Health Education Research, Vol. 18, No. 3, 318-328,
June 2003
© 2003 Oxford University Press
Human papillomavirus and the value of screening: young women’s knowledge of cervical cancer
Trent Institute for Health Services Research, and Schools of 1 Nursing and 2 Economics, University of Nottingham, Nottingham NG7 2RD, UK.
* To whom correspondence should be addressed;E-mail: david.whynes{at}nottingham.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
The study reports a questionnaire survey of female university students intended (1) to delineate their knowledge of cervical cancer and screening, and (2) to impute their valuation of the introduction of human papillomavirus (HPV) testing. It was found that almost 80% of respondents thought cervical cancer was a leading cause of cancer death amongst women. Most subjects consistently over-estimated the incidence of cervical cancer, consistent with the social amplification hypothesis. The major risk factors were accurately identified by the majority, although family history was emphasized to a degree unwarranted by epidemiological evidence. Subjects’ knowledge of the screening programme was accurate in some respects, but not in others, e.g. the majority under-estimated the abnormality rate and over-estimated the age at which abnormalities typically occurred. Subjects placed a premium on the increased accuracy which HPV testing might potentially offer, predictable from their beliefs about the performance of the existing screening programme. Policy makers can be reassured that female university students appear unconcerned about undergoing a test for a sexually transmitted disease, although their beliefs about cervical cancer and the power of screening are distorted, and their enthusiasm for HPV testing relates to an, as yet, unconfirmed belief in its accuracy.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Screening for cervical cancer was introduced in England and Wales in the late 1950s, initially on a local and discretionary basis. Until the 1980s, the take-up rates were low and variable by region, although indirect evidence has been used to infer that even the limited screening of that period was successful in reversing the trend of an increasing risk of disease (Herbert, 2000
). It was decided to re-launch the screening programme in a more concerted fashion and, in 1988, a national, computerized, call–recall system was introduced. As a result, all women aged between 20 and 64 years are now contacted regularly and directly by mail, inviting them to make an appointment to undertake a Papanicolaou (Pap) smear test. Invitation typically occurs every 3 years, although some local health authorities invite over longer intervals, up to a maximum of 5 years (McGahan et al., 2001). The screening service has always been provided at no direct cost to women and is funded by the National Health Service (NHS).
Most cervical screens take place in primary care settings and are conducted either by the subject’s General Practitioner (GP) or by a qualified member of the practice staff. The system of GP remuneration for screening was revised in 1991. A simple fee-for-service for each screen undertaken was replaced by a differential incentive payment, triggered by the achievement of 50 or 80% compliance within the GP’s eligible population. The post-1988 initiatives clearly met with some success, because overall compliance with cervical screening doubled between 1988 and 1994, from 42 to 85% (Quinn et al., 1999). Cervical cancer mortality fell by 25% between 1992 and 1997, and mortality reductions attributable specifically to screening between 1988 and 1997 have been estimated at around 60% for those aged under 55 years (Sasieni and Adams, 1999).
Whilst conceding the technical success of the recent cervical screening programme in the UK, concern has been expressed that it has been achieved largely in the absence of ‘informed consent’ (Austoker, 1999). Informed consent is the process whereby subjects are made fully aware of the relevant medical condition and the proposed intervention, including its potential benefits and limitations, and current uncertainties surrounding either. On the basis of this information being available, the subject is in a position to judge whether or not to accept the proposed intervention. The low profile accorded to informed consent in cervical screening to date appears to have been repeated for screening more generally. A recent systematic review of 190 interventions intended to enhance screening compliance noted that the provision of information on risk and benefit was reported in only four cases: ‘whether informed choice affects actual levels of uptake, therefore, remains yet to be fully evaluated’ [(Jepson et al., 2000), p. 97].
The case for informed consent in cervical screening is generally argued on two grounds. First, informed consent has an ethical basis, i.e. the acceptance that the autonomy of the subjects of medical procedures, and a respect for that autonomy, is of significant value in itself (Entwistle et al., 1998). Second, and more practically, uninformed women may have unrealistic expectations of the screening programme. They may be falsely reassured by normal results and may ignore subsequent warning signs, resulting in heightened distress at the onset of unexpected disease and, on occasions, inappropriate litigation (Raffle et al., 1995; Anderson and Nottingham, 1999). Similar consequences have been reported for breast cancer screening (Barratt et al., 1999). A fuller knowledge of the capabilities of screening might also serve to reduce the negative psychosocial effects associated with the receipt of an abnormal result (Fylan, 1998), in particular the uncertainty and anxiety caused by follow-up and treatment (Bell et al., 1995; Kavanagh and Broom, 1997).
