What women want. Women’s preferences for the management of low-grade abnormal cervical screening tests: a systematic review

Authors


Dr M Rebolj, Department of Public Health, University of Copenhagen, Øster Farimagsgade 5, DK-1014 København K, Denmark. Email mare@sund.ku.dk

Abstract

Please cite this paper as: Frederiksen M, Lynge E, Rebolj M. What women want. Women’s preferences for the management of low-grade abnormal cervical screening tests: a systematic review. BJOG 2012;119:7–19.

Background  If human papillomavirus (HPV) testing will replace cytology in primary cervical screening, the frequency of low-grade abnormal screening tests will double. Several available alternatives for the follow-up of low-grade abnormal screening tests have similar outcomes. In this situation, women’s preferences have been proposed as a guide for management decisions.

Objectives  To determine women’s preferences for the follow-up of low-grade cervical screening abnormalities.

Search strategy  Using Medical Subject Headings (MeSH) terms, PubMed was searched for articles published up to December 2010. The reference lists of the retrieved studies were consulted.

Selection criteria  Studies asking women to state a preference between active follow-up and observation for the management of low-grade abnormalities on screening cytology or HPV tests.

Data collection and analysis  Information on study design, participants and outcomes was retrieved using a prespecified form. Studies were sorted by design.

Main results  Thirteen studies were included in the review. In all five studies that surveyed women with abnormal tests before any management had started, two-thirds preferred active follow-up, predominantly as immediate colposcopy, to observation, predominantly as repeated Pap smears. In all but two studies testing other situations, women more often expressed a preference for active follow-up than for observation; however, women appeared to be somewhat more willing to accept observation if reassured of the low risk of cervical cancer.

Conclusions  Even for low-grade abnormal cervical tests, women tend to prefer active management strategies. It may be a challenge to meet their expectations of optimal follow-up when HPV testing is used in primary screening.

Introduction

Cytology-based screening and treatment of cervical intraepithelial neoplasia (CIN) have reduced the incidence of and mortality from cervical cancer.1 Cervical cells span from normal to low-grade (mild) and high-grade (severe) abnormalities. Low-grade abnormalities are typically found in 2–5% of screened women.2 They represent an important reservoir of CIN, although many are not associated with CIN and may normalise spontaneously. The management of low-grade abnormalities therefore represents a clinical challenge.

Low-grade cytological abnormalities are managed with repeated cytology testing, immediate referral for colposcopy or, more recently, with triage to colposcopy based on human papillomavirus (HPV) testing. These strategies all have distinct advantages and disadvantages. Although repeated Pap smears allow some low-grade abnormalities to normalise on their own, they rely on high compliance and expose women to long periods3,4 of uncertainty with regard to the final diagnosis. Immediate referral for colposcopy may lead to the detection of more ≥CIN2 but, to some degree, this may also reflect overdiagnosis rather than superior effectiveness.5,6

In a review in which the outcome was ‘remaining free of cervical cancer’, the two methods appeared to be equally effective.7 Based on data from the UK, there also seemed to be no reason to favour either of the management strategies when cost-effectiveness was considered.8 The choice of management strategy therefore remains debatable.9,10 As the management of an abnormal screening outcome has a psychological impact on the woman,11,12 it has been proposed to use women’s preferences to guide management decisions.13,14

In the future, HPV DNA testing may replace cytology as the primary screening test. Primary HPV testing is expected to be more effective than cytology in preventing cervical cancer.15–17 In recent randomised controlled trials, however, about twice as many women had a positive HPV test than an abnormal cytology.18 This extra group of women with HPV-positive tests and normal cytology will add to the problem of the management of women with low-grade abnormalities.

To shed light on the choice of management of low-grade abnormal cervical screening outcomes, we performed a systematic review of the literature on women’s preferences.

Methods

Literature search

The systematic review was based on published data. We used Medical Subject Headings (MeSH) terms to search for relevant studies in PubMed. The details of the search are reported in Appendix S1. Two researchers (MEF, MR) independently screened the retrieved abstracts to identify all potentially relevant articles. Finally, we checked the reference lists of all included articles.

Inclusion criteria

To include the article in the review, the participating women had to have a low-grade abnormal cervical screening test, or be given a hypothetical situation in which they would receive such a screening result. We considered the following categories of screening outcomes as low-grade abnormal: borderline dyskaryosis, mild dyskaryosis, atypical squamous cells of undetermined significance (ASCUS), low-grade squamous intraepithelial lesions (LSIL or LGSIL) or an HPV infection with normal cytology. Furthermore, women had to be given a choice between different management strategies, and report either on their choice or on the relative value that they would allocate to a given strategy.

