Main

The prevalence of type 2 diabetes (T2DM) in Mexican Americans is two to three times higher than in the general population, predisposing them to develop diabetic nephropathy (DN) more than non-Hispanic whites.1 DN (occurring mainly in adults) clusters in families with T2DM, suggesting that there is a genetic susceptibility for its development and progression.2 The main goal of the Family Investigation of Nephropathy and Diabetes (FIND) study is to identify gene(s) involved in the development of DN (http://darwin.cwru.edu/FIND). As a Participating Investigation Center in FIND, we are recruiting Mexican-American DN probands (first affected family member recruited into the study) with a family history of T2DM and their siblings and parents living in the San Antonio metropolitan area. FIND inclusion criteria include recruiting probands who have advanced nephropathy secondary to T2DM and have both living, willing parents or at least one sibling with T2DM or DN.

The success of a genetic family study (GFS) depends on recruiters' ability to enroll a sufficient number of probands and their relatives. Enrollment of probands and select family members is influenced by a variety of factors including social, cultural, and ethical components.3 One important ethical issue is to ensure subjects' informed, voluntary consent prior to their participation in GFS.4 Understanding these factors will enhance enrollment and address ethical issues associated with participation in GFS. In this report, we introduce the Contextual Assessment Approach (CAA), a method that we found helpful while conducting 50 structured interviews with DN probands to evaluate and identify issues regarding family enrollment. The CAA allows systematic data collection that facilitates the interpretation of specific contextual patterns and meanings when attempting to enroll families. We will describe this new approach, explain how it enhances enrollment, and discuss its efficacy by presenting our experience and results based on interviews conducted with recruiters, DN patients, and their relatives.

METHODS

CAA was standardized to our recruitment procedures in the FIND study at the University of Texas Health Science Center at San Antonio (UTHSCSA). We conducted structured interviews using the CAA questionnaire (CAA-Q) with 50 low-income, Mexican-American probands and 69 relatives recruited from University Health System facilities in San Antonio, Texas. Administering the CAA-Q requires approximately 20 minutes. To ensure comparability of both English and Spanish questionnaires, translation–back translation methodologies were applied as described by Bernard.5 Three bilingual recruiters reviewed, discussed, and revised discrepancies. This version was then administered to five bilingual participants who indicated that they understood the questions in both languages. All subjects provided their informed consent before being interviewed; the institutional review board at UTHSCSA approved the study protocol and final versions of the CAA-Q.

Three focus group sessions were performed with four FIND recruiters at UTHSCSA to discuss their experience with the CAA, particularly their thoughts concerning efforts and time designated to enroll a family pre- and postadministration of CAA-Q.

CAA evaluation

We evaluated the efficacy of the CAA by using both qualitative and quantitative approaches, including (1) case examples (Appendix 1); (2) the use of weekly statistics forms to evaluate the effectiveness of the CAA in maximizing participation by comparing the number of families likely to be enrolled based on CAA evaluation to the number of families successfully enrolled in FIND (this form also supplies data on technical difficulties that might be encountered by recruiters, i.e., inaccuracies in reported telephone numbers, sample collections); and (3) calculating the enrollment rate and percent effort saved by the CAA as follows:

Percent effort saved by the CAA

= (no. of probands with unwilling families/

total no. of candidate probands) × 100%

Enrollment rate of families by using CAA

= (no. of families enrolled/

no. of probands recruited) × 100%

Data analysis

Content analysis was applied to qualitative data generated through participant and recruiter interviews. We established a method for standardizing and displaying interview data, as described by Miles and Huberman,6 to analyze text using techniques from qualitative analysis, as discussed by Bernard.5 We built initial matrices for each participant from blocks of text to identify trends among responses. Patterns were then summarized into higher-level matrices from which categories and themes were grouped together.

