Association of Social Workers; and 3) to provide data that lead to modification of community social norms and practices in the direction of health promotion in affected communities.

Five urban metropolitan areas were selected for inclusion in the project evaluation based on three factors: 1) the presence of an NBLIC coalition in the city; 2) a high concentration of African American residents; and 3) high rates of cancer among African Americans. The cities chosen for the study are Atlanta, Houston, Miami, New Orleans, and Washington, D.C. These sites serve as the focal points for technical assistance, local organizing, intervention and evaluation of behavioral strategies to prevent cancer.

The project is currently in its fifth year, and the information here is based on research conducted as part of an effort to support cancer coalitions and to explore and follow-up on previous research findings. These preliminary findings with respect to women offer some interesting possibilities for intervention and prevention strategies.


The evaluation includes four components, each of which was designed as a means to collect information on changes in cancer-reduction activities. The first component consists of the collection of information from participants in cancer prevention activities, establishing the characteristics of persons in the community who take action toward cancer prevention. The second component, the community demographic profile, provides baseline and follow-up descriptive information about cancer prevention resources for each of the five sites explored in this study. This component includes the collection of secondary data on cancer incidence, prevalence and mortality for each site. The third component is a survey, in which baseline and follow-up information from representative samples of African Americans in six targeted cities was collected in order to assess the extent to which NBLIC coalition activities have permeated the community. Finally, the fourth component, a study of a subset of survey participants, provides more explanatory information about individual decision-making related to cancer prevention. The methodology for each component is described below.

Community Demographic Profiles

These profiles provide information about the current infrastructure supporting health promotion and cancer prevention in each community. This formal picture of the community was provided to coalitions to evaluate the match with perceptions of resources reported by survey respondents.

Sample: Community Demographic Profiles were used to gather information for each of the five cities included in the sample.

Instrument: The Community Demographic Profile was developed in order to construct a description of the community in which NBLIC coalition activities occur. In combination with the initial survey data, this information was made available to coalitions for planning purposes.

Responses from participants in cancer prevention programs and survey respondents are analyzed, using bivariate analyses to identify the important relationships among survey variables as well as to determine the effectiveness of specific interventions for achieving change in health promotion attitudes, knowledge, and the adoption of screening and early intervention behaviors.

Coalition Profile Study

Data Analysis: The coalition profiles were used to choose control versus intervention cities.

Baseline KAB Surveys of African American Residents of Target Communities

Purpose: The project outcome evaluation focuses on identifying and evaluating the most effective strategies for cancer prevention in African American communities. The independent variables are community health promotion strategies that are pursued by NBLIC community coalitions. Intervening variables (such as health promotion knowledge, awareness and behavior) that research indicates should influence the dependent variable of cancer incidence are included. Only two sampling points - a baseline survey and a follow-up survey of a ten-percent sample - are also included.

Sample Selection: A representative sample of African American residents in each of the five target cities was drawn with the assistance of the sampling firm, Survey Sampling, Inc. (N=275; response rate=275/420). Prior to drawing the sample, 1990 Census data were used to determine the numbers of African Americans in census tracts in each metropolitan area. This information was used to decide which county/counties to include or whether to use only city boundaries.

For each site, probability samples based on random digit-dialing techniques were drawn from a pool of all telephone numbers within telephone exchanges that most nearly coincided with city and/or county boundaries. For New Orleans and Washington, D.C., the city and county/parish boundaries coincide. For Atlanta, Houston and Miami, the samples were drawn for the county that most nearly coincides with the city boundaries (Fulton, Harris and Dade Counties, respectively).


In each of the five cities, nearly two-thirds of those interviewed are women (Table 1). The 1990 Census indicates that between 53 percent (Houston and Miami) and 56 percent (Washington, D.C.) of the African American population over age 18 are women (A contact had to be 18 or older to be interviewed). Since one would expect higher percentages of the persons reached for telephone interviews to be women, the gender distribution of the sample does not seem unreasonable, and it is highly consistent across the sites.


The age distribution of the total sample (all sites together) matches the census age distribution for the combined sites fairly well. The sample is generally within one or two percentage points of the census distribution except among 18- and 19-year-olds, who are overrepresented (9.6 percent compared to 4.7 percent), and among those 60 years old or over, who are underrepresented (13.1 percent compared to 18.0 percent).

