You are here

Article Text

Article menu
Behaviour
Original article
Cessation of intravaginal practices to prevent bacterial vaginosis: a pilot intervention in Zimbabwean women
  1. Allahna Esber1,
  2. Precious Moyo2,
  3. Marshall Munjoma2,
  4. Shelley Francis3,
  5. Janneke van de Wijgert4,
  6. Tsungai Chipato2,5,
  7. Abigail Norris Turner6
  1. 1Division of Epidemiology, The Ohio State University, Columbus, Ohio, USA
  2. 2UZ-UCSF Collaborative Research Programme, Harare, Zimbabwe
  3. 3School of Health Sciences, College of Health Sciences, Walden University
  4. 4Institute of Infection and Global Health, University of Liverpool, Liverpool, UK
  5. 5Department of Obstetrics and Gynaecology, University of Zimbabwe, Harare, Zimbabwe
  6. 6Division of Infectious Diseases, Department of Internal Medicine, The Ohio State University, Columbus, Ohio, USA
  1. Correspondence to Dr Abigail Norris Turner, Division of Infectious Diseases, Department of Internal Medicine, The Ohio State University, 410 West 10th Ave, Columbus, OH 43210, USA; ant{at}osumc.edu

Abstract

Objectives Intravaginal practices—including behaviours such as washing with soap or other materials, using fingers or cloth, or insertion of herbs, powders or other products to dry, cleanse or ‘tighten’ the vagina—may increase women's risk of bacterial vaginosis by disrupting the vaginal microbiota. In Zimbabwe, intravaginal practices are common. The objective of this study was to assess the feasibility of an intervention based on the transtheoretical model of behaviour change (also called the ‘stages of change’ model) to encourage cessation of vaginal practices among a sample of Zimbabwean women.

Methods We conducted a 12-week behaviour change intervention to encourage cessation of intravaginal practices (other than cleansing with water) among 85 Zimbabwean women who reported these practices.

Results Self-reported intravaginal practices declined significantly over follow-up, with 100% of women reporting at least one intravaginal practice at enrolment compared with 8% at the final visit. However, we found no significant effect of this reduction on bacterial vaginosis prevalence in unadjusted or adjusted multivariable models (adjusted prevalence ratio for any practice vs none: 0.94, 95% CI 0.61 to 1.43).

Conclusions While the intervention was successful in reducing women's self-reported engagement in intravaginal practices, we observed no corresponding benefit to vaginal health.

  • INTERVENTION STUDIES
  • AFRICA
  • BACTERIAL VAGINOSIS

https://doi.org/10.1136/sextrans-2014-051764

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Worldwide, bacterial vaginosis (BV) is the most common abnormal vaginal condition among women of childbearing age. This dysbiosis (microbial imbalance) is associated with increased risks of spontaneous abortion, preterm birth and HIV acquisition and transmission.1–5

    Intravaginal practices—including internal vaginal washing with soap or other materials, using fingers or cloth, flushing out the vagina by inserting pressurised liquid (douching), or insertion of herbs, powders or other products to dry, cleanse or ‘tighten’ the vagina—have been shown to increase BV risk.6–9 A 2010 meta-analysis of six studies among African women found that those reporting any intravaginal practice had 20% increased risk of incident BV.3 Another study among Zimbabwean and Ugandan women reported that insertion of products to dry or tighten the vagina was associated with significantly increased BV risk.9 In contrast, studies have generally confirmed that cleansing with water alone does not increase BV risk.2

    Even though women may be told by healthcare providers that intravaginal practices may be harmful, many women choose to continue these behaviours for a range of cultural reasons related to personal hygiene, genital health and to increase sexual pleasure.3 ,10 In some communities, intravaginal cleansing is culturally normative and habitual.11 ,12 A cross-cultural study of vaginal practices found that women's decisions to engage in intravaginal practices are multifaceted and culturally dependent.13 For example in some rural African settings, women express fear that as they age, their partners will leave them for younger women. These women report using intravaginal practices to reduce partner infidelity and preserve their relationships.13 Intravaginal practices are also used to change the odour, appearance or size of the vagina.13 Previous work in Zimbabwe found that extensive intravaginal washing, drying and tightening behaviours are common.9 ,14

