Psychosocial Variables, External Barriers, and Stage of Mammography Adoption

Transcription

1 Health Psychology Copyright 2003 by the American Psychological Association, Inc. 2003, Vol. 22, No. 6, /03/$12.00 DOI: / Psychosocial Variables, External Barriers, and Stage of Mammography Adoption Diane Ruth Lauver and Jeffrey B. Henriques University of Wisconsin Madison Lori Settersten University of Wisconsin Milwaukee Mary Carson Bumann University of Wisconsin Madison Guided by H. Triandis s (1980) theory of behavior and the transtheoretical model, the study purpose was to examine differences in psychosocial variables and external conditions by stage of mammography adoption. Sampled from a statewide population, participants (N 509) were women aged They had been contacted by telephone, screened for eligibility (e.g., no history of cancer or recent mammogram), and interviewed 3 6 months later. Higher utility beliefs, social influences, and practitioner interactions about mammography were associated with improved stage of adoption, as were lower negative affect and external barriers regarding mammography. Higher decisional balance scores, with and without negative affect toward mammography, were associated with improved stage. Controlling for variables reflecting pros and cons, negative affect toward mammography further distinguished among stages. A richer set of pros and cons measures could explain screening more fully. Key words: Triandis s theory of behavior, transtheoretical model, mammography behavior Understanding women s adoption of mammography, especially among those who have not sought it as recommended, is important as a theoretical and clinical issue. Both Triandis s theory of behavior (1980) and the transtheoretical model (TTM) have been useful in guiding research on health-related behaviors and mammography (e.g., Brown et al., 1996; Lauver, 1994; Lauver & Kane, 1999; Prochaska & Velicer, 1997; Prochaska et al., 1994). The aims of this study were to examine whether explanatory variables proposed by Triandis differed by stage and whether affect toward mammography further distinguished among stages of adoption, after controlling for explanatory variables proposed in the TTM. From Triandis s theory, psychosocial variables (i.e., utility beliefs, social norms, and affect regarding a behavior), external Diane Ruth Lauver and Jeffrey B. Henriques, School of Nursing, University of Wisconsin Madison; Lori Settersten, School of Nursing, University of Wisconsin Milwaukee; Mary Carson Bumann, Institute on Aging, University of Wisconsin Madison. This project was supported by National Institutes of Health Grant 3 R01 CA/NR58978, an Office of Research on Women s Health supplement, and in part by the University of Wisconsin Center for Women s Health and Women s Health Research. Appreciation is due to Judy Egan, Victoria Shaw, Joan Nugent, Miriam Jacobson, Ellen Vander Galien, and Brenda Owen for data collection; to Polly Newcomb and her team for assistance with recruitment; to Patricia Feltskog for data entry and management; to Janet Kane for statistical consultations; to Jean E. Johnson for reading the early draft; and to Kimberly Nolet and Elizabeth First for manuscript preparation. Thanks also to all participants, without whom this project would not have been possible. Correspondence concerning this article should be addressed to Diane Ruth Lauver, School of Nursing, K6/350 CSC, University of Wisconsin Madison, 600 Highland Avenue, Madison, Wisconsin drlauver@wisc.edu conditions, and habit are explanatory variables of intention and behavior. Applied to mammography, utility beliefs refer to one s perceptions of worthwhile outcomes of mammograms. Norms refer to one s perceptions of peers and professionals beliefs about mammograms. Affect toward mammography refers to women s feelings associated with mammograms. External conditions refer to objective factors that inhibit or facilitate mammography, especially in the health care system. Habit refers to an established pattern or prior engagement in breast screening (Lauver, 1992; Triandis, 1980). Sociodemographic and clinical factors are reflected mostly in the psychosocial, external conditions, and habit variables and are not proposed as direct explanatory variables. Utility beliefs, norms, affect, external barriers, and habit regarding mammography have been shown to explain women s mammography use (e.g., Brown et al., 1996; Lauver, Kane, Bodden, Mc- Neel, & Smith, 1999). According to the TTM, individuals progress through a series of staged steps as they adopt behaviors. The steps have included precontemplation (neither engaging nor planning to engage in the behavior), contemplation (not engaging in the behavior, but considering it), preparation (taking steps towards engagement in the behavior), and action (having begun engagement in the behavior), among others (Prochaska et al., 1994; Rakowski, Dube, & Goldstein, 1996). A central finding, based on the TTM, involves decisional balance. Individuals whose pros perceived positive aspects of engaging in a recommended health behavior outweighed their cons perceived negative costs of a behavior have been more likely to adopt that behavior (Prochaska et al., 1994). When women s pros have exceeded their cons for mammography, they have been more likely to engage in mammography (e.g., Rakowski et al., 1996, 1997). 649

