Mujeres felices por ser saludables: the association between attendance and outcome in a dietary/breast health randomized clinical trial for Latinas.

Abstract: Method. Two hundred fifty-six 20-40 year-old Latinas were randomly assigned to the intervention (n=127) or control group (n=129). Data were collected at a baseline and eight-months after randomization. Results. Compared to women who attended 0-2 sessions, those who attended at least three of the 16 sessions reported lower dietary fat (p<.001) and improved BSE proficiency (p<.001). However, there were no differences in fiber intake at eight months. Women who attended at least three sessions were more likely to be lower in acculturation. Conclusions. The optimal length for health-risk reduction interventions is unknown. Our results suggest that, while participation is associated with improved outcomes, fewer sessions than traditionally offered in most behavioral interventions may be sufficient to effect initial behavior change.

**********

In the United States, the Latino population growing five times faster than that of the general population (U.S. Bureau of the Census, 2001). Thus, the health needs of this ethnic group are of considerable public health concern. In particular, although breast cancer incidence and mortality rates are lower among Latina women compared to white and black women, it is the most commonly diagnosed cancer and the most common cause of cancer mortality among Latinas (Miller et al., 1996). It has been suggested that interventions targeting behavioral change consistent with breast cancer risk reduction with a focus on dietary and early detection practices are needed (Hiatt & Rimer, 1999). However, the most effective strategies to engage Latina and other underserved populations in screening and prevention programs are unclear (Carter, Pugh, & Monterrosa, 1996; Coates et al., 1999; Kumanyika, Morssink, & Agurs, 1992; Verschuren et al., 1995).

In order to most accurately assess the outcomes of interest, a primary goal of intervention trials is to motivate participants to attend as many sessions as possible (Bowen, Raczynski, George, Feng, & Fouad, 2000). Indeed, increased attendance will potentially improve outcomes (Bowen et al., 2000). Results from a recent large randomized trial of a diverse group of older women showed an association between increased intervention participation and reduction in dietary intake of percentage energy from fat (Bowen et al., 2000). In that study, however, Latinas participated in fewer intervention sessions than African-American or white women (Bowen et al., 2000). To our knowledge, no studies have specifically examined the dose-response relationship between attendance and outcome among Latino women. Demonstrating the optimal level of participation needed to affect specific behavioral changes could have implications for the design of intervention trials in this target population.

Mujeres Felices por ser Saludables (heretofore referred to as Mujeres Felices) was a randomized intervention trial designed to reduce fat intake, increase fiber intake and increase proficiency of breast self exam (BSE) among young, low-acculturated Latinas (Fitzgibbon, Gapstur, & Knight, 2003). Results of that study showed significantly lower fat intake and greater proficiency of BSE but no difference in fiber intake among the intervention versus the control group following the eight-month intervention (Fitzgibbon, Gapstur, & Knight, in press). The primary purpose of this paper is to describe the relationship between number of intervention sessions attended (i.e. level of attendance) and intervention outcome (i.e. dietary fat and fiber intake and BSE proficiency at the eight-month follow-up interview).

METHOD

PROCEDURES

Mujeres Felices was a randomized controlled intervention trial designed to evaluate the efficacy of a combined dietary and breast health intervention targeting young, low-acculturated Latinas. All recruitment, data collection and intervention activities were conducted at the Erie Family Health Center in Chicago, Illinois. This center primarily serves an immigrant Latino population.

The baseline characteristics of the study population, recruitment strategies, and main outcomes are described in detail elsewhere (Fitzgibbon & Gapstur, in press; Gapstur & Fitzgibbon, in press). Briefly, participants were screened for willingness to participate and pre-eligibility. Eligibility criteria included self-identification as Latina, between the ages of 20 and 40-years-old, a body mass index less than or equal to 35 kg/[m.sup.2], and willingness to be randomized and attend 16 sessions over an eight-month period. Following the signing of informed consent documents, data were collected at a baseline Health Center Visit (HCV), which included BSE proficiency and completion of three, 24-hour recalls within one month of the baseline HCV. A total of 256 women were randomized to either the classroom intervention (n=127) or mail (control, n=129) group. Eight months post-randomization, data were collected at the eight-month HCV. Institutional Review Board (IRB) approval was obtained from Northwestern University and Erie Family Health Center.

