Two human papillomavirus (HPV) vaccines have recently been developed, Gardasil and Cervarix. Although Gardasil is licensed in North America, the European Union, and worldwide, Cervarix is not yet licensed in North America, but is licensed for use in the European Union, Mexico, Australia, and more than 60 countries worldwide. Both vaccines provide protection against HPV Types 16 and 18, which cause over 70% of cervical cancer cases.1 However, Gardasil provides additional protection against Types 6 and 11, which cause over 90% of genital warts.2 A further difference between the vaccines is that Cervarix incorporates a new adjuvant, AS04, and there is some evidence that this is associated with enhanced B cell priming and perhaps duration of effective antibody responses.3 Therefore, it is possible or likely that different healthcare providers may recommend or implement the use of different HPV vaccines, which may either give, or not give protection against HPV 6/11 disease. In order to evaluate the cost-effectiveness of different HPV vaccination strategies, the costs and benefits of each vaccine needs to be assessed.
HPV is the most common viral sexually transmitted infection in many countries worldwide, including the United Kingdom. Data from the United Kingdom suggest approximately 500 new presentations of genital warts per 100,000 people per year.4,5 Recent European data suggest even higher rates, with steadily increasing incidence over time.6 Young adults are most at risk and peak incidences occur in 16 to 19 year old females and 20 to 24 year old males.4 Treatments are either home- or clinic-based, and often require several repeat visits. After initial resolution, warts recur in a high proportion of cases.7
In order to incorporate data on costs and benefits of reducing genital wart incidence through vaccination, data on 3 features of the disease are required: (1) the cost of treating an episode of genital warts, (2) the impact of genital warts on quality of life (QoL), and (3) the period of time for which QoL is affected. These data can be used to estimate the quality-adjusted life years (QALYs) gained, and the potential resources saved, by preventing genital warts through HPV vaccination.
The study described here was carried out to investigate the burden (in both financial and QoL terms) of genital warts in a sexually transmitted disease (STD) clinic by addressing the 3 points above. We have previously reported the baseline QoL results,8 but in this article we present the cost of treatment data, estimate the period of time for which QoL is affected, and calculate QALYs lost per episode of genital warts. These results were used to parameterize a recently published cost-effectiveness analysis of bivalent and quadrivalent HPV vaccination in the United Kingdom.9 Other cost-effectiveness analyses of HPV vaccination strategies in the United States and the United Kingdom have also incorporated estimates of costs and QALY losses attributed to genital warts10–15 relying on a number of published studies for these estimates.12,16–22 We collected data from a UK-specific context as most of the previous cost estimates have been based on data from the United States and there was a lack of information on QALY loss attributable to genital warts collected directly from people with genital warts.
MATERIALS AND METHODS
The study was carried out amongst 2 patient populations from the York STD clinic. The first group consisted of patients with genital warts aged 18 years or older attending the clinic during a 3-month period in 2007 (April 16 to July15), who consented to complete QoL questionnaires during their clinic visit.
The second group consisted of patients diagnosed with either first-attack or a new episode of recurrent genital warts during the corresponding 3-month period in 2006. Data on this second group were collected to increase the amount of follow-up time available. Subjects were included in the duration and cost analyses only if their first study visit fell in one of the two 3-month study periods to avoid prevalent case bias. Informed consent was not sought from the second population, as because of the particular constraints regarding confidentiality in STD services, it was not appropriate to contact this group by phone or letter. The study was approved by the South Humber Research Ethics Committee.
Definition of Episode of Care
Both groups’ case notes were reviewed by a member of the research team. The episode of care (EOC) was defined as the period from the first visit to the last visit associated with this clinical presentation. A visit was considered as the final one for that EOC when the patient either (a) was declared free of warts by a clinician, or (b) had warts and was given either a clinic- or home-based treatment and did not subsequently reattend the clinic, or (c) had warts and did reattend with genital warts at a later date, but stated explicitly that their warts had resolved after the previous visit.
Data were collected for each visit (for genital warts) associated with the study EOC until the end of the data collection period in December 2007. For each visit the following data were collected: date of visit; KC60 code (statutory UK reporting codes for sexually transmitted infections (STIs): C11A = incident; C11B = recurrent; C11C = persistent case lasting over 3 months); member(s) of staff attending the patient; whether it was a first or follow-up visit and treatments and investigations received. Subjects were coded as either “warts cleared” or “warts not cleared” at the last visit of the EOC. It was also noted where a subject’s EOC was ongoing at the end of the data collection period.
Duration of EOC and Time QoL Affected by Genital Warts
The duration of an EOC was measured as the time between the first and last visits of the study episode. Where patients still had warts and received treatment at their last visit, it is unlikely that they cleared the day after leaving the clinic, or that their QoL was only affected while attending for treatment. In light of this, we carried out Kaplan-Meier survival analysis to estimate the period of time from first presentation at the clinic until likely clearance of warts. As a high proportion of the sample still had warts at their last clinic visit, 2 models were considered:
- Lower-bound scenario: only subjects whose EOC was ongoing at the end of the data collection period were right censored (even if the subject still had warts at their last visit of that EOC we assumed, there was no further impact on QoL).
- Upper-bound scenario: those coded as warts not cleared were also right censored (it was assumed that there was an impact on QoL after the end of the EOC for those who still had warts at their last clinic visit).
