Trope, Lee A. BS*; Chumworathayi, Bandit MD†; Blumenthal, Paul D. MD, MPH‡
Although frequently repeated cytology screening has led to an 80% decline in cervical cancer mortality in the developed world, cervical cancer remains an important public health problem among adult women in developing countries. Although it is generally agreed that a cytology-based approach fulfills the criteria of an effective screening program,1 which includes cost effectiveness, reduction of incidence of a disease, and reduction of morbidity and mortality from a disease,2 such cytology-based cervical cancer-screening programs have largely failed within the developing country context in part because of an impoverished health care infrastructure, too few trained and skilled professionals, uninformed women who get lost to follow-up, and a lengthy turn-around time. Vaccines that prevent or treat human papillomavirus (HPV) acquisition are recently introduced or under development and, ultimately, may help resolve this important public health problem. However, large-scale vaccine implementation is still several years away in the countries that need it the most.3 The investigation of screening tests that use fewer resources and offer rapid results has therefore been a priority.
Among these tests, visual inspection with acetic acid (VIA) has proven to be a promising candidate for screening in low-resource settings as it fulfills the basic criteria of a satisfactory screening test and has demonstrated effectiveness in reducing incidence of and mortality from cervical cancer.4 Visual inspection with acetic acid testing is inexpensive and simple and can be provided at all levels of the health care system by nurses and midwives. Another key advantage of VIA testing is that the results are immediately available. This means that management decisions, especially whether to offer outpatient treatment if the cervix is found to be abnormal, can be made during a single visit, a significant benefit in countries where health care facilities are not easily accessible.5 However, there are disadvantages to VIA. First, it is subjective; providers must interpret what they see on the cervix, which can be problematic for nurses trained never to provide treatment without certainty. Moreover, a long period (∼2 weeks) of provider training is crucial to be able to perform VIA, and additional workshops to sustain skills are also necessary, thus decreasing the cost-effectiveness of the approach. Recent studies have stressed that without high-quality training of providers, continuous quality assurance, and extensive monitoring, VIA programs may fail.6
Human papillomavirus DNA testing, another test with promise in low resource settings, offers some advantage compared with VIA. Human papillomavirus testing is more objective and reproducible than VIA and does not depend on the visibility of the squamo-columnar junction of the cervix. It is less demanding in terms of training and quality assurance. Heretofore, the overarching drawback has been that the HPV DNA technology has been (1) too expensive for use in developing countries and (2) involves a lengthy turnaround time; results are not available in a single visit.7 In an attempt to address these issues, Digene (Gaithersburg, MD) has developed CareHPV, an HPV DNA test that potentially costs less than US$5 per kit, can be administered in a static or mobile clinic with no refrigeration, and can get results (for batches of either 24, 48, or 96 samples) in less than 2.5 hours. The swab can be performed by a clinician (cervical swab) or by the women herself (vaginal self-swab), and the test characteristics are felt to be ready for field implementation. Despite indication that the test may be clinically acceptable, no field-based surveys or studies have been published assessing attitudes toward the acceptability of integrating an HPV-based approach into current cervical cancer prevention programs in low resource settings.
Therefore, before a scheduled pilot implementation of the CareHPV DNA technology in Roi-et Province, Thailand, to assess the most appropriate implementation strategy for this test, this study was conducted to explore preferences and perceived benefits among likely users of the new test in this setting.
All data were collected by in-person surveys in Roi-Et Province, Thailand. Survey instruments were translated from English into Thai and back translated to assess accuracy. Surveys were provided to participants at face-to-face meetings in each district of Roi-et Province, Thailand between December 2007 and January 2008. Roi-et Province-a Northeastern province of Thailand with a population of 1.2 million-was selected because, since 2000, it has implemented a successful "single-visit" cervical cancer prevention approach (combining VIA and cryotherapy) and has improved coverage from 4.7% in 2000 to the highest coverage in Thailand with over 60% screened. Cervical cancer prevention services had not been successful there in the past, and the current incidence of cervical cancer in Roi-et is 20/100,000 women per year. The province is 70% rural, and 95% of people are Buddhist.1
Questionnaires were administered to health care providers, trainers, district medical directors, and district health officers in their personal district offices and hospitals throughout Roi-et Province. All 8 VIA trainers, 16/16 district medical directors, 20/20 district health officers, and 4/4 colposcopists were surveyed. A total of 80 health care providers have been trained to perform VIA in Roi-et Province since 2000. Since somewhat of a professional "revolving door" among health care providers in Roi-et Province exists (many leave because they get married, get a higher paying job in the city, etc), only 40 health care providers were surveyed in this study. It is unknown what percentage of current VIA providers this number is, but it is likely well over 50% (Table 1).