THEORETICAL MODEL OF MUJERES FELICES.

The Mujeres Felices intervention was grounded in behavior change theories, including the Transtheoretical Model (Prochaska & DiClemente, 1992) and Social Cognitive Theory (SCT; Bandura, 1986) providing strategies for tailoring the intervention to the individual. Briefly, according to the Transtheoretical Model, individuals will perceive and be willing to change behavior as it relates to their personal level of motivation (Prochaska & DiClemente). In the delivery of the intervention, we were sensitive to the potentially diverse levels of readiness to change among participants and allowed group time to discuss the pros and cons of behavior change related to both diet and breast health. Social Cognitive Theory (SCT) emphasizes the use of modeling, as well as enhancing self-efficacy and reducing barriers to change in the physical and social environment (Bandura). We were aware that in order for this intervention to be effective, we would need to apply concepts that would be particularly appropriate for a low-acculturated Latina population. Thus, we also incorporated tenets of culturally competent research (American Psychological Association, 1992; Porche-Burke, 2000). The development and implementation of culturally competent research requires recognition and respect for the specific age, social practices, and ethnicity of those for whom the intervention is being developed.

Culturally competent research also emphasizes the influence these factors exert in the lives of the participants and incorporates these factors into the intervention (American Psychological Association, 1992; Pachter, 1994; Porche-Burke, 2000). As part of the development of the intervention, we conducted a series of focus groups composed of project staff, dietary consultants with expertise in Latino dietary patterns, and Erie Family Health Center staff. Staff members included social workers, physicians, and administrators. Representative women from the target community were also included. The information obtained from these sessions helped us design an intervention that acknowledged and incorporated the notion of collective identity (Triandis, Leung, Villareal, & Clack, 1985). In other words, we were aware that our participants would likely acknowledge the needs of their family over their own (Marin & Triandis, 1985). The importance of their health for the continued care and stability of the family was emphasized to counteract this tendency. In tailoring the intervention to make it most relevant for young Latinas, we spent a substantial portion of the sessions addressing changes in diet and breast health practices that might affect the family unit.

INTERVENTION ELEMENTS

All intervention elements were piloted and refined with input from individuals comparable to prospective participants, community members, and minority health experts. During the eight-month intervention, women in the classroom group were invited to attend sixteen 90-minute sessions that provided information on diet, early detection of breast cancer, and breast health. The mail group received general health information (e.g. dental care, seat belt safety) on a schedule identical to the classroom group. The frequency of both the intervention sessions and the control mailings was once per week for eight weeks, followed by biweekly for two months, and then once per month for four months. The specific concepts covered in the intervention sessions and the control mailings are shown in Table 1. A research nutritionist and a trained breast health educator who were bilingual and bi-cultural led all groups, received weekly supervision from the investigators, and were thoroughly trained in the content areas, as well as the delivery of culturally proficient interventions. All intervention groups except one were delivered in Spanish.

STATISTICAL ANALYSES.

This analysis focuses on the 92 women randomized to the intervention who also completed the eight-month HCV. One-way analysis of covariance (ANCOVA) was used to relate the level of attendance to the eight-month values for the outcomes of interest. Three ANCOVA tests assessed the effects of attendance for initial sessions (eight weekly session), middle sessions (four biweekly sessions), and late sessions (four monthly sessions). For each analysis, participants were categorized according to attendance within the intervention period of interest. For the initial eight weekly sessions, attendance was divided between those who attended 0 sessions, 1-2 sessions, 3-4 sessions, 5-6 sessions, and 7-8 sessions. For the middle biweekly sessions, attendance was divided between those who attended 0 sessions, 1-2 sessions, and 3-4 sessions. For the late four monthly sessions, attendance was divided between those who attended 0 sessions, 1-2 sessions, and 3-4 sessions. Dependent variables were eight month follow-up values for percentage energy from fat, fiber intake (grams per day), and BSE proficiency. Along with baseline status on the dependent variable, other covariates included age (continuous), BMI (continuous), and education status (high school graduate versus not). Differences in eight-month unadjusted mean values were compared using Bonferroni tests for multiple comparisons (Keppel, 1991). Differences in sociodemographic and cultural factors between women who engaged (i.e. attended at least three sessions) compared to those who did not engage (attended 0-2 sessions) were compared using Student's t-test and xhi-square analyses.

