Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Developing a Global Cancer Stigma Index:

Developing a Global Cancer Stigma Index: Despite increasing recognition about the stigma associated with cancer screening, diagnosis, and treatment-seeking behaviors, there has been relatively little attention paid to how to assess and intervene to reduce that stigma. An index to measure cancer stigma could empower health program developers and policymakers by identifying the key areas in which a population could benefit from education to change perceptions and address misinformation. The index also could be used to rank countries and communities based on their level of cancer stigma to assess where interventions are needed. We used structured literature review and expert review to generate a cancer stigma item pool. The item pool was subject to cognitive interviews for cultural appropriateness and comprehension; and data from initial pilot testing were used to reduce the pool of items for translation and field testing. The field test was conducted using a web-based survey in four samples representing two regions and three languages—English and Arabic speakers in Jordan and Egypt, and English and Mandarin Chinese speakers in China. Factor analyses and item response theory were applied to finalize the index. The analyses resulted in a 12-item cancer stigma index (CSI) that was reliable across all four samples. The CSI scores were highly correlated with a general illness stigma scale, and operated as expected noting higher cancer stigma among men and those with lower income. The CSI can be used to inform initial cancer education efforts, identifying overall stigma levels in a country or community and particular issue areas requiring intervention. Keywords cancer, stigma, health, chronic disease, help-seeking, health disparities Specific types of cancer may also carry disease-specific Introduction stigma. For example, cervical and lung cancer are often cited Over the last several years, stigma associated with health- because each is linked to behavior that may be deemed unde- seeking behaviors has received increasing attention. sirable or marginal. In the case of lung cancer, individuals Health-related stigma (or principally, disease-related may feel guilt and shame attributed to their diagnosis, due to stigma) is unique and presents significant challenges and the link between smoking and cancer. Guilt may lead to barriers for the global health community to overcome. The denial of the diagnosis until such a point that treatment may framing of health-related stigma has begun to advance a not be successful (Batson et al., 1997). Cervical cancer, more complex discussion of stigma, one that encompasses breast cancer, and uterine cancer may also carry a particular both the internalization of stigma by the individual and the stigma as these cancers are often linked to sexual health, public reaction and potential marginalization that may regardless of the actual disease pathway. In some patriarchal occur. First, the word “disease” alone can induce a sense of societies, women are considered to be the property of their stigma (Green, Davis, Karshmer, Marsh, & Straight, 2005; spouses, and must comply with their spouse’s wishes to not Pettit, 2008). There are a number of diseases that have his- seek treatment. Religious and cultural beliefs may prohibit torically been highly stigmatized, including mental health disorders, HIV/AIDS, sexually transmitted diseases, lep- rosy, and skin diseases (Greene & Banerjee, 2006; Sartorius, RAND Corporation, Boston, MA, USA 2007). An emerging body of data in the stigma literature RAND Corporation, Arlington, VA, USA indicates that cancer is also often among the diseases that is RAND Corporation, Santa Monica, CA, USA highly stigmatized (LIVESTRONG Foundation, 2007), yet Livestrong Foundation, Austin, TX, USA it is less researched than other health issues. Individuals Corresponding Author: often react to that stigma by making decisions about Maria Orlando Edelen, RAND Corporation, 20 Park Plaza, Suite 920, whether or not to disclose their condition or seek treatment Boston, MA 02116, USA. (Joachim & Acorn, 2000). Email: Maria_Edelen@rand.org This article is distributed under the terms of the Creative Commons Attribution 3.0 License Creative Commons CC BY: (http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access page (http://www.uk.sagepub.com/aboutus/openaccess.htm). 2 SAGE Open seeking medical attention for parts of the body having a sex- comprehensible, (b) understand how respondents interpreted ual connotation (Brewster & Moradi, 2010). the items, and (c) ensure that the item content and wording Despite this emerging recognition that stigma related to were culturally appropriate. Because our initial pilot test was cancer can deter critical health-seeking behaviors, there has to be conducted in the Middle East, cognitive interview par- been comparatively little effort to measure the level of can- ticipants, although U.S.-based, were of Arab heritage, had cer stigma in a given population or community to intervene. been raised in the region (at least up to age 18), considered Most stigma scales have focused in other areas, principally themselves culturally Arab, and spoke both English and HIV/AIDS and mental health (Brohan, Slade, Clement, & Arabic. All respondents received a US$25 gift card for their Thornicroft, 2010; Evans-Lacko et al., 2010; Uys et al., time, and all procedures were approved by the RAND 2009). This study sought to develop a cancer stigma index Institutional Review Board (IRB). Following review of the (CSI) to measure perspectives on cancer, specifically atti- cognitive interview transcripts, we made changes to the item tudes about cancer screening and treatment, and to help pool based on respondent feedback including dropping items inform awareness and education programs. We proceeded in that were considered redundant and generating items to two phases. First, we conducted an initial study to gather reflect stigma and religion, a theme that was considered stigma measures and test a preliminary item pool (Study 1). important by interviewees. Then, we conducted a full field test in two regions (Study 2). Pilot Test Study 1: Creation of a Cancer Stigma Method Item Pool and Initial Pilot Test Measures. Cancer stigma item pool. A set of 59 candidate We used established methods for item pool development fol- items, formatted using Likert-type response options (1 = not lowing the National Institutes of Health Patient-Reported at all to 5 = very much), were administered to all respon- Outcomes Measurement Information System (NIH PROMIS®) dents. initiative blueprint (Reeve et al., 2007). The main goals of PROMIS® (http://www.nihpromis.org/default.aspx), which is General illness stigma. As there were no items available part of the NIH’s Roadmap initiative, are to standardize a set of on general illness stigma that corresponded well to the can- assessment tools and to use item response theory (IRT) tech- cer stigma constructs we are trying to measure, we used niques and advances in computer technology to create brief yet information in the literature to develop five items reflecting highly reliable and flexible assessment tools to measure patient- attitudes of stigma toward general illness (e.g., If I had an reported outcomes (Ader, 2007; Cella et al., 2007; Fries, Bruce, illness, I would feel left out of things). These items also used & Cella, 2005). Due to its scope and success, the rigorous the five-category Likert-type response format. approach utilized by PROMIS has become something of a standard for modern instrument development (DeWalt, Demographics. All respondents provided basic demo- Rothrock, Yount, & Stone, 2007). graphic information, including gender, age, education, reli- Briefly, we first conducted a structured literature search gion, country of origin, country of residence, and language for stigma as it relates to cancer, HIV/AIDS, mental health, spoken at home. Six additional items indicated the respon- and other health issues. The initial search summarized litera- dents’ connections to cancer (e.g., I know someone with can- ture in these areas known to underlie stigma: information and cer, I am a caregiver of someone with cancer). myths, fear, shame and labeling, concerns about diagnosis, concerns about treatment-seeking, concerns about peer and Sample and Procedures family disclosure, and concerns about public disclosure. We identified and organized a total of 553 items from 29 mea- All study procedures were approved by the RAND IRB. We sures into four broad domains: characterization about those contracted with Harris Interactive to pilot the items via the with the disease, self-stigma among those with the disease, Internet to English-speaking adult respondents residing in expectations of what others may think of those with the dis- Egypt and Jordan. We collected a total of N = 1,016 com- ease, and positive views on those with the disease. After pleted web surveys. The majority of respondents were male review by five technical advisors, we reduced the number of (72%), and the sample was also skewed toward younger and items based on item scope and quality of item wording, and more educated individuals than is representative of the gen- reworded them to be appropriate for the cancer stigma con- eral population in the region (52% were <30 years old, 94% text and in comparable format. were <50 years old; 71% had university education). We addressed this imbalance by selecting a subset of the data that had the same number of men and women (294 each; Cognitive Interviews Total N = 588). All women from the original sample were Using the reduced set of items, we conducted seven cogni- retained, and we used a random sampling approach to select tive