To assess perceptions about the CareHPV-inclusive protocols relative to the current cervical cancer screening protocol, participants were provided with 5 diagrams of 5 potential CareHPV-inclusive screening plans, labeled plans A to E. Figure 1 details the differences between the 5 plans.8 To explore how these plans compared relative to the current plan, participants were asked to assign each plan a ranking from 1 (much less preferable than the current plan) to 5 (much more preferable than the current plan). They were then asked to assign each plan a ranking from 1 (much less beneficial than the current plan) to 5 (much more beneficial than the current plan). To further assess reactions to these CareHPV-inclusive protocols, participants were asked to circle which plan of the 5: (1) he/she found most or least preferable and (2) which plan he/she thought would be most or least beneficial in terms of a reduction in incidence and mortality from cervical cancer in his/her district.
Quantitative Data Analysis
The data were quantitatively analyzed using Statistical Package for the Social Sciences (SPSS, Chicago, IL). Participants categorized based on district, occupation, sex, and distance of district to the assigned referral colposcopy clinic. Responses to survey questions, particularly differences in attitudes toward the various approaches, were tested for statistical significance (P < 0.05) using 2-tailed t tests, 1-way analysis of variance tests, and 2-tailed χ2 tests.
Comparing Plans A to E With Roi-et Province's Current Cervical Cancer Screening Program
When asked how preferable the new plan was in comparison with the current cervical cancer screening protocol, the most common response for the village-based, self-swab plan (E) overall was "5," with 39.8% of respondents viewing plan E as much more preferable than the current plan (Table 2). Combining this with the 18.2% that viewed plan E as slightly more preferable, 58% of all respondents found plan E to be more preferable than the current plan. Yet, 28.4% of respondents viewed plan E as slightly (15.9%) or much (12.5%) less preferable than the current cervical cancer screening program leaving plan E with a mean score of 3.58 (still the highest of the 5). Plan E was preferred relative to the current plan significantly more than the clinic-based, clinician swab, double-testing plan (A) (P < 0.05), the clinic-based, self-swab, double-testing plan (B) (P < 0.05), and the clinic-based, self-swab, HPV+ only-testing plan (D) (P < 0.05), but not significantly more than clinic-based, clinician swab, HPV+ only-testing plan (C) (P > 0.05) (Table 2). Although the most common response for plan C was "3," indicating that the plan was equally as preferable as the current plan, 23.9% of respondents viewed plan C as much more preferable than the current screening plan, leaving the plan with a mean score of 3.34. Notably, plan B received the lowest mean score of 2.99-the only plan viewed as less preferable than the current screening plan.
When asked to compare each individual plan to the current cervical cancer screening program and to indicate if the plan would be more or less beneficial in decreasing incidence and mortality from cervical cancer, participants expressed a preference for plan E. The mean score for plan E was the highest of the 5, at 3.70, and the most common response overall for Plan E was "5," with 42.0% of all respondents predicting Plan E to be much more beneficial than their current plan (Table 3). The most common response for Plan D was also "5," with 25.0% of respondents believing the plan to be much more beneficial than their current plan, but 23.9% believing it to be equally as beneficial as the current plan, leaving the plan with an average score of 3.32. The most common rank for all other plans was "3," indicating that the plans were predicted to be no more or no less beneficial than the current screening plan. Plan E was perceived as more beneficial than the current plan significantly more than plan B (P < 0.05) and plan D (P < 0.05), but not significantly more than plan A (P > 0.05) or plan C (P > 0.05) (Table 3).