RESULTS

Among the 127 women randomized to the intervention, 92 (72.4%) completed the eight-month HCV. Among these 92 women, the mean number of intervention sessions attended was 6.5 (range = 0 - 16). Approximately 12% attended 0 sessions, 34.7% attended 1-8 sessions, and 53.3% attended more than eight sessions. The baseline sociodemographic and cultural characteristics of these women are shown in Table 2.

Table 3 presents the baseline, eight-month, and adjusted eight-month values for percent energy from fat intake by number of sessions attended. At baseline, there were no meaningful differences in percent energy from fat intake according to level of attendance. For the adjusted analysis of percent energy from fat intake at eight months, results indicate that attending 3-4 of the initial sessions was associated with a significantly lower percent energy from fat (F(4,83)= 6.042, p<.001) than attending 0 or 1-2 sessions. However, for women who attended more than three sessions, there was no difference in fat intake at the eight-month HCV compared to those who attended 3-4 sessions. As shown in Table 4, there was no conclusive effect of level of attendance on fiber intake.

Table 5 presents the baseline, eight-month HCV, and adjusted eight-month HCV scores for BSE proficiency. Similar to percent energy from fat, results indicate that women attending 3-4 sessions showed significantly higher BSE proficiency (F (4,82) = 22.154, p<.001) compared to those who attended fewer sessions. Women who attended more than three sessions did not show a further improvement in proficiency. Table 6 presents differences between women who did not engage in the intervention (attended 0-2 sessions) versus those who did engage (attended at least three sessions). Results suggest that those women who were lower in acculturation and had fewer years living in the United States were more likely to engage in the intervention.

DISCUSSION

As mentioned previously, there is a paucity of information available from randomized clinical trials on changes in dietary and breast health behaviors in young Latinas. In addition, the relationship between participation in trials and the outcomes of interest are not routinely reported (Brownstein, Cheal, Ackermann, Bassford, & Campos-Outcalt, 1992; Navarro, Senn, McNicholas, Kaplan, Roppe, & Campo, 1998; Zapka, Harris, Hosmer, Costanza, Mas, & Barth, 1993). Even though the association between attendance and outcome may not be analyzed or reported, a primary goal of intervention trials is to motivate participants to attend as many sessions as possible (Bowen et al., 2000). The rationale for intervention length is usually not empirically determined. It is likely that the length is based on a combination of the number of sessions considered essential to teach content and motivate behavior change that is also acceptable to participants.

In this study, participants randomized to the intervention group did not have to attend the majority of the sessions to demonstrate behavior change. Indeed, participants who attended 3-4 sessions showed significantly lower percent energy from fat and higher BSE proficiency compared to women attended 0 or 1-2 sessions. More importantly, they showed similar outcomes to those who attended more than 4 sessions. No differences in dietary fiber intake were observed, but the high baseline intake may have precluded any substantial increases even following a dietary intervention. In addition, women who reported a lower level of acculturation (i.e., the specific target of the intervention) were more likely to engage in the intervention.

In the WHT:FSMP, 354 postmenopausal Latinas were randomized to 18 sessions over a one-year period. At the follow-up visit, the intervention group had a greater reduction in percent of energy from fat compared to the control group (Coates et al., 1999) and overall participation was linked to change in percent energy from fat. The Stanford Nutrition Action Program, which included a large number (n = 351) of English-speaking Latinas, was a dietary fat intervention for low-literacy adults. Participants were randomized to a dietary intervention or control group conducted over 20 months (Howard-Pitney, Winkleby, Albright, Bruce, & Fortmann, 1997). The intervention consisted of a six-week (twice weekly) classroom intervention followed by a 12-week (phone and mail contact) maintenance intervention. There were a total of 24 sessions. The program adapted lessons to address cultural appropriateness for the large number of Latinas, and was successful in lowering percentage of calories from fat from 37.1% to 33.2%. In this study, no data were reported on the association between level of intervention attendance and behavior change or a rationale for the number of intervention sessions (Howard-Pitney et al., 1997).