interviews to (a) assess whether the items were a subset of 294 males that was stratified according Edelen et al. 3 to education and an indicator of having lived outside the of the two-factor model was adequate (CFA = .905, TLI = country. Despite this stratification, the analytic sample was .901, RMSEA = .057). The final solution had 31 items in still relatively young (49% below 30 years of age) and fairly Factor 1 and 19 items in Factor 2. After consultation, we well-educated (68% completed some post-secondary educa- elected to set aside all the Factor 2 items as our ultimate goal tion). Nineteen percent of respondents were from Jordan, was to produce a unidimensional index, and the items in 60% lived in urban areas, 95% were Muslim, 22% had lived Factor 1 reflected the content of primary interest. outside the country for more than 5 years, and 59% reported some personal cancer connection. IRT Analyses Results from a series of IRT calibrations of the 31 remain- Evaluation and Reduction of Item Pool ing items from Factor 1 led to removal of 6 additional items For analyses, we randomly split the pilot sample of N = 588 based on poor fit to the IRT model and excess local depen- into two analytic samples for exploratory (n = 400) and con- dence. After examining results from an IRT calibration of firmatory (n = 188) analyses. Using the exploratory sample, the remaining 25 items, we elected to retain all of these we conducted exploratory factor analysis (EFA) with Mplus items for the larger field test. However, 3 were reworded software (Muthén & Muthén, 2007) modeling the 59 stigma for clarification and a new item was created to represent items as categorical with the weighted least squares means conflating cancer with death, as this was of particular inter- and variance adjusted [WLSMV] estimator. The main goals est. Finally, although we wished to arrive at a final index of this analysis were to identify the structure of the item set measuring only a single dimension of cancer stigma, we and remove items that were not performing well. Following elected to reword and retain 5 items from the original Factor this item reduction, we conducted a confirmatory factor 2 that had desirable content; all other items from Factor 2 analysis (CFA) using the confirmatory sample (n = 188), were discarded. and evaluated model fit with standard diagnostic fit indices (root mean square error of approximation [RMSEA] ≤ .08, Study 2: Translation, Review, and Field Tucker–Lewis Index [TLI] ≥ .95, comparative fit index Testing in Jordan/Egypt and China [CFI] ≥ .95; Browne, Cudeck, Bollen, & Long, 1993; Hu & Bentler, 1999). Based on strategic priorities, the revised 31-item index was Further item reduction was achieved based on consider- prepared to be fielded among English- and Arabic-speaking ation of the overall goals of the index. In addition to examina- respondents in the Middle East (Jordan/Egypt), and among tion of CFA results, we examined results from a series of IRT English- and Mandarin-speaking respondents in China. calibrations (conducted using IRTPRO; Cai, du Toit, & Thus, the item set was translated into Arabic and simple Thissen, 2011), including item properties, item fit, and local Mandarin Chinese, and cognitive interviews were conducted dependence indices to identify redundant or poorly perform- for each translation of the instrument. ing items either in content or in terms of item properties. After discussion of results among study team members, the refined Cognitive Interviews index was finalized for the next phase of field testing. English- and Arabic-speaking, Arab respondents. We conducted interviews with eight respondents of Arab heritage, who Factor Analyses spoke both English and Arabic. All respondents reviewed the Results from the initial EFA of the 59 items indicated that a English version, and six reviewed both the English and Ara- two-factor solution was most appropriate. In this solution, bic versions. We mirrored the recruitment strategy used in the first factor consisted of 36 items reflecting negative Study 1. Respondents identified a few instances where the stigma (fear, lack of understanding, negativity) and the sec- Arabic translation did not adequately capture the original ond factor contained 23 items reflecting more positive content; these translated items were modified. statements (compassion for cancer, understanding, pragma- tism). At this stage, we elected to remove a total of 7 items English- and Mandarin Chinese-speaking, Chinese respon- that either did not load cleanly on a single factor (double- dents. We utilized professional, family, and peer networks to loaders, 3 items) or loaded weakly on their respective fac- identify a demographically diverse set of individuals to par- tors (4 items). ticipate in the cognitive interviews to review English and The 52 remaining items were subject to a two-factor CFA. Mandarin Chinese versions of the cancer stigma item set. We The fit of the initial model was not quite acceptable (CFA = conducted interviews via Skype which allowed us to include .874, TLI = .869, RMSEA = .063). Model fit diagnostics sug- eight respondents in China as well as three in the United gested removal of two items from Factor 2, and this sugges- States. The sample included seven men and four women who tion was supported by the fact that these were the only were fairly well-distributed by age. In the Chinese interviews, reverse-keyed items in that factor. After their removal, the fit respondents expressed confusion about items regarding 4 SAGE Open Identify any CALIBRATED ITEM BANK – differences due to Country: Jordan and REFERENCE GROUP language Egypt Country: Jordan and Egypt Language: Arabic Language: English Number of items: 31 Number of items: 31 Sample Size: N=633 Sample Size: pilot N=588 with follow- up N=324 Identify any differences due to culture Country: China Country: China Language: English Language: Mandarin Identify any differences due to Number of items: 31 Number of items: 31 language Sample Size: N=500 Sample Size: N=500 Figure1. Flowchart of cancer stigma index development design. isolation and being an outcast. Thus, we modified the transla- the 31 items represent a single dimension. Based on results tion to more clearly communicate these terms. from these analyses, we considered items for removal to improve unidimensionality across the four samples. Once a set of items was identified that appeared to be Field Test sufficiently unidimensional in all samples, we used Method IRTPRO to conduct differential item functioning (DIF) Measures. All Study 2 measures were identical to Study analyses within an IRT framework. DIF, also referred to as 1 measures with the exception that the reduced 31 cancer measurement bias, occurs when people from different stigma items were administered as opposed to the initial 59 groups (e.g., gender or ethnicity) with the same level of the items. latent trait (in this case cancer stigma) have a different probability of giving a certain response to an item. Thus, Sample and procedures. As can be seen in Figure 1, the we compared performance of (a) the Arabic and English 31-item field test was administered in two regions, Jordan/ items from the two Jordan/Egypt samples (JE–JA), (b) the Egypt and China, and in three languages, English, Arabic, Mandarin and English items from the two China samples and Mandarin, to produce four distinct samples. Respon- (CE–CM), and (c) the English items from Jordan/Egypt dents residing in Jordan/Egypt were administered an English and China samples (JE–CE) to determine the comparabil- (JE, n = 324) or an Arabic version (JA, n = 633); and respon- ity of cancer stigma items across language and region (this dents residing in China were administered an English (CE, process is depicted in Figure 1). n = 500) or a Mandarin version (CM, n = 500). As in Study 1, DIF analysis used three steps. First, two-group chi-square we contracted with Harris Interactive to administer the field tests from IRTPRO were evaluated across comparison test via the Internet in both regions, and all study procedures groups and the significance tests for all comparisons were were approved by the RAND IRB. The characteristics of the adjusted using the Benjamini–Hochberg procedure four field test samples are displayed in Table 1. (Benjamini & Hochberg, 1995) at p < .05 to identify candi- date items for removal. Next, to evaluate the magnitude of Evaluation of item set. Our first analytic step was to con- DIF, items demonstrating significant DIF after p value cor- duct CFAs with each sample to evaluate the extent to which rection were further evaluated by computing the weighted Edelen et al. 5 Table 1. Characteristics of Study 2 Field Test Sample, by Region and Language. Jordan/Egypt: Jordan/Egypt: China: English China: Mandarin Variable Value English (n = 324) Arabic (n = 633) (n = 500) (n = 500) Gender Male 224 (69.1%) 351 (55.5%) 251 (50.2%) 261 (52.2%) Female 100 (30.9%) 282 (44.6%) 249 (49.8%) 239 (47.8%) Age 18-29 150 (46.3%) 244 (38.6%) 146 (29.2%) 136 (27.2%) 30-49 147 (45.3%) 331 (52.3%) 289 (57.8%) 264 (52.8%) 50+ 27 (8.3%) 58 (9.2%) 65 (13.0%) 100 (20.0%) Urbanicity City/large town 251 (77.5%) 453 (71.6%) 404 (80.8%) 402 (80.4%) Suburb 28 (8.6%) 88 (13.9%) 53 (10.6%) 47 (9.4%) Small town 31 (9.6%) 25 (3.9%) 33 (6.6%) 37 (7.4%) Village/rural 14 (4.3%) 67 (10.6%) 10 (2.0)% 14 (2.8%) Low income (≤800 JOD, Yes 80 (24.7%) 200 (31.6%) 197 (39.4%) 184 (36.8%) ≤3,000 EGP, ≤¥84,999) No 244 (75.3%) 433 (68.4%) 303 (60.6%) 316 (63.2%) Post-secondary Yes 295 (91.6%) 547 (87.9%) 429 (87.9%) 424 (85.7%) education No 27 (8.4%) 85 (13.5%) 59 (12.1%) 71 (14.3%) Religion Islam 279 (86.1%) 577 (91.2%) 10 (2.0%) 3 (.6%) Buddhism 0 (0%) 0 (0%) 54 (10.8%) 89 (17.8%) Christianity 34 (10.5%) 27 (4.3%) 48 (9.6%) 26 (5.2%) Other/multiple/ 11 (3.4%) 29 (4.6%) 122 (24.4%) 47 (9.4%) decline None 0 (0%) 0 (0%) 266 (53.2%) 335 (67.0%) area between the expected score curves (wABC; Edelen, Table 2. Items Comprising the CSI. Stucky, & Chandra, 2013). 