Comparing Plans A to E With Each Other
The village-based, self-swab plan (E) was chosen as the most preferable and most beneficial plan by far, with 50.6% and 58.3% of respondents choosing plan E in their respective responses. The clinic-based, clinician swab, double-testing plan (A) came in at a distant second with 24.7% of participants (less than half of plan E) indicating that it was their most preferred plan, and 27.4% (again less than half of plan E) predicting it would be the most beneficial plan. Plans B, C, and D were infrequently chosen as the most preferable or most beneficial plans of the 5 (Fig. 2).
When asked which plan was the least preferable and least beneficial of the 5, respondents most frequently chose plan A, with 48.3% of respondents believing it to be least preferable and 45.2% of respondents predicting it would be least beneficial in reducing incidence and mortality from cervical cancer in their districts. Plan E came in second, with 23.0% of respondents choosing it as their least preferred plan of the 5. Plan E also tied with plan D as the plan predicted to be the second least beneficial plan. Notably, plan E was ranked higher as the "least preferred" plan (23.0%) than as the "least beneficial" plan (17.9%) (Fig. 3).
Occupations were split into the broader categories of field-oriented (officers, providers, trainers) versus hospital-oriented (colposcopists, directors) professionals (Table 4). Hospital-oriented professionals were more evenly split between plans A and E for most preferred plan, with 30.0% choosing plan A and 45.0% choosing plan E, whereas 52.2% of field-oriented respondents chose plan E as the most preferred plan, and only 23.2% chose plan A. For most beneficial plan, hospital-oriented professions were most evenly split again, with 40% choosing plan A, and 40.0% choosing Plan E as the most beneficial plan of the 5. In contrast, 60.9% of field-oriented respondents thought plan E would be most effective, and only 21.2% preferred plan A. Hospital-oriented respondents were no more definitive about which plan they found least preferable and least beneficial. Forty-five percent chose plan A, and 30.0% chose plan E as the least preferable plan, in contrast to a 49.3% (plan A) to 20.3% (plan E) difference for field-oriented respondents. In addition, 35.0% chose plan A, and 30.0% chose plan E as the least beneficial plan, in contrast to a 46.4% (plan A) to 13.0% (plan E) difference for field-oriented respondents (Table 4).
Self-Swab Versus Clinician Swab
Plans B, D, and E called for the self-swab version of the CareHPV DNA test, whereas plans A and C called for the clinician swab test. When asked which of the 5 plans they found most preferable, 59.6% of participants chose a plan that incorporated the self-swab version of the CareHPV DNA test, whereas 40.4% chose a clinician swab-inclusive plan. A 2-tailed χ2 test was performed and indicated that this was not a significant difference. Moreover, 62.9% of participants responded that the most beneficial plan in reducing incidence and mortality from cervical cancer was a self-swab-inclusive plan, a significantly higher number (P < 0.05) than the remaining 37.1% of respondents who chose a clinician swab-inclusive plan.
When broken into occupation category, field-oriented respondents definitely chose a self-swab-based plan as their most preferred and most beneficial plan. Hospital-oriented respondents were less decisive, with 50% choosing a self-swab-based plan, and 50% choosing a clinician swab-based plan as the most preferred plan. Moreover, 45.0% of these hospital-oriented respondents choose a self-swab-based plan, and 55.0% choose a clinician swab-based plan as the most beneficial plan.
Comparing Plans A to E With Roi-et Province's Current Cervical Cancer Screening Program
Currently, Roi-et Province uses a dual-track cervical cancer screening program, which combines VIA (for those between 30 and 45 years and a visible squamocolumnar junction) in conjunction with cryotherapy in a single visit and a Papanicolaou smear policy (for the other groups of women). Why did participants see the village-based, self-swab plan (E) as an improvement from the current plan, whereas plans A to D were neither more nor less favored? The answer likely stems from Plan E's core difference from each other plan: it is proactive. When asked, what do you think is the biggest challenge to a successful cervical cancer screening program here in Roi-et Province? The most common response-given by participants of every occupation, sex, and district-was "achieving high screening coverage." Many others cited "a lack of public education on the value and purpose of cervical cancer screening," as an element needing more attention. Plan E would actively improve the likelihood of overcoming both of these obstacles. Coverage would likely increase considerably if Plan E were implemented, as screening would come to the women, instead of women traveling far distances to come to the clinic or hospital. Knowledge about the value and purpose of cervical cancer screening could also increase with the implementation of plan E, for the plan would involve substantial outreach from health care educators and/or providers, whose presence in the community would remind women about the importance of cervical cancer screening and prevention. Thus, plan E likely seemed an enticing option to correct the shortcomings persistent in the current plan.