    Several interventions have been developed to encourage the cessation of intravaginal practices, including some projects that incorporate a theory-based behaviour change model.8 ,15–17 One of the most commonly used models in public health interventions is the transtheoretical model (TTM),18 which posits that behaviour change occurs through a series of five stages, starting with ‘precontemplation’ and progressing through ‘contemplation,’ ‘preparation,’ ‘action’ and concluding with ‘maintenance’ of the behaviour change.19 A small study in Kenya among female sex workers used a combination of information motivation behavioural skills, TTM and harm reduction models to encourage cessation of vaginal practices.15 An information motivation behavioural skills model was also used successfully in a pilot intervention among HIV-positive women in Zambia.17 To our knowledge, no other vaginal practice cessation intervention has been developed or tested in general population African women (eg, not specialised groups such as sex workers or HIV-positive women).

    Methods

    The primary objective of this prospective pilot study among urban women in Harare, Zimbabwe, was to assess the feasibility and preliminary effectiveness of a TTM-based behaviour change intervention on cessation of intravaginal practices (other than cleansing with water) to reduce BV prevalence.

    Design

    Women attended an enrolment visit and three follow-up visits at 4-week intervals, for a total follow-up period of 12 weeks per participant.

    Study setting and population

    We recruited women from two family planning clinics: the Harare Central Hospital Family Planning Unit and the Zimbabwe National Family Planning Council's Spilhaus Family Planning Centre. Staff explained the study to women waiting for family planning services. Eligibility was assessed through a face-to-face questionnaire. Eligible women spoke Shona or English, were at least 18 years of age, lived within 1 h's journey of the study site, were not pregnant or planning pregnancy in the next 6 months, and reported engaging in at least one of nine intravaginal practices in the past 3 months: intravaginal cleansing, drying or tightening with soap; household cleaners; antiseptic solutions; vinegar, lemon or lemon juice; douching; inserting cloth, tissue, paper, newspaper or cotton wool; inserting alum or other powder, herbs, sticks, stones, crushed rock, leaves, soil or other vaginal product from a traditional healer; or swallowing products to clean, dry or tighten the vagina. The list of practices was developed from our own prior research as well as others’.3 ,14 During the screening interview, using an anatomical model to increase clarity, women were asked about practices which involved internal cleansing of the vagina and not simply cleaning of the external genitalia.

    Data collection

    After the participant provided informed consent, an interviewer administered the baseline questionnaire in a private room (see online supplementary appendix, baseline questionnaire). The survey captured demographic, reproductive and sexual behaviour, and motivations, frequency, timing, and stage of behaviour change with respect to each intravaginal practice of interest. At each follow-up visit, women were again interviewed to capture recent sexual and reproductive behaviour that could influence intravaginal practices or BV (see online supplementary appendix, follow-up questionnaire).

    At each visit participants also self-collected a vaginal sample using a sterile Dacron swab. Self-collection took place in a clinic examination room. The swab was immediately smeared onto a glass slide. Gram staining and Nugent scoring for BV detection occurred according to standard protocols at the end of the study.20 Two trained laboratory technicians separately scored each stained slide; discrepant findings were adjudicated by the laboratory director. As Nugent scoring was not performed until all participants had completed follow-up, treatment was not provided to BV-positive women. Women with symptoms suggestive of BV were encouraged to seek clinical care, but treatment was not given as part of the study.

    At enrolment and each follow-up visit, women were asked a series of questions related to each of the nine intravaginal practices, to assess stage of behaviour change for each behaviour: precontemplation, contemplation, preparation, action or maintenance.19 A woman who did not intend to stop in the next 6 months was classified in the ‘precontemplation’ stage. Women who intended to stop within the next 6 months were in the ‘contemplation’ stage; those intending to stop within the next 30 days were in the ‘preparation’ stage. Women in the ‘action’ stage had done the specific practice less than 6 months ago, but not in the last 30 days, and women who had stopped the practice more than 6 months prior to the interview were in the ‘maintenance’ stage.