At present, the cervical screening programme is unable to distinguish between minor abnormalities that are likely to progress to more severe disease and those which, most probably, will not. It is now understood that certain strains of the human papillomavirus (HPV) are central to the aetiology of cervical cancer; indeed, HPV has been shown to be present in ‘virtually all cervical cancers’ [(Walboomers et al., 1999), p. 12]. Accordingly, there has emerged an interest in using supplementary HPV testing within the screening programme as a means of improving its overall accuracy. It is believed by some that the use of HPV testing in parallel with conventional Pap smears could enable the programme’s sensitivity to be raised from around 70%, as at present, to virtually 100% (Watson, 2001). In other words, the current proportion of false-negative results should be reducible from around 30% to zero. A recent UK review of HPV testing within the cervical screening programme (Cuzick et al., 1999) was more circumspect, however. It warned that, were HPV testing to be applied to a general population, superior sensitivity would be accompanied by lower specificity, owing to the high prevalence of purely transitory HPV infection which resolves without consequence. The false-negative rate, in other words, would rise, but so too would the false-positive rate, implying unnecessary alarm to many and an increased cost of follow-up investigations. Moreover, the psychosocial implications of an HPV diagnosis remained to be established. The review expressed more confidence that HPV testing might usefully provide evidence of likely disease progression in the case of mild/borderline abnormal smears. This conclusion has led to observational studies (Rebello et al., 2001) and the initiation of pilot projects across the UK (Wise, 2000), specifically with respect to women receiving mild/borderline results. It is hoped that HPV triage in this group might reduce the number of repeat smears and colposcopy referrals, thereby reducing worry and unnecessary treatment (Lerman and Rimer, 1993), and securing cost economies.
Were HPV testing to be introduced as a standard screening procedure in the future, even if only for women recording mild/borderline smear results, the nature of this virus would presumably form an important additional component of informed consent. However, given that the principal transmission mechanism of HPV is sexual, women receiving HPV testing as an adjunct to screening would, in effect, also be being tested for a sexually transmitted disease (STD). Knowledge that this would be occurring might be expected to impact upon the acceptability of the procedure (Sevin, 1999).
Possibly because screening has such a long history in England and Wales, and the fact that the informed-consent-to-screen debate is relatively recent in origin, there are few primary studies of women’s knowledge of cervical cancer and their understanding of the characteristics of the screening programme. We present here the results of a survey of young women, aimed to assess their knowledge base of both. This we take to be informative in itself but, thereafter, we use the knowledge base to explore individual valuations of screening using the contingent valuation or willingness-to-pay (WTP) method (Mitchell and Carson, 1989).
WTP is being used with increasing frequency in health economics studies (Olsen, 1997; Klose, 1999) and the rationale is as follows. Any individual’s decision about future behaviour is likely to be complex and idiosyncratic. It will follow from the person’s personal circumstances, values, tastes and preferences, dispositions and beliefs with respect to the potential behaviour in question. Despite this nominal complexity in decision making, however, most people seem to be able to resolve many such problems with comparative ease. In purchasing everyday commodities, such as items of food or clothing, individuals resolve all relevant or salient decision factors into a single, subjective, monetary value or ‘reservation price’, which may be above or below the price requested by the retailer. If their reservation price is equal or higher, they purchase the commodity and, if below, they will not.
The WTP concept relies on the property of familiarity, by suggesting that, if individuals can identify value-for-money for day-to-day purchases, they are able to identify a value or reservation price even in cases where the commodity is not normally traded. The WTP valuation which a person offers thus embodies, in a single, holistic total, all the independent decision factors which that person has deemed relevant, weighted as has been deemed appropriate. In this study, we explore the extent to which young women’s WTP valuations can be explained with respect to their beliefs and perceptions of cervical cancer and screening.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Our sample was drawn from female students at the University of Nottingham, east-central England. The low average age of the sample implied that only a minority had actual screening experience, but we feel that analysing responses from this group is of particular value, for three reasons. First, the sample is, in comparison with the national population, homogenous and many potentially confounding variables, such as age, income and educational level, are being implicitly controlled. Second, were routine HPV testing to be introduced at some point in the future, it is currently young people who would be principally affected. Third, studying this group makes our findings directly comparable with those of other studies of young (pre-screening) university and college women.
A questionnaire was constructed, piloted for ease of completion and subsequently sent to a random sample of 500 students. All students were resident on the University campus, and the questionnaire was delivered and returned through the (free) internal mail system. Each questionnaire was anonymous and confidentiality was thus assured. All respondents were requested to provide routine socio-demographic information, including smoking habits, screening history (if any) and perceived risk of contracting cervical cancer.
Knowledge of cervical cancer and of screening was assessed principally by means of multiple-choice questions. Subjects were required to identify from a list of six options an annual population risk of the disease, decreasing from ‘1 in 100’ to ‘1 in 15 000’. They were provided with an un-labelled bar chart and told that the bars indicated the proportion of annual female cancer deaths arising from breast, bowel, cervical and lung cancers. They were then required to label each of the bars on the basis of their beliefs as to the identity of each. The bar chart to be labelled is reproduced as Figure 1, and actually displays, from left to right, the contribution to total female cancer mortality from cervical, breast, lung and bowel cancer (Office of National Statistics, 1999). The young women were offered a list of 11 possible risk factors for cervical cancer, in the following order: number of pregnancies, multiple sexual partners, weight, early first intercourse, smoking, height, use of oral contraception, early first pregnancy, Chlamydia infection, family history of cervical abnormality and HPV infection—and asked to identify those which represented major risks.
|
Knowledge of the screening programme was assessed from multiple-choice questions, specifically:
- When does the Department of Health recommend you have a cervical smear? (options: at least every 5 years from the age of 20; at least every 10 years from the age of 20; after the age of 40; after the age of 50).