2 650 BRIEF REPORTS Whereas the TTM includes a concept of decisional balance measured by subtracting cons from pros (Marcus, Rakowski, & Rossi, 1992; Rakowski et al., 1997) Triandis s theory (1980) does not. Incorporating such a concept into Triandis s theory and research could yield greater understanding of the steps taken in adoption of a behavior. Whereas researchers guided by Triandis have measured breast cancer detection behaviors with a dichotomy (e.g., Brown et al., 1996; Lauver, 1994; Lauver & Kane, 1999), TTM researchers have measured mammography with stage as a category (e.g., Prochaska et al., 1994). By measuring a behavior with multiple levels, researchers could achieve greater sensitivity in explaining behavior. We examined the proposed explanatory variables in these two theories. Pros about mammography have focused on individuals perceptions of positive outcomes and benefits identified by professionals. Cons for mammography have focused on undesirable aspects of the procedure, most often the perceived undesirable consequences (Rakowski et al., 1992, 1996, 1997). Affect toward mammography could reflect either pros or cons, depending on whether feelings are positive or negative about the behavior. Utility beliefs refer only to perceived positive aspects of engaging in a behavior and could reflect pros. Because norms could reflect support or lack of support for a behavior, norms could be considered either pros or cons. External conditions include costs of screening (e.g., financial and time), so these factors are consistent with the concept of cons. No researchers have examined whether the explanatory variables proposed by Triandis differ by stage as defined in the TTM. In prior TTM research on mammography, measures of pros and cons may not have reflected adequately other salient dimensions. Although Rakowski, Dube, et al. (1996) suggested that pros or cons may be based on feelings, affect toward mammography rarely has been measured in TTM research (e.g., Marcus et al., 1992; Rakowski et al., 1996). Pros and cons measures have had only one or two items to reflect peer or professional norm (e.g., Marcus et al., 1992; Rakowski, 1996, 1997). Measures of cons about mammography have had only one item to reflect such external conditions (e.g., Rakowski et al., 1996, 1997). Including measures that substantively and adequately reflect salient affect, external conditions, as well as peers and professional norms regarding screening, could improve explanations of screening behavior and designs of tailored interventions to promote such behavior (Champion & Huster, 1995; Rakowski et al., 1998). Our study hypotheses were that: (a) utility beliefs, social influences, and practitioner interactions about mammography will differ by stage and be higher with improved stage of adoption; (b) negative affect and external barriers regarding mammography will differ by stage and be lower as stage of adoption improves; (c) decisional balance scores will differ by stage and be higher as stage of adoption improves; and (d) after controlling for utility beliefs, social influences, practitioner interactions, and external barriers, affect toward mammography will provide additional information by distinguishing between stages of adoption. Sample Method A cross-sectional design and secondary data analyses from a longitudinal study (Lauver, Settersten, Kane, & Henriques, 2003) were used. Women aged who had not had mammograms in the previous 13 months and who had never had cancer were sought; this was done to focus on those for whom mammography was needed and to control partly for habit. Women (N 509) were sampled randomly from a statewide population in the midwest through driver s license and Medicare lists. Participants average age was 66.4 (SD 7.66); 86% had graduated from high school; 98.8% were Caucasian, 0.8% African American, 0.2% Native American, and 0.2% Other; 46% had both Medicare and private insurance, 31% had only private insurance, 16% had only Medicare, and 5% had no health care coverage. Procedure After approval from institutional review