    Intervention

    After the interview, at enrolment and each follow-up visit, a counsellor (a different person than the interviewer) delivered a 15-min semiscripted counselling message on cessation of intravaginal practices targeted to the participant's stage. For example, for women in contemplation phase, counselling topics included a discussion of the risks associated with intravaginal practices, the challenges and benefits of changing behaviour, and a discussion of a woman's perceptions of her male partner's views on intravaginal practices. (During all counselling sessions, cleansing with water was portrayed as a safe alternative to more abrasive practices). We assessed stage of change for all practices individually, but counselling was directed at the earliest stage relevant to the specific participant.

    We had conducted qualitative research with women and men prior to initiating the study, which demonstrated that male partners were overwhelmingly in favour of women's continued engagement in intravaginal practices and also that women feared disruption of their relationships if they stopped these behaviours (unpublished findings). This qualitative work also revealed that men were considerably more supportive of cessation of intravaginal practices after the associated risks were explained. As a result of this formative research, negotiation tactics for communicating with male partners was included in all counselling sessions, including role playing of discussions with male partners and how to respond to partner concern about cessation of intravaginal practices.

    Statistical analysis

    All data were double-entered into Microsoft Access. Data were converted to Stata 11.2 for analysis (College Station, Texas, USA). The analysis data set included time-varying data on vaginal practices, BV and other sexual and reproductive health variables. We did not perform power calculations when selecting the target sample size for this pilot study.21

    Our primary outcome was prevalent BV at each follow-up visit. Any woman who had a Nugent score of 7–10 was classified as BV-positive.20 At each visit, we asked participants about their intravaginal practices over the preceding 4-week interval. We first assessed the effect of each practice on BV prevalence. We also examined BV prevalence comparing women who reported any intravaginal practice with those who reported none. We examined the prevalence of BV comparing women who reported no abrasive intravaginal practices during the preceding period with those reporting one or two or more practices. Finally, we examined trends in stage of behaviour change by visit.

    Multivariable models

    To examine the effect of intravaginal practices on BV prevalence, we ran Poisson regression models with generalised estimating equations and an unstructured correlation matrix to account for within-woman clustering. (We used Poisson regression because the prevalence of BV was higher than 10%, the cut-off value above which logistic regression is not recommended.22) We computed unadjusted prevalence ratios (PRs) and 95% CIs and then ran two adjusted models. First, we adjusted only for BV status at the previous visit. Second, we adjusted for BV status at the previous visit, self-reported HIV status, and hormonal contraceptive use, as each of these may impact the development and persistence of BV. Missing data were not imputed, rather, complete-case analysis was performed.

    Sensitivity analysis

    Because molecular studies in the last decade have shown that an ‘intermediate’ vaginal microbiome—Nugent score of 4–6—is more similar to BV than to a normal microbiome,23 we conducted a sensitivity analysis of the robustness of our primary findings when the outcome was coded as Nugent score of 4–10 (intermediate plus BV) vs Nugent score 0–3 (normal). We ran all unadjusted and adjusted models again using this more expansive outcome definition.

    Ethical approval

    All participants provided written informed consent. The project received ethical approval from the Ohio State University Institutional Review Board (protocol 2012H0091), the Joint Parirenyatwa Hospital and College of Health Sciences Research Ethics Committee of the University of Zimbabwe (protocol JREC/47/11), the Medical Research Council of Zimbabwe (MRCZ/A/1612), and the Research Council of Zimbabwe.

    Results

    The study was conducted between June and September 2012. Only women who self-reported intravaginal practices were screened for inclusion. Eighty-five women enrolled in the study. The intervention proved feasible, with high retention: 82 women (96%) attended at least two visits, 80 (94%) attended at least three visits and 78 (92%) attended all four visits. Participants’ median age was 33 years (IQR: 27–40 years). Nearly all had a primary partner (97%) and reported a prior pregnancy (95%), and a majority had used hormonal contraception in the past 3 months (69%) (table 1). At enrolment, 47% of women reported vaginal itching, more than normal vaginal discharge, or malodorous or fishy discharge. By visit 4, only 9% reported having any of these symptoms. Approximately a quarter of participants (27%) were HIV-positive by self-report.

