Por La Vida Model Intervention Enhances Use of

Cancer Screening Tests Among LatinasAna M. Navarro, PhD, Karen L. Senn, EdD, MPH, Lori J. McNicholas, MA, RD, Robert M. Kaplan, PhD,Beatriz Ropp_, BS, Mary C. Campo

Objectives: To describe the short-term impact of the intervention known as Por La Vida (PLV) oncancer screening for Latinas in San Diego, California.

Methods: Thirty-six lay community workers (consejeras) were recruited and trained to conducteducational group sessions. Each consejera recruited approximately 14 peers from thecommunity to participate in the program. The consejeras were randomly assigned to eithera twelve-week cancer screening intervention group or a control group in which theyparticipated in an equally engaging program entitled "Community Living SkillsY' Pre- andpost-intervention self-report information was obtained from project participants on the useof cancer screening examinations. Outcome measures were changes in the percentages ofwomen who had breast and cervical cancer screening tests within the past year before andafter the intervention occurred. Experimental and control groups were compared usingt-tests. Analyses were conducted using both consejeras and participants as the unit of analysis.

Results: The increase in the use of the cancer screening tests was higher in the PLV cancerintervention group in comparison to women in the community living skills control group.

Conclusions: Key to the PLV intervention model is the identification of natural helpers in the Latinocommunity and their subsequent training in interventions based on social learning theoryusing culturally appropriate educational materials. The model is an effective and viableapproach for increasing the use of cancer screening tests in Latinas of low socioeconomiclevel and low level of acculturation.

Medical Subject Headings (MESH): screening (cancer), community health aides, Latinas,intervention studies, community, (Am J Prev Med 1998;15:32-41) © 1998 AmericanJournal of Preventive Medicine

arge discrepancies in the health status of various However, it is vcj'dely recognized that improving thesocioeconomic and ethnic groups in the United health status of special populations requires communi-States have been of great concern to health ty-based interventions in the target community. 1-4 Fur-

professionals. Closing the gap in health status in pop- ther, the use of lay health advisors is gaining recogni-ulations that are at especially high risk was one of the tion as a particularly promising strategy to increasethree major goals of the-Healthy People 2000 report. 1 access to health care and to improve the health status ofPopulation groups that have been identified as experi- underserved populations. 5 In this paper we present theencing above average rates of death, disease, and results of a study that focus on low-income Latinas. Thedisability are people with low incomes and people who study implements a community-based intervention in

are members of some racial and ethnic minority which the use of lay health advisors is central. The studygroups. These ethnic groups include African Ameri_ • was designed to examine the impact of the intervention

cans, American Indians, Asian and Pacific Islander, and on the use of breast and cancer screening exams.Hispanic/Latinos. Healthy People 2000 set specific The Healthy People 2000 Report set high objectivesobjectives to narrow the existing health gap. The strat- for the use of breast cancer and cervical cancer screen-egies to achieve the specific objectives are complex, ing: At the time the objectives were formulated, there

were substantial discrepancies in screening tests usebetween Hispanic and non-Hispanic women. 1 AmongFrom the Universityof California, San Diego, La Jolla, California

92093-0622USA(Navarro, Kaplan, Ropp_, Campo), the San Diego Hispanic women, the proportion of women over 50State University, San Diego, California 92182-1900USA (Senn, Mc- years who had received a clinical breast exam and a

Nicholas) mammogram within the past two years was only 18%,Address correspondence to: Ana M. Navarro, PhD, Department ofFamily and PreventiveMedicine,University of California, San Diego, and only 20% of Latinas 40 years old or older had ever9500 GilmanDrive,LaJolla, California 92093-0622 received a mammogram. The year 2000 goals for His-