1. I would feel uncomfortable talking to a person with cancer After removing items with problematic DIF, we fit a four- 2. Treatment and support are useless for someone with cancer group IRT model. Based on those results, we reconsidered 3. I would feel uncomfortable sitting next to someone with item content and properties in an effort to further reduce the cancer item set and arrive at a final index. 4. I would feel uncomfortable sending own child to school with another child with cancer Scoring and examining CSI scores. Once the CSI was final- 5. I would feel uncomfortable if someone with cancer lived ized, we calculated IRT-based scores (i.e., expected a pos- nearby teriori [EAP]) using a summed score conversion algorithm 6. If a close friend had cancer, I would avoid him or her (Thissen, Pommerich, Billeaud, & Williams, 1995). These 7. I would feel uncomfortable being friends with someone with cancer scores retain the benefits of the IRT model and are also prac- 8. People can only blame themselves for getting cancer tical for general use via score translation tables. Summed 9. If had cancer, I would be ashamed score EAPs were generated for the CSI using the final four- 10. I would feel isolated/alone if I received treatment for cancer group IRT model and were rescaled along a T-score metric 11. If my spouse had cancer, I would be ashamed of him or her with the Jordan English sample mean set to 50 and standard 12. If my spouse had cancer, I would consider leaving him or her deviation to 10. Finally, we conducted a set of descriptive analyses to exam- Note. CSI = cancer stigma index. ine the CSI scores according to demographic and personal characteristics of interest (e.g., gender, income, age, personal experience with cancer, attitudes about general health stigma). model fit with the same set of items across all four samples (JE: CFI = .955, TLI = .951, RMSEA = .063; JA: CFI = .922, Results TLI = .916, RMSEA = .078; CE: CFI = .903, TLI = .895, Identifying the 12-item CSI. The fit of the 31 items to a RMSEA = .087; CM: CFI = .946, TLI = .941, RMSEA = single-factor model in each sample was reasonable but did .078). IRT–DIF analyses identified 9 items with problematic not reach standard criteria for all four samples (JE: CFI = DIF that were also removed. .886, TLI = .878, RMSEA = .090; JA: CFI = .852, TLI = We used results from a four-group IRT calibration of the .841, RMSEA = .096; CE: CFI = .885, TLI = .877, RMSEA = remaining 18 items to reduce the item set further by remov- .084; CM: CFI = .921, TLI = .916, RMSEA = .083), indicat- ing items based on redundant content and/or poor psycho- ing the need for some modifications. In consultation with the metric properties. In all, we removed 6 items at this stage. larger study team, 4 items were removed to obtain acceptable Items comprising the final 12-item CSI are listed in Table 2. 6 SAGE Open Table 3. Cancer Stigma Scores Among Demographic Groups for Each Sample and Combined. Combined Jordan/Egypt: Jordan/Egypt: China: China: Sample English Arabic English Mandarin Gender M (SD ) 52.0 (10.3) 50.3 (9.9) 48.9 (7.5) 58.4 (12.0) 50.7 (8.6) M M M (SD ) 50.4 (9.4) 49.2 (10.3) 46.5 (6.1) 57.6 (10.6) 49.2 (7.4) F F p .0004 .35 <.0001 .46 .03 Income M (SD ) 50.7 (9.7) 48.6 (8.8) 46.9 (6.0) 58.5 (11.7) 50.0 (8.0) H H M (SD ) 52.4 (10.2) 54.2 (12.3) 49.1 (8.2) 57.3 (10.7) 49.9 (8.4) L L p .0003 .0003 .0007 .27 .88 Age M (SD ) 51.0 (9.5) 52.1 (11.2) 48.1 (7.1) 56.0 (9.9) 49.7 (8.1) L L M (SD ) 51.6 (10.4) 48.6 (8.9) 47.4 (6.9) 59.1 (11.9) 50.3 (8.4) M M M (SD ) 50.4 (9.0) 45.9 (5.4) 46.1 (5.0) 57.7 (11.1) 49.4 (7.1) H H p .18 .0009 .10 .03 .60 Post-secondary education M (SD ) 51.9 (11.3) 59.0 (15.4) 46.7 (7.1) 59.3 (12.1) 49.4 (8.3) N N M (SD ) 51.1 (9.5) 49.2 (9.0) 47.7 (6.8) 57.6 (11.0) 50.1 (8.0) Y Y p .27 .003 .21 .28 .52 Lived outside country >5 years M (SD ) 50.1 (8.8) 49.9 (10.0) 47.6 (6.5) 55.7 (10.3) 49.2 (7.5) N N M (SD ) 53.7 (11.5) 50.1 (10.1) 47.6 (8.1) 60.9 (12.0) 51.6 (9.0) Y Y p <.0001 .85 .97 <.0001 .005 Diagnosed with cancer: Self M (SD ) 50.6 (9.2) 49.6 (9.8) 47.5 (6.6) 56.7 (10.5) 49.9 (7.8) N N a a M (SD ) 64.4 (14.3) 57.1 (12.6) 52.1 (18.5) 67.4 (12.6) 67.4 (24.4) Y Y p <.0001 .005 .54 <.0001 .34 Diagnosed with cancer: Family/loved one M (SD ) 52.1 (10.3) 51.6 (11.0) 48.4 (7.7) 58.1 (11.4) 50.1 (8.3) N N M (SD ) 49.4 (8.6) 47.4 (7.5) 46.4 (5.2) 57.6 (11.1) 49.6 (7.5) Y Y p <.0001 <.0001 .0002 .63 .56 Ever paid caregiver for cancer patient M (SD ) 50.2 (8.9) 48.9 (8.9) 47.5 (6.8) 56.1 (10.6) 49.6 (7.6) N N M (SD ) 55.2 (12.2) 51.8 (11.2) 47.9 (7.6) 62.1 (11.8) 52.0 (10.7) Y Y p <.0001 .05 .64 <.0001 .09 Group size < 10. We ran a final four-group calibration to obtain item param- four samples, providing some preliminary validity evidence eters and generate CSI scores. Appendix A1 provides a for the CSI (range across samples r = .35-.50). The correla- score translation table for the 12-item CSI. The IRT-based tion of the CSI with the item likening cancer to a death sen- score reliability for the CSI varies by sample and ranges tence was less consistent and slightly lower on average from acceptable to excellent (JE = .79; JA = .73; CE = .91; (range across samples r = .20-.49). All correlations were sig- CM = .81). nificantly different from 0 at p < .05. Scores from the CSI were compared across various demo- Examining CSI scores. The JA sample had the lowest CSI graphic groups for each of the four samples to establish ini- score (M = 47.6, SD = 6.9), with JE (M = 50.0, SD = 10.0) tial validity evidence. A summary of these results is contained and CM (M = 50.0, SD = 8.1) both at 50. The CE sample had in Table 3. the highest mean CSI score (M = 58.0, SD = 11.3). Tests of significance between the four CSI sample means revealed Discussion that the JE and CM means are not different from one another, but all other group mean comparisons are statistically sig- This analysis of cancer stigma and development of the CSI nificant. will provide critical benefits to the cancer research and con- Correlation analyses revealed a moderate to strong corre- trol fields. Our process of integrating literature review with lation between the CSI and general illness stigma across all stakeholder input and measures analysis represents a robust Edelen et al. 7 method of developing a quality, user-friendly index that can of cognitive interviews to ensure terms and whole items are be used by cancer organizations to inform cancer stigma interpreted as intended. Furthermore, data from field tests in reduction initiatives and broader public awareness cam- new regions and languages must be analyzed to determine paigns. Moreover, the CSI can be added to cancer research the comparability of CSI scores back to the reference sample studies examining patient, family, and public perspectives (Jordan/Egypt—English). regarding cancer screening, diagnoses, and treatment. It is important and possible to show that CSI scores cor- While there were limited cancer stigma items or scales relate with other indicators of stigma. For example, we available for modification at the outset of this study, there should expect to find that higher CSI scores correlate with were several scales in the area of stigma and health care national or local policies that discriminate against people decision making as well as stigma and chronic health issues with cancer, and predict lower levels of treatment-seeking, (e.g., mental health) that were particularly useful. The less positive psychological well-being, and greater social stakeholder input we solicited through technical advisors isolation among people with cancer. Future research along and cognitive interviews was critical. Without that feed- these lines could provide valuable validity evidence for the back, the scale would not have included particular cultural CSI. Overall, the CSI can be used to inform initial cancer “pulse points” such as the role of religion or fate in driving education efforts, identifying overall stigma levels in a coun- cancer views. The moderate to strong correlation between try or community and particular issue areas requiring con- the CSI and general illness stigma across all four samples certed intervention. Over time, following careful data suggests that the CSI is reasonably robust and indicative of analyses and perhaps slight modifications to index scoring, general health-seeking stigma. As expected, male gender the CSI can be used as an index comparing countries or com- and low income were associated with higher CSI scores. munities on stigma levels, prioritizing where education Interestingly, CSI scores indicated that those who had a resources should be allocated, and helping to determine the loved one diagnosed with cancer reported lower stigma, impact of stigma reduction efforts. whereas those who had personally experienced cancer reported more stigma. Appendix A1 Our approach attempted to limit potential weaknesses in design where possible. But a few study limitations should be 12-Item Cancer Stigma Index Total Score to noted. First, our cognitive interviews, particularly for the T-Score Translation Table Arab origin samples, were not conducted in Egypt and Jordan. Although we attempted to find individuals who had strong cultural ties (e.g., using criteria about upbringing in Total Score T-Score Total Score T-Score the region), we may not have received the full complement 12 42 37 69 of cultural insights from those who were Americans or had 13 43 38 70 spent considerable time in the United States. Second, our 14 44 39 71 pilot and field tests endeavored to obtain diversity by age, 15 46 40 72 gender, income, and education. For the latter two categories, 16 47 41 73 we approached our goal but did not always meet it in terms 17 48 42 74 of education level, with a slightly higher education status 18 49 43 75 overall. We know that education may influence cancer stigma 19 50 44 76 perspectives. Furthermore, our mode of testing the cancer 20 51 45 77 stigma items was web-based, which may impede participa- 21 52 46 78 tion from those of lower socioeconomic status. Mode effects 22 53 47 80 of web-based administration could not be tested (given that 23 54 48 81 was the only mode used), but may affect the interpretation of 24 55 49 82 items. 