    View this table:
    Table 1

    Demographic and behavioural characteristics of participants (n=85)*

    By design, all women (100%) reported at least one intravaginal practice at enrolment. The most common practices were inserting cloth, tissue paper, newspaper or cotton wool (89%) and cleaning, drying or tightening the vagina using vinegar, lemon or lemon juice (79%). Use of household cleaners (2%) and douching (21%) were the least commonly reported practices at enrolment. Over the follow-up period, women reported fewer practices, with most participants ceasing all intravaginal practices by the final visit (figure 1). By visit 4, only 8% reported any practice. BV prevalence remained stable across visits, with 42% of women BV-positive at enrolment, 43% at visit 2, 46% at visit 3 and 44% at visit 4 (figure 1).

    Figure 1
    Figure 1

    Prevalence of intravaginal practices and bacterial vaginosis (BV), by visit. Douche=flushing out the inside of the vagina by inserting pressurised liquid; Cloth=cloth, tissue, paper, newspaper or cotton wool; Alum=alum or other powder, herbs, sticks, stones, crushed rocks, leaves, soil or other vaginal product from a traditional healer or herbalist; Swallow= swallow any product by mouth. 1BV is defined as Nugent score 7–10.

    In unadjusted models examining each individual practice and BV, we found no significant associations (figure 2). We also observed no association between report of any intravaginal practice and BV (PR comparing women reporting any practice with those reporting none: 0.94, 95% CI 0.63 to 1.39) (table 2). Measures of effect were unchanged after adjusting for only BV status at last visit (PR: 0.94, 95% CI 0.62 to 1.44) and after adjusting for BV status at last visit, hormonal contraceptive use and self-reported HIV status (PR: 0.94, 95% CI 0.61 to 1.43). Findings were very similar when we compared women reporting none, one, and two or more intravaginal practices (table 2). In the sensitivity analysis using the expanded outcome definition, we also observed no significant effect of intravaginal practices on BV (PR 1.05, 95% CI 0.78 to 1.42).

    View this table:
    Table 2

    Unadjusted and adjusted associations between intravaginal practices and BV prevalence

    Figure 2
    Figure 2

    Unadjusted prevalence ratios for BV by vaginal practice. Douche=flushing out the inside of the vagina by inserting pressurised liquid; Cloth=cloth, tissue, paper, newspaper or cotton wool; Alum=alum or other powder, herbs, sticks, stones, crushed rocks, leaves, soil or other vaginal product from a traditional healer or herbalist; Swallow=swallowed any product by mouth. 1BV is defined as Nugent score 7–10. 2Vaginal practice use was allowed to vary by visit in the model.

    Behaviour change

    We next examined women's progression through the stages of change for each practice. We describe two examples here, representing an intravaginal practice that was commonly reported at enrolment (insertion of cloth, tissue paper, newspaper or cotton wool) and a practice much less common at enrolment (intravaginal use of antiseptics). At enrolment, only 5% of women were in the action stage for insertion of cloth, tissue paper, newspaper or cotton wool, and 86% were in the action stage by visit 3 (web appendix, figure 3). By visit 3, all women were at least in the preparation stage with the majority in the action stage. At enrolment, 32% of participants were using antiseptics intravaginally with 4% in the precontemplation phase, 3% in the contemplation phase and 25% in the preparation phase (web appendix, figure 4). Only 29% were in action phase at enrolment. By visit 2, 47% were in action phase and by visit 4 all women were in either the action phase (33%) or the maintenance phase for antiseptics (67%).

    Discussion

    In this study of urban Zimbabwean women, we demonstrated that an intravaginal practice cessation intervention targeted to stage of behaviour change was effective in reducing self-reported intravaginal practices. Yet while participants reduced their engagement in intravaginal practices, we did not observe a corresponding reduction in BV prevalence. A sensitivity analysis using a more expansive outcome definition also did not show a significant association between cessation of intravaginal practices and BV.