With the exception of Washington, D.C., however, the sample differs slightly from census reports of age distributions for each city. Atlanta, Miami, and New Orleans have substantially larger percentages of respondents under 30 years old than their respective metropolitan areas. This age group is underrepresented, however, in the Houston sample and partially offsets the differential in the other sites. Washington, D.C. is the only site to approximate the percentage of persons 60 years and older. The reasons for the site variations in age distributions are not immediately obvious and merit further investigation. The high percentage of 18- and 19-year-olds at all five sites is of particular concern as it might suggest unevenness in following the screening protocols.

Each site's sample has considerably higher educational attainments than the African American population as a whole. Less than 15 percent in each city had not completed high school, compared to 35 percent or more in each population.


Thirty to 33 percent of the Atlanta, New Orleans and Washington, D.C. samples are high school graduates. These percentages are even higher in Houston (37 percent) and Miami (46 percent). From 22 to 25 percent of respondents in Atlanta, Miami, and Washington had attended college or obtained Associates degrees. This figure reached about 30 percent in Houston and New Orleans. About 16 percent of the Houston and Miami samples, 20 percent of New Orleans and Washington, and 27 percent of Atlanta respondents are college graduates, and over 1-in- 10 Washington respondents have a graduate degree. The percent of college graduates in each sample is at least twice that in the African American population in the 1990 Census, except in Washington, the percentages with graduate degrees are comparable. The overrepresentation of young adults in the sample and the underrepresentation of cohorts over 60 would contribute to the higher educational attainments of the sample.

Health Outcomes

Almost 4 of every 5 respondents (79 percent) indicated that they were doing things to improve or maintain their health. Atlanta respondents (76 percent) and Miami respondents (77 percent) were slightly less likely to take actions towards these ends, and residents of Houston and New Orleans (82 percent) were slightly more likely. High school graduates were less likely (73 percent) to try actively to maintain or improve their health than either high school dropouts (77 percent) or college graduates (89 percent). Respondents under 30 years old were somewhat less likely (76 percent), and those over 60 years old (82 percent) were somewhat more likely to undertake such efforts. The elderly were, however, not quite as likely to be taking actions to maintain or improve their health as 41- to 50-year-old respondents (83 percent).

Exercise and diet were each cited by about 70 percent of the African Americans responding to the survey as something they were doing or had done to maintain or improve their health. Much smaller percentages of the sample sought to maintain or improve their health by reducing or quitting smoking (13 percent), changing their lifestyles (11 percent) or medical practices (10 percent), and reducing alcohol intake (7 percent).

Less than 5 percent of the sample sought to improve their health by changing work habits, reducing exposure to sunshine, or taking actions concerned with the environment or estrogens (each less than 5 percent). Over 80 percent of our respondents who made efforts other than exercise and diet also indicated that they were exercising or changing their diet to stay or become healthy. About a quarter of those who had undertaken three activities to maintain or improve their health cited smoking, lifestyles, or medical practices as one of their efforts.

At the same time, respondents to the survey were very likely to take precautions against exposure to sunshine. Over three-quarters (76 percent) said they tried to avoid the sun on a sunny day, and nearly 6-in-10 respondents wore protective clothing on sunny days. A much smaller percentage (22 percent) used sunscreen or sunblock lotion. Over 90 percent of those who took two or more precautions against exposure to sunshine both avoided the sun and wore protective clothing on a sunny day.

About 3 in every 4 Atlanta respondents (76 percent) and Miami respondents (77 percent) indicated they were taking actions to improve their health. The proportion climbed to nearly 80 percent or more in Washington, D.C. (79 percent), and in Houston and Miami (82 percent).

Virtually everyone in the sample (97 percent) had heard of fiber, and this percentage fell to 95 percent only in Miami, among respondents with high school education or less and among those over 60 years old. The majority of respondents (56 percent) felt that their diet was medium in fiber, and another 17 percent characterized their diets as high in fiber. A little less than one-quarter (23 percent) indicated that their diets were low in fiber. This percentage was higher in New Orleans (27 percent) and Houston (30 percent) but fell to 20 percent in Atlanta and Washington, D.C., and to 18 percent in Miami. Houston respondents were the most likely (20 percent) to report that their diets were high in fiber.

The percentage of diets low in fiber decreased with greater educational attainment. Nearly one-third (32 percent) of those who did not complete high school reported low-fiber diets, while one-quarter of respondents who completed high school or some college did, and the percentage fell to 13 percent among college graduates. The latter, as well as respondents with some college, was notably more likely to report having diets high in fiber (about 20 percent) than high school graduates (15 percent) and dropouts (13 percent).

