Factors affecting enrollment in literacy studies for English- and Spanish-speaking cancer patients

Center on Outcomes, Research and Education, Evanston Northwestern Healthcare, Evanston, IL 60201, USA.
Statistics in Medicine (Impact Factor: 1.83). 09/2008; 27(20):4119-31. DOI: 10.1002/sim.3259
Source: PubMed


Study participation bias can affect inferences regarding outcomes.
The objective is to compare characteristics of participants and non-participants of two literacy studies.
Two studies of literacy and health-related quality of life were conducted in English- and Spanish-speaking cancer patients. Patients had a range of literacy skills, and each enrolled patient received $20.
Nine hundred and twenty-two English-speaking patients were approached. Among the 651 who met eligibility criteria, 420 were enrolled (64.5 per cent). Four hundred and eighty-seven Spanish-speaking patients were approached. Among the 455 who met eligibility criteria, 414 were enrolled (91.0 per cent) (p<0.001). Multiple imputations were performed to impute missing data. Multivariable logistic regression revealed that recruiting site was the only factor predictive of enrollment in Spanish-speaking patients. Age, education, and recruiting site were important predictors in English-speaking patients. Sensitivity analysis using patients with complete data generated similar results.
Spanish-speaking patients enrolled at a much higher rate than English-speaking patients, which is encouraging for future research in this underserved population. One important literacy-related factor (education) did not affect enrollment in Spanish-speaking patients, suggesting that there was no selection bias in this study. Recruiting sites with more indigent patients and long clinic waiting times had higher enrollment, suggesting that monetary compensation and time availability may be important considerations in study participation.

9 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: Health-related quality of life (HRQL) assessment is frequently used in comparative effectiveness research, but low-literacy patients are often excluded. Appropriately translated and user-friendly HRQL measures are essential to ensure inclusion of low-literate and non-English-speaking patients in comparative effectiveness research. To compare HRQL responses across literacy levels in Spanish-speaking patients with cancer using a multimedia touch screen program. A total of 414 adult patients with cancer (213 with low literacy and 201 with high literacy). The touch screen system administered 3 questionnaires: The Functional Assessment of Cancer Therapy-General, the Short Form-36 Health Survey, and the Standard Gamble Utility Questionnaire. Measurement bias was evaluated using item response theory. Effects of literacy on HRQL were evaluated using regression models. Patients rated the touch screen easy to use and commented favorably on the multimedia approach. There was statistically significant item response theory measurement bias in 6 of 10 HRQL subscales; however, only 3 showed meaningful bias. Low-literacy patients had significantly lower mean scores on 3 of 4 Functional Assessment of Cancer Therapy-General subscales, before and after adjustment for patient characteristics. Low-literacy patients also had significantly lower mean scores on 5 of 6 Short Form-36 subscales; adjustment for patient characteristics attenuated or eliminated differences. Similar proportions of low- and high-literacy patients valued their current health as equivalent to perfect health. This study demonstrates the feasibility of this multimedia touch screen program for low-literacy patients. The program will provide opportunities to evaluate the effectiveness of interventions in more diverse patient populations.
    Medical care 06/2010; 48(6 Suppl):S75-82. DOI:10.1097/MLR.0b013e3181d6f81b · 3.23 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To describe recruitment and retention strategies of a psychosocial intervention with African American and Latina American breast cancer survivors (BCSs). Prospective design with pre- and post-testing. A mailed survey and assignment to telephone counseling or education booklet only. 587 African American and Latina American BCSs were recruited. The sample was drawn from the population-based California cancer and hospital registries, as well as community agencies. Mailed self-report health-related quality-of-life assessments were at baseline and 4-6 months follow-up. Accrual outcomes; recruitment and retention strategies. A total of 375 (64%) completed the baseline survey and 320 (55%) completed both baseline and follow-up assessments. The recruitment outcomes suggest that very special attention must be paid to the initial recruitment of Latina Americans to engage their interest and participation. For African Americans, particular attention must be devoted to their retention to address potential attrition. Findings suggest that the inclusion of lower-income and ethnic minority cancer survivors in a longitudinal intervention study is doable. The results indicate that recruitment outcomes are influenced by participant and study characteristics. Successful enrollment requires investigations that attend to culturally and socioecologically informed recruitment and retention strategies, from staff selection, training, and supervision to overall study approach protocol, to address barriers to participation. Nursing research and practice have championed survivorship care, including psychosocial care. This article outlines practical strategies to recruit and retain population-based samples, ethnic minorities, and underserved survivors.
    Oncology Nursing Forum 09/2012; 39(5):E434-42. DOI:10.1188/12.ONF.E434-E442 · 2.79 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Purpose: To better understand research participation among hard-to-reach populations, this exploratory investigation examined characteristics of enrollees and non-enrollees from a population-based longitudinal study with African-American and Latina-American breast cancer survivors. Methods: A mixed-method recruitment approach was utilized to enroll participants from cancer registries and community groups who were 1-6 years post-diagnosis. Results: Four hundred and sixty-eight participants agreed to participate constituting an 81% participation rate; 65 and 55% completed Time-1, and both Time-1 and Time-2 assessments, respectively. African-Americans were more likely to agree to participate and complete the T1 assessment (73%) than Latinas (62%) (p < 0.05). Participation was influenced by educational attainment and comorbidities (p < 0.05) for African-Americans. Among Latinas, language proficiency, comorbidities and psychological difficulties (p < 0.01) influenced participation. Conclusions: Our findings suggest that enrollment in research studies may be influenced by complex and multi-dimensional factors stemming from subjects' characteristics including ethnicity, culture, language proficiency and literary, and socioeconomic status, as well as medical characteristics including co-occurring chronic illness and psychological status. Thus, comprehensive, multi-method research studies are urgently needed to better understand and address the challenge of minority recruitment in biomedical research. To increase research participation among cancer survivors, it is imperative to implement focused strategies that will support and encourage individuals' enrollment and continued participation in studies.
    Quality of Life Research 07/2014; 24(2). DOI:10.1007/s11136-014-0758-9 · 2.49 Impact Factor