    Other interventions on intravaginal practices using a theoretical framework have been conducted among adolescents in the USA,16 HIV-positive women in Zambia17 and female sex workers in Kenya.15 These interventions ranged in size from 22 women15 to 275 women.16 The US-based study, a randomised controlled trial, had 15-min individualised counselling sessions at enrolment, 1 month and 3 months.16 The Zambia study had one 8-week follow-up visit after a 20–30 min individual counselling session at enrolment.17 The Kenya study had four weekly visits with structured group sessions.15

    Despite differences in length, intensity and setting, our findings largely align with previous intervention research on this topic. In the study of Kenyan sex workers, all women who completed the study (n=22, 96%) were in the action stage and no longer engaging in any intravaginal practice, yet the authors reported no change in odds of BV (OR: 1.00; 95% CI 0.42 to 2.38).15 In the study of HIV-positive Zambian women, the intervention reduced intravaginal practices by 22% in the intervention arm compared with only 6% in the control arm; BV was reduced by 59% in the intervention arm and 47% in the control arm.17 In the trial of young US women the intervention was successful in reducing douching (OR: 1.34; 95% CI 1.03 to 1.73), but BV was not assessed.16 In a small case cross-over study in the USA, 77% of women did not resume vaginal practices after a cessation intervention, however, this change did not significantly reduce BV prevalence (OR: 0.76, 95% CI 0.33 to 1.76).8 In contrast, non-intervention studies have generally reported that intravaginal practices are associated with increased BV,2 and that women who stop vaginal practices have healthier vaginal microbiota.7 ,9 ,15

    Several reasons may explain why we observed a profound reduction in self-reported intravaginal practices but no corresponding reduction in BV prevalence. First, while it has been shown that these practices can increase BV risk, the converse—that stopping intravaginal practices decreases BV risk—may not be true. Second, other factors also contribute to BV persistence. For example, women with BV biofilms have very persistent disease, and the effect of the biofilm may overwhelm the benefit of stopping intravaginal practices, particularly over the short time frame of this study.24 Third, although participants reported stopping intravaginal practices, some may have continued but been unwilling to reveal this information in interviews.25 We attempted to mitigate this social desirability bias by employing separate staff to conduct interviews and counselling. In addition, although they were repeatedly instructed otherwise, women may have felt that their participation was contingent on reporting cessation of vaginal practices. Fourth, as a short-term pilot study, we may not have followed enough participants for a long enough time period to detect changes in BV owing to cessation of intravaginal practices. While some women reported stopping all practices after the first visit, others continued engaging in various practices even after two or three counselling sessions. Women who did not stop until later visits may not have been followed long enough to see a corresponding reduction in BV.

    Some authors have criticised the TTM because it presents behaviour change as a linear function, with steady progression from precontemplation through maintenance, whereas actual behaviour change may not be linear.26 Our data support this non-linearity: many participants moved directly from precontemplation or contemplation to action stage. Another limitation of the TTM is that categorisation of individual stages is overly subjective. We minimised this bias by using a standardised questionnaire and algorithm to properly stage women, for each practice, at each visit.

    In summary, this pilot study provides important preliminary data about a behaviour change intervention encouraging cessation of intravaginal practices. Our findings are highly generalisable to family planning clinic patients in regions, like Zimbabwe, where intravaginal practices are common. We determined that a counselling intervention targeted at the stage of behaviour change is feasible and effective for reducing reported engagement in intravaginal practices. However, although promotion of the cessation of intravaginal practices led to rapid reductions in engagement in these behaviours, behaviour change alone was not sufficient to reduce BV prevalence. We believe that additional interventions are likely needed to meaningfully impact BV. For example, increasing use of male condoms,27 together with reductions in intravaginal practices, may have greater impact. Various combinations of interventions could be evaluated in a factorial design to assess the effects of each component intervention alone, as well as their impact in combination. Larger, longer-term studies will be necessary to determine the role of reduced engagement in intravaginal practices on BV prevalence.