Descriptive statistics were performed (i.e., frequency distributions, mean, standard deviation, and range) on variables such as participants' age and years of education. Associations between variables in the form of contingency tables were examined by using a χ2 test. A Fisher exact test was used when the expected cell frequencies were small. We cross-checked all phases of analysis to determine how to apply coding categories and resolve any anomalies or discrepancies. Interrater reliability was established by validating consistency in coding and classification procedures by a second researcher recoding 50% of the case materials and checking for discrepancies.6 All data analyses were performed with SPSS.7

RESULTS

A. CAA: Concept development

CAA entails conducting structured interviews to explore the circumstances influencing subjects' opinions and attitudes toward participating in GFS. In structured interviews, all subjects are asked to respond to identical sets of questions to ensure that the output can be reliably compared.5 CAA-Q is composed of closed- and open-ended questions directed to explore health beliefs, family dynamics, and ethical issues associated with participation in FIND. Family dynamics are a special focus of the CAA because the family is the basic unit of analysis in GFS. Therefore, for the purpose of this report, family dynamics are regarded as the ongoing interpersonal interactions and relationships between family members (i.e., siblings and parents) that influence decisions regarding enrollment.8,9

The concept of integrating the CAA into regimens used in GFS resulted from performing several ethnographic studies.10,11 Previously, we conducted in-depth, semistructured interviews of open-ended questions with DN patients and their relatives (a subset of FIND participants) to explore their health beliefs about T2DM and DN and the associated ethical issues toward participation in GFS. Beliefs and attitudes are important distinct aspects of each culture and are essential for analyzing health-related behaviors.12,13 On the basis of participants' responses that emerged from these interviews, we quantified the interview data by developing closed-ended questions using theme-based coding categories which were representative of four or five response options. Participants selected the choice that best reflected their answer from which we classified several rising patterns that influenced probands' decisions to invite relatives to GFS. These patterns included (1) reporting close family relationships, (2) understanding the nature of the study, (3) enhancing awareness of the genetic nature of T2DM and DN, and (4) relatives' concerns about privacy and confidentiality. These patterns were tested, modified, and tailored to response options that best captured participants' answers in the CAA-Q. We then tested the questionnaire for contextual relevance (i.e., clarity, logic, and degree of importance for questions) by administering it to a select group of 10 ethnically diverse respondents and modifying it according to questions and responses that were best understood by FIND participants (mainly Mexican Americans). Several focus group interviews were also conducted with recruiters as an additional measure of instrument validity.6

CAA-Q is composed of three parts. The first set of questions is directed at collecting basic demographic data on participants such as gender, age, and educational level. These questions are important in the characterization of the study population to generate hypotheses regarding enrollment. The second part of the questionnaire is composed of 15 closed-ended and 4 open-ended questions formulated to gather data on subjects' health beliefs about T2DM and DN and to allow recruiters to establish trust and rapport with subjects. It also prompts relatives to consider family history as an important risk factor for developing DN. The third part of the questionnaire (12 closed-ended and 2 open-ended questions) examines two important themes: family dynamics and ethical issues associated with participants' decisions to participate in GFS (i.e., awareness of risks, disclosure of health information; see Appendix 2).

B. The recruitment process

The recruitment process with the CAA-Q is summarized in Table 1. Initial screening begins with a list of patients from dialysis units, renal clinics, and transplant clinics that recruiters screen for FIND inclusion criteria (i.e., DN diagnosis and a family history of diabetes;Fig. 1). Contrary to conventional recruitment regimens used in GFS (Table 1), probands meeting FIND inclusion criteria and agreeing to participate provide CAA consent to interview them in person or by telephone. The contextual analysis of subjects' perspectives, attitudes, and opinions concerning participation in the FIND study provides an understanding of factors influencing enrollment and allows for efficient reporting of information to determine what data are relevant and appropriate to the inference process. CAA evaluation will determine whether recruiters attempt to enroll potential probands and their relatives. Probands who indicate “family unwillingness” will not be enrolled in FIND even if they meet FIND inclusion criteria. On the other hand, we will attempt to enroll probands who reveal “family willingness” or who are “undecided” regarding their relatives' participation. In such cases, recruiters first ask probands to contact their relatives to assess whether they are willing to participate. If relatives agree, recruiters contact relatives and enroll them in FIND. Steps 3, 4, 5, and 8 are integrated into conventional GFS recruitment regimens as shown in Table 1.

Table 1 Recruitment and enrollment of families at the San Antonio Center of FIND and application of the CAA
Full size table
Fig. 1
figure 1

Pedigree screening sheet. Initial screening for all diabetic nephropathy patients to identify whether they meet FIND inclusion criteria (i.e., have family history of diabetes and have a sibling with diabetes or diabetic nephropathy).