32 AmJ Prev Med 1998;15(1) 0749-3797/98/$19.00© 1998 American Journal of Preventive Medicine PII S0749-3797(98) 00023-3

panic women 50 years old or older included 60% usage helpers (i.e., consejeras) in their existing linkship socialof both professional breast exam and mammography network 17is central to the intervention model. Further-over the last 2 years. Among women 40 years old or more, principles of social learning theory are used toolder, the year 2000 objective was an 80% rate for ever structure the educational program. A third key compo-having obtained both professional breast exam and nent of the intervention model is the identification andmammogram. 1 Data from the National Health Inter- development of culturally appropriate educational ma-view Surveys indicate that the rate of mammography terials to implement the programJ 6use nearly doubled between 1987 and 1990 among Previous projects utilizing the Por La Vida interven-

women 40 years old or older. Specifically, Breen and tion model had suggested its potential for promotingKessler 6 found that the percentage of women 40 years health in the Latino community. However, this studyor older who had had a mammogram in the previous represents the first randomized field experiment de-year had increased from 17% in 1987 to 33% in 1990. signed to evaluate the impact of the PLV interventionThere has als0 been progress in the percentage of on health-seeking behaviors.Hispanic women who are getting screening mammog-raphy. 6'7 Breen and Kessler found that the percentageof Latinas over 40 who had had a mammogram in the Method1990 NHIS was 31% compared to 13% in 1987. Kaplan Interventionand colleagues 7 also found percentages of Lafinas thathad ever had a mammogram to be higher than the The study was implemented in the Southeast area of

San Diego County where more than one fifth of the1987 baseline used in the Healthy People 2000 report.Percentages of Latinas 50 years old or older that had population is Latino. 18The intervention was developedever had a mammogram varied between 34.6% and specifically to target low-income Latinas whose access to79.5% in six samples collected in Arizona, Texas, health care service and cancer screening rates wasColorado, and Southern California. 7 However, His- expected to be particularly low in comparison to other

panic women remain overall significantly behind other socioeconomic groups. Key to the intervention is theethnic and racial groups in the use of cancer screening use of lay health workers. Women from the Latinotests, community were identified as consejeras based on their

A review by Rimer s provides an excellent summary of behaviors associated with the role of "natural helper"the determinants of mammography use and the bene- traditional in the Latino community. Personnel atfits of intervention. A variety of intervention strategies schools and community centers as well as other organi-have been used to increase the use of cancer screening zafions serving the Lafino community were asked totests. These include mass media campaigns, individual- identify women who were perceived as natural helpers.directed interventions, interventions directed at a phy- Individuals identified by this process were invited to ansician or a health care system, the use of mobile vans, interview in which project staff explored personal,changes in reimbursement, social network interven- performance, and situational characteristics relevant totions, and multi-strategy interventions. Rimer's review the nature of the project. Personal characteristics referindicates that single-intervention trials have produced to attributes such as sense of humor, self-esteem, andmodest results. In particular, single interventions may being dedicated, respectful, and realistic. Performancenot be effective for low-income women who face mul- characteristics refer to the ability of the candidates to

tiple obstacles to screening. 9 To date, the most effective read, write, and communicate and their willingness tointerventions have been those that use multiple strate- participate in meetings and to complete documenta-

gies. Successful strategies must also combine informa- tion needed for the project. Situational characteristicstion on the benefits of mammography with specific refer to time availability and the existence of an ex-instruction on how to obtain low-cost services) °'11 tended network of family and friends. A standardized

This paper reports the results of an evaluation of the rating scale was completed by project staff at the end ofPor la Vida model of community health promotion to the interview) 9 Eng and colleagues 5 have described aincrease the use of cancer screening tests. This study continuum in the various lay health advisors interven-

was one of the five cooperative projects funded by the tion strategies from natural helping to paraprofessionalNational Cancer Institute that targeted breast and helping. The Por La Vida intervention model relies on

cervical cancer control in Hispanic women. 12-15 A individuals identified as natural helpers. As describeddetailed description of the Por La Vida model as well as by Eng and colleagues, 5 natural helpers are individualsthe development and implementation of the interven- who have a reputation in their community for goodtion has been published elsewhere, a6 Briefly, the PorLa judgment, sound advice, a caring ear, and being dis-Vida intervention model capitalizes on the use of exist- creet. Further, the Por La Vida intervention model wasing social networks in the Latino community. The developed in the theoretical framework of naturalidentification and training of women who are natural networks in HispanicsJ 7'2° Valle distinguishes the ag-