Definition of active follow-up and observation

The identified studies presented women with several different management strategies for the follow-up of low-grade cervical screening abnormalities, which we categorised as either active follow-up or observation. As active follow-up, we considered immediate colposcopy, repeated testing in less than the standard 6–12 months and HPV triage. HPV triage was placed in this category because it typically includes an immediate referral for colposcopy if the HPV test is positive. This was not always specified in the studies, but we assumed that it was the case. As observation, we considered repeated Pap smears or HPV testing (typically in 6–12 months) and less frequent colposcopy compared with the standard 6–12 months.

Statistical analysis

We abstracted the data using a prespecified list of study characteristics, systematised in Tables 1–3, S1 and S2. As the study designs and outcomes were highly variable, we did not attempt to proceed with a formal meta-analysis and, instead, opted for a separate presentation of the results from each trial in a structured format.

Table 1.   Overview of studies included in this systematic review
ReferenceYear of data collectionLocationInclusion criteria
  1. ASCUS, atypical squamous cells of undetermined significance; CIN, cervical intraepithelial neoplasia; GP, general practitioner; HPV, human papillomavirus; LGSIL/LSIL, low-grade squamous intraepithelial neoplasia.

  2. *Atypia, HPV effects, LGSIL.

  3. **The goal was to recruit approximately similar numbers of White, Latina, Asian and African American women, as well as a substantial number of participants with limited English proficiency.58

  4. ***Additional information was provided in an earlier report.58

  5. ****Women who replied ‘yes’ to the question: ‘If the HPV test was available to you, would you want to be tested?’.

  6. *****Specified as 18–70 years in the accompanying decision aid (http://www.psych.usyd.edu.au/cemped/docs/08_IMAP_decision_aid.pdf, accessed 2 May 2011).

  7. ******Nonspecific minor changes with or without HPV effect (possible LSIL and LSIL according to the Australian classification revised in 2005). Women with CIN1 were not included in the study.

  8. *******At recruitment (primary testing).

Jones et al.11NAUKManaged, either by repeated Pap smears or an early colposcopy, for first-time mild dyskaryosis over the previous 30 months at three hospital settings
Ferris et al.14NAUSAAge ≥16 years, recruited from waiting rooms of the participating family practice or obstetrics and gynaecology clinics
Peters et al.191994–1995UKFirst-time mildly dyskaryotic cervical smear in a participating GP practice
Meana et al.26NACanadaEnglish-speaking women with a first-time mildly abnormal Pap smear* referred for colposcopy
Melnikow et al.24NAUSAAge ≥18 years or emancipated if minors, English or Spanish speaking, previously screened, recruited from participating family planning clinics
Le et al.30NACanadaWomen with ASCUS/LGSIL smears seen at colposcopy clinic
Kitchener et al.291998–1999UKAge 20–60 years, mild dyskaryosis or recurrent borderline cytological changes; not: pregnant, abnormal vaginal bleeding, impending migration from area, previous cytological abnormality or treatment for CIN
Philips et al.28NAUKWomen eligible for cervical screening, attending routine (nonscreening) consultations with GP
Howard et al.232003–2004AustraliaWomen eligible for cervical screening (≥18 years), not had a Pap smear in the last 6 months, not due for a Pap smear, no hysterectomy, not pregnant
Huang et al.202002–2006USA50–80 years, self-identified as White, Latina, African American or Asian, **,*** English/Spanish/Cantonese/Mandarin speaking, seen at one of the participating continuity practices at least once in the previous 2 years, no current cancer diagnosis, not cognitively impaired, had not undergone a hysterectomy, willing to be tested for HPV****
Whynes et al.22NAUKMild or borderline cervical abnormalities detected at routine screening
McCaffery et al.212004–2006AustraliaAged 16–70 years,***** attending routine cervical screening at family planning clinics, minor abnormal Pap smear,****** no other abnormal result or genital condyloma in ≤2 years, not pregnant, English speaking
Kitchener et al.252001–2003*******UKRandomised into the intervention arm, HPV positive/cytology normal at recruitment, invited for repeated HPV testing after 12 months, HPV positive at repeated HPV testing
Table 2.   Study sizes, screening history and socio-demographic characteristics of the women included in the studies
ReferenceStudy size*Number excluded/refused (% of total number)Mean age in years (SD)Smear/treatment historySocio-demographic characteristics**
  1. CIN, cervical intraepithelial neoplasia; NA, not available.