    Key messages

    • Intravaginal practices—commonly performed by women worldwide—may increase risk of bacterial vaginosis.

    • A four-visit, monthly, counselling intervention using the transtheoretical model of behaviour change was effective in reducing women's reported engagement in intravaginal practices during the study period.

    • Despite the decrease in self-reported intravaginal practices, we observed no corresponding reduction in prevalence of bacterial vaginosis among study participants.

    References

    1. ↵
      1. Hillier S,
      2. Marrazzo J,
      3. Holmes K
      . Bacterial vaginosis. In: Holmes K, Sparling P, Mardh P. eds. Sexually transmitted diseases. New York: McGraw-Hill, 2008:737–68.
    2. ↵
      1. Low N,
      2. Chersich MF,
      3. Schmidlin K, et al
      . Intravaginal practices, bacterial vaginosis, and HIV infection in women: individual participant data meta-analysis. PLoS Med 2011;8:e1000416.
      OpenUrlCrossRefPubMed
    3. ↵
      1. Hilber AM,
      2. Francis SC,
      3. Chersich M, et al
      . Intravaginal practices, vaginal infections and HIV acquisition: systematic review and meta-analysis. PLoS ONE 2010;5:e9119.
      OpenUrlCrossRefPubMed
    4. ↵
      1. Cohen CR,
      2. Lingappa JR,
      3. Baeten JM, et al
      . Bacterial vaginosis associated with increased risk of female-to-male HIV-1 transmission: a prospective cohort analysis among African couples. PLoS Med 2012;9:e1001251.
      OpenUrlCrossRefPubMed
    5. ↵
      1. Atashili J,
      2. Poole C,
      3. Ndumbe PM, et al
      . Bacterial vaginosis and HIV acquisition: a meta-analysis of publisehd studies. AIDS 2008;22:1493–501.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Brown JM,
      2. Hess KL,
      3. Brown S, et al
      . Intravaginal practices and risk of bacterial vaginosis and candidiasis infection among a cohort of women in the United States. Obstet Gynecol 2013;121:773–80.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Ness RB,
      2. Hillier SL,
      3. Richter HE, et al
      . Douching in relation to bacterial vaginosis, lactobacilli, and facultative bacteria in the vagina. Obstet Gynecol 2002;100:765.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Brotman RM,
      2. Ghanem KG,
      3. Klebanoff MA, et al
      . The Effect of Vaginal Douching Cessation on Bacterial Vaginosis: A Pilot Study. Am J Obs Gynecol 2008;198:e1–7.
      OpenUrlCrossRefPubMed
    9. ↵
      1. Van De Wijgert JH,
      2. Mason PR,
      3. Gwanzura L, et al
      . Intravaginal practices, vaginal flora disturbances, and acquisition of sexually transmitted diseases in Zimbabwean women. J Infect Dis 2000;181:587–94.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Alcaide ML,
      2. Chisembele M,
      3. Mumbi M, et al
      . Examining targets for HIV prevention: intravaginal practices in Urban Lusaka, Zambia. AIDS Patient Care STDS 2014;28:121–7.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Lees S,
      2. Zalwango F,
      3. Andrew B, et al
      . Understanding motives for intravaginal practices amongst Tanzanian and Ugandan women at high risk of HIV infection: the embodiment of social and cultural norms and well-being. Soc Sci Med 2014;102:165–73.
      OpenUrlCrossRefPubMed
    12. ↵
      1. Allen CF,
      2. Desmond N,
      3. Chiduo B, et al
      . Intravaginal and menstrual practices among women working in food and recreational facilities in Mwanza, Tanzania: implications for microbicide trials. AIDS Behav 2010;14:1169–81.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Martin Hilber A,
      2. Hull TH,
      3. Preston-Whyte E, et al
      . A cross cultural study of vaginal practices and sexuality: implications for sexual health. Soc Sci Med 2010;70:392–400.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Turner AN,
      2. Morrison CS,
      3. Munjoma MW, et al
      . Vaginal practices of HIV-negative Zimbabwean women. Infect Dis Obstet Gynecol 2010;2010:1–7.
      OpenUrl
    15. ↵
      1. Masese L,
      2. McClelland RS,
      3. Gitau R, et al