AmJ Prev Med 1998;15(1) 33

gregate, linkship, and kinship natural networks. The Experimental Design

PLV intervention model relies primarily on the linkship Each established group was randomly assigned in annetworks, which are comprised of mutually linked

individuals bound by ties of friendship based on reci- experimental-control group design with pretest, post-procity and exchange behaviors. 17 The consejeras have test, and follow-up. The unit of randomization was the

consejera. Pretest and posttest measures were collectedan established relationship of trust and confidence before and after the twelve weekly educational sessions.among Hispanics within their own social environment. In addition, two follow-up interviews were scheduledThis relationship can facilitate the dissemination of one year and two years after the pretest, respectively. Inhealth promotion information in modes that are more this paper we report the comparisons between pretestacceptable and adaptable to the needs of the Hispanic and posttest. Posttest was scheduled upon graduationc°mmunity"2° of the group (approximatelythree months after pre-

Recruited consejeras were trained following the conse- test) and no later than six months after the pretest.jera manual specifically designed to guide the weekly A total of 36 consejeras conducted group educationaleducational sessions in the topic focus of the interven- sessions. Each consejera recruited on average 14 womention. Upon completion of the training, each consejera to participate in the group sessions. Half of the conse-invited women from their naturally occurring social jeras were randomly assigned to a control group innetworks to participate in small group educational which they participated in an equally engaging pro-sessions. Following the consejera manual, each consejera gram entitled "Community Living Skills." The experi-conducted twelve weekly educational sessions with her mental group attended sessions on breast and cervicalrespective group of women. More details about the cancer early detection, the importance of screeningnature of the intervention have been presented else- tests, nutrition, skills training in breast self-examina-where. 16 tion, and obtaining services. 16

Subjects Measures

As described above, participants were recruited through The effects of the PLV intervention were assessed

consejeras. The consejeras were responsible for forming through extensiye face-to-face interviews conducted ingroups of approximately 10 to 15 participants. Conse- either Spanish or English. Ninety-seven percent of thejeras were instructed to recruit project participants before participants preferred to be interviewed in Spanish. Athey knew whether they had been assigned to the 178-item questionnaire was developed that covers infor-intervention or control group. More than 500 Latinas mation on access to health care services, cancer knowl-were initially recruited to participate in their respective edge, preventive measures, and previous cancer screen-

groups and 512 were interviewed at baseline. However, ing examinations. The questionnaire also included theMarin's Short Scale of Acculturation, 21 and the Sociala total of 609 Latinas were regular program participantsat the end of the educational sessions. In addition, 141 Support Questionnaire. 22 Average duration of the in-

women participated occasionally in the educational terview was 45 minutes at pretest and 42 minutes at

sessions as guests of regular program participants, posttest, with 95% of the interviews being completed inmore than 20 minutes and less than 80 minutes.The sociodemographic characteristics of the partici- Variations in the duration of the interview were due

pants are as follows. The average age was 34 years mostly to the proper application of the skip patterns of(range 18-72). Socioeconomic level of participants was the interview protocol and the needs of the interviewee.low on average as indicated by the years of formal In this paper, we report pretest and posttest measure-education (median = 7) and the yearly gross family ments on the use of the most common breast andincome (median = $12,000). The average family size cervical cancer screening tests: breast self-exams, pro-was 5 persons. The majority of the women were married fessional breast exams, mammography, and Pap test.and full-time homemakers. Ninety-two percent of the These variables constitute the main outcome measuresinterviewees had been born in Mexico, 5% in the of the study.United States and 3% in other Spanish-speaking coun-

tries. Women not born in the United States had been Statistical _,_uy_A--1-'-isliving in this country approximately eight years on

average. Average acculturation, as measured by the Preliminary Analysis. As described above, 512 LatinasMarin's Short Scale of Acculturation 21 (range l-low completed the baseline survey. However, 147 failed to

level to 5-high level of acculturation), was 2. Regarding complete the posttest survey. The percentage of womenaccess to health care services, over 60% of the women who did not complete the posttest interview was similarhad no health insurance and more than 40% did not in the experimental and control groups (27.4% and

have a regular health care provider. 31.9%, respectively). An attempt was made to posttest