25 56 50 83 26 57 51 84 More robust validity tests of the CSI will require addi- 27 58 52 85 tional use in the countries in which we developed the first 28 59 53 86 versions of the scale—Egypt, Jordan, and China. This testing 29 60 54 87 may include using the CSI in diverse communities and with 30 61 55 88 a wide variety of subpopulations (e.g., setting, age). The 31 62 56 89 scale was developed for use initially in these countries based 32 64 57 90 on strategic plans and investments. However, the intention is 33 65 58 91 for the scale to ultimately be used worldwide. Thus, as the 34 66 59 92 CSI is translated and used in new contexts, it will be impor- 35 67 60 93 tant to step through all of the phases used in this study, prin- 36 68 cipally review of the translation by experts and some version 8 SAGE Open Acknowledgments Evans-Lacko, S., Little, K., Meltzer, H., Rose, D., Rhydderch, D., Henderson, C., & Thornicroft, G. (2010). Development We acknowledge the data collection support from Sandra and psychometric properties of the mental health knowledge Applebaum and Roz Pierson at Harris Interactive. We thank Patrick schedule. Canadian Journal of Psychiatry/Revue Canadienne Orr for his efforts in manuscript formatting, and researchers Omar De Psychiatrie, 55, 440-448. Al-Shahery, Samuel Berkowitz, Dolly Dahdal, and Chaoling Feng Fries, J. F., Bruce, B., & Cella, D. (2005). The promise of PROMIS: at RAND for their efforts to review the Arabic and Chinese versions Using item response theory to improve assessment of patient- of the scale. Finally, we thank the many individuals who have par- reported outcomes. Clinical and Experimental Rheumatology, ticipated in the cognitive interviews for their insight, which 23(5), S53-S57. strengthened our research. Green, S., Davis, C., Karshmer, E., Marsh, P., & Straight, B. (2005). Living stigma: The impact of labeling, stereotyping, separa- Declaration of Conflicting Interests tion, status loss, and discrimination in the lives of individuals The author(s) declared no potential conflicts of interest with respect with disabilities and their families. Sociological Inquiry, 75, to the research, authorship, and/or publication of this article. 197-215. Greene, K., & Banerjee, S. C. (2006). Disease-related stigma. Journal of Homosexuality, 50(4), 185-209. Funding Hu, L.-t, & Bentler, P. M. (1999). Cutoff criteria for fit indexes in The author(s) disclosed receipt of the following financial support for covariance structure analysis: Conventional criteria versus new the research and/or authorship of this article: This work was funded alternatives. Structural Equation Modeling: A Multidiscipli- by an unrestricted grant from the LIVESTRONG Foundation. nary Journal, 6, 1-55. Joachim, G, & Acorn, S. (2000). Living with chronic illness: The References interface of stigma and normalization. The Canadian Journal Ader, D. N. (2007). Developing the patient-reported outcomes mea- of Nursing Research/Revue canadienne de recherche en sci- surement information system (PROMIS). Medical Care, 45(5), ences infirmieres, 32(3), 37-48. S1-S2. LIVESTRONG Foundation. (2007). Cancer stigma and silence Batson, C. D., Polycarpou, M. P., Harmon-Jones, E., Imhoff, H. around the world. Austin, TX: LIVESTRONG Foundation. J., Mitchener, E. C, Bednar, L. L., . . .Highberger, L. (1997). Muthén, L. K, & Muthén, B. O. (2007). 1998-2007. Mplus user’s Empathy and attitudes: Can feeling for a member of a stig- guide. Los Angeles, CA: Author. matized group improve feelings toward the group? Journal of Pettit, M. L. (2008). Disease and stigma: A review of literature. Personality and Social Psychology, 72, 105-118. Health Educator, 40(2), 70-76. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false dis- Reeve, B. B., Hays, R. D., Bjorner, J. B., Cook, K. F., Crane, P. K., covery rate: A practical and powerful approach to multiple Teresi, J. A., . . . Cella, D. (2007). Psychometric evaluation and testing. Journal of the Royal Statistical Society: Series B calibration of health-related quality of life item banks: Plans (Methodological), 57, 289-300. for the patient-reported outcomes measurement information Brewster, M. E., & Moradi, B. (2010). Perceived experiences of system (PROMIS). Medical Care, 45(5 Suppl. 1), S22-S31. anti-bisexual prejudice: Instrument development and evalu- doi:10.1097/01.mlr.0000250483.85507.04 ation. Journal of Counseling Psychology, 57, 451-468. Sartorius, N. (2007). Stigma and mental health. The Lancet, 370, doi:10.1037/A0021116 810-811. Brohan, E., Slade, M., Clement, S., & Thornicroft, G. (2010). Thissen, D., Pommerich, M., Billeaud, K., & Williams, V. S. L. Experiences of mental illness stigma, prejudice and discrimi- (1995). Item response theory for scores on tests including poly- nation: A review of measures. BMC Health Services Research, tomous items with ordered responses. Applied Psychological 10, Article 80. doi:10.1186/1472-6963-10-80 Measurement, 19, 39-49. Browne, M. W, Cudeck, R., Bollen, K. A., & Long, J. S. (1993). Uys, L. R, Holzemer, W. L, Chirwa, M. L, Dlamini, P. S, Greeff, Alternative ways of assessing model fit. Sage Focus Editions, M., Kohi, T. W, . . . Naidoo, J. R. (2009). The development 154, 136. and validation of the HIV/AIDS stigma instrument-nurse Cai, L., du Toit, S. H. C, & Thissen, D. (2011). IRTPRO version 2: (HASI-N). AIDS Care, 21, 150-159. Flexible, multidimensional, multiple categorical IRT modeling. Chicago, IL: Scientific Software International. Author Biographies Cella, D., Yount, S., Rothrock, N., Gershon, R., Cook, K., Reeve, Maria Orlando Edelen, PhD, is a senior Behavioral Social B., . . .Rose, M. (2007). The Patient-Reported Outcomes Scientist scientist and psychometrician at the RAND Corporation. Measurement Information System (PROMIS): Progress of an Her research focuses primarily on the use of item response theory NIH Roadmap cooperative group during its first two years. for development, refinement and evaluation of bahavioral health Medical Care, 45(5 Suppl. 1), S3-S11. measures. DeWalt, D. A., Rothrock, N., Yount, S., & Stone, A. A. (2007). Evaluation of item candidates: The PROMIS qualitative item Anita Chandra, PhD is a senior policy researcher and director of review. Medical Care, 45(5 Suppl. 1), S12-S21. the Behavioral and Policy Sciences Department at the RAND Edelen, M. O., Stucky, B., & Chandra, A. (2013). Quantifying Corporation. Her background is in public health, child and adoles- “problematic” DIF within an IRT framework: Application to cent development, and community-based participatory research a Cancer Stigma Index. Quality of Life Research. Advance and evaluation. She currently leads or co-leads studies on com- online publication. munity well-being; deployment and military families; community Edelen et al. 9 resilience and long-term disaster recovery; and child health and Claire Neal, MPH is the Executive Director of the Triangle Global development. Health Consortium, a non-profit member organization represent- ing institutions and individuals from the pharmaceutical and bio- Brian Stucky is an Associate Behavioral Social Scientist at RAND technology industry, the international health development commu- Corporation, where he specializes in psychometrics and health nity, and academia – all dedicated to improving the health of the research. He is a trained psychometrician whose research interests world’s communities. Claire has over 13 years of experience in in focus on developing and applying new test theory models (includ- the development and delivery of innovative health programs. For ing factor analytic and item response theory-based) that serve to the last decade, Claire has worked at the LIVESTRONG Foundation; better understand patient response data in settings where multiple most recently as the Vice President of Global Strategy where she dimensions are present. provided the leadership and vision for LIVESTRONG’s global work. Rebekkah M. Schear, MIA has spent her career working at the intersection of global health and international development. Over Ruth Rechis, PhD, is the Vice President of Programs & Strategy the last five years at the LIVESTRONG Foundation, Rebekkah led at the LIVESTRONG Foundation. As a member of the Executive the Foundation’s implementation of six international programs Leadership Team, she oversees the day to day operations of the throughout Latin America, Africa and Asia focusing on reducing Foundation’s programmatic work and shepherds the strategic cancer stigma and empowering cancer survivors to become advo- planning process for the Foundation. As an adolescent cancer sur- cates. Currently, she is working to create a new, innovative model vivor, Dr. Rechis has a personal connection to the mission of the of patient-centered cancer care with the goal of improving delivery Foundation. of patient-centered cancer care globally. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png SAGE Open SAGE

Loading next page...
 
/lp/sage/developing-a-global-cancer-stigma-index-smMAY496Fp
Publisher
SAGE
Copyright
Copyright © 2022 by SAGE Publications Inc, unless otherwise noted. Manuscript content on this site is licensed under Creative Commons Licenses.