- In most cases an abnormal cervical smear result indicates: (options: cancer; pre-cancerous changes).
- A normal smear result means: (options: low risk of developing disease; no risk of developing disease; I don’t need to have any further smear tests).
- What proportion of women who have smears each year have an abnormal smear result? (five options: 1; 5; 10; 20; 30%).
- At what ages are women most likely to have an abnormal cervical smear result? (options: 18–25; 25–35; 36–50; over 50 years).
- How many cases of cervical cancer in the UK do you think are prevented by the NHS cervical screening programme each year? (options: less than 1000; 1000–4000; 4000–9000; more than 9000).
- How accurate do you think the cervical smear test is? (options: 99% or better; between 50 and 98%; between 25 and 50%; less than 25%).
To obtain a baseline WTP valuation, subjects were requested to supply a value for a single Pap smear test, using a payments scale from which one of the following was to be selected: £0, £10, £20, £30, £40, £50, £75, £100, more than £100 (specify). Thereafter, they were required to augment this valuation under two sets of circumstances. As it is hoped that HPV testing will be able to target the screening programme more effectively, we asked respondents to indicate how much extra they would be willing to pay for a 10% increase in screening accuracy. They were then invited to revise this estimate if such increased accuracy necessarily entailed testing for an STD.
We hypothesized that subjects’ valuations of increased accuracy would be influenced, amongst other things, by their knowledge both of cervical cancer and of the capabilities of the screening test. We accordingly formulated a regression model including six independent variables, with WTP for increased accuracy as the dependent variable. Four of the independents were knowledge or perception related and chosen because of intuitively plausible associations with test accuracy. First, women perceiving themselves to be at higher-than-average risk of cervical cancer would be expected to value more accurate screening more highly than would women perceiving themselves at average or lower risk. A positive sign on the regression coefficient would therefore be predicted. Second, we should expect that women who under-estimate the proportion of abnormal smears routinely detected by the Pap test would place a lower incremental value on increased accuracy, because a belief that abnormalities are rare implies that the test’s information content would be, to them, insubstantial. Third, women who over-estimate the accuracy to the existing Pap test, i.e. perceive it to be high already, would presumably attach lower values to further increases in accuracy, other things remaining equal. Fourth, young women who believe that most abnormalities occur at a younger age should place a higher premium on accuracy than those who believe that abnormalities typically occur in an age group distant from them, as the risk proximity would be perceived as being greater.
The remaining two variables included in the model were characteristics. Smokers are likely to possess a lower health motivation generally and thus a reduced interest in screening. Indeed, such a negative association has already been identified in other forms of cancer screening (Vernon et al., 1990; Vernon, 1997). A negative association was thus predicted between smoking and incremental WTP. Finally, the baseline WTP value was included as an independent, to represent all other factors contributing to the individual’s judgement about cancer, screening and valuation. The sign on this coefficient should be positive, on the grounds that those with a higher initial value would be expected to have a correspondingly higher incremental value for the same test with improved characteristics.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Of the 500 distributed, 222 completed questionnaires were returned, implying a response rate of 44.4%. The sample’s age range was 18–23 years (mean 18.9 years, SD 0.79), with 60.4% aged 19 years. The majority (93.7%) declared their ethnicity as white. None had children and only one was married, although around half (50.9%) reported as being in a relationship. The majority (79.6%) received annual incomes of less than £5000 per annum, with most of those remaining (18.5%) reporting incomes of £5–10 000. Virtually all (96.8%) were studying for an undergraduate university degree, the remainder having already achieved a degree qualification. Good or excellent health (as opposed to fair or poor) was reported by 85.6%. Most (77.9%) declared they had never smoked cigarettes and 8.6% reported themselves to be current smokers.
Around one-third of respondents (37.8%) had experience of cancer in a close family member and most (93.7%) felt their risks of contracting cervical cancer were similar to, or less than, women of their own age. Around a quarter (26.6%) declared they were not at all worried about getting cervical cancer. The majority (59.0%) admitted they were ‘a bit’ worried, with the remainder being ‘quite’ or ‘very’ worried. Perhaps unsurprisingly, those perceiving themselves to be at higher risk of cervical cancer expressed higher degrees of worry (
2 = 23.3, P < 0.01). Those reporting a close-family history of cancer were less likely to perceive themselves as being at below-average risk of cervical cancer than those reporting no such history (10.7 versus 25.4%, 2 = 8.5, P = 0.01) and more likely to express at least some worries about developing the disease (79.8 versus 69.6%, 2 = 8.6, P = 0.04).
