Kevin McKenna

How Many Focus Groups Are Enough? Building an Evidence Base for Nonprobability Sample Sizes:

Few empirical studies exist to guide researchers in determining the number of focus groups necessary for a research study. The analyses described here provide foundational evidence to help researchers in this regard. We conducted a thematic analysis of 40 focus groups on health-seeking behaviors of African American men in Durham, North Carolina. Our analyses revealed that more than 80% of all themes were discoverable within two to three focus groups, and 90% were discoverable within three to six focus groups. Three focus groups were also enough to identify all of the most prevalent themes within the data set. These empirically based findings suggest focus group sample sizes that differ from many of the “rule of thumb” recommendations in the existing literature. We discuss the relative generalizability of our findings to other study contexts, and we highlight some methodological questions about adequate sample sizes for focus group research.

The science of being a study participant: FEM-PrEP participants' explanations for overreporting adherence to the study pills and for the whereabouts of unused pills.

Background:FEM-PrEP was unable to determine whether once-daily, oral emtricitabine/tenofovir disoproxil fumarate reduces the risk of HIV acquisition among women because of low adherence. Self-reported adherence was high, and pill-count data suggested good adherence. Yet, drug concentrations revealed limited pill use. We conducted a follow-up study with former participants in Bondo, Kenya, and Pretoria, South Africa, to understand factors that had influenced overreporting of adherence and to learn the whereabouts of unused pills. Methods:Qualitative, semistructured interviews were conducted with 88 participants, and quantitative, audio computer-assisted self-interviews were conducted with 224 participants. We used thematic analysis and descriptive statistics to analyze the qualitative and quantitative data, respectively. Results:In audio computer-assisted self-interviews, 31% (n = 70) said they had overreported adherence; the main reason was the belief that nonadherence would result in trial termination (69%, n = 48). A considerable percentage (35%, n = 78) acknowledged discarding unused pills. Few acknowledged giving their pills to someone else (4%, n = 10), and even fewer acknowledged giving them to someone with HIV (2%, n = 5). Many participants in the semistructured interviews said other participants had counted and removed pills from their bottles to appear adherent. Conclusions:Despite repeated messages that nonadherence would not upset staff, participants acknowledged several perceived negative consequences of reporting nonadherence, which made it difficult to report accurately. Uneasiness continued in the follow-up study, as many said they had not overreported during the trial. Efforts to improve self-reported measures should include identifying alternative methods for creating supportive environments that allow participants to feel comfortable reporting actual adherence.

A descriptive analysis of perceptions of HIV risk and worry about acquiring HIV among FEM-PrEP participants who seroconverted in Bondo Kenya and Pretoria South Africa.

Risk perception is a core construct in many behaviour change theories in public health. Individuals who believe they are at risk of acquiring an illness may be more likely to engage in behaviours to reduce that risk; those who do not feel at risk may be unlikely to engage in risk reduction behaviours. Among participants who seroconverted in two FEM‐PrEP sites – Bondo, Kenya, and Pretoria, South Africa – we explored perceived HIV risk and worry about acquiring HIV prior to HIV infection.

Evaluating Bang for the Buck A Cost-Effectiveness Comparison Between Individual Interviews and Focus Groups Based on Thematic Saturation Levels

Evaluators often use qualitative research methods, yet there is little evidence on the comparative cost-effectiveness of the two most commonly employed qualitative methods—in-depth interviews (IDIs) and focus groups (FGs). We performed an inductive thematic analysis of data from 40 IDIs and 40 FGs on the health-seeking behaviors of African American men (N = 350) in Durham, North Carolina. We used a bootstrap simulation to generate 10,000 random samples from each data set and calculated the number of data collection events necessary to reach different levels of thematic saturation. The median number of data collection events required to reach 80% and 90% saturation was 8 and 16, respectively, for IDIs and 3 and 5 for FGs. Interviews took longer but were more cost-effective at both levels. At the median, IDIs cost 20–36% less to reach thematic saturation. Evaluators can consider these empirically based cost-effectiveness data when selecting a qualitative data collection method.