boards, age-eligible women were identified. Women received a letter explaining the study, notice of an upcoming phone contact, and a name to contact if they wished not to be called. In an initial phone contact, we confirmed eligibility about prior mammography use, cancer history, and ability to communicate in English. From 3 to 6 months after identification, eligible participants were interviewed regarding subsequent mammography behavior. They first were asked questions about their engagement in breast screening (i.e., mammography, clinical breast examination, and breast self-examination), then about psychosocial variables and external conditions regarding mammography, clinical factors, and demographic factors. Measures Stage of mammography adoption was based on women s intentions and behaviors; our questions were based on theory and prior research (e.g., Prochaska et al., 1994; Rakowski et al., 1996). For intention, women were asked, What are the chances of you seeking and having a mammogram, within the next year? Responses were from not at all likely (0) to definitely likely (4). Because the distribution of these scores was skewed, we dichotomized them at low ( 3) or high (4); 44% of participants had intention scores of 4. For behavior, women were asked, Since we spoke with you last [month of contact], have you talked about, or done anything about, mammograms? Interviewers clarified women s responses regarding recent mammography-related behavior to classify them by stage. Interviewers asked whether participants had discussed mammograms with others and whether they had made an appointment for one. Stage categories were based on theoretical descriptions and prior research (Prochaska et al., 1994; Rakowski, Dube, et al., 1996; Stoddard et al., 1998), as follows: (a) precontemplation, neither having had a prior mammogram nor having high intentions to have one; (b) relapse, having had a prior mammogram, now off-schedule, with low intentions to seek one; (c) risk of relapse, having had a recent mammogram (in the prior 3 6 months), with low intentions to seek one; (d) contemplation, neither having a recent mammogram nor an appointment for one, with high intentions to seek one; (e) preparation, not having had a recent mammogram, with an appointment for one, and high intentions to seek one; and (f) action, having had a recent mammogram with high intentions to seek. Measures of utility beliefs, norms, affect, and external barriers regarding mammography were short-answer questions using scales from 0 to 4 to reflect either strongly disagree to strongly agree or not at all to extremely. Utility beliefs were measured by 14 items reflecting positive outcomes of having mammograms (e.g., Montano & Taplin, 1991). One dimension of norms, social influences, was measured with 5 items reflecting perceived normative beliefs of friends or family regarding mammography (Lauver, 1996). One dimension of norms, interpersonal interactions with health practitioners about screening, was measured with 5 items (Lauver, Henriques, Owen, Egan, & Lovejoy, 2003). Affect toward mammography was measured with a 13-item mood adjective checklist about having a routine mammogram (Lauver & Kane, 1999; McNair, Lorr, & Droppleman, 1971; Sandra Underwood, personal communication, May 1993). External, objec-

3 BRIEF REPORTS 651 tive barriers were measured with 11 items (Melnyk, 1990) reflecting difficulties with affordability and accessibility of screening. Barriers scores were low and skewed. A median split was used to create a dichotomy ( 2 and 3 to reflect either lesser or greater barriers). For analyses based on Triandis s theory, selected items were reverse coded so that utility beliefs and norms scales reflected support for mammograms, whereas affect and external barriers reflected lack of support for mammograms. For utility beliefs, norms, and affect regarding mammography, items were averaged to create scale scores; for barriers, items were summed. Cronbach alphas for utility beliefs, social influences, interpersonal interactions with professionals, affect, and barriers regarding mammography were.85,.89,.89,.89, and.78, respectively (cf. Lauver, Settersten, et al., 2003, for details on measures). A decisional balance score was computed. Utility beliefs and social and professional norms items supportive of mammography were transformed to Z scores and summed to reflect pros. Transformed barriers and norms items not supportive of mammography were summed to reflect cons. Positive affect toward mammography items were conceptualized as pros, and negative affect toward mammography items as cons (Rakowski et al., 1996). Decisional balance scores were computed by subtracting the cons from the pros (Marcus et