Although 30.6% stated that they had heard of HPV, only a minority (17.6%) had undergone a cervical smear test, as might be expected in a young population. Screening experience, however, appears to be an insensitive variable—for none of the analyses relating to knowledge of cervical cancer and screening presented below was there a significant difference in response between those with screening experience and those without (2, P = 0.1 or higher).
Cervical cancer incidence is approximately 1 in 10 000 women per annum in England and Wales at present (NHS Cancer Screening Programmes, 2001). Only 5.9% of the sample correctly identified this figure from the list provided. The vast majority (92.3%) over-estimated the incidence, with 13.1% believing the incidence to be 1 in 100. The bar chart (Figure 1
) permitted 24 possible labelling combinations, with only one, of course, being correct. All 24 combinations were observed amongst the responses, although not all were equally represented. The two most commonly selected combinations were both incorrect in all respects, i.e. bowel, lung, cervical and breast (24.0%), and bowel, lung, breast and cervical (17.8%), from left to right, respectively. Only a small minority (12.5%) correctly labelled the first bar as cervical cancer and two-thirds (67.3%) labelled it as bowel cancer. The labelling of all four bars was completely correct in only seven cases (3.4% of the sample).
Although the question posed by Figure 1 was related to mortality risks for women generally, it seems worthy of conjecture that respondents might actually have made their judgements on the basis of perceived risks amongst women of around their own age. In other words, the young women might have attributed their perception of their own relative risks to the female population as a whole. Accepting this possibility, and given relative mortality in the 15–29 year age group, the ‘correct’ labelling for Figure 1 would become lung, cervix, breast and colorectal, from left to right. Even allowing for this interpretation, however, the accuracy of identification remained equally poor. Only 9.6% of the sample identified cervical cancer as the biggest cause of cancer mortality (second bar) and only one respondent offered the ‘correct’ combination for the15–29 year age group.
Table I reports young women’s knowledge of cervical cancer risk factors. For those offering four or fewer risk factors, multiple partners and family history of cervical cancer were those most frequently chosen, and 46.1% of this group identified both. When women identified more than four factors, these first two continued to be chosen, although others were reported with increasing frequency. Of those on our list, the major risk factors for cervical cancer accepted by authorities such as the American Cancer Society (American Cancer Society, 2001) and the NHS Screening Programme (NHS Cancer Screening Programmes, 2001) are smoking, age at first intercourse, multiple sexual partners and HPV infection. These were all in the ‘top five’ identified by the full sample. Understandably, identifying HPV as a risk factor was significantly more common amongst those who had heard of it (2 = 59.1, P < 0.01); 89.7% of the latter cited HPV. Family history of cervical abnormalities is not cited by the above authorities as a major risk factor, although it was the factor most commonly cited by our sample.
|
Despite the age of the women being investigated and their consequent inexperience with the cervical screening programme, virtually all respondents (96.8%) were correct in believing that women between the ages of 20 and 64 are expected to attend for a cervical smear test at least every 5 years. Just seven women believed that wider intervals of screening were recommended. The vast majority were also aware of the meaning of normal and abnormal results, i.e. that a normal result means a low risk of developing cervical cancer (90.5%) and that an abnormal smear result usually means that pre-cancerous cell changes are present (91.9%). Eighteen women believed that an abnormal smear result usually meant cancer and 20 thought that a normal result meant no risk of disease.
The smear abnormality rate is a decreasing function of age and thus highest in the 18–25 year range (Department of Health, 2000). In response to being asked the age range where abnormal smear results are most likely to occur, only 7.7% of our sample selected that range from amongst the choices available and 34.2% selected the 25–35 years range. The majority thus believed that abnormal smears were more common amongst age groups distant from their own, with 49.1% selecting the 35–50 years range and 9.0% selecting the ‘over 50’ option. In practice, approximately 8% of adequate screening tests yield abnormal results and 28.8% of the sample correctly identified this. The majority (50.9%) under-estimated the abnormality rate, with 9.0% believing it to be as low as 1%. The remainder (24.3%) over-estimated the rate, 5.9% believing it to be as high as 30%.
The true accuracy of the Pap smear test cannot be identified with precision, as it is governed by human variability in pathological diagnosis and the practical variability of the pattern of repeated smears. Sensitivities as low as around 50% have been reported for single tests, up to the ‘high 90s’ for a battery of repeated tests (Cox, 1999). We were thus constrained to offer our sample a wide nominal range with respect to assessing test accuracy, and 76.5% selected the option of between 50 and 98% accurate in detecting disease. However, one in five women (21.3%) believed the test to be 99% accurate or better, with only a small minority (2.2%) believing the accuracy to be lower than 50%. Almost half (48.6%) believed that between 1000 and 4000 cases of cervical cancer are prevented each year by the screening programme, consistent with current estimates (Sasieni et al., 1996). The remainder, however, over-estimated this number, with 11.3% believing that more than 10 000 cases would be prevented each year. Subjects who over-estimated the number of cases prevented by screening were also more likely to over-estimate the accuracy of the smear test than those giving correct or under-estimates (29.0 versus 15.7%, 2 = 5.6, P = 0.02). By implication, those who perceive the greatest threat from cervical cancer appear to place the greatest faith in the effectiveness of screening.