      . A pilot study of the feasibility of a vaginal washing cessation intervention among Kenyan female sex workers. Sex Transm Infect 2013;89:217–22.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Grimley DM,
      2. Oh MK,
      3. Desmond RA, et al
      . An intervention to reduce vaginal douching among adolescent and young adult women: a randomized, controlled trial. Sex Transm Dis 2005;32:752–8.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Alcaide ML,
      2. Mumbi M,
      3. Chitalu N, et al
      . An intervention to decrease intravaginal practices in hiv-infected women in Zambia: a pilot study. J Assoc Nurses AIDS Care 2013;24:219–26.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Glanz K,
      2. Bishop DB
      . The role of behavioral science theory in development and implementation of public health interventions. Annu Rev Public Health 2010;31:399–418.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Prochaska JO
      . Systems of psychotherapy: a transtheoretical analysis. Homewood: Dorsey Press, 1984.
    20. ↵
      1. Nugent RP,
      2. Krohn MA,
      3. Hillier SL
      . Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991;29:297–301.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Leon AC,
      2. Davis LL,
      3. Kraemer HC
      . The role and interpretation of pilot studies in clinical research. J Psychiatr Res 2011;45:626–9.
      OpenUrlCrossRefPubMed
    22. ↵
      1. McNutt L,
      2. WU C,
      3. Xue X, et al
      . Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003;157:940–3.
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Van de Wijgert JHHM,
      2. Borgdorff H,
      3. Verhelst R, et al
      . The vaginal microbiota: what have we learned after a decade of molecular characterization? PLoS ONE 2014;9:e105998.
      OpenUrlCrossRefPubMed
    24. ↵
      1. Verstraelen H,
      2. Swidsinski A
      . The biofilm in bacterial vaginosis: implications for epidemiology, diagnosis and treatment. Curr Opin Infect Dis 2013; 26:86–9.
      OpenUrlCrossRefPubMed
    25. ↵
      1. Francis S,
      2. Lees S
      . Vaginal practices diary: development of a pictorial data collection tool for sensitive behavioural data. Sex Transm Dis 2012;39:614–21.
      OpenUrlCrossRefPubMed
    26. ↵
      1. Weinstein ND,
      2. Rothman AJ,
      3. Sutton SR
      . Stage theories of health behavior: conceptual and methodological issues. Health Psychol 1998;17:290–9.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Guédou FA,
      2. Van Damme L,
      3. Deese J, et al
      . Behavioural and medical predictors of bacterial vaginosis recurrence among female sex workers: longitudinal analysis from a randomized controlled trial. BMC Infect Dis 2013;13:208.
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:

    Footnotes

    • Handling editor Jackie A Cassell

    • Contributors AE cleaned and analysed the data and drafted and revised the paper. PM implemented the study, collected the data and reviewed the draft paper. SF assisted with the development of the data collection tools and reviewed the draft paper. MM performed all laboratory diagnostics and reviewed the draft paper. JvdW assisted with the development of the data collection tools and with interpretation of the data. TC assisted with the development of the data collection tool, oversaw data collection, and revised the paper. ANT conceived of the study, designed the data collection tools, monitored data collection, wrote the statistical plan and drafted and revised the manuscript.

    • Funding The study was funded by a grant from the Public Health Preparedness in Infectious Diseases (PHPID) programme at Ohio State University.

    • Competing interests None.

    • Ethics approval Ohio State University Institutional Review Board (Columbus, OH), the Joint Parirenyatwa Hospital and College of Health Sciences Research Ethics Committee of the University of Zimbabwe (JREC), the Medical Research Council of Zimbabwe, and the Research Council of Zimbabwe.

    • Provenance and peer review Not commissioned; externally peer reviewed.