  2. *Number of women who stated their choice.

  3. **Education: only the lowest and highest levels of education reported in the articles are listed in this table.

  4. ***’Just under one-third’ of women managed with repeated Pap smears met exclusion criteria for colposcopy review clinic invitation taking place about 30 months after the mildly dyskaryotic study entry smear, i.e. already had colposcopy as a result of persistent smear abnormality during the 30-month study period.

  5. ****Main study women, n = 351.

  6. *****Weighted averages.

  7. ******Participation rate reported in the article: 75%.

  8. *******Choice (intervention) arm.

  9. ********The questionnaire was anonymous and reminders could not be sent.

  10. *********Based on a related sample of 1244 respondents.43

  11. **********All participants (n = 865), including those who would not want to be tested for HPV and were excluded from the analysis of management preferences: 53% aged 50–60 years, 31% aged 60–70 years, 16% aged 70–80 years.

  12. ***********At recruitment; women were followed up for 3 years after randomisation, and the questionnaire was offered at trial exit.

  13. ************Reported as 96 in a related article.41

  14. *************57% > 30 years.

  15. **************Age range at recruitment: 20–64 years.

Jones et al.11345NA***• Women managed with repeated Pap smears: 33.1 (NA)****NA• 48% married****,*****
   • Women managed with early colposcopy: 32.4 (NA)**** • 30% social class 1–2, 5% social class 4–5****,*****
Ferris et al.149687 (<1%)• Civilian women: 31.8 (NA)• 29% abnormal Pap smear*****• 11% <high-school, 37% college or postgraduate education*****
   • Military women: 33.8 (NA)• 6% cervical neoplasia***** 
Peters et al.19240333 (58%)• Intervention arm: 35 (11)No previous mildly dyskaryotic Pap smear• 15% no educational qualifications, 37% A-level or higher educational qualifications*****
   • Control arm: 32 (10)  
Meana et al.261369 (6%)33.4 (10.8)No previous mildly abnormal Pap smear• 1% primary school, 77% postsecondary school or above
     • 44% married/cohabiting
Melnikow et al.2414822 (13%)26 (NA)14% prior colposcopy• 34% less than high school, 20% some college
Le et al.30100NA******33.6 (11.3)86% previous abnormal Pap smear, 67% previous colposcopy• 66%≥office workers with community college-level education, 20% manual labourers without college-level education
Kitchener et al.29233*******126 (35%)*******29.6 (10.3)No previous cytological abnormality or CIN treatmentNA
Philips et al.2811414695 (80%)********41.3 (11.7)84% always attended for screening when called*********• 77% married or cohabiting
     • 56% left full-time education aged ≤16 years, 22% continued beyond age 18 years
Howard et al.23676 (8%)41.3 (12.9)36% previous abnormal Pap smear, 24% prior colposcopy• 67% married/living together
     • 6% some high school only, 73% completed university
Huang et al.205483975 (88%)NA**********7% prior HPV test, 49% last Pap <1 year ago**********• 57% married**********
     • 39% less than high school graduate, 34% college graduate**********
Whynes et al.221901 (<1%)• Colposcopy arm: 35.5 (NA)***********NA• Colposcopy arm: 25% no training or further education, 25% degree level***********
   • Observation arm: 35.9 (NA)*********** • Observation arm: 23% no training or further education, 17% degree level***********
McCaffery et al.2194************12 (11%)NA*************26% previous abnormal tests• 55% Married/de facto
     • 30% high school or lower education, 42% university degree
Kitchener et al.2542712 (3%)NA**************NANA
Table 3.   Outcomes of the studies included in the systematic review
ReferenceWoman’s situationDegree of abnormalityManagement choice revealed before or after management*Managed according to chosen strategy*Measure of choiceManagement strategies to choose fromRevealed preference for
Active follow-up (type)Observation (type)
  1. ASCUS/ASC, atypical squamous cells of undetermined significance; CI, confidence interval; CV, contingent valuation, the maximum amount of money a woman would be willing to spend on an evaluated option; HPV, human papillomavirus; LSIL (LGSIL), low-grade abnormal squamous intraepithelial neoplasia; NA, not applicable; NOS, not otherwise specified in article; NR, not reported/not relevant; Type, type of management strategy.

  2. *If the woman herself had abnormal test, otherwise not applicable.

  3. **Before being asked to state their management preferences, women were managed either with immediate colposcopy or repeated Pap smears.

