Cohen, Jessica A. MHS*; Steele, Matthew S. PhD*; Ureña, Francisco I. Cáceres PhD†; Beksinska, Mags E. MSc‡
WOMEN NOW ACCOUNT FOR OVER 50% of all new HIV infections worldwide with approximately 7000 women becoming infected daily. The majority of these infections are occurring among women in developing countries. The Dominican Republic (DR) and South Africa (SA) reflect the epidemiologic trends visible in many parts of the world where HIV rates are rising among women, and heterosexual transmission is the primary mode of HIV infection. In the Dominican Republic, the HIV prevalence among adults is 1.7% compared with South Africa, where the adult prevalence is 21.5%.1,2
In this third decade of HIV/AIDS, male and female condoms are still the only methods available for HIV prevention among sexually active people. In many settings, however, gender inequalities, social norms, and economic disparities severely limit women’s ability to negotiate condom use with a partner. Microbicides, products which can substantially reduce the risk of HIV when applied vaginally or rectally, are considered an urgently needed option for women who are not able to negotiate condom use with their partners.
The need for acceptability research on microbicides has been widely acknowledged3–5 and is considered important for developing products that are both accessible and culturally appropriate for target populations. Among microbicide acceptability studies already conducted, several have explored user needs related to microbicide applicators among women in developing country settings.6–10 A review of applicator research in 2002, however, identified the issues of price, reuse, and perceived safety of microbicide applicators as warranting further research among women in low-resource settings.12 Although several applicator studies have noted users’ acceptance and/or preference for single-use applicators,7,8,10,12 these studies have not asked about preferences in the context of price. To date, prefilled, single-use applicators such as those being used in most clinical trials of microbicides will likely be more costly than user-filled, single-use or reusable applicators.13 Understanding women’s preferences, including the tradeoff between applicator price and applicator design (single-use vs. reusable), as well as the perceived safety associated with each design, will have implications for future delivery and acceptability of microbicides.
The primary objective of this study was to prioritize and characterize preferences for applicator attributes as determined by potential female microbicide users in the DR and SA. A secondary objective was to explore how women’s preferences differed based on demographic and perceived disease risk variables in each setting.
Materials and Methods
PATH conducted this study in collaboration with the Institute of Population and Development Studies (IPDS) at Profamilia in the Dominican Republic and the Reproductive Health and HIV Research Unit (RHRU) in Durban, South Africa between September and December 2003. The study protocol was reviewed and approved by the PATH Human Subjects Protection Committee and the Institutional Review Boards of both Profamilia and RHRU.
We administered a survey to 449 women attending 4 public clinics specializing in reproductive and sexual health in the DR (Profamilia clinics) and 446 women attending 3 primary health care clinics in the region of KwaZulu Natal, South Africa. Profamilia clinics charge set fees for services (depending on ability to pay), whereas the clinics in the SA sample provide services free of charge. We used random sampling techniques with the clinic waiting room attendees as our study population and a random number generator and predetermined intervals (10 or 20) to determine which attendees to approach for participation. In this study, clinic attendees were eligible to participate if they were 1) female, 2) 18 or older, and 3) willing and able to provide informed consent. Clinic attendees were chosen as the study population because they were considered to be likely users of a future microbicide (i.e., women currently seeking care or prevention methods for pregnancy and/or reproductive health issues). In the DR, women visiting the clinic for the sole purpose of purchasing contraceptives, and who were therefore not spending significant time in the waiting room, were excluded from the study sample. However, women visiting the clinic to begin a contraceptive method or for a contraceptive checkup visit were included in the study sample. Female interviewers conducted these surveys in Spanish and Zulu, respectively.
Although women in this study sample may not represent those likely to buy microbicides through the private-sector marketplace, they do represent a large sector of the population in developing countries who are likely to purchase microbicides through social marketing programs that offer family planning and HIV prevention services to the public sector at greatly reduced prices.