Comparing focus groups and individual interviews: findings from a randomized study

Abstract Qualitative researchers often have to decide whether to collect data using focus groups or individual interviews. We systematically compare these two methods on their ability to generate two types of information: unique items in a brainstorming task and personally sensitive disclosures. Our study sample consisted of 350 African-American men living in Durham, North Carolina. Participants were randomized into either a focus group arm or individual interview arm, and were asked the same open-ended questions about their health-care seeking behavior. For the item-generating task, we compared data at two levels of analysis – the event and the individual. At the event level, focus groups and individual interviews generated similar numbers of unique items in a free-listing task. When compared on a per-person basis, however, individual interviews were more effective at generating a broad range of items. We also compared the number of data collection events in which personal, sensitive information was disclosed. We found that several types of sensitive and personal disclosures were more likely in a focus group setting, and that some sensitive themes only occurred in the focus group context. No sensitive themes emerged exclusively, or more often in, an individual interview context. Researchers may use these findings to help align their choice of qualitative data collection method with research objectives.

Preferences for a potential longer-acting injectable contraceptive: perspectives from women, providers, and policy makers in Kenya and Rwanda

High effectiveness, predictable return to fertility, and a single, prepackaged, disposable delivery system ranked high. Side effects were generally acceptable to women if they did not last long or disrupt daily activities. Cost was considered important for providers but not so much for most potential users. High effectiveness, predictable return to fertility, and a single, prepackaged, disposable delivery system ranked high. Side effects were generally acceptable to women if they did not last long or disrupt daily activities. Cost was considered important for providers but not so much for most potential users. Background: Between 1995 and 2005, injectable use doubled worldwide. However, discontinuation rates remain high, partly because of side effects but also because of missed appointments for reinjection. A longer-acting injectable (LAI) may improve compliance by reducing the required number of reinjection visits, thereby reducing unintentional discontinuation. This study examined acceptability of LAI characteristics comprising the target product profile (TPP). Methods: In 2012, we conducted qualitative case studies in Kenya and Rwanda, consisting of 19 focus group discussions (FGDs) with 177 current, previous, or never users of injectables and 46 in-depth interviews (IDIs) with providers, program implementers, and policy makers. FGDs and IDIs assessed current injectable experiences; attitudes toward potential LAI products; and perceptions of TPP attributes, including ranking preferences for the most and least important characteristics. In addition, we obtained completed electronic surveys from 28 international family planning opinion leaders about the perceived need for an LAI, important product characteristics, and challenges to LAI development or introduction. Results: Many FGD participants and interviewees spontaneously expressed strong interest in an LAI, but there was some variation in TPP preferences. The majority of participants ranked effectiveness as the most important TPP attribute. Providers were generally more concerned about side effects than potential users; some potential users suggested that side effects were related less to the product than to their own body chemistry and that side effects were acceptable as long as they did not last a long time or disrupt daily activities. Women and providers, especially in Kenya, preferred a method with a predictable return to fertility. Some participants associated amenorrhea with delayed or reduced fertility. Most women and providers preferred delivery of the LAI in a single, prepackaged, disposable injection system to facilitate injections by providers and to reduce the risk of pain or discomfort for women. While providers and policy makers ranked cost as one of the most important issues, it was among the least important issues for most potential users. Many Kenyan, but few Rwandan, participants appeared willing to pay for an LAI, with some presuming cost savings from reduced menstruation and fewer clinic visits. Conclusions: Some TPP preferences for an LAI have implications for product development decisions about formulation, delivery mechanism, or presentation, while others point to the need for tailored communication and counseling approaches to ensure acceptability and adherence within clinical trials and beyond.

HPTN 062: A Feasibility and Acceptability Pilot Intervention to Reduce HIV Transmission Risk Behaviors Among Individuals with Acute and Early HIV Infection in Lilongwe, Malawi

Acute HIV infection (AHI) is a relatively brief period of time when individuals are highly infectious and the opportunity to intervene to prevent forward transmission is extremely limited. HPTN 062 partnered with CHAVI 001 to evaluate the feasibility and acceptability of a motivational interviewing (MI)-based counseling intervention to reduce HIV-transmission risk behaviors among individuals with acute and early HIV infection in Lilongwe, Malawi. Participants were randomized to receive either (1) brief education sessions about HIV and AHI; or (2) the same brief education sessions plus an MI-based counseling intervention called Uphungu Wanga. Although Uphungu Wanga was determined to be feasible and acceptable, few major differences existed between the two arms with regard to acceptability, feasibility, and self-reported sexual behaviors. We therefore conclude that an additional MI-based counseling intervention may not be needed during the short period of AHI. Instead, we recommend that individuals with AHI receive frequent, but brief, counseling immediately after diagnosis and then transition to receiving counseling at less frequent intervals until they can initiate antiretroviral therapy. Other recommendations are provided.