The twelve week (once weekly) Por La Vida intervention was designed to enhance compliance with breast and cervical cancer screening among Latinas in southern California and delivered by community workers (consejeras). Por La Vida researchers reported an increase in screening (Navarro et al., 1998). The average number of sessions attended was not reported nor was a rationale for a 12-week intervention. However, the women who completed the follow-up assessment were more likely to have attended at least half of the sessions. Additionally, Mishra and colleagues (1998) demonstrated increased BSE proficiency among 88 low-acculturated Latinas randomized to a cancer control program. In this study, 78% of the women attended all of the four sessions, but a rationale for delivering a four-session intervention was not reported (Mishra, Chavez, Magana, Nava, & Burciaga Valdez, 1998).

A strength of our intervention was that it was grounded in a solid theoretical framework. Significant time and effort were also afforded to assessment of the target population and tailoring the intervention to low-acculturated Latinas. Although we did not test the efficacy of the inclusion of a theoretical framework or tailoring the intervention to this specific population, other research has supported grounding interventions in a solid theoretical framework (Ammerman, Lindquist, Lohr, & Hersey, 2002) and emphasizing the importance of detailed assessment and planning for the specific needs of the target population (Howard-Pitney et al., 1997; Mishra et al., 1998; Navarro et al., 1998; Navarro, Rock, McNicholas, Senn, & Moreno, 2000).

A primary limitation of a study that is developed for and delivered to a very specific group is a potential lack of generalizability to other populations. Most of the Mujeres Felices participants were patients of the Erie Family Health Center, and ap proximately 85% of the women in our study were of Mexican descent. Therefore, they may not be representative of all young, low-acculturated Latinas, and the results of our study cannot be readily generalized to other Latino subgroups. Failure to recognize the heterogeneity within ethnic groups has been referred to as ethnic glossing (Trimble, 1990) and can impact the cultural proficiency and ultimate effectiveness of an intervention.

Additionally, our study only showed the impact of attendance on short-term behavioral changes. We know little about the long-term effectiveness of this type of an intervention. Future studies should explore long-term maintenance of change in low-acculturated Latinas. These studies should also assess sustained behavior change among a larger and more diverse sample. Studies of this nature could have a greater impact on disease prevention. Another area of study might be the cost-effectiveness of implementing interventions of differing levels of intensity. A study of this nature would be important in assessing wider dissemination of this type of intervention.

In summary, our results provide evidence that participation in intervention sessions is critical to alter and improve health risk reduction behaviors. Results also suggest that, in contrast to a longer intervention, it may be possible to effect change in diet and breast health proficiency with fewer sessions. Although the optimal length for behavior change may be dependent on the specific behavior addressed and the target population, this study demonstrates the potential importance of establishing an empirically driven rationale for intervention length. Indeed, future research should test the association between attendance and behavior change in other populations and behaviors.

ACKNOWLEDGEMENTS: The Mujeres Felices por Ser Saludables project was supported by the US Army Medical Research and Material Command under DAMD 17-96-2-6013. We would like to thank the Carol Gollob Foundation for their generous support of our work, our collaborators at the Erie Family Health Center and the women who participated in this project. We also would like to thank Ms. Sharon Sugarman and Mr. Bruce Briscoe for their scientific expertise in the design, conduct and analysis of this project, and Jamie Gayle, BS and Allison Thompson, MA for their technical assistance. Finally, we would like to thank the staff for their dedication and belief in the importance of improving health behaviors in an underserved population.

REFERENCES

American Psychological Association. (1992). Ethical principles and code of conduct. American Psychologist, 48, 1597-1611.

Ammerman, A. S., Lindquist, C. H., Lohr, K. N., & Hersey, J. (2002). The efficacy of behavioral interventions to modify dietary fat and fruit and vegetable intake: A review of the evidence. Preventive Medicine., 35(1), 25-41.

Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall.

Bowen, D., Raczynski, J., George, V., Feng, Z., & Fouad, M. (2000). The role of participation in the women's health trial: Feasibility study in minority populations. Preventive Medicine, 31(5), 474-480.

Brownstein, J. N., Cheal, N., Ackermann, S. P., Bassford, T. L., & Campos-Outcalt, D. (1992). Breast and cervical cancer screening in minority populations: A model for using lay health educators. Journal of Cancer Education, 7(4), 321-326.

Carter, J. S., Pugh, J. A., & Monterrosa, A. (1996). Non-insulin-dependent diabetes mellitus in minorities in the United States. Annals of Internal Medicine, 125(3), 221-232.