al., 1992; Rakowski et al., 1997). Two sets of pros and cons and decisional balance scores were computed, one did not include affect toward mammography and one did. Cronbach alphas for pros and cons without affect were.87 and.73, respectively; with affect, they were.89 and.86, respectively. Demographic factors assessed were age, education, occupation, type of health insurance, marital status, and race. Clinical factors assessed were personal history of breast problems and of breast biopsy and family history of cancer or breast cancer. Results The Statistical Packages for Social Sciences (Version 9.0) were used for analyses. Overall, 19.8% (101) of participants were categorized in precontemplation, 27.3% (139) in relapse, 9.2% (47) in risk of relapse, 28.1% (143) in contemplation, 5.5% (28) in preparation, and 10% (51) in action. Associations of clinical and demographic factors with stage were examined. Age, education, and personal history of benign breast biopsy differed by stage, F(5, 499) 3.34, p.01, F(5, 501) 2.76, p.05; F(5, 500) 2.76, p.001, respectively, and were selected as covariates for subsequent analyses. Also, family history of breast cancer was selected as a covariate because it has been associated with mammography use (e.g., Lauver et al., 1999). A multivariate analysis of covariance (MANCOVA) was run with stage as the grouping variable. Scores for beliefs, social influences, practitioner interactions, negative affect toward mammography, and external barriers were entered as dependent variables. The effect of stage was significant in the overall MANCOVA, after controlling for clinical and demographic factors, F(25, 1,807) 10.06, p.001. To test our first and second hypotheses, to determine the significance of each proposed variable, and to follow up on the MANCOVA, we ran one-way analyses of covariance (ANCOVAs). As we proposed, beliefs, social influences, and practitioner interactions about mammography differed by stage, with the lowest scores being among precontemplators. Furthermore, negative affect toward mammography and external barriers about mammography also differed by stage, with the lowest scores being among those in the action group (see Table 1). Post hoc tests of these ANCOVAs were conducted to determine whether the adjusted means differed among stage categories using Tukey s b procedure. Means that differed ( p.05) by stage category are shown by subscripts in Table 1. Because the pattern and direction of Triandis s (1980) explanatory variables differed by stage, as we proposed, our findings support the importance of Triandis s variables in differentiating among stages of mammography adoption. To test our third hypothesis, we conducted MANCOVAs on pros, cons, and decisional balance scores, with stage as the grouping variable. The first set of MANCOVAs did not incorporate affect toward mammography, and the second set of MANCOVAs did. In both sets of analyses, pros, cons, and decisional balance scores differed (see Table 2). Pros and decisional balance were highest with improved stage, but cons were lower with improved stage. In these analyses, precontemplators differed from relapsers, and both of these groups differed from all others. As hypothesized from Triandis s theory and the TTM, we observed differences in pros, cons, and decisional balance scores by stage. To test our fourth hypothesis, we ran a series of MANCOVAs to examine the unique contributions of the variables from Triandis s (1980) theory. In the first MANCOVA, beliefs and barriers differed by stage, F(5, 490) 15.13, p.001, and F(5, 490) 3.62, p.05, respectively, consistent with our prior analyses. A second MANCOVA revealed that social influences did not differ by stage, but that practitioner interactions did, F(5, 488) 1.51, p.05, and F(5, 488) 11.24, p.001, respectively, after controlling for variables entered in the first MANCOVA. The failure of social influences to distinguish between stages was unexpected given the greater breadth of items used to reflect social influences as compared with prior research guided by the TTM (Rakowski et al., 1996, 1997). Yet, social influences may not have explained stage because these influences were correlated with practitioner interactions (r.25, p.001) as well as with utility beliefs (r.25, p.001). Finding that practitioner interactions differed by stage after controlling for beliefs, barriers, and social norms underscores the importance of having adequate measures of practitioner interactions when explaining mammography adoption (Lauver, Henriques, et al., 2003; Skinner, Champion, Gonin, & Hanna, 1997). A final MANCOVA revealed that affect toward mammography distinguished between stages even after controlling for variables entered in