:$4 American Journal of Preventive Medicine, Volume 15, Number 1

all women who had been interviewed at baseline, even reported not doing monthly breast self-exam at pretestif they had not been regular participants in the educa- but indicated she was conducting breast self-exam oncetional sessions, a month at posttest. Otherwise a valueof 0 wasassigned

Before additional analyses were conducted, we calcu- to the variable for this woman.

lated chi-square statistics to examine whether, on a wide For mammography, Pap test, and professional breastrange of variables, there were statistically significant exam, the respective dependent variable was assigned adifferences between those who completed and those value of 1 if .the woman had not had the test done

who did not complete the posttest survey. These mea- within the past year at pretest but had the test donesures included demographic variables, access to health within the past year at posttest. Otherwise a value of 0care services, and use of cancer screening tests at was assigned to the respective variable for this woman.pretest. Demographic variables included in these anal- We compared change in use of cancer screening tests

yses were the following: age, number of people in the between intervention groups and community livinghousehold, annual household income, years of formal skills control groups, calculating two-tailed t-tests foreducation, marital status, employment status, number each of the main outcome dependent variables. Be-of years residing in the United States, and language- cause the unit of randomization in the study was the

based acculturation as estimated by the Marin's Short group and not the participant, we conducted compar-Acculturation Scale. 21 Variables relevant to access to isons between experimental and control groups usinghealth care were: difficulties experienced in the past both consejeras and participants as the unit of analysis.year accessing health care services, health insurance When relying on participants as the unit of analysis, astares, having a regular health care provider, general rectangular data file was utilized in which change in thehealth status. A third group of variables analyzed was use of cancer screening tests had been computed fordirectly related to the use of breast and cervical cancer each of the participants following the description out-screening tests at baseline: knowledge about how to lined above. For computations with consejeras as the unitperform breast ser-examinafion, the use of monthly self- of analysis, we prepared an aggregated data file with 36examination, and the recency of physical breast exam by cases (i.e., consejeras). In the aggregated data file, the

a health professional, mammography, and Pap test. value of the outcome variable for each particular conse-Furthermore, we explored potential statistically sig- jera was the arithmetic average of the outcomes for the

nificant (P < .05) differences between the Cancer persons participating in this consejera's group.intervention group and the Community Living Skills

control group at baseline. We computed chi-square Resultsstatistics to examine differences between experimentaland control group on demographic variables as well as Table 1 shows comparisons between program partici-in the main outcome variables on the use of screening pants who completed both pretest and posttest andtests at pretest. These variables included the same set of those who failed to complete the posttest. We found ademographic variables, access to health care services, statistically significant (P < .05) difference on onlyand use of cancer screening tests at pretest used in the one of the variables. A higher proportion of peoplepreliminary analyses described above, who failed to complete the posttest interview (45.7%)

had health insurance compared to program partici-

Statistical Analysis Assessing Impact of' pants who completed both the pretest and posttestsurveys (34.6%). We do not regar d this difference as an

Intervention important threat to the validity of the results becauseChanges in the use of breast and cervical cancer no statistically significant differences were found on anyscreening tests were the main outcome measures of the other variable related to access to health care andstudy. For women 18 years old or older, the statistical because of the ample variety of variables examined for

analyses included the following outcome variables as potential differences. Specifically, with regard to vari-dependent variables: monthly breast self-exam, profes- ames indicators of access to health care, the percentagesional breast exam in the past year, and Pap test in the of women who had a regular health care provider was

pastyear. lower among women who had completed only theFor women 40 years old or older, we :conducted pretest (51.7%) than in women who had completed

analyses in which mammography screening in the past both the pretest and posttest (60.1%, P = 0.077). Inyear was the dependent variable. The independent fact, the difference in the percentage of women whovariable in all analyses was group assignment, that is, had a regular health care provider as an indicator of

cancer intervention group versus community living access to health care, was in the opposite direction asskills control group. Change in the use of screening test the statistically significant difference in percentage ofwas defined as follows. For breast self-exam, the depen- women who had health insurance, since individualsdent variable was assigned a value of 1 if the woman with health insurance tend to have a regular health care