Full size image

C. CAA-Q implementation (subjects' structured interviews)

Of the 200 DN patients screened at our local clinics, 50 (25%) met FIND inclusion criteria and completed the CAA-Q. Twenty-eight of the 50 were willing probands who reported having willing relatives, 6 were willing probands but undecided concerning their relatives' willingness to participate, and 16 were willing probands with unwilling relatives (Table 2).

Table 2 Probands' responses concerning family participation based on CAA evaluation
Full size table

We enrolled two families from the six willing probands who were undecided concerning their relatives' participation. Consequently, the total number of families enrolled was 30 (28 + 2). The total number of probands recruited for FIND was 34 (28 + 6).

Enrollment rate of families by using CAA

= (no. of families enrolled/

no. of probands recruited) × 100%

(30/34) × 100% = 88%

By using CAA, we excluded 16 families and saved time by not allocating efforts to recruit unwilling relatives.

Our percent effort effort saved by using the CAA

= (no. of probands with unwilling families/

total no. of candidate probands) × 100%

(16/50) × 100% = 32%

CAA saved 32% of our recruiting efforts by excluding this subset of participants. Recruiters saved the time needed to enroll willing probands and time required to recruit unwilling relatives.

Multiple attempts to recruit members from the 16 “unwilling” families were carried out. We planned to follow up with 8 of the 16 willing probands who gave us permission to contact their relatives to validate their responses concerning relatives' participation. After numerous contacts with four probands (six follow-ups) and eight of their relatives (seven follow-ups), we were unsuccessful at recruiting family members. We elected not to follow any more cases for the following ethical and practical reasons: (1) we wished to avoid exerting unnecessary pressure on probands and their relatives to the extent that participation becomes involuntary, (2) the CAA has clearly identified important factors affecting family participation, and (3) recruiting unwilling relatives is very tedious, time-consuming, and unlikely to materialize based on our previous experience.

D.1. Sociodemographic characteristics of subjects

The CAA was administered to 119 subjects, 50 FIND Mexican-American DN probands on dialysis and 69 relatives. Fifty-eight percent of the subjects were female, with an average age of 54.2 years (range: 33–76, SD = 9.04). The majority of our study population reported having at least two generations of Mexican ancestry (90%), with only 10% born in Mexico but permanently residing in the United States. Two thirds had a high school education or less. Eighty-four percent of the probands had a yearly household income ≤$20,000 compared with 48% of their relatives (Table 3). Analysis of association revealed a higher percentage of females (73%) among enrolled relatives (χ2 test, P ≤ 0.05;Table 3). This observation infers that there is a significant gender difference in relatives' attitudes toward enrollment that warrants further investigation (manuscript in preparation). We have also studied the relationship between relatives' attitudes toward participating in FIND and their age, level of education, and income. No significant differences were observed. This may be due to the fact that our sample size is relatively small. An increased variation in responses may result from a larger sample of subjects of different ethnicities.

Table 3 Sociodemographic characteristics of probands and relatives
Full size table

D.2. Health beliefs

CAA-Q explored DN patients' and their relatives' health beliefs concerning the clustering of T2DM and DN. Data revealed many similarities between patients' and relatives' thoughts about T2DM and DN clustering (χ2 test;P > 0.05). The majority of participants (n = 76, 64%) perceived T2DM to run in their family; only 36% (n = 43) believed that DN ran in their family. Ninety-two percent (n = 110) of the subjects indicated that they were willing to provide information about the clustering of T2DM in their families, whereas 8% (n = 9) did not agree.

D.3. Ethical issues

Family relationship was one important factor that influenced probands' decisions regarding relatives' enrollment in FIND. Sixty-three percent of willing probands were reluctant to contact their relatives because they felt they did not have good relationships with them (Table 4). CAA promotes relatives' voluntary participation by exploring and asking specific questions related to family dynamics and relationships. For example, the CAA was administered to a 28-year-old proband who reported having a brother with T2DM and four healthy siblings. The proband stated, “Although I would like to contact them [siblings], they will not participate.” Further probing revealed that his siblings would feel that researchers would test them for kidney compatibility for a donation. The proband disclosed that kidney donation was a sensitive issue within the family and that it was a source of tension. Through administering the CAA-Q and building rapport with this proband, recruiters were able to identify this sensitive issue. CAA identifies and addresses the degree to which researchers should pursue recruiting relatives to ensure voluntary participation.