Yet, not all participants preferred the village-based, self-swab plan (E) over what they do today-responding that plan E would be "too cumbersome" and involve "too much work" for the health system. Interestingly, more participants responded that they found plan E "much more beneficial" (42.0%) than the current plan than responded that they found plan E "much more preferable" (39.8%) (Tables 1 and 2). This slight disparity suggests that even those respondents who found plan E personally burdensome could still see its potential benefits in the effort to reduce incidence and mortality from cervical cancer.
Overall, with the exception of the clinic-based, self-swab, double-testing plan (B), each plan's mean score suggested that it was slightly to substantially more preferable to participants than the current screening program (Table 2). Moreover, each plan's mean score (including plan B) indicated that the plan was seen as slightly to substantially more beneficial than the current screening program (Table 3). These results suggest that although participants do not believe that the mere addition of the CareHPV test in any capacity would be a huge leap forward in successful cervical cancer prevention, neither was incorporating the CareHPV test seen as a step back in the way the community approaches cervical cancer screening.
Comparing Plans A to E With Each Other
The vast majority of participants (81.8%) cited either clinic-based, clinician swab, double testing (A) or village-based, self-swab (E) as the most preferable, least preferable, most beneficial, and least beneficial plans of the 5 in terms of each plan's likelihood to reduce cervical cancer incidence and mortality. In contrast, Plans B, C, and D were cited only 18.2% of the time in any of these 4 questions (Figs. 1 and 2).
Why did most respondents favor Plan E, but others most passionately disapprove of it? Why did the largest proportion of respondents disapprove of plan A, but others most fervently favor it? Put differently, why did Plans A and E elicit such strong response? Insight comes from the written responses of participants when asked to justify their choices to these questions. Those who cited Plan E as the most preferred and/or most beneficial plan wrote of the benefits of a "proactive" plan that is "convenient for women." They cited the need for "increased coverage" and "better access for those who do not have the transportation to get to the clinic," as justification for their choice. These pro-plan E participants were often the same people who most fervently disliked and distrusted plan A (in fact, 77.8% of those who chose plan E as the most preferred and/or most beneficial plan, chose plan A as the least preferred and/or least beneficial plan). Plan A was disliked because the use of both tests (regardless of the results of the HPV DNA test) was seen as "redundant" and "a waste of time." In addition, participants wrote that women are too "shy and embarrassed" to have a pelvic examination as called for in plan A. In contrast, pro-plan A participants cited it as the most preferred and/or most beneficial plan because it ensured a "higher probability for detecting precancerous lesions" and would "increase the confidence in the results of screening." Again, the participants that favored plan A were most often the same participants who disliked and distrusted plan E (60.0% of those who chose plan A as the most preferred and/or beneficial plan chose plan E as the least preferred and/or least beneficial plan). Plan E was disliked because clients would "not be able or be confident in a self-swab test" and because "a clinician swab would be more accurate." Moreover, some respondents cited the "burden" on the providers or health educators to travel around their district as a reason for opposing plan E.
As illuminated by Table 4, there was an interesting divide between occupations that are field-oriented (have a primary health center as a base of activities and work closely with patients at the primary level) and those that are hospital-oriented (have hospital as a base of activities and have more clinical outlook) in their responses. This dichotomy may also be looked at as a physician, nonphysician divide. It is interesting that field-oriented, nonphysician respondents (officers, providers, trainers) were decisively pro-plan E and anti-plan A, whereas hospital-oriented, physicians (colposcopists, directors) were much more ambivalent-with many pro-plan E and anti-plan A like their field-oriented colleagues, but a sizable proportion pro-plan A and anti-plan E.
Notably, those hospital-oriented, physicians who were pro-plan A and anti-plan E often justified their choice by lauding the added "credibility" and "accuracy" when both a clinician CareHPV test and VIA are administered together, as in plan A. They then disparaged plan E because they would not be "confident in a woman's ability to swab herself"because they thought "a woman would not want to be in charge of her own screening" or because they feared "missing a chance for the clinician to observe the cervix if a woman swabs incorrectly." Field-oriented nonphysicians, on the other hand, were far less worried about the women's confidence and ability to perform the self-swab, often citing the uncomfortable pelvic examination as an overwhelming deterrent for woman to come in and get screening.
