Comparison of 3 Modes of Genetic Counseling in High-Risk Public Hospital Patients

Specific Aims.

The specific aims of this mixed methods study are:

Aim 1. Compare the effectiveness of 3 modes of genetic counseling in a diverse sample of patients at high risk for HBOC in 3 public hospitals. Conduct a multicenter partially randomized preference noninferiority trial with high-risk patients assigned by (a) randomization to three counseling modes: in-person, phone, video conference; or (b) patients´ preference. Utilize validated measures of study outcomes adapted as needed for literacy and language. Recognizing that some potential participants may have a strong preference for one counseling mode, after explaining the study design and obtaining informed consent, participants will be asked if they have such a preference. Those who do will be offered that mode, and those who do not will be randomized. Randomization will be stratified according to hospital and personal history of breast cancer in order to ensure that there is no imbalance in important factors that may be associated with outcome.

Aim 2. Explore inductively and qualitatively variation in patients' genetic counseling experiences and understandings, genetic counselor satisfaction and perceptions, counseling session similarities and differences, and implications of organizational context across three modes of genetic counseling.

Aim 3 employs inductive, qualitative methods to explore in depth the cases of 30 patients using their pre- and post-counseling survey responses, audio tapes of their counseling session, and in-depth interviews to explore questions common to all respondents and those specific to what was learned about the individual from the other data sources. In-depth interviews will also be conducted with genetic counselors.

Study Design:

Using mixed methods, investigators will conduct a multicenter partially randomized preference noninferiority trial with high-risk English-, Spanish-, and Cantonese-speaking patients assigned by (1) patients´ preference or (2) randomization to three counseling modes: (a) in-person; (b) phone; or (c) video conference. A total of 600 patients will complete counseling and 540 will complete the final survey. Baseline and post-counseling surveys will use validated measures (adapted for literacy and language) of study outcomes. All counseling sessions will be audio-taped. A sample of 90 tapes will be analyzed for counseling content and to identify 30 participants for in-depth interviews and analysis triangulating all forms of data. Genetic counselors will be interviewed in depth to elicit their perceptions of the strengths and limitations of each counseling mode.

Investigators' mixed methods combine a multicenter partially randomized preference noninferiority trial with inductive methods that embed this research in the real world of public health system patients. This is practice-based research, designed to emphasize external validity, (relevance and generalizability that enhance translation into actual use), as well as internal validity. The purpose of a non-inferiority trial is to compare an intervention to an active control or standard treatment when the intervention is not expected to have superior efficacy, but to have other benefits, e.g., greater convenience or fewer side effects. In the case of genetic counseling, in- person counseling is the standard of care, with well-documented efficacy.4 Since video and telephone counseling do not offer a more personalized approach or more pertinent content than in-person counseling, it seems unlikely that either mode would produce superior psychosocial outcomes or greater knowledge gains; similarly, investigators do not expect telephone counseling to be more efficacious than video counseling.

The gold standard for assessing the effectiveness of interventions is the randomized clinical trial; yet patients who have a strong preference for one of the intervention conditions may decline to participate, which is a threat to external validity, or (if randomized) participate half-heartedly or drop out, threatening internal validity. Investigators will address this issue with a partially randomized preference trial in which patients with a strong preference are assigned to their preferred treatment (i.e., intervention condition) and those without a strong preference are randomized. This study design enables comparison of treatment outcomes among patients who receive their preferred treatment (the desired real world situation) and ascertainment of the effects of preference-as well as the evaluation of treatment outcomes in a randomized trial. However, comparisons involving preference participants are subject to confounding, since patients who prefer a particular treatment may differ in ways that affect outcome. In analyses involving preference participants it is possible to reduce confounding substantially using covariate adjustment. Nevertheless, because residual confounding may be present, preference participant outcomes should be considered observational data.

Investigators will blend qualitative and quantitative, deductive and inductive methods using varied forms of data from in-depth interviews, surveys, and audio taped observations. This will allow us to address our central question from different perspectives triangulated in the analysis for a rich understanding of patient-counselor- institution relationships most of which are too complex to describe using one dimension alone (e.g., cognitive understanding captured in surveys). While the randomized trial is the gold standard for comparative effectiveness, it cannot answer questions such as what it is about the counseling interaction that was reassuring or anxiety-provoking to a patient? what techniques enabled a patient to recall important points? or what about the conversation precluded such recall? It is only by embedding the survey data in open-ended inductive exploration and audio observations that investigators can determine if a counselor provided too much information to a low-literacy patient, or alternatively used plain language to emphasize key points, checking frequently for patient comprehension. Importantly, was the counselor able to do this as well by phone or video? Thus, mixed methods illuminates important dynamics that informants may not be consciously aware of.