Coates, R. J., Bowen, D. J., Kristal, A. R., Feng, Z., Oberman, A., Hall, W. D., et al. (1999). The Women's Health Trial feasibility study in minority populations: Changes in dietary intakes. American Journal of Epidemiology, 149(12), 1104-1112.

Fitzgibbon, M. L., & Gapstur, S. M. (in press). Mujeres Felices por ser Saludables: Process evaluation for a dietary/breast health randomized clinical trial for Latinas. Annals of Behavioral Medicine.

Fitzgibbon, M. L., Gapstur, S. M., & Knight, S. J. (2003). Mujeres Felices por ser Saludables: A Breast Cancer Risk Reduction program for Latino Women. Preventive Medicine, 36(5), 536-546.

Fitzgibbon, M. L., Gapstur, S. M., & Knight, S. J. (in press). Results of Mujeres Felices por Ser Saludables: A dietary/breast health randomized clinical trial for Latino women. Annals of Behavioral Medicine.

Gapstur, S. M., & Fitzgibbon, M. L. (in press). Recruitment strategies in a low-acculturated sample of Hispanic women. Journal of Women's Cancer.

Hiatt, R. A., & Rimer, B. K. (1999). A new strategy for cancer control research. Cancer Epidemiology, Biomarkers & Prevention, 8(11), 957-964.

Howard-Pitney, B., Winkleby, M. A., Albright, C. L., Bruce, B., & Fortmann, S. P. (1997). The Stanford nutrition action program: A dietary fat intervention for low-literacy adults. American Journal of Public Health, 87(12), 1971-1976.

Keppel, G. (1991). Design and analysis: A researcher's handbook (3rd ed.). Englewood Cliffs, NJ: Prentice Hall.

Kumanyika, S. K., Morssink, C., & Agurs, T. (1992). Models for dietary and weight change in African-American women: identifying cultural components. Ethnicity & Disease, 2(2), 166-175.

Marin, G., & Triandis, H. C. (1985). Allocentrism as an important characteristic of the behavior of Latin Americans and Hispanics. In R. Diaz-Guerrero (Ed.), Cross-cultural and national studies in social psychology (pp. 85-104). Amsterdam: Elsevier Science Publishers.

Miller, B. A., Kolonel, L. N., Bernstein, L., Young, J. J. L., Key, C. R., Swanson, G. M., et al. (1996). Racial/ethnic patterns of cancer in the United States: 1988-1992 (NIH Pub. No. 96-4104). Bethesda, MD: National Cancer Institute.

Mishra, S., Chavez, L., Magana, J., Nava, P., & Burciaga Valdez, R. (1998). Improving breast cancer control among Latinas: Evaluation of a theory-based educational program. Health Education and Behavior, 25(5), 653-670.

Navarro, A. M., Rock, C. L., McNicholas, L. J., Senn, K. L., & Moreno, C. (2000). Community-based education in nutrition and cancer: the Por La Vida Cuidandome curriculum. Journal of Cancer Education, 15(3), 168-172.

Navarro, A. M., Senn, K. L., McNicholas, L. J., Kaplan, R. M., Roppe, B., & Campo, M. C. (1998). Por La Vida model intervention enhances use of cancer screening tests among Latinas. American Journal of Preventive Medicine, 15(1), 32-41.

Pachter, L. M. (1994). Culture and clinical care. Folk illness beliefs and behaviors and their implications for health care delivery. Journal of the American Medical Association, 271(9), 690-694.

Porche-Burke, L. (2000). Guidelines for research in ethnic minority communities: Introduction. Council of National Psychological Associations for the Advancement of Ethnic Minority Interests.

Prochaska, J. O., & DiClemente, C. C. (1992). Stages of change in the modification of problem behaviors. In M. Hersen, R. M. Eisler & P. M. Miller (Eds.), Progess in behavior modification (pp. 184-214). Sycamore, IL: Sycamore Press.

Triandis, H. C., Leung, K., Villareal, M. J., & Clack, F. L. (1985). Allocentric versus idiocentric tendencies: convergent and discriminant validation. Journal of Research in Personality, 19, 395-415.

Trimble, J. E. (1990). Ethnic specification, validation prospects, and the future of drug use research. International Journal of the Addictions, 25(2A), 149-168; discussion 168-170.