The mean baseline WTP offered by the subjects was £24.8 (SD 22.1) and the mean increment offered for increased accuracy was £15.3 (SD 18.0). As noted earlier, the regression model explaining WTP for increased accuracy included six independent variables. Several were entered as binary variables, necessitating re-coding the data. First, we dichotomized the proportion of smears testing normal simply as under-estimation versus the remainder, given our hypothesis that the under-estimators would see little value resulting from increased accuracy. Second, we re-coded the accuracy data to distinguish specifically those over-estimating accuracy, for which a negative association with increased WTP had been predicted. Third, we dichotomized abnormality age into the two age bands closest to our subjects (18–25 and 25–35 years) to distinguish these from the two more distant age bands, hypothesizing that valuations would be lower if greater risk was perceived to be occurring amongst, in effect, a different generation. Table II displays the regression results. The signs for all coefficients are as hypothesized, although only three of the coefficients achieve statistical significance. Uniformly low variance-inflating factors suggests an absence of collinearity.
|
With respect to the second WTP augmentation—valuation for a more accurate test which also identified an STD—over half of the sample (55.2%) provided precisely the same value as for the first augmentation. Around a third (36.5%) were actually willing to pay more, whilst only a minority (8.4%) revised their estimate downwards. The mean increment in the WTP for increased accuracy entailing an STD test was £18.3 (SD 20.8). This was significantly higher than for the mean value of increased accuracy alone (t = -4.0, P < 0.01) and implies a modest average positive valuation on testing for an STD.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Four main conclusions can be derived from this study. First, most of the young women consistently over-estimated the significance of cervical cancer as a cause of female mortality, many of them to a considerable degree. The major risk factors were correctly identified by most, although family history was emphasized to a degree unwarranted by the epidemiological evidence. Second, the young women’s knowledge of the management of screening was broadly accurate, despite the fact that the majority had yet to participate in the programme. Virtually all appeared to be familiar with the call–recall protocol and the meaning of a normal result, for example. Knowledge of the more technical characteristics of screening, however, was less well developed, e.g. the majority under-estimated the abnormality detection rate and over-estimated the age at which abnormalities typically occurred. Third, the young women placed a premium on increased screening accuracy. Although not completely robust statistically, the regression result lends support to the view that individual valuation of an accuracy premium is logically consistent with beliefs about the screening programme, even if those beliefs themselves are inaccurate. Finally, testing for an STD, when coupled with superior cervical screening performance, appeared to pose acceptability problems to only a small minority, when judged by manifested WTP. Indeed, a sizeable proportion valued such testing its own right.
As noted earlier, the fact that our sample was confined to university students makes generalizations about the female population as a whole problematic. However, the restriction makes our findings comparable with those of studies of college and university populations in the US (Ramirez et al., 1997; Hasenyager, 1999; Yacobi et al., 1999; Baer et al., 2000; Gerhardt et al., 2000). Broadly speaking, the conclusions of these studies accord with those of our own, i.e. that young women are relatively familiar with screening procedures, such as local protocols and the meaning of test results, but less familiar with specific details, such as disease prevalence and risk factors. They also show that only a minority of women has any knowledge or awareness of HPV in relation to cervical cancer.
Only one of these studies (Ramirez et al., 1997) addressed the HPV valuation issue specifically and discovered that most subjects would be unfavourably disposed towards the HPV test, feeling that such a test would reveal information about their sexual behaviour. In our study, implicit STD testing within cervical screening appeared to be far less of a problem, with most young women being either unconcerned or positively valuing this additional information. This having been said and despite their apparent confidence in responding to our valuation question, it would seem probable that even our subjects would experience a degree of distress as a result of a positive HPV diagnosis, were one actually to occur. There is no reason to suppose, however, that such distress would be different from that engendered by any STD test result (Conaglen et al., 2001) nor that anticipated distress, in itself, is necessarily a source of negative value in screening.
Published assessments of knowledge of cervical cancer or of screening are rare, except in cases where the introduction of a new screening programme is being contemplated (Buga, 1998; Cotter et al., 1999) or where non-compliance in existing programmes is to be explained (Larsen and Olesen, 1998; Neilson and Jones, 1998). One previous UK study (Slater, 2000) which included women of all ages, produced, where similar questions were being asked, findings essentially similar to ours. As with our women, interpretations placed on normal and abnormal results were largely correct, although, in that study, a smaller proportion over-estimated the number cases being prevented each year by screening. The majority of subjects in that study were unaware that the mortality rate from cervical cancer was declining.