The acceptability survey included 3 main sections with questions on: 1) sociodemographic characteristics and HIV risk perception, 2) applicator attribute preferences, and 3) price/design tradeoffs.
The main sociodemographic variables explored in our analysis included a categorical age variable (≤23, 24–29, and 30 years and above), educational attainment (<secondary school vs. secondary school or higher), urban or rural residence, current marital status, living children versus no living children, and income (a median split variable ≤160 rand and above in SA and ≤2000 pesos and above in the DR). We also asked participants about their perceptions of HIV risk (belief that their risk of acquiring HIV was moderate or great vs. no risk) and concern about germs when reusing microbicide applicators.
In the applicator attribute section, we asked participants about the desirability of different prices, single-use design, reusable design, prefilled or user-filled design, plastic and paper material. We also explored issues related to applicator storage, portability, cleaning, disposal, and perceived safety as a function of design.
To characterize the importance of applicator design and price in relation to one another, we asked a series of choice-based, tradeoff questions using different combinations of design and price attribute levels (Table 1). Specifically, for each tradeoff question, the respondent was asked the exact same question: “If these microbicide applicators were identical in all other ways, which would you prefer?” We then asked participants to give their preference between 2 pairs of variables with each pair representing a random combination of a design variable (single-use or reusable) and a price variable (low, medium, higher price). They were asked to give their preference for one of the pairs using a 9-point scale (ranging from “strongly prefer option A” to “strongly prefer option B”). Price levels were set equivalent to condom prices in each country with the lowest level set to the price of a socially marketed condom and the highest level set to the price of a pharmacy-bought condom.
At the beginning of each interview, the interviewer explained the concept of a microbicide, an applicator, and the different attributes of the applicator using pictorial illustrations and standardized descriptions. Interviewers also referred to the applicator illustrations for each applicator-related question in the survey. These illustrations were developed at PATH to provide a standardized representation of study terms across study sites and to facilitate respondent comprehension of each attribute. All study instruments, including the preference scale, were pretested in the DR and SA sites and revised before initiation of the study.
Descriptive statistics were used to characterize sociodemographic and applicator attribute data. Adaptive conjoint analysis was used to prioritize preferences for applicator attributes from the choice-based questions, and logistic regression was used to identify associations between applicator preferences and sociodemographic variables.
Conjoint analysis theory is founded on the assumptions that individuals make tradeoffs within resource constraints based on valuation and need, and that all decision-making processes are based on the weighing of specific product attributes against one another. By analyzing these tradeoffs, one can calculate the relative value (and thereby importance) of each product attribute in relation to one another.14
The sample used for the conjoint analysis includes all participants for which the choice-based tradeoff questions were complete (DR n = 449, SA n = 446). For this part of the analysis, conjoint utilities were scaled to an arbitrary additive constant within each attribute. Utilities were scaled to sum zero within each attribute and then averaged for mean utility scores. Higher utility scores demonstrate greater preferences. For a thorough review of conjoint analysis and other public preference elicitation techniques, see Ryan et al.14
Study participants’ data were included in the logistic regression analysis if they provided complete responses for all sociodemographic and applicator attribute questions. (DR n = 396, SA n = 443). We used preference for single-use applicator as a dependent variable and demographic and risk perception factors as independent variables in a series of bivariate and multivariate logistic regression models. After evaluating bivariate associations (data not shown here), we entered all demographic and risk perception variables into multivariate logistic regression models. We performed backward stepwise elimination removing variables from the model one at a time until reaching a final model containing variables associated (at the P ≤0.1 level) with the outcome of interest.
To understand the possible effect on study results from removing the 53 incomplete data points from the DR sample, we reran the analyses with the full dataset and the reduced sample. There were no significant differences in demographic characteristics, univariate results, or results for the multivariate models.