Table 4 Factors affecting family participation in FIND
Full size table

Furthermore, CAA-Q explores subjects' awareness of some of the risks associated with participation in GFS to ensure that subjects are informed when they consent to participate. Examples of such risks include (1) the disclosure of health and identifiable information regarding relatives without their prior knowledge and (2) confidentiality and privacy of genetic information. Seventy-six percent (n = 90) of our participants stated that there were no risks associated with participation compared with only 10% (n = 12) who felt that there were. Fourteen percent (n = 17) of our participants reported that they did not know whether or not there were any risks (manuscript in preparation). Therefore, when enrolling subjects in FIND, recruiters spend additional time to explain thoroughly the risks associated with participation in GFS/FIND as they obtain subjects' consent.

DISCUSSION

An understanding of subjects' opinions and attitudes toward participation in GFS will maximize enrollment and address subjects' needs.14 Incorporating the CAA into regimens used in GFS entails conducting structured interviews to explore the circumstances influencing subjects' attitudes toward participating in GFS. The CAA focuses on participants' perspectives because the initiation, implementation, and evaluation of the enrollment process are based on the contextual analysis of participants' health beliefs, family dynamics, and ethical issues associated with the enrollment. Once probands meet inclusion criteria, recruiters can apply the CAA to assess probands' attitudes toward participation and inviting their relatives. Correspondingly, Trauth et al.14 examined public attitudes regarding willingness to participate in medical research studies. They found that almost half of the respondents were willing, while one third were undecided. They suggested that researchers enhance participation by inviting those subjects identified as willing to participate in studies.

Enrolling probands in GFS generally requires less time and effort than enrolling their relatives.10 As illustrated in Case 3 (Appendix 1), without the CAA, a significantly larger amount of time is spent on recruitment of probands who are unsuccessful at inviting their relatives. Our results indicate that at least one third of our probands fit this profile. By excluding this subset of participants, the CAA saves nearly 32% of recruitment efforts and ensures voluntary participation. CAA enhances and addresses ethical issues associated with subjects' participation in GFS for the following reasons:

1. CAA explores factors and concerns affecting participation such as family dynamics and probands' relationships with their relatives (Case 1, Appendix 1). Our participants' concerns were consistent with those raised in research related to GFS, yet different from concerns associated with research studies that are clinical in nature. The inherent difference in GFS results from the shift in focus from the traditional model of an individual to that of a family model. In the family studies model, (1) subjects have ties to other research participants through shared genetic heritage, (2) information learned from the research may affect the entire family, and (3) family members may become part of the study without their consent.15 Some of our probands were reluctant to contact relatives because they (1) do not have a good relationship or (2) have concerns about privacy and confidentiality. Probands indicated that if they supply lists of family members for researchers to contact, relatives might feel that their privacy is invaded. Therefore, by applying the CAA before enrolling subjects into GFS, recruiters spend additional time to thoroughly explore and address concerns and risks associated with participation in GFS/FIND as they recruit and obtain consent for all members of a family. Similarly, Parker16 indicated that enrollment in GFS can involve researchers pressuring probands to contact their relatives, resulting in their involuntary participation. Consequently, researchers should aim to minimize pressure on participants by applying a preventive ethics approach to recognize potential ethical conflicts and prevent them from arising.

2. Another ethical issue identified in this report is our subjects' pervasive lack of awareness regarding risks associated with participation in GFS. While discussing potential risks such as discrimination in the workforce and health insurance, subjects did not consider these issues to influence their decision to participate in GFS. Individuals who are unaware of their risks cannot provide informed consent when participating in a study. Potential enrollment in genetic research is affected by subjects' knowledge about the risks associated with their participation.4,16 Unlike research in which the risks are primarily physical, in our study the primary physical risk, drawing blood, is negligible in comparison with the potential magnitude of social, economic, and psychological harms. For this reason, effective communication and broadening the perspective to include the family unit are crucial components in ensuring that informed decisions to participate in GFS are made. This report stresses the importance of incorporating CAA to GFS recruitment regimens used among a low-income minority Mexican-American population. This assessment enables recruiters to explore potential risks associated with subjects' participation by asking specific questions about the disclosure of health and identifiable information on relatives and issues related to confidentiality and privacy of genetic information to ensure informed consent before enrolling them in FIND. Future publications will discuss such ethical issues in more detail (Arar et al., American Society of Human Genetics, October 2002).