Surveys in both the US (Breslow et al., 1997) and Australia (Smith et al., 1999) have demonstrated that the accuracy of knowledge of cancer risks is positively associated with educational attainment. Assuming such a finding to be transferable across countries, our sample could be taken to represent the better-informed element of the local population, with the inference that knowledge of cervical cancer would be poorer in the wider community. As with our study, these surveys showed that family history was the most commonly identified risk factor for cervical cancer, followed by multiple sexual partners.
It is worth observing that the family history as a risk factor is qualitatively different from, say, sexual behaviour and smoking. The former implies images of genetic predisposition, whereas the latter are potentially controllable behavioural factors. Health education, in the form of the provision of information, can influence risky sexual behaviour, at least in the short term. Very little research, however, appears to have been undertaken specifically with respect to the sexual transmission aspect of cervical cancer, although it might be concluded that a positive information effect with respect to sexual behavioural modification can carry through to the transmission of HPV (Shepherd et al., 2000). According to our findings, however, health information needs to counter strong presuppositions on the part of young women, who presently perceive their cervical cancer risks to be both inherited and behavioural. The message to be imparted is that risks are, in reality, primarily behavioural, significantly shifting responsibility for cervical cancer risk onto individuals.
The general over-estimation of cervical cancer risk exhibited by our sample is consistent with the social amplification hypothesis, i.e. that repeated attention paid to cervical screening via the media and primary care makes individuals believe that the likelihood of an event occurring is considerably higher than it is in reality (Smith and McCloskey, 1998). Precisely the same phenomenon has been observed in the case of cigarette smoking’s contribution to lung cancer and the related mortality risk, with smokers typically believing their habit to be far more damaging than the epidemiological evidence warrants (Viscusi, 1999). The hypothesis also explains the failure of our sample of young women to identify bowel cancer as a major mortality risk. Although the large bowel is one of the most common cancer sites in countries throughout Europe, mass screening for the disease has only been publicly contemplated in England and Wales over the past few years (Brooks, 1998). Social amplification is, of course, particularly valuable in maintaining high compliance with screening, because individuals will be better disposed towards screening if they believe the disease to be widespread and the screening method effective (Raffle, 2001).
As noted earlier, the evidence base for the relation between knowledge and screening behaviour remains slight (Jepson et al., 2000), although studies are slowly emerging. Two investigations of subjects’ response to increased information in the case of compliance with colorectal cancer screening, for example, indicated either a small positive (Pignone et al., 1999) or non-existent (Wolf and Schorling, 2000) effect. As an additional complication, it is becoming increasingly clear that the provision of information is actually not neutral, in that individuals appear to respond differently to the manner in which the ‘facts’ are conveyed to them (Sarfati et al., 1998). In a study of screening for pancreatic cancer, subjects appeared far less willing to participate when they were told about the test’s low accuracy rate, the need for hospital investigation of positive results and poor post-treatment prognosis (Domenighetti et al., 2000).
Although most were yet to participate in cervical cancer screening, the positive WTP valuations offered by our sample of young women presumably indicate a likelihood of participation in the future. However, we speculate that, were our respondents to be correctly informed about cervical cancer and screening, many of their individual valuations of screening would alter, and this change in valuation could exert a subsequent impact on screening participation. Unfortunately, the direction of the change is difficult to predict in our particular case. On the one hand, informing young women that their cervical cancer risk is actually far lower than most currently appear to believe, and that the test is less accurate than some appear to believe, should lower the overall propensity to comply with screening in the future, other things remaining equal. On the other, the realization that they, and not the older generation, are in the highest-risk group for abnormalities, might be expected to stimulate higher compliance in a young population.
|
Acknowledgments |
|---|
This research was conducted within the Trial of Management of Borderline and Other Low-grade Abnormalities (TOMBOLA) randomized control trial, funded by the UK Medical Research Council. The authors are grateful for valuable comments from three anonymous referees.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
American Cancer Society (2001) Cervical Cancer: Prevention and Risk Factors [Online]. Available: http://www3.cancer.org/cancerinfo/load_cont.asp?st=pr&ct=8&language=english
Anderson, C. M. and Nottingham, J. (1999) Bridging the knowledge gap and communicating uncertainties for informed consent in cervical cytology screening: we need unbiased information and a culture change. Cytopathology, 10, 221–228.[CrossRef][Web of Science][Medline]
Austoker, J. (1999) Gaining informed consent for screening. British Medical Journal, 319, 722–723.