  4. ***These are the outcomes of the first choice only; women could also state their preferences for second choice. As first choice, 20.6% (ASCUS) and 27.2% (LSIL) would choose cervicography. Cervicography was not classified as an active or delayed strategy because it was not clear from the article what the recommendations would be if the result was positive.

  5. ****A questionnaire sent 6 weeks after the cytological diagnosis; the outcomes of a questionnaire sent at 4 months were not reported separately by type of management strategy.

  6. *****By the time of the questionnaire sent at 4 months after the abnormal Pap smear, the difference between the two trial arms had decreased from 21% to 7%.

  7. ******Preferences of two women not reported.

  8. *******Utilities per scenario (three scenarios were evaluated) and per management strategy (two management strategies were evaluated) were derived using the standard gamble method. Outcome scenarios for Pap smear abnormalities: spontaneous resolution of abnormality vs. cryotherapy needed vs. cone biopsy needed. Management strategies: observation with repeated Pap smears vs. early colposcopy. Probabilities of outcomes of a low-grade Pap smear (spontaneous resolution vs. cryotherapy vs. cone biopsy) were derived from literature and expert opinion.

  9. ********0.93 ± 0.20 if it is assumed that the abnormality will regress spontaneously without requiring treatment, 0.95 ± 0.11 if it is assumed that cryotherapy would be needed (low-grade abnormalities) and 0.92 ± 0.16 if it is assumed that cone biopsy would be needed (more severe abnormalities).

  10. *********0.96 ± 0.13 if it is assumed that the abnormality will regress spontaneously without requiring treatment, 0.93 ± 0.17 if it is assumed that cryotherapy would be needed (low-grade abnormalities) and 0.91 ± 0.21 if it is assumed that cone biopsy would be needed (more severe abnormalities).

  11. **********Two interviews were undertaken with each respondent to assess the stability of the preferences, and the responses from the second interview were used for the analysis. Mean utilities per scenario were derived using the standard gamble method. Scenarios spanned a time period of 18 months and did not involve any plausible risk of death. Participants evaluated one repeated Pap smear scenario and one HPV triage scenario, whereby the assumption was made that the clinical outcomes were the same.

  12. ***********Additional 16% of women would agree to having more frequent Pap smears if the doctor recommended it.

  13. ************Authors argued that, as a result of self-selection for participation in their trial, ‘high expectations for colposcopy have been over-represented, whilst prior preferences for observation have been under-represented’, and that ‘the minority opinions which ostensibly favour colposcopy are suspect, on the grounds of preference bias in sample collection’.22

  14. *************Twelve women who dropped out of the study may have opted for repeated Pap smears or a colposcopy elsewhere.21 In a later article from the same study (n = 96), the reported proportions were 64% for HPV triage and 34% for repeated Pap smears.41