AmJ Prey Med 1998;15(1) 35

Table 1. Comparisons between program participants who completed the posttest (n = 361) and those who failed tocomplete the posttest (n = 151)

Pretest and

Variable posttest (%) Pretest only (%) Chi-square P value

Age

Table 2. Comparisons between experimental (n = 274) and control (n = 238) groups at pretest

Control Cancer

Variable group (%) intervention (%) Chi-square P value

Age

t=3.23 t=2.43 t=2.22 t=1.96p

examined the reasons for failure to complete the days after the patient contacts the health care provider.posttest. As indicated above, at posttest an attempt was It is possible that some project participants had alreadymade to interview all women who had completed the made appointment for the screening tests at the time ofpretest. However, as would be expected, a high propor- posttest but the appointment was scheduled after thattion (88.4%) of the women completing both the pretest time. We compared intervention and control groups inand the posttest surveys had attended at least half of the the increase in the percentage of women who had hadeducational sessions. In contrast, only 44.0% of women the test in the previous year as an estimate of the impactwho completed only the pretest attended at least half of of the intervention. The outcome measure was chosen

the sessions. In an interview with consejeras, we exam- to detect changes in the screening behavior that couldined the reasons for participant drop-out. The most have only occurred between pretest and posttest. Ascommon reasons were related to personal circum- described in Methods, a success (value of 1) wasstances and not to the nature of the program. Specifi- possible only in women who (1) had not had a cancercally, the reasons mentioned most frequently were that screening test in the year before the pretest and (2) hadthe person moved out of the area or that they had completed the specified cancer screening test in thefound jobs with schedules that made it impossible to year before posttest. That is, by definition, none of thecontinue attending the educational sessions. Moreover, women who had had a particular screening test in thethe primary reason for failing to complete the posttest year before pretest had a positive outcome at posttestinterview was that the participants had moved and no (i.e., they were assigned a value of 0). Since breast and

forwarding address was available at the time the posttest cancer screening tests performed by health profession-was scheduled (86%). The posttest was scheduled upon als are not routinely recommended more than once acompletion of the twelfth weekly educational session by year, it is unlikely that women who had had a particularthe group (approximately three months after pretest) screening test in the year before the pretest would needand no later than six months after the pretest. Our to have an additional exam between pretest and post-analyses indicated that those who remained in the test. In the analyses presented in this paper, the per-program were similar in most ways to those who centages were calculated including all project partici-dropped out. Thus, exposure to the program remains pants in the denominator. We also conductedthe best explanation for the observed differences in additional analyses in which the denominator includedchange between the intervention and control groups, only women who had not had the screening tests the

One additional limitation of our findings is that the year before pretest. Since screening rates were similaroutcomes are based exclusively on self-report. We ex- between intervention and control groups, the results ofplored the possibility of contacting the health care these analyses are consistent with the results presentedproviders where the women indicated that they had in this paper. Furthermore, we also conducted multi-had their last cancer screening test. However, the variate analyses in which logistic regression modelsoption was not feasible within our limited budget. We were used to estimate the impact of the interventionfound that resources necessary to collect information after adjusting for health insurance status, level offrom clinical records to validate self-report data would education, and acculturation. The results of thesehave been especially high considering (1) the high analyses also confirm the results presented in thisproportion of women who did not have regular health paper.care providers (i.e., 42.6%), and (2) a fifth of the Our results indicate that there were no statisticallywomen with a regular health care provider received significant differences between experimental and con-health care services in Mexico. Future efforts will be trol groups in the increase of the percentage of womenneeded to confirm the validity of self-report in this who had physical breast exams, but differences in thepopulation. In a similar population in E1 Paso, Texas, increase of mammography use were statistically signifi-Suarez and colleagues 25found that self-reports of mare- cant. These results seem contradictory at first glance.mograms and Pap tests overestimate the prevalence of However, note that the analyses on the use of mare-screening. Further, we cannot completely rule out the mography were restricted to women who were 40 yearspossibility that the overestimation of the actual screen- old or older. Analyses on the use of physical breasting rates is higher in the intervention group than in the exam by a health professional in women over 40 yearscontrol groups at posttest, of age indicate that the rate of improvement is twice the