U.S. Bureau of the Census. (2001). Profiles of general demographic characteristics: 2000 census of population and housing. Retrieved June 20, 2001, 2001

Verschuren, W. M., Jacobs, D. R., Bloemberg, B. P., Kromhout, D., Menotti, A., Aravanis, C., et al. (1995). Serum total cholesterol and long-term coronary heart disease mortality in different cultures. Twenty-five-year follow-up of the seven countries study. Journal of the American Medical Association, 274(2), 131-136.

Zapka, J. G., Harris, D. R., Hosmer, D., Costanza, M. E., Mas, E., & Barth, R. (1993). Effect of a community health center intervention on breast cancer screening among Hispanic American women. Health Services Research, 28, 223-235.

CHES AREA

Responsibility I--Assessing Individual & Community Needs for Health Education

Marian L Fitzgibbon, PhD is affiliated with the Institute for Health, Research and Policy at the University of Illinois at Chicago. Zoran Martinovich, PhD is affiliated with the Department of Psychiatry and Behavioral Sciences of Feinberg School of Medicine at Northwestern University. Susan M. Gapstur, PhD is affiliated with The Robert H. Lurie Comprehensive Cancer Center and the Department of Preventive Medicine of Feinberg School of Medicine at Northwestern University. Address all correspondence to Marian L. Fitzgibbon, PhD, Institute for Health, Research and Policy (MC 275), 1747 W. Roosevelt Rd. 4th Floor, Chicago, IL 60608; PHONE: [phone]; FAX: [phone]; E-MAIL: [e-mail].

Table 1. Concepts Covered in the Intervention Curriculum and in
the Control Mailings by Session

Intervention Curriculum

Session   Breast Health                 Nutrition

   1      Introduction to BSE           Nutrition and cancer
   2      Breast cancer survivor talk   Good food, less fat &
                                          food labels
   3      Barriers to BSE               Food and your family

   4      Learning BSE                  Fiber
   5      Timing of BSE                 Fruits and vegetables
   6      Communicating with            More on fruits and
            your provider                 vegetables
   7      Presentation by breast        Trimming the fat
            cancer surgeon
   8      Feedback on doctor's          Recipe makeovers &
            presentation                  your family
   9      Breast cancer myths & risks   Building healthy recipes
  10      Mammogram specialist          Potluck lunch
  11      Barriers to detection         Importance of breakfast
            techniques
  12      Breast health reminders       Healthy lunches
  13      Breast health resources in    Healthy meals on the go
            the community                 Services
  14      Review BSE technique          Maintaining traditional
                                          diets
  15      Review breast health
            information
  16      Changes in breast             Fat & fiber review/
            health behavior               Awarding of
                                          certificates

Intervention Curriculum

Session   Breast Health                 Control Mailings

   1      Introduction to BSE           Adult, infant and child
                                          lifesaving
   2      Breast cancer survivor talk   Choking, how to get
                                          medical help fast
   3      Barriers to BSE               Chicago Transit Authority
                                          map, sightseeing guide,
                                          fare card info
   4      Learning BSE                  Child safety in the home
   5      Timing of BSE                 Drug facts
   6      Communicating with            Blue bag recycling guide
            your provider
   7      Presentation by breast        Immunization
            cancer surgeon
   8      Feedback on doctor's          Depression
            presentation
   9      Breast cancer myths & risks   Kid bike/skating safety
  10      Mammogram specialist          Owning your home
  11      Barriers to detection         Sexually transmitted
            techniques                    diseases
  12      Breast health reminders       HIV risks at your job
  13      Breast health resources in    Department of Consumer
            the community
  14      Review BSE technique          Car safety
  15      Review breast health          Neighborhood watch
            information                   information
  16      Changes in breast             Child allergies
            health behavior

Table 2. Baseline Characteristics

Demographic measure                   M      SD    %      N

Age (years)                           30.4   5.4
Acculturation (1=lowest, 5=highest)    1.6   0.9
Years living in the US                11.2   8.4
Number of children                     2.3   1.3
High school graduate
  No                                               59.8   76
  Yes                                              40.2   51
Married
  No                                               27.6   35
  Yes                                              72.4   92
BMI (kg/[m.sup.2])
  <25                                              29.9   38
  25-29.9                                          40.2   51
  [greater than or equal to] 30                    29.9   38
Read and speak Spanish only
  No                                               44.9   57
  Yes                                              55.1   70