ISSN
2158-2440
eISSN
2158-2440
DOI
10.1177/2158244014547875
Publisher site
See Article on Publisher Site

Abstract

Despite increasing recognition about the stigma associated with cancer screening, diagnosis, and treatment-seeking behaviors, there has been relatively little attention paid to how to assess and intervene to reduce that stigma. An index to measure cancer stigma could empower health program developers and policymakers by identifying the key areas in which a population could benefit from education to change perceptions and address misinformation. The index also could be used to rank countries and communities based on their level of cancer stigma to assess where interventions are needed. We used structured literature review and expert review to generate a cancer stigma item pool. The item pool was subject to cognitive interviews for cultural appropriateness and comprehension; and data from initial pilot testing were used to reduce the pool of items for translation and field testing. The field test was conducted using a web-based survey in four samples representing two regions and three languages—English and Arabic speakers in Jordan and Egypt, and English and Mandarin Chinese speakers in China. Factor analyses and item response theory were applied to finalize the index. The analyses resulted in a 12-item cancer stigma index (CSI) that was reliable across all four samples. The CSI scores were highly correlated with a general illness stigma scale, and operated as expected noting higher cancer stigma among men and those with lower income. The CSI can be used to inform initial cancer education efforts, identifying overall stigma levels in a country or community and particular issue areas requiring intervention. Keywords cancer, stigma, health, chronic disease, help-seeking, health disparities Specific types of cancer may also carry disease-specific Introduction stigma. For example, cervical and lung cancer are often cited Over the last several years, stigma associated with health- because each is linked to behavior that may be deemed unde- seeking behaviors has received increasing attention. sirable or marginal. In the case of lung cancer, individuals Health-related stigma (or principally, disease-related may feel guilt and shame attributed to their diagnosis, due to stigma) is unique and presents significant challenges and the link between smoking and cancer. Guilt may lead to barriers for the global health community to overcome. The denial of the diagnosis until such a point that treatment may framing of health-related stigma has begun to advance a not be successful (Batson et al., 1997). Cervical cancer, more complex discussion of stigma, one that encompasses breast cancer, and uterine cancer may also carry a particular both the internalization of stigma by the individual and the stigma as these cancers are often linked to sexual health, public reaction and potential marginalization that may regardless of the actual disease pathway. In some patriarchal occur. First, the word “disease” alone can induce a sense of societies, women are considered to be the property of their stigma (Green, Davis, Karshmer, Marsh, & Straight, 2005; spouses, and must comply with their spouse’s wishes to not Pettit, 2008). There are a number of diseases that have his- seek treatment. Religious and cultural beliefs may prohibit torically been highly stigmatized, including mental health disorders, HIV/AIDS, sexually transmitted diseases, lep- rosy, and skin diseases (Greene & Banerjee, 2006; Sartorius, RAND Corporation, Boston, MA, USA 2007). An emerging body of data in the stigma literature RAND Corporation, Arlington, VA, USA indicates that cancer is also often among the diseases that is RAND Corporation, Santa Monica, CA, USA highly stigmatized (LIVESTRONG Foundation, 2007), yet Livestrong Foundation, Austin, TX, USA it is less researched than other health issues. Individuals Corresponding Author: often react to that stigma by making decisions about Maria Orlando Edelen, RAND Corporation, 20 Park Plaza, Suite 920, whether or not to disclose their condition or seek treatment Boston, MA 02116, USA. (Joachim & Acorn, 2000). Email: Maria_Edelen@rand.org This article is distributed under the terms of the Creative Commons Attribution 3.0 License Creative Commons CC BY: (http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access page (http://www.uk.sagepub.com/aboutus/openaccess.htm). 2 SAGE Open seeking medical attention for parts of the body having a sex- comprehensible, (b) understand how respondents interpreted ual connotation (Brewster & Moradi, 2010). the items, and (c) ensure that the item content and wording Despite this emerging recognition that stigma related to were culturally appropriate. Because our initial pilot test was cancer can deter critical health-seeking behaviors, there has to be conducted in the Middle East, cognitive interview par- been comparatively little effort to measure the level of can- ticipants, although U.S.-based, were of Arab heritage, had cer stigma in a given population or community to intervene. been raised in the region (at least up to age 18), considered Most stigma scales have focused in other areas, principally themselves culturally Arab, and spoke both English and HIV/AIDS and mental health (Brohan, Slade, Clement, & Arabic. All respondents received a US$25 gift card for their Thornicroft, 2010; Evans-Lacko et al., 2010; Uys et al., time, and all procedures were approved by the RAND 2009). This study sought to develop a cancer stigma index Institutional Review Board (IRB). Following review of the (CSI) to measure perspectives on cancer, specifically atti- cognitive interview transcripts, we made changes to the item tudes about cancer screening and treatment, and to help pool based on respondent feedback including dropping items inform awareness and education programs. We proceeded in that were considered redundant and generating items to two phases. First, we conducted an initial study to gather reflect stigma and religion, a theme that was considered stigma measures and test a preliminary item pool (Study 1). important by interviewees. Then, we conducted a full field test in two regions (Study 2). Pilot Test Study 1: Creation of a Cancer Stigma Method Item Pool and Initial Pilot Test Measures. Cancer stigma item pool. A set of 59 candidate We used established methods for item pool development fol- items, formatted using Likert-type response options (1 = not lowing the National Institutes of Health Patient-Reported at all to 5 = very much), were administered to all respon- Outcomes Measurement Information System (NIH PROMIS®) dents. initiative blueprint (Reeve et al., 2007). The main goals of PROMIS® (http://www.nihpromis.org/default.aspx), which is General illness stigma. As there were no items available part of the NIH’s Roadmap initiative, are to standardize a set of on general illness stigma that corresponded well to the can- assessment tools and to use item response theory (IRT) tech- cer stigma constructs we are trying to measure, we used niques and advances in computer technology to create brief yet information in the literature to develop five items reflecting highly reliable and flexible assessment tools to measure patient- attitudes of stigma toward general illness (e.g., If I had an reported outcomes (Ader, 2007; Cella et al., 2007; Fries, Bruce, illness, I would feel left out of things). These items also used & Cella, 2005). Due to its scope and success, the rigorous the five-category Likert-type response format. approach utilized by PROMIS has become something of a standard for modern instrument development (DeWalt, Demographics. All respondents provided basic demo- Rothrock, Yount, & Stone, 2007). graphic information, including gender, age, education, reli- Briefly, we first conducted a structured literature search gion, country of origin, country of residence, and language for stigma as it relates to cancer, HIV/AIDS, mental health, spoken at home. Six additional items indicated the respon- and other health issues. The initial search summarized litera- dents’ connections to cancer (e.g., I know someone with can- ture in these areas known to underlie stigma: information and cer, I am a caregiver of someone with cancer). myths, fear, shame and labeling, concerns about diagnosis, concerns about treatment-seeking, concerns about peer and Sample and Procedures family disclosure, and concerns about public disclosure. We identified and organized a total of 553 items from 29 mea- All study procedures were approved by the RAND IRB. We sures into four broad domains: characterization about those contracted with Harris Interactive to pilot the items via the with the disease, self-stigma among those with the disease, Internet to English-speaking adult respondents residing in expectations of what others may think of those with the dis- Egypt and Jordan. We collected a total of N = 1,016 com- ease, and positive views on those with the disease. After pleted web surveys. The majority of respondents were male review by five technical advisors, we reduced the number of (72%), and the sample was also skewed toward younger and items based on item scope and quality of item wording, and more educated individuals than is representative of the gen- reworded them to be appropriate for the cancer stigma con- eral population in the region (52% were <30 years old, 94% text and in comparable format. were <50 years old; 71% had university education). We addressed this imbalance by selecting a subset of the data that had the same number of men and women (294 each; Cognitive Interviews Total N = 588). All women from the original sample were Using the reduced set of items, we conducted seven cogni- retained, and we used a random