An additional limitation is that the impact of the size in the cancer group compared to the controlintervention was evaluated only in the short term. As group. Specifically, the percentage of improvement isdescribed above, posttest data were collected upon 30.4% in the cancer group versus 17.5% in the controlcompletion of the twelve weekly educational sessions group when using participant as the unit of analysis.and no later than six months after the first educational The consejera-based analyses indicate an improvementsession was conducted. Appointments for routine can- of 33.1% in the cancer group compared to 16.8% in the

cer screening tests are not always readily available a few control group. The differences approach statistical

AmJ Prev Med 1998;15(1) 39

significance at the. 10 level (P values are. 113 and .097 3. Breckon DJ, Harvey JR, Lancaster RB. Community health

for participant and consejera analyses, respectively), education: settings, roles, and skills for the 21st century.

One of the advantages of the Por La Vida model is Rockville, MD: Aspen Publications; 1994.

that it uses existing social networks within the commu- 4. Vega WA, Amaro H. Latino Outlook: Good health, un-certain prognosis. Annu Rev Public Health. 1994;15:39-

nity. Consejeras are selected based on their characteris- 67.tics as natural helpers in their respective social net-

5. Eng E, Parker E, Harlan C. Lay health advisor interven-works. In particular, the PLV intervention model uses tion strategies: A continuum from natural helping tothe link-person natural support networks, which are paraprofessional helping. Health Educ Behav 1997;24:most common among Latinos in general and Mexican- 413-17.

Americans in particular. 17 Link-person networks are 6. Breen N, Kessler L. Changes in the use of screening

comprised of mutually linked individuals bound by ties mammography: evidence from the 1987 and 1990 Na-

of friendship based on reciprocity and exchange behav- tional Health Interview Surveys. Am J Public Health

iors. 17 During the educational sessions, the consejeras 1994;84:62-7.

become models who promote and reinforce behaviors 7. Kaplan RM, Navarro AM, Castro FG, ElderJP, Mishra SI,

among their peers. Social learning theory suggests that et al. increased use of mammography among Hispanicsimilar and admired models are most effective in women: baseline results from the NCI cooperative group

on cancer prevention in Hispanic communities. Am Jenhancing behavior change. 26 Furthermore, previous Prev Med 1996;12:467-71.

studies provide evidence that positive health behavior is 8. Rimer BK_Mammography use in the U.S.: trends and themore likely to occur when reinforced by members of a impact of interventions. Ann Behav Med 1994;16:317-26.

social network. 27 9. Skinner CS, Strecher VJ, Hospers H. Physicians' recom-

The results of this paper reinforce the importance of mendations for mammography: do tailored messagesutilization of intervention models created from and for make a difference? Am J Public Health 1994;84:43-9.

the Latino community. Future demonstration studies 10. Rimer BK, Resch N, King E, Ross E, Lertuan C, Boyce A,

will be needed to test the feasibility and the impact of Kessler H, Engstrom PF. Multistrategy health education

the intervention model when implemented in a wider program to increase mammography use among womenages 65 and older. Public Health Rep t992;107:369-80.

geographic area with other populations. This could 11. Fletcher SW, Harris RP, GonzalezJJ, Degnan D, Lannininclude Latino populations that are not primarily Mex- DR, et al. Increasing mammography utilization: a con-ican American, Latinos of different socioeconomic trolled study. J Natl Cancer Inst 1993;85:112-20.levels or acculturation, or individuals residing in rural 12. Ruiz E, Caban CE. Introduction: cancer research in

populations. One step further, future studies could Hispanic populations. J Natl Cancer Inst Monogr 1995;examine whether the intervention model can be uti- 18:ix-xi.

lized as an integral part of mainstream existing health 13. Hubbell F, Chavez L, Mishra S, Magana R, Valdez R.

care services to improve cancer screening rates and From ethnography to intervention: developing a breast

other health-related behaviors. Although cost-effective- cancer control program for Latinas. J Natl Cancer Inst

ness data are not currently available, theoretical and Monogr 1995;18:109-15.

practical considerations suggest that the implementa- 14. Castro F, Elder J, Coe K, Tafoya-Barraza H, Moratto S, ettion of the intervention model has modest costs and the al. Mobilizing churches for health promotion in Latino

communities: compa_eros en la salud. J Natl Cancer Instbenefits to health promotion in the low-socioeconomic Monogr 1995;18:127-356.