Table 3. Intake of Percent Energy from Fat at Baseline and 8-Months
by Level of Attendance for Each of the Stages of Intervention

                              Level of Attendance

                        0             1-2            3-4
                     Sessions       sessions       sessions

    Stage of            M              M              M
  intervention         (SD)           (SD)           (SD)

Initial sessions
  N                     14             3              18
  Baseline             29.6           28.3           26.7
                      (7.7)          (7.0)          (4.8)
  8 months             32.5           30.3           25.2
                      (5.8)          (6.0)          (4.6)
  8 months *       [32.8.sub.a]   [29.3.sub.a]   [25.7.sub.b]

Middle sessions
  N                     26             27             39
  Baseline            29.5           27.9           28.9
                      (7.8)          (5.4)          (6.0)
  8 months             29.1           25.8           26.1
                      (7.0)          (4.5)          (6.2)
  8 months *           28.4           26.5           26.2

Late sessions
  N                     36             27             29
  Baseline             28.4           28.8           29.1
                      (6.2)          (7.6)          (5.4)
  8 months             28.2           26.0           26.1
                      (6.4)          (7.0)          (5.1)
  8 months *           28.3           25.9           26.0

                       Level of attendance

                       5-6            7-8
                     sessions       sessions

                                                             P value
                                                  P value      for
    Stage of            M              M         for Stage   linear
  intervention         (SD)           (SD)       variable     trend

Initial sessions
  N                     31             26
  Baseline             30.0           28.3
                      (6.5)          (6.3)
  8 months             27.1           24.5
                      (5.9)          (5.7)
  8 months *       [26.2.sub.b]   [25.0.sub.b]   0.000 (1)    0.000

Middle sessions
  N                     --             --
  Baseline              --             --

  8 months              --             --

  8 months *            --             --        0.264 (2)    0.117

Late sessions
  N                     --             --
  Baseline              --             --

  8 months              --             --

  8 months *            --             --        0.151 (3)    0.109

Note. ANCOVA = analysis of covariance. * Estimates are adjusted
for baseline age (years), baseline BMI (kg/[m.sup.2]), education
(less than high school graduate, high school graduate or greater)
and baseline percent fat intake. Means with different subscripts
differ significantly at p < 0.05. (1) F[4,83]=6.042. (2) F[2,85]=1.352.
(3) F[2,85]=1.933. Adjustment for multiple comparisons: Bonferroni.

Table 4. Grams of Fiber Intake at Baseline and at 8 Months by
Level of Attendance for Each of the Stages of Intervention

                              Level of Attendance

                      0         1-2        3-4        5-6
                   Sessions   sessions   sessions   sessions

    Stage of          M          M          M          M
  intervention       (SD)       (SD)       (SD)       (SD)

Initial sessions
  N                   14         3          18         31
  Baseline           15.8       15.6       19.4       22.3
                    (4.5)      (2.2)      (9.1)      (7.2)
  8 months           16.9       18.6       19.8       22.9
                    (7.1)      (6.5)      (9.9)      (7.5)
  8 months *        19.8       21.8       20.6       22.0

Middle sessions
  N                   26         27         39         --
  Baseline          16.8       22.8       21.7         --
                    (5.7)      (6.9)      (8.7)
  8 months           17.3       24.0       21.7        --
                    (6.7)      (7.4)      (8.2)
  8 months *         19.5       23.0       20.9        --

Late sessions
  N                   36         27         29         --
  Baseline           18.9       20.8       22.7
                    (7.2)      (8.5)      (7.5)        --
  8 months          19.6       20.0       24.1         --
                    (8.1)      (7.5)      (7.5)
  8 months *        20.7       19.8       22.9         --

                    Level of
                   Attendance

                      7-8
                    sessions

                                            P value
                                 P value      for
    Stage of           M        for Stage   linear
  intervention        (SD)      variable     trend

Initial sessions
  N                    26
  Baseline            22.7
                     (8.1)
  8 months            22.5
                     (6.9)
  8 months *         21.1       0.888 (1)    0.653

Middle sessions
  N                    --
  Baseline             --

  8 months             --

  8 months *           --       0.171 (2)    0.418

Late sessions
  N                    --
  Baseline
                       --
  8 months             --

  8 months *           --       0.203 (3)    0.206

Note. ANCOVA = analysis of covariance. * Estimates are adjusted for
baseline age (years), baseline BMI (kg/[m.sup.2]), education (less
than high school graduate, high school graduate or greater) and
baseline fiber. Means with different subscripts differ significantly
at p < 0.05. (1) F[4,83]=0.283. (2) F[2,85]=1.804. IF[2,85]=1.627.