Description of Participants From the Dominican Republic and South Africa
The majority of participants in both the DR and SA were from urban settings, had secondary or higher education, had between one and 2 children, and perceived themselves to be at moderate to great risk of HIV infection (Table 2). The median ages were 27 and 24, and the median monthly incomes were US $72 and US $26, respectively, in the DR and SA study populations. In both settings, the majority of women were either single or not living with their partners.
Applicator Attribute Preferences
Prefilled applicators were preferred in SA for storage and portability, whereas in the DR, both prefilled and user-filled designs were considered acceptable. Plastic was preferred to paper as applicator material in both sites. For disposal of applicators, 96% participants in SA considered burning to be somewhat to very desirable, whereas in the DR, burning and throwing in a garbage can were both considered desirable options (75% and 82%, respectively). When asked about the number of times it would be important to be able to wash and reuse a reusable applicator, over half of participants in both sites considered it somewhat to very important to be able to wash and reuse an applicator 5 times (58% in DR, 64% in SA), whereas reuse 20 or more times was considered less important in both sites. For cleaning of reusable applicators, participants in both sites considered boiling and cleaning with disinfectant as the most desirable cleaning methods.
Conjoint Analysis Findings: Preference for Single-Use Applicators
Applicator design was more important than price in the DR (Fig. 1). In SA, these 2 attributes were of almost equal importance to participants. Mean utility ratios reflecting the relative importance of each attribute are shown in Table 3. In both populations, single use was the most valued attribute and reusable design and low price were the least valued attributes. Similarly, with each combination of attributes, single use and high price was the most valued combination, whereas reusable and low price was the least valued combination in both populations (Table 3). Although the order of attribute preference is the same between sites, there is a notable difference between DR and SA in the relative strength of their preference for reusable applicators. In the DR, there is very little value attributed to reusable applicator designs (at any price), whereas in SA, the value associated with reusable (at high price) and single use (at low price) is almost equal.
Predictors of Desirability for Single-Use Design
Several sociodemographic and risk perception variables were associated with the belief that it was somewhat to very important for microbicide applicators to be single use (Table 4). In both the DR and SA, women who were somewhat to very concerned that reusable applicators would spread germs were approximately 3 to 10 times more likely to consider single-use applicators important when compared with women who were not concerned about spreading germs with reusable applicators. Additionally, in the DR, secondary or higher education and older age (>30) were associated with single-use applicator preference, whereas in SA, women considering themselves at moderate or high risk of acquiring HIV having secondary or higher education and having one or more children were approximately 2 to 3 times more likely than their counterparts to prefer a single-use applicator. In the DR, women who considered themselves at risk for HIV were less likely to consider single-use applicators important.
Microbicides, products which can substantially reduce the risk of HIV when applied vaginally or rectally, represent an important new HIV prevention option for women. Because microbicides are still in development, acceptability research that can inform the design of both microbicides and their applicators will have important implications for future acceptability and use of these products. We conducted this study to characterize preferences for applicator attributes among potential microbicide users in the DR and SA, and to explore how these preferences might differ based on demographic and disease risk variables in each setting.
In both the DR and SA, women valued single-use design above all else and, conversely, attributed the least value to a reusable design. In SA, however, there appeared to be a direct tradeoff between design and price. One hypothesis for this tradeoff is that single use and price are indirect indicators for safety and quality, whereby the perception of increased safety is attributed to single-use products and the perception of increased quality is attributed to higher-priced products. Within this framework, the equal valuation of single-use products at a low price with reusable products at a high price might be better understood. Comments noted during the interviews in SA illustrate these perceptions: “She prefers a single-use applicator like a condom—to avoid infecting others. Reusable is not really safe.” “The single-use applicator is safer than the reusable applicator because it will not spread germs. The single-use applicator should be cheaper than the reusable.” In SA, where there was more price sensitivity reflected in the mean utility scores and more value placed on the middle- and low-price variables, there were also numerous comments such as this one supporting the need for very inexpensive microbicide products: “Microbicide applicators should be [made] available soon, but it should be sold at low cost because many people are unemployed.” This greater price sensitivity in SA might also reflect the economic status of the participants whose median monthly earnings were one third of those in the DR.