sampling approach to select tive interviews to (a) assess whether the items were a subset of 294 males that was stratified according Edelen et al. 3 to education and an indicator of having lived outside the of the two-factor model was adequate (CFA = .905, TLI = country. Despite this stratification, the analytic sample was .901, RMSEA = .057). The final solution had 31 items in still relatively young (49% below 30 years of age) and fairly Factor 1 and 19 items in Factor 2. After consultation, we well-educated (68% completed some post-secondary educa- elected to set aside all the Factor 2 items as our ultimate goal tion). Nineteen percent of respondents were from Jordan, was to produce a unidimensional index, and the items in 60% lived in urban areas, 95% were Muslim, 22% had lived Factor 1 reflected the content of primary interest. outside the country for more than 5 years, and 59% reported some personal cancer connection. IRT Analyses Results from a series of IRT calibrations of the 31 remain- Evaluation and Reduction of Item Pool ing items from Factor 1 led to removal of 6 additional items For analyses, we randomly split the pilot sample of N = 588 based on poor fit to the IRT model and excess local depen- into two analytic samples for exploratory (n = 400) and con- dence. After examining results from an IRT calibration of firmatory (n = 188) analyses. Using the exploratory sample, the remaining 25 items, we elected to retain all of these we conducted exploratory factor analysis (EFA) with Mplus items for the larger field test. However, 3 were reworded software (Muthén & Muthén, 2007) modeling the 59 stigma for clarification and a new item was created to represent items as categorical with the weighted least squares means conflating cancer with death, as this was of particular inter- and variance adjusted [WLSMV] estimator. The main goals est. Finally, although we wished to arrive at a final index of this analysis were to identify the structure of the item set measuring only a single dimension of cancer stigma, we and remove items that were not performing well. Following elected to reword and retain 5 items from the original Factor this item reduction, we conducted a confirmatory factor 2 that had desirable content; all other items from Factor 2 analysis (CFA) using the confirmatory sample (n = 188), were discarded. and evaluated model fit with standard diagnostic fit indices (root mean square error of approximation [RMSEA] ≤ .08, Study 2: Translation, Review, and Field Tucker–Lewis Index [TLI] ≥ .95, comparative fit index Testing in Jordan/Egypt and China [CFI] ≥ .95; Browne, Cudeck, Bollen, & Long, 1993; Hu & Bentler, 1999). Based on strategic priorities, the revised 31-item index was Further item reduction was achieved based on consider- prepared to be fielded among English- and Arabic-speaking ation of the overall goals of the index. In addition to examina- respondents in the Middle East (Jordan/Egypt), and among tion of CFA results, we examined results from a series of IRT English- and Mandarin-speaking respondents in China. calibrations (conducted using IRTPRO; Cai, du Toit, & Thus, the item set was translated into Arabic and simple Thissen, 2011), including item properties, item fit, and local Mandarin Chinese, and cognitive interviews were conducted dependence indices to identify redundant or poorly perform- for each translation of the instrument. ing items either in content or in terms of item properties. After discussion of results among study team members, the refined Cognitive Interviews index was finalized for the next phase of field testing. English- and Arabic-speaking, Arab respondents. We conducted interviews with eight respondents of Arab heritage, who Factor Analyses spoke both English and Arabic. All respondents reviewed the Results from the initial EFA of the 59 items indicated that a English version, and six reviewed both the English and Ara- two-factor solution was most appropriate. In this solution, bic versions. We mirrored the recruitment strategy used in the first factor consisted of 36 items reflecting negative Study 1. Respondents identified a few instances where the stigma (fear, lack of understanding, negativity) and the sec- Arabic translation did not adequately capture the original ond factor contained 23 items reflecting more positive content; these translated items were modified. statements (compassion for cancer, understanding, pragma- tism). At this stage, we elected to remove a total of 7 items English- and Mandarin Chinese-speaking, Chinese respon- that either did not load cleanly on a single factor (double- dents. We utilized professional, family, and peer networks to loaders, 3 items) or loaded weakly on their respective fac- identify a demographically diverse set of individuals to par- tors (4 items). ticipate in the cognitive interviews to review English and The 52 remaining items were subject to a two-factor CFA. Mandarin Chinese versions of the cancer stigma item set. We The fit of the initial model was not quite acceptable (CFA = conducted interviews via Skype which allowed us to include .874, TLI = .869, RMSEA = .063). Model fit diagnostics sug- eight respondents in China as well as three in the United gested removal of two items from Factor 2, and this sugges- States. The sample included seven men and four women who tion was supported by the fact that these were the only were fairly well-distributed by age. In the Chinese interviews, reverse-keyed items in that factor. After their removal, the fit respondents expressed confusion about items regarding 4 SAGE Open Identify any CALIBRATED ITEM BANK – differences due to Country: Jordan and REFERENCE GROUP language Egypt Country: Jordan and Egypt Language: Arabic Language: English Number of items: 31 Number of items: 31 Sample Size: N=633 Sample Size: pilot N=588 with follow- up N=324 Identify any differences due to culture Country: China Country: China Language: English Language: Mandarin Identify any differences due to Number of items: 31 Number of items: 31 language Sample Size: N=500 Sample Size: N=500 Figure1. Flowchart of cancer stigma index development design. isolation and being an outcast. Thus, we modified the transla- the 31 items represent a single dimension. Based on results tion to more clearly communicate these terms. from these analyses, we considered items for removal to improve unidimensionality across the four samples. Once a set of items was identified that appeared to be Field Test sufficiently unidimensional in all samples, we used Method IRTPRO to conduct differential item functioning (DIF) Measures. All Study 2 measures were identical to Study analyses within an IRT framework. DIF, also referred to as 1 measures with the exception that the reduced 31 cancer measurement bias, occurs when people from different stigma items were administered as opposed to the initial 59 groups (e.g., gender or ethnicity) with the same level of the items. latent trait (in this case cancer stigma) have a different probability of giving a certain response to an item. Thus, Sample and procedures. As can be seen in Figure 1, the we compared performance of (a) the Arabic and English 31-item field test was administered in two regions, Jordan/ items from the two Jordan/Egypt samples (JE–JA), (b) the Egypt and China, and in three languages, English, Arabic, Mandarin and English items from the two China samples and Mandarin, to produce four distinct samples. Respon- (CE–CM), and (c) the English items from Jordan/Egypt dents residing in Jordan/Egypt were administered an English and China samples (JE–CE) to determine the comparabil- (JE, n = 324) or an Arabic version (JA, n = 633); and respon- ity of cancer stigma items across language and region (this dents residing in China were administered an English (CE, process is depicted in Figure 1). n = 500) or a Mandarin version (CM, n = 500). As in Study 1, DIF analysis used three steps. First, two-group chi-square we contracted with Harris Interactive to administer the field tests from IRTPRO were evaluated across comparison test via the Internet in both regions, and all study procedures groups and the significance tests for all comparisons were were approved by the RAND IRB. The characteristics of the adjusted using the Benjamini–Hochberg procedure four field test samples are displayed in Table 1. (Benjamini & Hochberg, 1995) at p < .05 to identify candi- date items for removal. Next, to evaluate the magnitude of Evaluation of item set. Our first analytic step was to con- DIF, items demonstrating significant DIF after p value cor- duct CFAs with each sample to evaluate the extent to which rection were further evaluated by computing the weighted Edelen et al. 5 Table 1. Characteristics of Study 2 Field Test Sample, by Region and Language. Jordan/Egypt: Jordan/Egypt: China: English China: Mandarin Variable Value English (n = 324) Arabic (n = 633) (n = 500) (n = 500) Gender Male 224 (69.1%) 351 (55.5%) 251 (50.2%) 261 (52.2%) Female 100 (30.9%) 282 (44.6%) 249 (49.8%) 239 (47.8%) Age 18-29 150 (46.3%) 244 (38.6%) 146 (29.2%) 136 (27.2%) 30-49 147 (45.3%) 331 (52.3%) 289 (57.8%) 264 (52.8%) 50+ 27 (8.3%) 58 (9.2%) 65 (13.0%) 100 (20.0%) Urbanicity City/large town 251 (77.5%) 453 (71.6%) 404 (80.8%) 402 (80.4%) Suburb 28 (8.6%) 88 (13.9%) 53 (10.6%) 47 (9.4%) Small town 31 (9.6%) 25 (3.9%) 33 (6.6%) 37 (7.4%) Village/rural 14 (4.3%) 67 (10.6%) 10 (2.0)% 14 (2.8%) Low income (≤800 JOD, Yes 80 (24.7%) 200 (31.6%) 197 (39.4%) 184 (36.8%) ≤3,000 EGP, ≤¥84,999) No 244 (75.3%) 433 (68.4%) 303 (60.6%) 316 (63.2%) Post-secondary Yes 295 (91.6%) 547 (87.9%) 429 (87.9%) 424 (85.7%) education No 27 (8.4%) 85 (13.5%) 59 (12.1%) 71 (14.3%) Religion Islam 279 (86.1%) 577 (91.2%) 10 (2.0%) 3 (.6%) Buddhism 0 (0%) 0 (0%) 54 (10.8%) 89 (17.8%) Christianity 34 (10.5%) 27 (4.3%) 48 (9.6%) 26 (5.2%) Other/multiple/ 11 (3.4%) 29 (4.6%) 122 (24.4%) 47 (9.4%) decline None 0 (0%) 0 (0%) 266 (53.2%) 335 (67.0%) area between the expected score curves (wABC; Edelen, Table 2. Items Comprising the CSI. Stucky, & Chandra, 2013). 