the second MANCOVA. This test of Triandis s proposition demonstrates that affect about mammography differentiates among stages, after controlling for the clinical, demographic, beliefs, norms, and barriers variables. This finding extends prior research on mammography based on the TTM (e.g., Marcus et al., 1992; Rakowski et al., 1996) in which variables reflecting affect toward mammography had not been included. Discussion New contributions of our study include that we demonstrated that explanatory variables proposed by Triandis differed by stage of mammography adoption. Also, we computed sets of pros, cons, and decisional balance scores that did and did not incorporate affect toward mammography and demonstrated that both sets of scores differed by stage and in directions as we proposed. Although our decisional balance scores were derived from variables based on Triandis s (1980) theory, rather than from measures used in research based on the TTM, they differed across categories of

4 652 BRIEF REPORTS Table 1 Adjusted Means and Standard Errors for Negative Affect Toward Mammography, Beliefs, Barriers, Social Influences, and Health Care Practitioner Interactions by Stage of Adoption Precontemplation (n 101) Relapse (n 139) Risk relapse (n 47) Contemplation (n 143) Preparation (n 28) Action (n 51) Measure M SE M SE M SE M SE M SE M SE F(5, 490) Beliefs a b c c c c 30.90** Social influences a a a, b b a, b b 4.41** Practitioner interactions a b b b b b 19.27** Negative affect a b c, d b, c b, c, d d 17.68** Barriers a a a, b a, b a, b b 3.62* Note. N 509. Means in the same row that do not share subscripts differ at p.05 (Tukey s honestly significant difference test). * p.01. ** p.001. Table 2 Adjusted Means and Standard Errors for Pros, Cons, and Decisional Balance by Stage of Adoption Precontemplation (n 101) Relapse (n 139) Risk relapse (n 47) Contemplation (n 143) Preparation (n 28) Action (n 51) Measure M SE M SE M SE M SE M SE M SE F(5, 490) Without affect toward mammography Pros a b c c c c 27.45* Cons a b b, c b, c a, b, c c 8.87* Decisional balance a b c c c c 31.53* With affect toward mammography Pros a b c c c c 31.25* Cons a b b, c b, c b, c c 15.17* Decisional balance a b c, d c c, d d 38.74* Note. N 509. Means in the same row that do not share subscripts differ at p.05 (Tukey s honestly significant difference test). * p.01.

5 BRIEF REPORTS 653 stage consistent with propositions in the TTM. Finding that the decisional balance scores differed by stage also extends prior research on Triandis s theory, because this theory does not incorporate a concept of or propositions about decisional balance. We identified the unique contributions of variables proposed by Triandis (1980) to stage. One new contribution of our study is that affect toward mammography differed significantly by stage after controlling for measures of clinical, demographic factors, utility beliefs, social influences, practitioner interactions, and barriers. This suggests that measures in prior research, based on the TTM, have been less than adequate by failing to include affect toward the target behavior (e.g., Marcus et al., 1992; Rakowski et al., 1996, 1997). Our study extends prior TTM research on mammography by examining the differences in stage by social influences in multivariate, rather than in bivariate analyses (e.g., Pearlman, Rakowski, & Ehrich, 1995). After utility beliefs and external barriers were controlled for, we found that professional interactions differed by stage but social influences did not. Researchers can distinguish between types of such influences. Future intervention research could determine the effects of addressing selected variables at particular stage categories on mammography adoption. Researchers who design tailored interventions could test whether promoting utility beliefs and reducing external barriers or negative affect about mammography promotes stage progression in general and examine at what particular stage such interventions would be most effective (Champion & Huster, 1995; Rakowski et al., 1998). Researchers could test whether social influences from peers or professionals promote progression later, rather than earlier, in stage progression as proposed by Prochaska and Velicer (1997). In summary, our findings support assumptions both in Triandis s (1980) theory of behavior and in the TTM (e.g., Prochaska & Velicer, 1997). Overall, utility beliefs, supportive social influences, and encouraging practitioner interactions were associated with a higher stage of adoption whereas negative affect toward mammography and external barriers were associated inversely with stage. This research extends research based on Triandis s (1980) theory by demonstrating that