Jones et al.11Abnormal testMildly dyskaryotic cytologyAfterNoPreferred type, % per type• Immediate colposcopy**(colposcopy)NR
      • Repeated Pap smears**• 84% of women managed with early colposcopy 
       • 84% of women managed with repeated Pap smears 
Ferris et al.14HypotheticalASCUS or LSIL cytologyNANAPreferred type, % per type• Repeated Pap smears(colposcopy or HPV testing)(repeated Pap smears)
      • HPV testing• If ASCUS – 21.1%***• If ASCUS – 58.4%***
      • Cervicography• If LSIL – 58.6%***• If LSIL – 14.2%***
      • Colposcopy  
Peters et al.19Abnormal testMildly dyskaryotic cytologyBeforeNoPreferred interval for repeated Pap smears, % per interval• Sooner than 6 months(sooner than 6 months)NR
      • At 6 months• Intervention arm: 53%****,***** 
      • No preference• Control arm: 74%****,***** 
Meana et al.26Abnormal testMildly abnormal cytology (atypia, HPV effects, LGSIL)BeforeNoPreferred type, % per type• Histological evaluation(histological evaluation) 64%(repeated Pap smears) 18%
      • Repeated Pap smears(no strong preference) 17%******
      • No strong preference  
Melnikow et al.24HypotheticalLow-grade cytology (ASC, LSIL)NANAPreferred type, mean utility*******• Early colposcopy(early colposcopy) 0.940********(repeated Pap smears) 0.932*********
      • Repeated Pap smears  
Le et al.30Abnormal testASCUS, LGSILUnknownNoPreferred interval for colposcopic observation, % per interval• Colposcopy every 6 months, twiceNR(less frequent colposcopy if HPV negative)
      • Colposcopy once per year if HPV negative 64%
Kitchener et al.29Abnormal testMild dyskaryosis or recurrent borderline cytologyBeforeYesPreferred type, % per type• Colposcopy(colposcopy) 56%(repeated Pap smear) 44%
      • Repeated Pap smears  
Philips et al.28HypotheticalMinor abnormalities on cytology (NOS)NANAPreferred type of screening programme, mean CV• Screening programme including repeated Pap smears for minor abnormalities(screening programme including HPV triage)NR
      • Screening programme including HPV triage for minor abnormalitiesAverage CV of a screening programme increased by 47% 
Howard et al.23HypotheticalMinor atypia on cytology (ASCUS)NANAPreferred type, mean utility**********• HPV triage(HPV triage)(repeated Pap smears)
      • Repeated Pap smears• if cytological abnormality resolves spontaneously: 0.9967 (95% CI: 0.9957–0.9978)• if cytological abnormality resolves spontaneously: 0.9972 (95% CI: 0.9964–0.9980)
       • if treatment is needed for cytological abnormality: 0.9354 (95% CI: 0.8544–1.0)• if treatment is needed for cytological abnormality: 0.9656 (95% CI: 0.9081–1.0)
Huang et al.20HypotheticalHPV positive/cytology negativeNANAPreferred frequency of repeated Pap smears, % preferring less than annual testing• Repeated Pap smears more often than once a year(more than once a year) 78%***********NR
Whynes et al.22Abnormal testBorderline or other low-grade abnormal cytologyAfterNo•% Satisfied with, and CV for a screening programme including own management strategy• Immediate colposcopy(immediate colposcopy)************(repeated Pap smears)************
     • CV for a screening programme including alternative management strategy• Repeated Pap smears• 94% satisfied or very satisfied with own strategy• 90% satisfied or very satisfied with own strategy
     •% to recommend own management strategy to a close friend • Mean CV own: £409.9• Mean CV own: £469.5
       • Mean CV alternative: £397.9, 14% valued repeated Pap smears less than colposcopy, and 7% more• Mean CV alternative: £514.3, 1% valued colposcopy less than repeated Pap smears, and 16% more
       • 82% would advise own strategy• 62% would advise own strategy
McCaffery et al.21Abnormal testMinor abnormal cytology (nonspecific minor changes with or without HPV effect)BeforeYesPreferred type, % per type• HPV triage(HPV triage) 65%*************(repeated Pap smears) 35%*************
      • Repeated Pap smears  
Kitchener et al.25Abnormal testHPV positive (1st)/cytology negative + repeated HPV positive (2nd)BeforeYesPreferred type, % per type• Immediate colposcopy(colposcopy) 62%(repeated HPV testing) 38%
      • Repeated (3rd) HPV testing  

Results

Search results

The PubMed search identified 2822 articles. We retrieved 26 full-text articles, and 14 were found to be relevant for the systematic review.11,14,19–30 No additional study was identified via the reference lists of the 14 articles. Of the 12 articles retrieved in full text but not considered to be relevant for the review, in one article women had histology-confirmed low-grade neoplasia rather than low-grade abnormalities on cytology,31 in eight articles women were not given a choice between active follow-up and observation, or the grade of cytological abnormality was not reported separately,4,32–38 and three articles39–41 were considered to be duplicate publications or included subsamples of the studies covered in the articles included in the review.21,24,25 Finally, one of the 14 articles that matched the inclusion criteria did not state the management choices unequivocally, and was subsequently excluded.27 In total, 13 articles were included in the review.

The variation in study designs and participants

Six studies were undertaken in the UK, three in the USA, two in Australia and two in Canada (Table 1). The inclusion criteria varied. Studies varied in size and the attained response rates (Table 2). In seven studies, women were relatively young, with a mean age not exceeding 35 years; in four studies, the mean age of women was above 35 years; and two studies did not report the mean ages. The overall age range in the studies with reported data was 14–80 years. The variations in women’s screening history and the levels of attained education were substantial. In studies with reported data, most women were married and/or cohabiting.

All studies presented women with some information regarding the nature of the cervical abnormality and/or the management strategies (Tables S1 and S2). In seven studies, the exact wording of this information was not reported (Table S1). Among the six studies that reported the exact wording (Table S2), all explained the relationship of cytological abnormalities and/or an HPV infection with cervical cancer. In five of these studies, it was explained that HPV is a very common and sexually transmitted infection. The characteristics and potential side effects of Pap and/or HPV tests, other triage tests, colposcopy/biopsy and treatment were not described systematically in all studies.