Diets low in fiber also declined with age, dropping from about 26 percent of those under 40 years old, to 20 percent of 41- to 50-year-olds, 18 percent of 51- to 60-year-olds, and to 15 percent of those over 60 years old. About 24 percent of the elderly, and 21 percent of those between 41 and 60 years old, but only 17 percent of 31- to 40-year-olds, and 12 percent of those under 30 years old reported their diets were high in fiber.

About 35 percent of the respondents felt that their diets were low in fat. This ranged from one-third of Atlanta respondents to 38 percent of those in Miami. Respondents in Miami were much less likely (14 percent) than Atlanta (22 percent) and Houston (24 percent) respondents to report having high-fat diets. College-educated respondents were especially likely (41 percent) to have low-fat diets, while high school graduates (32 percent) were the least likely to report diets low in fat. Respondents who did not complete high school were more likely than other respondents to have high-fat diets (24 percent), and this declined with increasing education to about 23 percent of high school graduates, about 20 percent of those with some college, and only 12 percent of college graduates.

Similarly, older respondents were more likely to have low-fat diets and less likely to report high-fat diets than younger respondents. The percentage of low-fat diets increased from about 27 percent of those under 30 years old to 32 percent of 31- to 40-year-olds, 38 percent of 41- to 50-year-olds, 43 percent of 51- to 60-year olds and over half (54 percent) of those over 60 years old. Conversely, only 10 percent of the elderly reported high-fat diets, and this grew to 13 percent of 51- to 60-year-olds, 17 percent of 41- to 50-year-olds, and about 23 percent of respondents under 40 years old.

Almost l-in-5 respondents to the survey (18 percent) indicated that they had been on a special diet for medical reasons during the prior year. About 22 percent of New Orleans and Washington, D.C. respondents reported special diets, while only 11 percent of Miami respondents and 15 percent of Atlanta did so. Special diets were especially prevalent among respondents who did not complete high school (25 percent) and among those over 60 years old (34 percent). They were less frequent among high school graduates (15 percent) and those under 40 years old (13 percent or less). About 20 percent of 41- to 50-year-olds and 27 percent of 51- to 60-year-olds were on special diets for medical reasons, suggesting that this increases steadily with age.

Exactly half of the sample indicated that in the past five years they made lasting and major changes in their diets for health reasons. A smaller percentage of Miami respondents (42 percent) had done so, with larger percentages of Atlanta (53 percent) and Washington (55 percent) respondents. College-educated respondents were considerably more likely (60 percent) and high school graduates less likely (44 percent) to have made major changes in their diets in the past five years for medical reasons. The likelihood of such changes increased steadily with the age of respondents, growing from about 41 percent of those under 30, to 48 percent of 31- to 40-year-olds, 58 percent of 41- to 50-year-olds, and 61 percent of 51- to 60-year-olds, before peaking at 69 percent of 61- to 70-year-olds. Those over 70 years old were far less likely (42 percent) to have made major and lasting changes in diet in the past five years, perhaps because they are likely to have done so earlier in their lives.

For almost three-quarters (74 percent) of the respondents who made major changes in diet, one change involved cooking food differently. While cooking food differently was cited by 77 percent of Atlanta respondents and 79 percent of Washington respondents, it was cited by only 65 percent of those in the New Orleans sample. Cooking food differently was more frequently mentioned by respondents with some college (78 percent) and with college diplomas (80 percent) who had made major changes in their diets during the past five years, and was less frequently mentioned by comparable high school graduates (65 percent). Respondents between 40 and 60 years old were more likely (about 82 percent) to have changed the way they cooked food than those over 60 or between 31 to 40 years old (75 percent) and respondents under 30 (63 percent).

These findings suggest greater attention to primary intervention - i.e., efforts to promote healthy diets and behaviors generally, particularly in the earlier ages. Consistent patterns in the results reported in this section suggest that respondents often adapt healthier diets or behaviors (e.g., exercise, changes in diet) only after having been advised to do so by a doctor and seemingly in response to some specific health problem that has already become visible. Advice from doctors or health professionals to do things differently also seems to increase thought about and concern with getting cancer, interest in cancer prevention, and use of additional sources of information on cancer, including such "narrow-cast" media as pamphlets, magazines, and billboards. It seems that most respondents undertake improvements in their diets and lifestyles and increase their attention to cancer prevention primarily when advised to do so by a doctor or health professional.