decisional balance measures differed by stage. Advancing TTM research, decision balance scores, computed both with and without affect toward mammography, differed by stage of mammography adoption. This study provides new findings that practitioner interactions varied positively with stage after controlling for beliefs, barriers, and social influences, that negative affect about mammography varied inversely with stage, and that this association persisted, after controlling for utility beliefs, norms, and external barriers. Thus, explanations of breast screening may be enhanced by including a richer set of items in future measures of pros and cons, particularly including affect toward mammography and practitioner interactions about breast screening. References Brown, R., Baumann, L., Helberg, C., Han, Y., Fontana, S., & Love, R. (1996). The simultaneous analysis of patient, physician and group practice influences on annual mammography performance. Social Science & Medicine, 43, Champion, V., & Huster, G. (1995). Effect of interventions on stage of mammography adoption. Journal of Behavioral Medicine, 18, Lauver, D. (1992). A theory of care-seeking behavior. Image: A Journal of Nursing Scholarship, 24, Lauver, D. (1994). Care-seeking behavior with breast cancer symptoms in Caucasian and African-American women. Research in Nursing & Health, 17, Lauver, D. (1996). Understanding barriers to mammography use among women of low socioeconomic status: Comparisons from quantitative and qualitative data. Journal of Women s Health, 5, Lauver, D., Henriques, J., Owen, B., Egan, J., & Lovejoy, L. (2003). Relationships of practitioner communications and characteristics with women s mammography use. Patient Education & Counseling, 51, Lauver, D., & Kane, J. (1999). A motivational message, external barriers, and mammography utilization. Cancer Detection and Prevention, 23, Lauver, D., Kane, J., Bodden, J., McNeel, J., & Smith, L. (1999). Engagement in breast screening behaviors. Oncology Nursing Forum, 26, Lauver, D., Settersten, L., Kane, J., & Henriques, J. (2003). Tailored messages, external barriers, and women s utilization of professional breast cancer screening over time: A randomized clinical trial. Cancer, 97, Marcus, B., Rakowski, W., & Rossi, J. (1992). Assessing motivational readiness and decision making for exercise. Health Psychology, 11, McNair, D., Lorr, M., & Droppleman, L. (1971). Profile of mood states. San Diego, CA: Educational and Industrial Testing Service. Melnyk, K. (1990). Barriers to care, operationalizing the variable. Nursing Research, 39, Montano, D., & Taplin, S. (1991). A test of theory to predict mammography. Social Science & Medicine, 32, Pearlman, D. N., Rakowski, W., & Ehrich, B. (1995). The information environment of women and mammography screening: Assessing reciprocity in social relationships. Journal of Women s Health, 4, Prochaska, J., & Velicer, W. F. (1997). The transtheoretical model of health behavior change. American Journal of Health Promotion, 12, Prochaska, J., Velicer, W., Rossi, J., Goldstein, M., Marcus, B., Rakowski, W., et al. (1994). Stages of change and decisional balance for 12 problem behaviors. Health Psychology, 13, Rakowski, W., Dube, C., Goldstein, M. (1996). Considerations for extending the transtheoretical model of behavioral change to screening mammography. Health Education Research: Theory & Practice, 11, Rakowski, W., Dube, C., Marcus, B., Prochaska, J., Velicer, W., & Abrams, D. (1992). Assessing elements of women s decisions about mammography. Health Psychology, 11, Rakowski, W., Ehrich, B., Dube, C., Pearlman, D., Goldstein, M., Peterson, K., et al. (1996). Screening mammography and constructs from the theoretical model: Associations using two definitions of the stages-ofadoption. Annals of Behavioral Medicine, 18, Rakowski, W., Ehrich, B., Goldstein, M., Rimer, B., Pearlman, D., Clark, M., et al. (1998). Increasing mammography among women aged by use of a stage-matched, tailored intervention. Preventive Medicine, 27, Rakowski, W., Stoddard, A., Rimer, B., Fox, S., Andersen, M., Urban, N., et al. (1997). Confirmatory analysis of opinions regarding the pros and cons of mammography. Health Psychology, 16, Skinner, C., Champion, V., Gonin, R., & Hanna, M. (1997). Do perceived barriers and benefits vary by mammography stage? Psychology, Health, & Medicine, 2, Stoddard, A., Rimer, B., Lane, D., Fox, S., Lipkus, I., Luckmann, R., et al. (1998). Underusers of mammogram screening: Stage of adoption in five U.S. subpopulations. Preventive Medicine, 27, Triandis, H. (1980). Values, attitudes and interpersonal behavior. In M. Page (Ed.), 1979 Nebraska Symposium on Motivation (pp ). Lincoln, NE: University of Nebraska Press.