Latino community can be substantial. 15. Ramirez A, McAlister A, Gallion K, Ramirez V, Garza I, et

al. Community level cancer control in a Texas barrio: Part

Supported by grant U01-52948 from the National Cancer I. Theoretical basis, implementation, and process evalua-Institute, National Institutes of Health. tion. J Natl Cancer Inst Monogr 1995;18:117-22.

16. Navarro AM, Senn KL, Kaplan RM, McNicholas L, CampoMC, Ropp_ B. Por La Vida intervention model for cancerprevention in Latinas. J Natl Cancer Inst Monogr 1995;

References 18:137- 45.1. u.s. Department of Health and Human Services, Healthy 17. Valle R. A natural resource system for health-mental

People 2000: national health promotion and disease health promotion in the Latino/Hispano communities.prevention objectives. Washington, DC: U.S. Department In: Valle R, Vega W, Eds. Hispanic natural supportof Health and Human Services; 1991. systems. Sacramento, CA: Department of Mental Health;

2. Latino Task Force for the Unity in Health, Diversity in 1980:35-43.

Culture Conference. Latino task force report on the year 18. Sourcepoint. The Census Data Center for the San Diego

2000 health promotion objectives and recommendations Region. 1990 census, Hispanic population by race: majorfor California. Sacramento, CA: California Department of statistical areas, subregional areas, jurisdictions, censusHealth Services, Health Promotion Section; 1992. tracts. San Diego, CA: Sourcepoint Corp.; 1991.

40 American Journal of Preventive Medicine, Volume 15, Number 1

19. Senn KL, Roppe B, Rupp JW, McNicholas LJ, Vega WA, Palo Alto, California: Stanford Cefiter for Research inNader PR. Por La Vida: accessing the comn_unity. Palo Disease Prevention; 1989.

Alto, California: Stanford Center for Research in Disease 24. Senn KL, McNicholas LJ, Ropp6 B, Lozada P. Por La VidaPrevention; 1990. luchando contra el uso de tabaco/the fight against tobacco: a

20. Mendoza L. Hispanic helping networks: Techniques of Spanish/English facilitator's guide to a tobacco preven-cultural' support. In: Vega W, Valle R, eds. Hispanic tion program. San Diego, California: Center for Behav-natural support systems: mental health promotion per- ioral and Community Health Studies, San Diego Statespectives. Sacramen.to, California: Department of Mental University Foundation; 1991.Health; 1980:55-63. 25. Suarez L, Goldman DA, Weiss NS. Validity of Pap smear

21. Marin G, Sabogal F, Marin BV, Otero-Sabogal R, Pfirez- and mammogram self-reports in a low-income HispanicStable E. Development of a short acculturation scale for population. Am J Prey Med 1995;11:94-8.Hispanics. Hispanic J Behav Sci 1987;9:183-205. 26. Bandura A. Social foundations of thought and action.

22. Sarason IG, Levine HM, Basham RB, Sarason BR (1983) Englewood Cliffs, New Jersey: Prentice-Hall; 1986.Assessing social support: the Social Support Question- 27. Kaplan RM, Toshima MT. The functional effects of socialnaire. J Personality Social Psychol 1983;44:127-39. relationships on chronic illnesses and disability. In: Sara-

23. Senn KL, Roppe B, Rupp JW, McNicholas LJ, Vega WA, son BR, Sarason IG, Pierce GR, eds. Social support: anNader PR. Por la Vida: a Spanish/English facilitator's interactional view. Wiley series on personality processes.guide to a community health improvement program. New York:John Wiley & Sons; 1990:427-453.

AmJ Prev Med 1998;15(1) 41