Women in our sample provide an interesting pivotal position for intervention within families and across generations. The data suggest that many women undertake activities that contribute to the improvement of general health without a doctor necessarily having told them to do so. Such suggestions from a doctor are not significantly associated with many health improvement behaviors for women, including exercising, changes in food, diet, lifestyle or work habit, or reducing smoking, alcohol, or exposure to sunshine. It seems that such health improvement activities, and especially changes in diet and ways of cooking food, reflect self-initiated health measures rather than changes taken under a doctor's advice. It is important to note that this contrasts with the full sample, where health improvement behaviors are significantly associated with doctor's suggestions and hence do not represent self-initiated efforts. Thus, it seems that African American women are more likely than men to undertake some measures to improve their health on their own, while men are more likely to require a doctor's suggestion to do so.

In contrast to these self-initiated activities to improve health in general, women's participation in cancer screening programs was higher when a doctor suggested that the respondent should do things differently to prevent cancer. Women who received such suggestions from their doctors were more likely to participate in early screening programs for breast cancer (gamma=.24, p=.03), and for oral, skin and rectal cancers (gamma=.23, p=.001). The suggestions may have followed the screening activities, or the screening may have reflected prevention and health practices of the doctors, and again the analysis from survey follow-up will help shed light on the causal ordering of the relationships. There is evidence, however, that women's health awareness is higher in these cases: they are more likely to ask for screening tests (gamma=.15, p=.000), undertake to pay for such tests themselves or lose time from work if they were sick (gamma=15, p=.000). They were particularly likely to follow the doctor's advice (gamma=.49, p=.000). Again, it is possible that general health awareness is the key independent variable in these relationships. Women who expressed an interest in preventing cancer were less likely to have doctors suggest they do things differently (gamma=.14, p=.04). At the same time, however, women who specifically asked a doctor about cancer prevention were more likely to be told to do something differently (gamma=.27, p=.002). The suggestions to change may, of course, have triggered the specific questions.

Having a regular doctor had stronger and more consistent relationships with participation in preventive screening for breast cancer (gamma=.38, p=.000), oral, skin and rectal cancers (gamma=.21, p=.004), and taking precautions about sun exposure (gamma=.19, p=.024). It also had important relationships with changes in diet for women, including changes in cooking food (gamma=.37, p=.003), changing diet within the past five years (gamma=.25, p=.004), and having special diet for medical reasons (gamma=.46, p=.000). Having a regular doctor made it much more likely that a woman had spoken to a doctor about cancer (gamma=.40, p=.004) and that the doctor had suggested doing things differently (gamma=.28, p=.002). More complete insurance coverage also made it more likely that a woman had discussed cancer with a doctor (gamma=.25, p=.04) and that the doctor had suggested changes (gamma=.26, p=.004). Insurance coverage also increased participation in the screening activities and diet changes associated with having a regular doctor, but it seems likely that these outcomes reflected the relationships with doctors just cited.

Women with more complex sources of information on cancer (i.e., resources included such narrowest media as pamphlets, magazines, and newspaper) and who had received their most recent information on cancer from such sources thought more often about the possibility of getting cancer (gammas=-.17 and -.14, p=.000 and .002, respectively). They were also more likely to express concern about getting cancer in the future (gammas=.11 and .12, p=.002 and .000, respectively) and to be interested in learning more about how to prevent cancer (gammas=-.17 and -.13, p=.001 and .03, respectively). Women with more complex and narrowcast sources of information on cancer are also more likely to disagree that people cannot prevent cancer (gamma=.16 and .12, p=.003 and .03, respectively). They were also more likely to participate in cancer prevention activities such as meetings, screening programs, or donating funds (gammas=.27 and .20, p=.000 and .001, respectively).

Women whose latest information on cancer was a narrowest medium were more likely to have spoken to a doctor about cancer (gamma=-.12, p=.008). Again, the data does not permit one to determine whether the consultation with the doctor led to more narrowest sources of information, or vice versa.

Finally, knowledge of such cancer prevention programs as NBLIC is associated with greater knowledge and more favorable attitudes towards cancer issues, including greater familiarity with the causes of cancer, greater interest in preventing it, and greater belief in its curability. It is also associated with greater involvement in cancer prevention programs, greater likelihood of attending meetings or talks, and greater participation in screening programs. While one cannot determine from the data what is responsible for the greater knowledge of programs like NBLIC, the results do suggest that such programs play a vital role in these important aspects of cancer prevention among African Americans.