3. CAA also explores participants' health beliefs and enhances their awareness of the genetic nature of DN and T2DM, especially among at-high risk healthy relatives. Contrary to common beliefs about Mexican Americans' use of alternative treatment and curanderos (traditional healers) for the treatment of diabetes, we found that in our study population, biomedical treatments were the preferred treatment. Consequently, traditional attitudes and beliefs did not present any barriers for treatments.12 The CAA-Q encourages participants to learn about T2DM and DN while stimulating them to think about family history as an important risk factor to promote participation in GFS and consider preventive measures. This aspect of the CAA is very important because 53% of FIND diabetic relatives have unknowingly already developed kidney problems and about 30% of “healthy” participants are not aware of having diabetes. Similar to our investigation, Kreiger et al.17 found that potential participants' knowledge about the genetic nature of a disease influences their decisions toward participation. However, willingness to participate in FIND was not only influenced by participants' beliefs about the clustering of T2DM and DN. Other factors such as family relationships seemed to be equally important in determining family participation (refer to Cases 1 and 2, Appendix 1). While our study addresses a population affected by a common complex disease (T2DM) caused by both environmental and genetic factors, health beliefs regarding monogenic diseases (i.e., cystic fibrosis) caused by genetic factors might have a greater impact on participation and the disclosure of health information by family members.

4. In addition, CAA enhances the interaction between recruiters and participants by allowing recruiters to establish rapport and trust with subjects (for more information about recruiters' perspective refer to http://darwin.cwru.edu/FIND). Recruiters learn to identify the “key informant” of a family, who is usually responsible for issues related to his/her family. Identifying a “key informant” facilitates contact between recruiters and families, thus enhancing the study's enrollment outcome. In addition, probands feel that recruiters are involving them in the study by exploring their opinions regarding diabetes and kidney disease as opposed to enrolling them and providing financial compensation for their samples. By establishing rapport with subjects, recruiters better understand factors influencing subjects' participation, evaluate probands' attitudes toward enrollment, and direct recruitment efforts toward willing families (see Case 2, Appendix 1). Similarly, Foster et al.18 suggested that understanding the relevant decision-making units in a society will permit investigators to identify appropriate representatives who can promote discourse with the community and enhance subjects' participation in GFS. Using the CAA before enrolling subjects in GFS will build trust and facilitate rapport with subjects. Initial contact with probands will, to a large extent, dictate the ability to contact relatives. Similarly, Kreiger et al.17 conducted several focus group sessions to build an understanding of concerns that probands and their relatives have about participating in GFS. Subjects were more likely to participate in GFS when rapport was established with recruiters. These results substantiate previous findings about enrollment strategies that stress the importance of establishing communication between recruiters and study participants.10

5. Finally, successful enrollment is established from interactions between three domains: participants, recruiters, and technical difficulties. Figure 2 shows how these domains interact to enhance enrollment. The CAA improves enrollment by expanding the overlap area among domains and enhancing family participation, monitoring recruiters, and identifying technical difficulties encountered in the field. In particular, CAA improves the interaction between recruiters and participants and allows systematic data collection to reveal the factors that influence participation in GFS.

Fig. 2
figure 2

How the CAA enhances enrollment in GFS. This illustration shows how three main domains—participants, technical difficulties, and recruiters—interact to enhance enrollment. The CAA enhances enrollment by expanding the overlap area among these three domains, particularly to improve the interaction between recruiters and participants.

Full size image

While the broader implications of this approach remain to be verified, we believe CAA will prove to be especially useful when conducting GFS among low-income minority Mexican-American populations to ensure informed decisions to participate and successful enrollment of families. This approach is generalizable to the extent that the questionnaire is tailored to capture ethical concerns within the cultural and socioeconomic context of the population being studied. Our intention in this report is not to generalize current findings to include all subjects participating in GFS but to present and describe an approach that we have found helpful when evaluating and identifying issues that influence enrollment. Additional studies among larger populations and other ethnic minority groups should be performed in the future.