1. I would feel uncomfortable talking to a person with cancer After removing items with problematic DIF, we fit a four- 2. Treatment and support are useless for someone with cancer group IRT model. Based on those results, we reconsidered 3. I would feel uncomfortable sitting next to someone with item content and properties in an effort to further reduce the cancer item set and arrive at a final index. 4. I would feel uncomfortable sending own child to school with another child with cancer Scoring and examining CSI scores. Once the CSI was final- 5. I would feel uncomfortable if someone with cancer lived ized, we calculated IRT-based scores (i.e., expected a pos- nearby teriori [EAP]) using a summed score conversion algorithm 6. If a close friend had cancer, I would avoid him or her (Thissen, Pommerich, Billeaud, & Williams, 1995). These 7. I would feel uncomfortable being friends with someone with cancer scores retain the benefits of the IRT model and are also prac- 8. People can only blame themselves for getting cancer tical for general use via score translation tables. Summed 9. If had cancer, I would be ashamed score EAPs were generated for the CSI using the final four- 10. I would feel isolated/alone if I received treatment for cancer group IRT model and were rescaled along a T-score metric 11. If my spouse had cancer, I would be ashamed of him or her with the Jordan English sample mean set to 50 and standard 12. If my spouse had cancer, I would consider leaving him or her deviation to 10. Finally, we conducted a set of descriptive analyses to exam- Note. CSI = cancer stigma index. ine the CSI scores according to demographic and personal characteristics of interest (e.g., gender, income, age, personal experience with cancer, attitudes about general health stigma). model fit with the same set of items across all four samples (JE: CFI = .955, TLI = .951, RMSEA = .063; JA: CFI = .922, Results TLI = .916, RMSEA = .078; CE: CFI = .903, TLI = .895, Identifying the 12-item CSI. The fit of the 31 items to a RMSEA = .087; CM: CFI = .946, TLI = .941, RMSEA = single-factor model in each sample was reasonable but did .078). IRT–DIF analyses identified 9 items with problematic not reach standard criteria for all four samples (JE: CFI = DIF that were also removed. .886, TLI = .878, RMSEA = .090; JA: CFI = .852, TLI = We used results from a four-group IRT calibration of the .841, RMSEA = .096; CE: CFI = .885, TLI = .877, RMSEA = remaining 18 items to reduce the item set further by remov- .084; CM: CFI = .921, TLI = .916, RMSEA = .083), indicat- ing items based on redundant content and/or poor psycho- ing the need for some modifications. In consultation with the metric properties. In all, we removed 6 items at this stage. larger study team, 4 items were removed to obtain acceptable Items comprising the final 12-item CSI are listed in Table 2. 6 SAGE Open Table 3. Cancer Stigma Scores Among Demographic Groups for Each Sample and Combined. Combined Jordan/Egypt: Jordan/Egypt: China: China: Sample English Arabic English Mandarin Gender M (SD ) 52.0 (10.3) 50.3 (9.9) 48.9 (7.5) 58.4 (12.0) 50.7 (8.6) M M M (SD ) 50.4 (9.4) 49.2 (10.3) 46.5 (6.1) 57.6 (10.6) 49.2 (7.4) F F p .0004 .35 <.0001 .46 .03 Income M (SD ) 50.7 (9.7) 48.6 (8.8) 46.9 (6.0) 58.5 (11.7) 50.0 (8.0) H H M (SD ) 52.4 (10.2) 54.2 (12.3) 49.1 (8.2) 57.3 (10.7) 49.9 (8.4) L L p .0003 .0003 .0007 .27 .88 Age M (SD ) 51.0 (9.5) 52.1 (11.2) 48.1 (7.1) 56.0 (9.9) 49.7 (8.1) L L M (SD ) 51.6 (10.4) 48.6 (8.9) 47.4 (6.9) 59.1 (11.9) 50.3 (8.4) M M M (SD ) 50.4 (9.0) 45.9 (5.4) 46.1 (5.0) 57.7 (11.1) 49.4 (7.1) H H p .18 .0009 .10 .03 .60 Post-secondary education M (SD ) 51.9 (11.3) 59.0 (15.4) 46.7 (7.1) 59.3 (12.1) 49.4 (8.3) N N M (SD ) 51.1 (9.5) 49.2 (9.0) 47.7 (6.8) 57.6 (11.0) 50.1 (8.0) Y Y p .27 .003 .21 .28 .52 Lived outside country >5 years M (SD ) 50.1 (8.8) 49.9 (10.0) 47.6 (6.5) 55.7 (10.3) 49.2 (7.5) N N M (SD ) 53.7 (11.5) 50.1 (10.1) 47.6 (8.1) 60.9 (12.0) 51.6 (9.0) Y Y p <.0001 .85 .97 <.0001 .005 Diagnosed with cancer: Self M (SD ) 50.6 (9.2) 49.6 (9.8) 47.5 (6.6) 56.7 (10.5) 49.9 (7.8) N N a a M (SD ) 64.4 (14.3) 57.1 (12.6) 52.1 (18.5) 67.4 (12.6) 67.4 (24.4) Y Y p <.0001 .005 .54 <.0001 .34 Diagnosed with cancer: Family/loved one M (SD ) 52.1 (10.3) 51.6 (11.0) 48.4 (7.7) 58.1 (11.4) 50.1 (8.3) N N M (SD ) 49.4 (8.6) 47.4 (7.5) 46.4 (5.2) 57.6 (11.1) 49.6 (7.5) Y Y p <.0001 <.0001 .0002 .63 .56 Ever paid caregiver for cancer patient M (SD ) 50.2 (8.9) 48.9 (8.9) 47.5 (6.8) 56.1 (10.6) 49.6 (7.6) N N M (SD ) 55.2 (12.2) 51.8 (11.2) 47.9 (7.6) 62.1 (11.8) 52.0 (10.7) Y Y p <.0001 .05 .64 <.0001 .09 Group size < 10. We ran a final four-group calibration to obtain item param- four samples, providing some preliminary validity evidence eters and generate CSI scores. Appendix A1 provides a for the CSI (range across samples r = .35-.50). The correla- score translation table for the 12-item CSI. The IRT-based tion of the CSI with the item likening cancer to a death sen- score reliability for the CSI varies by sample and ranges tence was less consistent and slightly lower on average from acceptable to excellent (JE = .79; JA = .73; CE = .91; (range across samples r = .20-.49). All correlations were sig- CM = .81). nificantly different from 0 at p < .05. Scores from the CSI were compared across various demo- Examining CSI scores. The JA sample had the lowest CSI graphic groups for each of the four samples to establish ini- score (M = 47.6, SD = 6.9), with JE (M = 50.0, SD = 10.0) tial validity evidence. A summary of these results is contained and CM (M = 50.0, SD = 8.1) both at 50. The CE sample had in Table 3. the highest mean CSI score (M = 58.0, SD = 11.3). Tests of significance between the four CSI sample means revealed Discussion that the JE and CM means are not different from one another, but all other group mean comparisons are statistically sig- This analysis of cancer stigma and development of the CSI nificant. will provide critical benefits to the cancer research and con- Correlation analyses revealed a moderate to strong corre- trol fields. Our process of integrating literature review with lation between the CSI and general illness stigma across all stakeholder input and measures analysis represents a robust Edelen et al. 7 method of developing a quality, user-friendly index that can of cognitive interviews to ensure terms and whole items are be used by cancer organizations to inform cancer stigma interpreted as intended. Furthermore, data from field tests in reduction initiatives and broader public awareness cam- new regions and languages must be analyzed to determine paigns. Moreover, the CSI can be added to cancer research the comparability of CSI scores back to the reference sample studies examining patient, family, and public perspectives (Jordan/Egypt—English). regarding cancer screening, diagnoses, and treatment. It is important and possible to show that CSI scores cor- While there were limited cancer stigma items or scales relate with other indicators of stigma. For example, we available for modification at the outset of this study, there should expect to find that higher CSI scores correlate with were several scales in the area of stigma and health care national or local policies that discriminate against people decision making as well as stigma and chronic health issues with cancer, and predict lower levels of treatment-seeking, (e.g., mental health) that were particularly useful. The less positive psychological well-being, and greater social stakeholder input we solicited through technical advisors isolation among people with cancer. Future research along and cognitive interviews was critical. Without that feed- these lines could provide valuable validity evidence for the back, the scale would not have included particular cultural CSI. Overall, the CSI can be used to inform initial cancer “pulse points” such as the role of religion or fate in driving education efforts, identifying overall stigma levels in a coun- cancer views. The moderate to strong correlation between try or community and particular issue areas requiring con- the CSI and general illness stigma across all four samples certed intervention. Over time, following careful data suggests that the CSI is reasonably robust and indicative of analyses and perhaps slight modifications to index scoring, general health-seeking stigma. As expected, male gender the CSI can be used as an index comparing countries or com- and low income were associated with higher CSI scores. munities on stigma levels, prioritizing where education Interestingly, CSI scores indicated that those who had a resources should be allocated, and helping to determine the loved one diagnosed with cancer reported lower stigma, impact of stigma reduction efforts. whereas those who had personally experienced cancer reported more stigma. Appendix A1 Our approach attempted to limit potential weaknesses in design where possible. But a few study limitations should be 12-Item Cancer Stigma Index Total Score to noted. First, our cognitive interviews, particularly for the T-Score Translation Table Arab origin samples, were not conducted in Egypt and Jordan. Although we attempted to find individuals who had strong cultural ties (e.g., using criteria about upbringing in Total Score T-Score Total Score T-Score the region), we may not have received the full complement 12 42 37 69 of cultural insights from those who were Americans or had 13 43 38 70 spent considerable time in the United States. Second, our 14 44 39 71 pilot and field tests endeavored to obtain diversity by age, 15 46 40 72 gender, income, and education. For the latter two categories, 16 47 41 73 we approached our goal but did not always meet it in terms 17 48 42 74 of education level, with a slightly higher education status 18 49 43 75 overall. We know that education may influence cancer stigma 19 50 44 76 perspectives. Furthermore, our mode of testing the cancer 20 51 45 77 stigma items was web-based, which may impede participa- 21 52 46 78 tion from those of lower socioeconomic status. Mode effects 22 53 47 80 of web-based administration could not be tested (given that 23 54 48 81 was the only mode used), but may affect the interpretation of 24 55 49 82 items. 