Diffusion and Adoption Strategies

Given women's propensity to believe that prevention of cancer in the future is possible and the likelihood that they will participate in programs or prevention activities, it is possible that diffusion strategies hold the greatest potential for prevention and intervention of cancer in African American communities. Diffusion research examines the flow of information through communication channels in a community in order to understand how to introduce an innovation and track the rate of dissemination and adoption. Health promotion campaigns focus attention on the development of the innovation as well as on the community in which it is to be introduced. A health innovation may be a program, service, or information that has not previously been known. Rogers (1983) developed a six-stage innovation-development process as diagrammed below:

Problem Recognition Research [right arrow] Basic & Applied [right arrow] Product Development [right arrow] Production [right arrow] Diffusion [right arrow] Consequences & Adoption

Recently, the National Institutes of Health have adopted this model for the development of preventive interventions, with some variations. Problems that may inhibit the successful diffusion of innovations using this model are that user systems are often not very involved in the development process, difficulties in translating research findings to applied settings, and research tends to focus on short-term effects rather than long-term integration of an innovative health practice in a community (Orlandi, et al., 1990).

The aim of diffusion strategies is to spread cancer prevention information through existing community channels. The success of the strategy is based on the rate of dissemination, the rate of adoption of an innovation and the extent to which the innovation is integrated into a system becomes normative and is sustained over time.


Since women are primarily responsible for shopping and food preparation for their families, and also often for the play and exercise activities of children, changing their understanding and decisions on diet and exercise would seem central to any lasting changes in these behaviors within families. Women are also often responsible for scheduling regular checkups for their spouses and children and for deciding what symptoms need attention. They are the pivotal group to reach in any effort to transmit and change health values and behaviors in families and across generations. Thus, the policy implication would underscore attention to primary intervention - i.e., efforts to promote healthy diets and behaviors generally, particularly in the earlier ages and particularly through women.

Additionally, since women in this study whose sources of information on cancer include "narrowcast" media showed greater knowledge and awareness of cancer prevention and had more positive perceptions and beliefs about treatment. As these women also had greater awareness of other health issues and problems, this might suggest the importance of diffusion strategies, such as "narrowcasting," for reaching them on health issues. The evidence at least points to a diffusion model as a strategy to reach more African Americans in general with women appearing to be the most potentially powerful conduit. The data have indicated that a regular physician plays a central role in the adoption of cancer prevention and health improvement behaviors for most respondents. Thus, the diffusion model strategies designed to improve doctors' promotion of cancer prevention and health improvement behaviors among women seem to be logical. So too are alternative diffusion structures for African Americans who may be less able to rely upon a regular physician for the diffusion of prevention or health improvement information. This would support the continued value of grassroots community involvement, and the Cancer Coalitions have the best potential for achieving positive results.


Blanton, J. and S. Alley

1975 Youth Participation in Program Development. Contract HSN-42-72-143, NIMH.

Bonner, F. and V. Ferguson

1995 Resource Guide to Health Promotion, National Black Leadership Initiative on Cancer Washington, D.C.: Cancer Education and Prevention Resource Office, Howard University.

National Cancer Institute

1986 Cancer Among African Americans and Other Minorities: Statistical Profiles. P. 6 in Publication No. 86-2785. Bethesda, MD: National Institute of Health.

Orlandi, M.A., C. Landers, R. Weston, and N. Haley

1990 "Diffusion of health promotion innovations." In K. Glanz, F.M. Lewis, and B.K. Rimer (eds.), Health Behavior and Health Education. San Francisco, CA: Jossey Bass.

Rogers, E.M.

1983 Diffusion of Innovations, Third Edition. New York: Free Press.

Rogers, E.M. and F.F. Shoemaker

1971 Communication of Innovations: A Cross Cultural Approach. New York: Free Press.

Winnett, R.A.

1986 "Diffusion from a behavioral systems perspective." In R.A. Winett (ed.), Information and Behavior: Systems of Influence. Hillsdale, NJ: Erlbaum.

U.S. Bureau of the Census

1990 Census of Population and Housing Summary Tape File 3A. National Black Leadership Initiative on Cancer: Cancer Education and Prevention Resource Office. Continuation Grant September 1992 - August 1995.