Baer, H., Allen, S. and Braun, L. (2000) Knowledge of human papillomavirus among young adult men and women: implications for health education and research. Journal of Community Health, 25, 67–78.[CrossRef][Web of Science][Medline]
Barratt, A., Cockburn, J., Furnival, C., McBride, A. and Mallon, L. (1999) Perceived sensitivity of mammographic screening: women’s views on test accuracy and financial compensation for missed cancers. Journal of Epidemiology and Community Health, 53, 716–720.[Abstract]
Bell, S., Porter, M., Kitchener, H., Fraser, C., Fisher, P. and Mann, E. (1995) Psychological response to cervical screening. Preventive Medicine, 24, 610–616.[CrossRef][Web of Science][Medline]
Breslow, R. A., Sorkin, J. D., Frey, C. M. and Kessler, L. G. (1997) Americans’ knowledge of cancer risk and survival. Preventive Medicine, 26, 170–177.[CrossRef][Web of Science][Medline]
Brooks, A. (1998) Pilot bowel screening scheme to start in Britain. British Medical Journal, 317, 769.
Buga, G. A. (1998) Cervical cancer awareness and risk factors among female university students. East African Medical Journal, 75, 411–416.[Web of Science][Medline]
Conaglen, H. M., Hughes, R., Conaglen, J. V. and Morgan, J. (2001) A prospective study of the psychological impact on patients of first diagnosis of human papillomavirus. International Journal of STD and AIDS, 12, 651–658.
Cotter, A., Greene, R. and Turner, M. (1999) What do Irish women know about cervical screening? Irish Medical Journal, 92, 428–429.[Web of Science][Medline]
Cox, J. T. (1999) Evaluating the role of HPV testing for women with equivocal Papanicolaou test findings. Journal of the American Medical Association, 281, 1645–1647.
Cuzick, J., Sasieni, P., Davies, P., Adams, J., Normand, C., Frater, A., van Ballegooijen, M. and van der Akker, E. (1999) A systematic review of the role of human papillomavirus testing within a cervical screening programme. Health Technology Assessment, 3, 1–204.
Department of Health (2000) Cervical Screening Programme, England: 1999–2000. Department of Health, London.
Domenighetti, G., Grillia, R. and Maggi, J. R. (2000) Does provision of an evidence-based information change public willingness to accept screening tests? Health Expectations, 3, 145–150.[CrossRef][Medline]
Entwistle, V. A., Sheldon, T. A., Sowden, A. and Watt, I. S. (1998) Evidence-informed patient choice. International Journal of Technology Assessment in Health Care, 14, 212–225.[Web of Science][Medline]
Fylan, F. (1998) Screening for cervical cancer: a review of women’s attitudes, knowledge and behaviour. British Journal of General Practice, 48, 1509–1514.
Gerhardt, C. A., Pong, K., Kollar, L. M., Hillard, P. J. and Rosenthal, S. L. (2000) Adolescent’s knowledge of human papillomavirus and cervical dysplasia. Journal of Pediatric and Adolescent Gynecology, 13, 15–20.[CrossRef][Medline]
Hasenyager, C. (1999) Knowledge of cervical screening among women attending a university health centre. Journal of American College Health, 47, 221–224.[Web of Science][Medline]
Herbert, A. (2000) Cervical screening in England and Wales: its effect has been underestimated. Cytopathology, 11, 471–479.[CrossRef][Web of Science][Medline]
Jepson, R., Clegg, A., Forbes, C., Lewis, R., Sowden, A. and Kleijnen, J. (2000) The determinants of screening uptake and interventions for increasing uptake: a systematic review. Health Technology Assessment, 4, 1–123.
Kavanagh, A. M. and Broom, D. H. (1997) Women’s understanding of abnormal cervical smear test results: a qualitative interview study. British Medical Journal, 314, 1388–1391.
Klose, T. (1999) The contingent valuation method in health care. Health Policy, 47, 97–123.[CrossRef][Web of Science][Medline]
Larsen, L. P. and Olesen, F. (1998) Women’s knowledge of and attitude towards organized cervical smear screening. Acta Obstetrica et Gynecologica Scandinavica, 77, 988–996.[CrossRef]
Lerman, C. and Rimer, B. (1993) Psychosocial impact of cancer screening. Oncology, 7, 67–79.
McGahan, C. E., Blanks, R. G. and Moss, S. M. (2001) Reasons for variation in coverage in the NHS cervical screening programme. Cytopathology, 12, 354–366.[CrossRef][Web of Science][Medline]
Mitchell, R. C. and Carson, R. T. (1989) Using Surveys to Value Public Goods: The Contingent Valuation Method. Resources for the Future, Washington, DC.
Neilson, A. and Jones, R. K. (1998) Women’s lay knowledge of cervical cancer/cancer screening: accounting for non-attendance at cervical screening clinics. Journal of Advanced Nursing, 28, 571–575.[CrossRef][Web of Science][Medline]
NHS Cancer Screening Programmes (2001) Cervical Cancer—Incidence, Mortality and Risk Factors [Online]. Available: http://www.cancerscreening.nhs.uk/cervical/risks
Office of National Statistics (1999) Mortality Statistics: General (Series DH1 30). The Stationary Office, London.