The definition of the low-grade screening abnormalities evaluated in the studies varied but, in 11 studies, the abnormalities were cytological, in one the abnormality was a positive HPV test with normal cytology and in one it was a repeated positive HPV test after an initial positive HPV test with normal cytology (Table 3). In eight studies, women themselves had this abnormality, whereas in five they did not, so that they revealed their management preferences for a hypothetical situation. In five studies, women revealed their preferences before undergoing follow-up for their own abnormality, whereas, in two studies, women had already gone through either immediate colposcopy or repeated Pap smears before they were asked to reveal their preferences, and, in one study, the timing was unknown. Subsequent to revealing their management preferences, women were managed with the preferred strategy in three studies. In two of the five studies that used hypothetical scenarios, women were told beforehand that the abnormality would be resolved successfully.

As active follow-up, six studies offered immediate colposcopy, three studies offered HPV triage, one study offered both and two studies offered repeated Pap smears at short intervals. As observation, 11 studies offered repeated testing with Pap smears and one study with HPV testing. One study offered colposcopy at six-monthly intervals as the active strategy, and less frequent colposcopy for HPV-negative women as the observational strategy. One study offered cervicography as the fourth option.

What women want

In the three studies in which the women could choose their management, about two-thirds opted for active follow-up (65,21 6225 and 56%29). In one study in which women with a first-time, low-grade, abnormal Pap smear were referred for colposcopy, 64% preferred this strategy, 18% preferred observation and 17% had no preference.26 When women were asked whether they would like to have repeated Pap smears earlier than recommended, i.e. earlier than in 6 months, 64% answered that they would prefer to be retested earlier than in the recommended 6 months (53% among women who had received an educational intervention, and 74% among women who had not),19 and 78% said that they would prefer to have repeated Pap smears more often than once per year, the recommended frequency.20 However, in another study in which women with an ASCUS/LGSIL smear had been referred immediately for colposcopy, 64% would have preferred less frequent colposcopy in case they were HPV negative.30

The two studies in which women with low-grade abnormalities were asked about their preferences after their follow-up had been completed both found a stronger preference for active follow-up. One study found that 84% of women would have preferred immediate colposcopy, regardless of whether they had gone through immediate colposcopy themselves, or were in fact managed with repeated Pap smears.11 In another study, women who had undergone repeated Pap smears were equally satisfied with their management as women who had undergone immediate colposcopy. However, significantly more women who had undergone repeated Pap smears (16%) placed a higher value on immediate colposcopy than the opposite way around (7%). The majority stated that they would recommend to their close friends the management they had received themselves. Of those recommending the alternative, significantly more women having undergone repeated Pap smears stated that they would recommend immediate colposcopy (38%) than the other way around (18%).22

In the three studies using hypothetical scenarios that asked women whether they would prefer active follow-up or observation, the preferred management appeared to depend on the severity of the abnormality being evaluated. In one study, 21% preferred active follow-up and 58% observation for ASCUS, whereas, for LSIL, 59% preferred active follow-up and 14% preferred observation.14 In another study, immediate colposcopy had a slightly higher overall average utility (0.940) than observation (0.932). However, if the scenario hypothesised that the abnormality would resolve spontaneously, women placed a higher value on observation than on active follow-up.24 In addition, in the third study, the initial abnormality was hypothesised to either resolve spontaneously or necessitate treatment, although no progression to cancer was foreseen. For the scenario with a spontaneous resolution, women placed a somewhat higher valuation on observation (utility 0.9972) than on active follow-up (utility 0.9967). The same was true when treatment with laser or wire loop was hypothesised to be required: utility 0.9656 for observation versus utility 0.9354 for active follow-up.23

Finally, HPV triage for low-grade cytological abnormalities increased the value of a screening programme by 47% relative to a screening programme with repeated Pap smears.28

Discussion

General findings

The most realistic scenario for the determination of women’s preferences is to ask women with abnormal tests about their choice before any management is started. Only five studies tested this situation. These studies unanimously showed that the majority of women, about two-thirds, preferred active follow-up (most often in the form of colposcopy) to observation (most often in the form of repeated Pap smears). One study showed that this proportion could be decreased somewhat if women received extra information on the distinction between precancer and cancer. Among women asked about their preference after their own management had been completed, the majority in one study were satisfied with the management they had received, but trended towards active follow-up when asked about what they would recommend to a friend. In another study, the majority would have preferred active follow-up. The hypothetical studies tested various choices, such as, for instance, repeated cytology versus HPV triage, and women, in general, preferred the most intensive management options. Our review therefore suggested that the management preference was dependent on several factors. For example, women may more often prefer observation when they are not having the abnormality themselves. Women, furthermore, appeared to be reassured when they were told that their risk of cervical cancer was low, e.g. in the form of a negative HPV triage test, and were then more willing to accept observation. Nevertheless, our review indicated overall that the majority of women preferred active follow-up, most often in the form of an immediate referral for colposcopy.