25 56 50 83 26 57 51 84 More robust validity tests of the CSI will require addi- 27 58 52 85 tional use in the countries in which we developed the first 28 59 53 86 versions of the scale—Egypt, Jordan, and China. This testing 29 60 54 87 may include using the CSI in diverse communities and with 30 61 55 88 a wide variety of subpopulations (e.g., setting, age). The 31 62 56 89 scale was developed for use initially in these countries based 32 64 57 90 on strategic plans and investments. However, the intention is 33 65 58 91 for the scale to ultimately be used worldwide. Thus, as the 34 66 59 92 CSI is translated and used in new contexts, it will be impor- 35 67 60 93 tant to step through all of the phases used in this study, prin- 36 68 cipally review of the translation by experts and some version 8 SAGE Open Acknowledgments Evans-Lacko, S., Little, K., Meltzer, H., Rose, D., Rhydderch, D., Henderson, C., & Thornicroft, G. (2010). Development We acknowledge the data collection support from Sandra and psychometric properties of the mental health knowledge Applebaum and Roz Pierson at Harris Interactive. We thank Patrick schedule. Canadian Journal of Psychiatry/Revue Canadienne Orr for his efforts in manuscript formatting, and researchers Omar De Psychiatrie, 55, 440-448. Al-Shahery, Samuel Berkowitz, Dolly Dahdal, and Chaoling Feng Fries, J. F., Bruce, B., & Cella, D. (2005). The promise of PROMIS: at RAND for their efforts to review the Arabic and Chinese versions Using item response theory to improve assessment of patient- of the scale. Finally, we thank the many individuals who have par- reported outcomes. Clinical and Experimental Rheumatology, ticipated in the cognitive interviews for their insight, which 23(5), S53-S57. strengthened our research. Green, S., Davis, C., Karshmer, E., Marsh, P., & Straight, B. (2005). Living stigma: The impact of labeling, stereotyping, separa- Declaration of Conflicting Interests tion, status loss, and discrimination in the lives of individuals The author(s) declared no potential conflicts of interest with respect with disabilities and their families. Sociological Inquiry, 75, to the research, authorship, and/or publication of this article. 197-215. Greene, K., & Banerjee, S. C. (2006). Disease-related stigma. Journal of Homosexuality, 50(4), 185-209. Funding Hu, L.-t, & Bentler, P. M. (1999). Cutoff criteria for fit indexes in The author(s) disclosed receipt of the following financial support for covariance structure analysis: Conventional criteria versus new the research and/or authorship of this article: This work was funded alternatives. Structural Equation Modeling: A Multidiscipli- by an unrestricted grant from the LIVESTRONG Foundation. nary Journal, 6, 1-55. Joachim, G, & Acorn, S. (2000). Living with chronic illness: The References interface of stigma and normalization. The Canadian Journal Ader, D. N. (2007). Developing the patient-reported outcomes mea- of Nursing Research/Revue canadienne de recherche en sci- surement information system (PROMIS). Medical Care, 45(5), ences infirmieres, 32(3), 37-48. S1-S2. LIVESTRONG Foundation. (2007). Cancer stigma and silence Batson, C. D., Polycarpou, M. P., Harmon-Jones, E., Imhoff, H. around the world. Austin, TX: LIVESTRONG Foundation. J., Mitchener, E. C, Bednar, L. L., . . .Highberger, L. (1997). Muthén, L. K, & Muthén, B. O. (2007). 1998-2007. Mplus user’s Empathy and attitudes: Can feeling for a member of a stig- guide. Los Angeles, CA: Author. matized group improve feelings toward the group? Journal of Pettit, M. L. (2008). Disease and stigma: A review of literature. Personality and Social Psychology, 72, 105-118. Health Educator, 40(2), 70-76. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false dis- Reeve, B. B., Hays, R. D., Bjorner, J. B., Cook, K. F., Crane, P. K., covery rate: A practical and powerful approach to multiple Teresi, J. A., . . . Cella, D. (2007). Psychometric evaluation and testing. Journal of the Royal Statistical Society: Series B calibration of health-related quality of life item banks: Plans (Methodological), 57, 289-300. for the patient-reported outcomes measurement information Brewster, M. E., & Moradi, B. (2010). Perceived experiences of system (PROMIS). Medical Care, 45(5 Suppl. 1), S22-S31. anti-bisexual prejudice: Instrument development and evalu- doi:10.1097/01.mlr.0000250483.85507.04 ation. Journal of Counseling Psychology, 57, 451-468. Sartorius, N. (2007). Stigma and mental health. The Lancet, 370, doi:10.1037/A0021116 810-811. Brohan, E., Slade, M., Clement, S., & Thornicroft, G. (2010). Thissen, D., Pommerich, M., Billeaud, K., & Williams, V. S. L. Experiences of mental illness stigma, prejudice and discrimi- (1995). Item response theory for scores on tests including poly- nation: A review of measures. BMC Health Services Research, tomous items with ordered responses. Applied Psychological 10, Article 80. doi:10.1186/1472-6963-10-80 Measurement, 19, 39-49. Browne, M. W, Cudeck, R., Bollen, K. A., & Long, J. S. (1993). Uys, L. R, Holzemer, W. L, Chirwa, M. L, Dlamini, P. S, Greeff, Alternative ways of assessing model fit. Sage Focus Editions, M., Kohi, T. W, . . . Naidoo, J. R. (2009). The development 154, 136. and validation of the HIV/AIDS stigma instrument-nurse Cai, L., du Toit, S. H. C, & Thissen, D. (2011). IRTPRO version 2: (HASI-N). AIDS Care, 21, 150-159. Flexible, multidimensional, multiple categorical IRT modeling. Chicago, IL: Scientific Software International. Author Biographies Cella, D., Yount, S., Rothrock, N., Gershon, R., Cook, K., Reeve, Maria Orlando Edelen, PhD, is a senior Behavioral Social B., . . .Rose, M. (2007). The Patient-Reported Outcomes Scientist scientist and psychometrician at the RAND Corporation. Measurement Information System (PROMIS): Progress of an Her research focuses primarily on the use of item response theory NIH Roadmap cooperative group during its first two years. for development, refinement and evaluation of bahavioral health Medical Care, 45(5 Suppl. 1), S3-S11. measures. DeWalt, D. A., Rothrock, N., Yount, S., & Stone, A. A. (2007). Evaluation of item candidates: The PROMIS qualitative item Anita Chandra, PhD is a senior policy researcher and director of review. Medical Care, 45(5 Suppl. 1), S12-S21. the Behavioral and Policy Sciences Department at the RAND Edelen, M. O., Stucky, B., & Chandra, A. (2013). Quantifying Corporation. Her background is in public health, child and adoles- “problematic” DIF within an IRT framework: Application to cent development, and community-based participatory research a Cancer Stigma Index. Quality of Life Research. Advance and evaluation. She currently leads or co-leads studies on com- online publication. munity well-being; deployment and military families; community Edelen et al. 9 resilience and long-term disaster recovery; and child health and Claire Neal, MPH is the Executive Director of the Triangle Global development. Health Consortium, a non-profit member organization represent- ing institutions and individuals from the pharmaceutical and bio- Brian Stucky is an Associate Behavioral Social Scientist at RAND technology industry, the international health development commu- Corporation, where he specializes in psychometrics and health nity, and academia – all dedicated to improving the health of the research. He is a trained psychometrician whose research interests world’s communities. Claire has over 13 years of experience in in focus on developing and applying new test theory models (includ- the development and delivery of innovative health programs. For ing factor analytic and item response theory-based) that serve to the last decade, Claire has worked at the LIVESTRONG Foundation; better understand patient response data in settings where multiple most recently as the Vice President of Global Strategy where she dimensions are present. provided the leadership and vision for LIVESTRONG’s global work. Rebekkah M. Schear, MIA has spent her career working at the intersection of global health and international development. Over Ruth Rechis, PhD, is the Vice President of Programs & Strategy the last five years at the LIVESTRONG Foundation, Rebekkah led at the LIVESTRONG Foundation. As a member of the Executive the Foundation’s implementation of six international programs Leadership Team, she oversees the day to day operations of the throughout Latin America, Africa and Asia focusing on reducing Foundation’s programmatic work and shepherds the strategic cancer stigma and empowering cancer survivors to become advo- planning process for the Foundation. As an adolescent cancer sur- cates. Currently, she is working to create a new, innovative model vivor, Dr. Rechis has a personal connection to the mission of the of patient-centered cancer care with the goal of improving delivery Foundation. of patient-centered cancer care globally.

Journal

SAGE OpenSAGE

Published: Sep 16, 2014

Keywords: cancer; stigma; health; chronic disease; help-seeking; health disparities

References