Olsen, J. A. (1997) Aiding priority setting in health care: is there a role for the contingent valuation method? Health Economics, 6, 603–612.[CrossRef][Web of Science][Medline]
Pignone, M., Bucholtz, D. and Harris, R. (1999) Patient preferences for colon cancer screening. Journal of General Internal Medicine, 14, 432–437.[CrossRef][Web of Science][Medline]
Quinn, M., Babb, P., Jones, J. and Allen, E. (1999) Effect of screening on incidence of and mortality from cancer of the cervix in England: evaluation based on routinely collected statistics. British Medical Journal, 318, 904–908.
Raffle, A. E. (2001) Information about screening—is it to achieve high uptake of to ensure informed choice? Health Expectations, 4, 92–98.[CrossRef][Medline]
Raffle, A. E., Alden, B. and Mackenzie, E. F. (1995) Detection rates for abnormal cervical smears: what are we screening for? Lancet, 345, 1469–1473.[CrossRef][Web of Science][Medline]
Ramirez, J. E., Ramos, D. M., Clayton, L., Kanowitz, S. and Moscicki, A.-B. (1997) Genital human papillomavirus infections: knowledge, perception and actual risk in a nonclinic population of young women. Journal of Women’s Health, 6, 113–121.[Web of Science][Medline]
Rebello, G., Hallam, N., Smart, G., Farquharson, D. and McCafferty, J. (2001) Human papillomavirus testing and the management of women with mildly abnormal smears: an observational study. British Medical Journal, 322, 893–894.
Sarfati, D., Howden-Chapman, P., Woodward, A. and Salmond, C. (1998) Does the frame affect the picture? A study into how attitudes to screening for cancer are affected by the ways benefits are expressed. Journal of Medical Screening, 5, 137–140.
Sasieni, P. and Adams, J. (1999) Effect of screening on cervical cancer mortality in England and Wales: analysis of trends with an age period cohort model. British Medical Journal, 318, 1244–1245.
Sasieni, P., Cuzick, J. and Lynch-Farmery, E. (1996) Estimating the efficacy of screening by auditing smear histories of women with and without cervical cancer. British Journal of Cancer, 73, 1001–1005.[Web of Science][Medline]
Sevin, B. U. (1999) Social implications of sexually transmitted cancer. Journal of Womens’ Health and Gender-based Medicine, 8, 759–766.[CrossRef]
Shepherd, J., Peersman, G., Weston, R. and Napuli, I. (2000) Cervical cancer and sexual lifestyle: a systematic review of health education interventions targeted at women. Health Education Research, 15, 681–694.
Slater, D. N. (2000) Are women sufficiently well informed to provide valid informed consent for the cervical smear test? Cytopathology, 11, 166–170.[CrossRef][Web of Science][Medline]
Smith, B., Sullivan, E., Bauman, A., Powell-Davies, G. and Mitchell, J. (1999) Lay beliefs about the preventability of major health conditions. Health Education Research, 14, 315–352.
Smith, D. and McCloskey, J. (1998) Risk communication and the social amplification of public sector risk. Public Money and Management, 18, 41–50.
Vernon, S. W. (1997) Participation in colorectal cancer screening: a review. Journal of the National Cancer Institute, 89, 1406–1422.
Vernon, S. W., Laville, E. A. and Jackson, G. L. (1990) Participation in breast screening programs: a review. Social Science and Medicine, 30, 1107–1118.
Viscusi, W. K. (1999) Commentary on public perception of smoking risks. In Jeanrenaud, C. and Soguel, N. (eds), Valuing the Cost of Smoking. Kluwer, Boston, MA, pp. 151–164.
Walboomers, J. M. M., Jacobs, M. V., Manos, M. M., Bosch, X., Kummer, A., Shah, K. V., Snijders, P. J. F., Peto, J., Meijer, C. J. L. M. and Munoz, N. (1999) Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. Journal of Pathology, 189, 12–19.
Watson, R. (2001) European women’s group calls for human papillomavirus testing. British Medical Journal, 323, 772.
Wise, J. (2000) UK pilot scheme for HPV testing announced. British Medical Journal, 320, 600.
Wolf, A. M. D. and Schorling, J. B. (2000) Does informed consent alter elderly patients’ preferences for colorectal cancer screening? Journal of General Internal Medicine, 15, 24–30.[CrossRef][Web of Science][Medline]
Yacobi, E., Tennant, C., Ferrante, J., Pal, N. and Roetzheim, R. (1999) University students’ knowledge and awareness of HPV. Preventive Medicine, 28, 535–541.[CrossRef][Web of Science][Medline]
Received on January 8, 2002; accepted on April 23, 2002
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. A. Tiro, H. I. Meissner, S. Kobrin, and V. Chollette What Do Women in the U.S. Know about Human Papillomavirus and Cervical Cancer? Cancer Epidemiol. Biomarkers Prev., February 1, 2007; 16(2): 288 - 294. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Anhang, A. Goodman, and S. J. Goldie HPV Communication: Review of Existing Research and Recommendations for Patient Education CA Cancer J Clin, September 1, 2004; 54(5): 248 - 259. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||