Some of the studies reviewed asked women to state their reasons for choosing a particular management. One main reason to choose active follow-up was that treatment was performed more rapidly.21,22 Another reason was that it gave a more definite result.22 A reason for choosing observation, however, was a smaller chance of having to undergo colposcopy and biopsy.21 Anxious women were more likely to choose active follow-up,21,26,28,29 as were women with children.21 Such follow-up might allow anxious women to cope better.29 The data were, to some degree, conflicting with regard to whether women with previous abnormalities preferred more active follow-up. Although, in three studies, this was the case,14,20,21 in another study, women with previous abnormalities preferred less frequent colposcopy, provided that they knew that they were HPV negative.30 These choices may reflect the variation in, and the dissatisfaction with, their previous management.21,30 One study suggested that preferences for the management strategy might not be related to knowledge on cervical abnormalities,26 whereas another study concluded that women had adequate knowledge when they made the choice.21 However, measurement of knowledge was not standardised among the studies.

Studies of women with histologically confirmed low-grade abnormalities, CIN1, also found a wide variation in management preferences. In one study, 43% of low-income women with histologically confirmed CIN1 chose immediate treatment with cryotherapy and 57% chose repeated Pap smears.42 In another study, women could choose between Pap smears at 6 and 12 months and an HPV test at 12 months as follow-up after histologically confirmed CIN1; 67% chose repeated Pap smears, 19% chose HPV testing and 13% were undecided.31 Furthermore, a study on women with any grade of Pap abnormality showed that they preferred to know their disease status immediately. They valued strategies with immediate diagnosis and treatment higher than strategies with immediate diagnosis and delayed treatment, and the latter were, in turn, valued higher than strategies with delayed diagnosis and delayed treatment.37

Limitations

The studies had several limitations. Firstly, most studies did not utilise random sampling. The results may therefore not be easily generalisable. Indeed, women in the study by Meana et al.26 had a high educational level because of the study site’s proximity to the university, whereas Whynes et al.22 suggested that the women included in their study may have been self-selected, to some degree, based on their motivation for participating in randomised trials. However, women included in the study by Philips et al.28,43 were comparable with the source population with regard to age distribution, smoking and screening behaviour, and various socioeconomic indicators, and, in the study by Howard et al.,23 the mean age of women included in the study was comparable with the mean age of women with low-grade cervical abnormalities and of women recommended for colposcopy. Secondly, in some studies, women were recruited from various health clinics, and two studies, for example, recruited women who had already been referred for colposcopy. This context could have affected women’s preferences towards active follow-up. Thirdly, some studies included very young girls. However, other studies also included older women, and the studies had an overall age range of 14–80 years. Fourthly, studies evaluated predominantly low-grade abnormalities on cytology, whereas we could find only two that evaluated the combination of a positive HPV test and normal cytology. Of these, one study only included women aged 50–80 years.20 Another study included women with a persistently positive HPV test;25 as HPV persistence is a stronger risk factor for cervical cancer than a single positive test,44 women’s perceptions may have changed accordingly.

The content of information could also have influenced the preferences for follow-up. The information presented to women prior to making a choice varied, although, in about one-half of the studies, the precise information was not published and could therefore not be examined. In several studies, women received the information that HPV is a sexually transmitted infection. Although testing HPV positive may not be associated with a higher degree of anxiety than testing positive on cytology,45 it appeared, in some studies, that testing positive for HPV was associated with psychological consequences relating to the sexually transmitted nature of the virus, and that HPV-positive women more frequently described feeling stigmatised and concerned about their sexual relationships.46–49 Knowing that HPV is a sexually transmitted infection could then make some women not prefer HPV triage. This could have been the case in the study by Howard et al.,23 where women valued active follow-up with HPV triage significantly less than observation with repeated Pap smears.

The 13 reviewed studies were all undertaken in English-speaking countries (six in the UK, three in the USA, two in Australia and two in Canada). As these countries have a common cultural heritage, the results might not be generalisable, although the values of women in these countries should reflect those of women in modern, westernised countries.