From a different perspective, perhaps this is not a plan E/plan A divide, but rather a self-swab versus clinician swab divide. As evident by Table 5, more field-oriented respondents chose a self-swab-based plan (B, D, E) than a clinician swab-based plan (A, C) as the most preferred and most beneficial plan. In contrast, hospital-oriented respondents were split 50-50 on which version of the CareHPV DNA test would be most preferable and-for the most beneficial plan-more actually believed in the clinician swab version of the CareHPV test. Perhaps, this divide stems from the fact that field-oriented participants work more closely with women on a day-to-day basis and are more in touch with their abilities and needs. In addition, perhaps, hospital-oriented participants (at least those who were pro-Plan A) are more concerned with the clinical accuracy of the diagnosis than with the woman's confidence and ability to self-swab. A survey of the target Roi-et women, which assesses both their acceptability of a self-swab screening method and the accuracy with which they can perform a self-swab examination, would be helpful in assessing the truth to these claims.
It is still important to remember that, despite this segment of hospital-oriented participants suggesting that women would be inaccurate and insecure performing a self-swab and despite a Thai physician's anecdotal remark (repeated by others) that "women here will never be able to figure out the self-swab CareHPV test," the overall majority of respondents had faith in the self-swab technology-with a significant proportion (62.9%) predicting 1 of the self-swab-inclusive plans (B, D, E) would be the most beneficial in reducing cervical cancer screening and mortality. This finding-along with the 59.6% of participants who preferred a self-swab-inclusive plan-suggests that the majority of screening stakeholders in Roi-et Province are not as shocked and appalled by the idea of a self-swab as the minority that is against such a method of screening.
Yet, the sizeable support given to the village-based, self-swab plan (E) may also be a result of the practical constraints inherent in plans A to D. As was previously mentioned, the new CareHPV DNA test can only be run in batches of 24, 48, or 96 samples and takes 2.5 hours to provide results. If plans A to D were implemented, a woman would have to wait for a minimum of 23 other women to get screened (either by a clinician swab or self-swab). This could take anywhere between hours and days. After enough women have been screened that a batch could be run (and still be cost-effective), a woman would still need to wait another 2.5 hours for the results. Thus, the original problem with cytology-based screening-a lengthy turnaround time, loosing women to follow-up-would essentially be brought back to the system. In addition, almost ironically, this is the very problem that the VIA-based program was introduced to eliminate. In plan E, on the contrary, the self-swab could be provided to 24, or even 48 or 96 women right in their community. The samples could then be collected, run on a car battery and in 2.5 hours-while women either waited in their homes, went to work, and so on-the results could be retrieved. This would still break the link between screening and treatment brought about with the VIA-based program, but in a way that was found to be even more favorable by participants than the current system.
Strengths and Weaknesses
This study was strong; the majority of the target sample were surveyed, and no participants refused the interview. Yet all research has its shortcomings and biases. This study's precision was limited by written and on-sight translation, the group interview format, investigator-participant cultural disparities, and an incomplete provider sample.
Overall, participants indicated strong support for an innovative plan in which women are screened in their homes and villages using the self-swab version of the CareHPV DNA test, and only those who score positive for HPV are screened with VIA. With a bit of nuance, plan E emerged out of the study as the "big winner"-and this overwhelming support for plan E was revealing. It suggests that by and large, opinion leaders give higher priority to increasing screening coverage of the target population than to providing a more sensitive and specific diagnosis. It means that, overall, stakeholders in Roi-et Province are ready and willing to accept self-screening by the females in their society. In addition, it indicates that the Roi-et health care community is tolerant of program reform.
If women feel similarly-and if the a program to instill confidence in the self-swab version of the CareHPV-test supplements its implementation-then there is significant justification for a Plan E-style cervical cancer screening program in Roi-et Province in the near future. Moreover, the study provides a good model for exploring the "best" CareHPV-inclusive program in other settings beyond Roi-et.