Prior research suggests that HIV-positive women and women concerned about cleaning reusable applicators might prefer single-use applicators.6,7 Data from this study support these findings and indicate an association in both populations between the perceived risk of spreading germs with a reusable applicator as well as the perceived risk of HIV acquisition (SA only) with the preference for single-use applicators. Although no studies have substantiated the concern that a used vaginal applicator could infect another person with HIV or a sexually transmitted infection through handling or use, this is a widely held sentiment in both study populations. In the DR and SA, 91% and 93% of participants were somewhat to very concerned about spreading germs to someone else with a reusable applicator. The high HIV prevalence rates in both areas where the study was conducted, especially in KwaZulu Natal province where the HIV prevalence is 40.7%15 may help explain women’s concern with spreading germs or infections.
In SA, the association between having children and preference for single-use design may also be linked to the perception noted previously, namely the fear that children could become exposed to germs through handling of a used applicator. The concern that children could find or play with reusable applicators has previously been expressed by women participating in a microbicide study in India and Thailand.6 In this same study, 31% of women sampled from Malawi, Zimbabwe, India, and Thailand reported that they disliked having to clean and reuse the reusable applicator.
In the DR, women who perceived themselves at moderate to high risk for HIV were less likely to prefer a single-use applicator than those women who perceived themselves at low risk for HIV. This association is the opposite of that found in SA. There are no clear explanations for this association given that the majority of this population preferred the single-use design and had concerns about spreading germs through reusable applicators. There were only 2 comments captured by interviewers that might illustrate the sentiment that women felt more confident with a reusable applicator because they could wash it themselves. One interviewer noted: “The most important thing to her is hygiene. She prefers to disinfect it herself.” Another noted: “She prefers the reusable because that way she can have control of her hygiene.”
Although reusable applicators will likely be one of the least expensive methods for microbicide delivery and may be very acceptable to certain segments of the population, single-use applicators appear to be a much more acceptable and the preferred delivery method for women with concerns about disease transmission. Any wide-scale introduction of reusable applicators with microbicides should address concerns about reuse and infection and should include information on proper cleaning methods.
In addition to the associations linked to the preference for single-use applicators, findings from this study suggest that both manually filled and prefilled applicators are acceptable for microbicide portability and storage. Findings also suggest that devices should be made of clean-burning materials, whether plastic or paper, because burning was named as a primary method of disposal. If reusable applicators are to be used with microbicides, they should be of appropriate design and materials to allow for either boiling or cleaning with disinfectant for at least 5 reuses.
Conjoint analysis survey designs have not often been applied in low-literacy settings. Although we developed and pretested the survey, illustrations, and visual scales, we did not have the benefit of using tools that had been validated through prior research. Also, although conjoint utility measures provide information on the priority and strength of attribute preferences, they may not be indicative of actual buying patterns when individuals are making real purchasing decisions. Although both populations placed more value on the higher-priced products, this may reflect the desire for high-quality products. There should still be an important emphasis on ensuring that microbicide products are as inexpensive as possible to ensure access among the widest range of users. Lastly, because this study did not include applicator use, it will be important for future studies to assess whether these acceptability trends, as well as product adherence, differ when linked to applicator and microbicide use within the context of sexual relations.
In summary, because the first microbicides are now entering phase III clinical trials, any changes to applicator products (changes in design or use of alternative applicators) that could enhance user acceptability and use effectiveness must be considered promptly to prevent delay in wide-scale introduction and use of first-generation microbicide products. This study clearly indicates that single-use applicators are preferable to reusable applicators in both study populations and that concern about spreading germs with reusable devices is an important factor in this decision process. Acceptability and usability issues related to microbicides and their delivery mechanisms should continue to be evaluated among potential microbicide users so that they can directly inform the process of product development and introduction of microbicides.