Future cervical screening

Until recently, the UK guidelines for women with borderline cytological changes recommended repeated Pap smears in 6–12 months, but immediate referral for colposcopy or repeated Pap smears after no more than 6 months for women with mild dyskaryosis.50 New guidelines recommend HPV triage for both cytological screening outcomes, with an immediate referral for colposcopy if the woman is HPV positive, and a return to routine screening if the woman is HPV negative.51 In the current US guidelines, women with ASCUS are recommended reflex HPV testing with subsequent colposcopy if they are HPV positive and repeated cytological testing in 12 months if they are HPV negative, although repeated Pap smears at 6 and 12 months, or immediate colposcopy, are also acceptable initial management options. Women with LSIL are recommended immediate colposcopy.52 Some management guidelines thus appear to be somewhat more conservative than the preferences reported by the majority of women. By not recommending repeated testing, however, the new UK guidelines appear to better match the preferences of the majority of women. However, it could be argued that this is a temporary solution, as cytology may, in the future, be replaced by HPV testing in primary screening. Nevertheless, it should be noted that most women with low-grade cervical abnormalities seem to be satisfied with their management regardless of the strategy.22 In the short term, low-grade screening abnormalities raise anxiety; however, this anxiety tends to decrease over time.11,22,29,53 Women may wish to participate in decision making,26,38 but may also be willing to let the physicians take the lead.20,26

If cytology is replaced by HPV testing in primary screening, twice as many women will have positive screening tests. Most, predominantly those with a positive HPV test and normal cytology, will be recommended repeated testing.54 However, judging from the reviewed studies of women with low-grade cytological abnormalities, the majority of women with positive HPV tests and normal cytology are expected to prefer immediate referral for colposcopy.

Immediate referral for colposcopy and repeated Pap smears have been shown to be equally effective in preventing cervical cancer7 and, based on UK data, the two management strategies are equally cost-effective.8 However, as immediate referral for colposcopy does not leave time for some abnormalities to regress, this management strategy might amplify the problem of overdiagnosis,5 and colposcopy may lead to after-effects, such as pain, bleeding and discharge.55 Furthermore, countries tend to have limited colposcopy capacities. It is therefore pertinent to find a solution as to how the expected doubling in the number of women with low-grade abnormal screening tests should be handled in future primary HPV screening. Increasing the colposcopy capacity will take resources away from other medical fields, and, moreover, an intensified colposcopy activity, from the point of view of overdiagnosis, might not be an optimal solution. As the guidelines tend to be somewhat more conservative than women’s preferences, better health education might be an option. Although this proved to be effective in one of the reviewed studies, it only changed women’s preferences for immediate follow-up from 74 to 53%. Furthermore, in the same study, the concern about waiting 6 months for repeated Pap smears was high regardless of whether women were counselled: 78% of women in the counselled arm versus 88% in the control arm described themselves as concerned.19 The best possible solution therefore seems to be to limit the number of women with positive primary HPV tests in need of follow-up. This could be achieved by increasing the standard cut-off points for positive HPV tests: for the Hybrid Capture 2 test, the cut-off could be increased to above the currently used ≥1 rlu/co (rlu/co, relative light unit/cut-off ratio);56 another promising option appears to be the use of HPV mRNA as a primary screening test,57 although further validation is needed before HPV mRNA can be considered for cervical screening.

Conclusion

Cytology screening is expected to be replaced by primary HPV screening. This will lead to a doubling of the number of women with abnormal screening tests. Under these circumstances, it may be a challenge to meet women’s expectations of optimal management.

Disclosure of interests

EL and MR are currently undertaking a comparative study of new-generation HPV tests, involving collaboration with Roche Diagnostics A/S, Genomica S.A.U., Qiagen Gaithersburg Ltd. and GenProbe Inc. Concerning the present article, there has been no collaboration with, or support from, any of the companies. EL has served as an unpaid scientific advisor to GenProbe and Norchip.

Contribution to authorship

MEF and MR designed the study, analysed the data, interpreted the results, drafted the manuscript and made the decision to submit. EL interpreted the results, drafted the manuscript and made the decision to submit. The guarantor was MR. All authors had full access to all of the data in the study.

Details of ethics approval

This is a systematic review of published studies, and ethical approval was not sought.

Funding

The Danish Strategic Research Council (case number 10-092793) provided funding, but had no role in the study design, collection of data, analysis, interpretation of the data, writing of the article and decision to submit the article for publication.

Acknowledgements

The authors would like to thank Dr Lene Borrits from the Danish Royal Library for operationalising the PubMed search using MeSH terms.

Ancillary