Table 5. Nurse Rating of BSE at Baseline and 8 Months by Class
Attendance for Each of the Stages of Intervention

                            Level of Attendance

                        0            1-2           3-4
                    Sessions      sessions      sessions

    Stage of            M             M             M
  intervention        (SD)          (SD)          (SD)

Initial sessions
  N                    14             3            17
  Baseline            2.8           1.3           3.7
                      (1.9)         (0.6)         (2.1)
  8 months            3.2           4.0           7.2
                      (2.6)         (2.0)         (2.4)
  8 months *       [3.3.sub.a]   [4.5.sub.a]   [7.1.sub.b]

Middle sessions
  N                    26            26            39
  Baseline            2.8           3.5           3.4
                      (1.8)         (2.0)         (2.2)
  8 months            4.6           7.8           8.5
                      (3.1)         (1.7)         (1.2)
  8 months *       [4.6.sub.a]   [7.7.sub.b]   [8.6.sub.b]

Late sessions
  N                    35            27            29
  Baseline            3.1           3.2           3.6
                      (1.9)         (2.0)         (2.3)
  8 months            5.0           8.3           8.6
                      (3.0)         (1.3)         (1.3)
  8 months *       [5.1.sub.a]   [8.2.sub.b]   [8.7.sub.b]

                     Level of Attendance

                       5-6           7-8
                    sessions      sessions

                                                           P value
                                                P value      for
Stage of                M             M        for Stage   linear
intervention          (SD)          (SD)       variable     trend

Initial sessions
  N                    31            26
  Baseline            3.6           3.1
                      (2.0)         (2.2)
  8 months            8.0           8.8
                      (1.7)         (0.8)
  8 months *       [7.9.sub.b]   [8.8.sub.b]   0.000 (1)    0.000

Middle sessions
  N                    --            --
  Baseline             --            --

  8 months             --            --

  8 months *           --            --        0.000 (2)    0.000

Late sessions
  N                    --            --
  Baseline             --            --

  8 months             --            --

  8 months *           --            --        0.000 (3)    0.000

Note. BSE = Breast self-exam (range is 0 to 10). ANCOVA = analysis
of covariance. * Estimates are adjusted for baseline age (years),
baseline BMI (kg/[m.sup.2]), education (less than high school
graduate, high school graduate or greater) and baseline BSE score.
Means with different subscripts differ significantly at p < 0.05.
(1) F[4,82]=22.154. (2) F[2,84]=29.038. (3) F[2,84]=28.261.
Adjustment for multiple comparisons: Bonferroni.

Table 6. Comparison of Engagement Subtypes on Sociodemographic
and Cultural Variables

                       Not engaged       Engaged
                         (n=39)           (n=88)

                                                         [X.
  Characteristics       N      (%)      N      (%)      sup.2]     p

High School Graduate
  No                    19    (48.7)    57    (64.8)     2.9       ns
  Yes                   20    (51.3)    31    (35.2)

Married
  No                    12    (30.8)    23    (26.1)     0.6       ns
  Yes                   27    (69.2)    65    (73.9)

BMI (kg/[m.sup.2])
  <25                   9     (23.1)    29    (33.0)     3.0       ns
  25 - 29.9             20    (51.3)    31    (35.2)
  [greater than or
    equal to] 30        10    (25.6)    28    (31.8)

Read & speak Spanish
    only
  No                    22    (56.4)    35    (39.8)     3.0       ns
  Yes                   17    (43.6)    53    (60.2)

                        M      (SD)     M      (SD)       t        p

Age (years)            30.2   (5.5)    30.6   (5.4)     -0.3       ns
Acculturation          1.9    (1.1)    1.5    (0.8)      2.2     < 0.05
Years living in the
  US                   13.7   (9.3)    10.0   (7.7)      2.2     < 0.05
Number of children     2.3    (1.6)    2.2    (1.2)      0.3       ns

Note. Not engaged participants attended 0 - 2 intensive sessions.
Engaged participants attended 3 - 8 intensive sessions.