HPTN 062: A Pilot Randomized Controlled Trial Exploring the Effect of a Motivational-Interviewing Intervention on Sexual Behavior among Individuals with Acute HIV Infection in Lilongwe, Malawi

Objective We pilot tested a Motivational Interviewing (MI) –based counseling intervention for individuals with Acute HIV Infection (AHI) to reduce risky sexual behavior in Lilongwe, Malawi. Methods Twenty-eight individuals diagnosed with AHI were randomized to receive either brief education alone, or the brief education plus the MI-based intervention, called Uphungu Wanga. Participants in Uphungu Wanga received four sessions delivered on the day of diagnosis, three days later and at weeks 1 and 2 with a booster session at week 8; participants were followed for 24 weeks from diagnosis. An interviewer administered quantitative questionnaire was conducted at baseline and at weeks 2, 4, 8, 12, 16, 20 and 24. Semi-structured qualitative interviews (SSI) were conducted at weeks 2, 8, 12, and 24. Results The majority of participants in both arms reported rapid and sustained behavior change following diagnosis with AHI. Very few participants reported having sex without a condom after diagnosis. Participants reported a trend towards fewer sex partners and abstaining from sex during study follow-up. Participants in the MI-based arm provided concrete examples of risk reduction strategies in the SSIs while those in the brief education arm primarily described reducing risk behavior, suggesting that the MI-based group may have acquired more risk reduction skills. Conclusions Individuals in both study arms reduced risky sexual behaviors after diagnosis with AHI. We found few major differences between study arms during the 6-month follow up period in self-reported sexual behaviors therefore a MI-based intervention may not be needed to trigger behavior change following AHI. However, comparing the MI-based intervention to repeated brief education sessions made it difficult to assess the potential benefit of an MI-based intervention in a setting where standard counseling often consists of one post-test session. Nevertheless, provision of counseling immediately following diagnosis with HIV to support behavior change should remain a priority. Trial Registration ClinicalTrials.gov NCT01197027

Policy and programmatic considerations for introducing a longer-acting injectable contraceptive: perspectives of stakeholders from Kenya and Rwanda

Unique attributes of a longer-acting injectable would likely appeal to both existing injectable users and new clients, both for spacing and for limiting births, and allow health systems to operate more efficiently. Considerations for enhancing successful introduction of this potential new method include keeping the cost low, expanding access through community-based distribution, and training providers to improve practices about injectables in general. Unique attributes of a longer-acting injectable would likely appeal to both existing injectable users and new clients, both for spacing and for limiting births, and allow health systems to operate more efficiently. Considerations for enhancing successful introduction of this potential new method include keeping the cost low, expanding access through community-based distribution, and training providers to improve practices about injectables in general. ABSTRACT Background: More than 40 million women use injectable contraceptives to prevent pregnancy, and most current or previous injectable users report being satisfied with the method. However, while women may find injectables acceptable, they may not always find them accessible due to stock-outs and difficulties with returning to the clinic for reinjections. FHI 360 is spearheading efforts to develop a longer-acting injectable (LAI) contraceptive that could provide at least 6 months of protection against pregnancy. This article addresses systems-level considerations for the introduction of a new LAI. Methods: We conducted qualitative case studies in Kenya and Rwanda—two countries that have high levels of injectable use but with different service delivery contexts. Between June and September 2012, we conducted in-depth interviews with 27 service providers and 19 policy makers and program implementers focusing on 4 themes: systems-level barriers and facilitators to delivering LAI services; process for introducing an LAI; LAI distribution approaches; and potential LAI characteristics. We also obtained electronic feedback from 28 international family planning opinion leaders. Results: Respondents indicated strong interest in an LAI and thought it would appeal to existing injectable users as well as new family planning clients, both for spacing and for limiting births. Providers appreciated the potential for a lighter workload due to fewer follow-up visits, but they were concerned that fewer visits would also decrease their ability to help women manage side effects. The providers also appreciated the 1-month grace period for follow-up LAI injections; some seemed unaware of the latest international guidance that had increased the grace period from 2 weeks to 4 weeks for the currently available 3-month injectable. The majority of policy makers and program implementers were supportive of letting community health workers provide the method, but many nurses and midwives in Kenya had reservations about the approach. At the policy level, respondents indicated that obtaining regulatory approvals before introducing the new method could be costly and time-consuming. Manufacturing and procurement decisions could also affect cost and availability. Conclusions: Successful introduction of a potential longer-acting injectable may be enhanced by considering broader systemic issues, including managing cost to the health system and users, expanding access through community-based distribution, and training providers on the latest service delivery guidelines.