Cost per EOC
Costs were measured from the healthcare provider perspective and only costs associated with clinic visits were included. The types of procedures undergone by each patient (consultation for a first visit, consultation for a follow-up visit, treatment using trichloroacetic acid, cryotherapy, curettage, electrosurgery, and health advice) were collected from the case note review. The time taken for each procedure was estimated by interviews with nine members of the clinical team. Costs of staff time and equipment were obtained from standard national sources and local purchasing data. Distributions and sources of the unit costs are shown in Table 1(2007 prices, converted from GB pounds sterling to US dollars using Her Majesty’s Revenue and Customs exchange rate for December 200723). These parameters were used to estimate the overall cost per EOC. Uncertainty in the overall cost per EOC incorporates variation in the types of procedures undergone in each EOC, as well as the uncertainty in unit costs and times (assumed to be normally distributed) applied to each type of procedure. Comparisons were made between groups (on the basis of sex, whether the EOC was for incident or recurrent warts, and treatment pattern parameters) using 2-tailed t tests.
Cost per EOC
The mean cost of an EOC was estimated at $286 (£139, 95% CI: $246–$327). The mean cost per EOC for females was higher than for males (P = 0.74) and first episodes had a higher mean cost per EOC than recurrent ones (P = 0.43) (Table 2), although these differences were not statistically significant. The mean number of visits per EOC was 2.8, and the median 2, the distribution in the number of visits being highly skewed with 46% attending for only 1 visit. Over 80% of those who attended for only 1 visit had been provided with home treatment (either podophyllotoxin or imiquimod). Only 6 (3%) cases overall had EOC lasting more than 6 months. The diagram in Figure 1 provides a schematic representation of treatment patterns per visit for this sample, and the resultant costs per visit. Visits to specialist physicians, and those where imiquimod was prescribed were associated with a higher cost than those with other staff or treatments. These data are, however, descriptive and should not be used to indicate cost-effectiveness of any of the different staff or therapy choices.
Time QoL Affected by Genital Warts
One hundred eighty-nine individuals (96 males, 93 females) were included in the duration and treatment cost analyses. Thirty-five (19%) subjects’ warts had cleared at their last visit for the study EOC, and 154 (81%) had not cleared. The mean duration of an EOC (with no censoring) was 41 days.
Using the models described in the methods section concerning the duration of time for which QoL is likely to be affected, upper- and lower-bound estimates were produced. The survival curves for both scenarios are shown in Figure 2. One subject had an EOC which was considered to be ongoing at the end of the data collection, and was thus subject to censoring in the lower-bound estimate. The mean duration (in days) for the lower bound was 42 (95% CI: 32–53), with a median of 11 days. The mean duration for the upper bound was 213 days (95% CI: 155–271), median 162 days.
Loss of QALYs
QALYs lost due to genital warts were calculated by multiplying the duration of time for which QoL is affected for by the QoL lost due to genital warts. We have previously reported baseline scores and comparison to a sample of the UK population, where we observed a difference of 0.039 (95% CI: 0.005–0.068) in EQ-5D index score in 18 to 30 year olds.8
Applying the 2 mean estimates of duration for which QoL is affected for provided here, the QALY loss attributable to genital warts was 0.0045 (95% CI: 0.0014–0.0078) for the lower-bound and 0.023 (95% CI: 0.0072–0.039) for the upper-bound scenario. The 95% confidence intervals represent a combination of the uncertainty distributions in both QoL detriment due to warts and in the duration of time for which QoL is likely to be affected (due to intersubject variation in survival times).
We have presented data that can be used as part of economic evaluation of HPV vaccination strategies; QALYs lost due to genital warts could potentially have a significant impact on cost-effectiveness.
Our estimate of the average cost of treatment for a genital warts episode is consistent with the previous study carried out by Langley et al.,19 who reported an average cost of $274.07 for males and $306.95 for females per EOC. Our estimate is somewhat lower than that of Brown et al.,20 who reported an average treatment cost per episode of $390.49 (all costs inflated to 2006/07 prices using hospital and community health services inflation index and converted from GB pounds sterling using Her Majesty’s Revenue and Customs exchange rate for December 200723,24). Differences may be due to variation in study design or staffing and treatment patterns in different clinic settings. Large variability in treatment patterns has been observed19 and hence the treatment patterns in this clinic described in Figure 1 should be considered if our estimates are used to inform the national picture about the burden of genital warts. Also, our study only measured costs associated with visits to an STD clinic. It has been estimated that up to 16.5% of all incident cases of genital warts among males and 20.9% of those among females in the United Kingdom are diagnosed and treated in primary care.5 It is therefore likely that treatment provided outside of STD clinics will be a significant additional economic burden in the UK context. As such, any national estimates of treatment costs should also include costs associated with visits to other healthcare providers.
Our estimate of length of EOC (41 days) was low in comparison with previous studies. Insinga et al.16 reported a mean length of EOC of 3 months and Wilson et al.25 just over 2.5 months (for 4–10 warts). These previous estimates should be interpreted in the context of their time and location: Insinga et al. carried out their study in the United States based on data from 1998, whereas Wilson’s study was conducted in the United Kingdom but relating to 1996/1997 data. Although our single-centre study is likely to be affected by local treatment patterns, our low estimate of length of EOC may reflect actual changes in clinical practice, with the widespread shift to home-based treatment modalities.
We produced upper- and lower-bound estimates of QALYs lost due to genital warts. These estimates have been used in a previously published economic evaluation of HPV vaccination in the United Kingdom.9 Other estimates of QALYs lost, which have also been used in economic evaluations of HPV vaccination are broadly consistent with our own, with alternative estimates having been based on a QoL loss of 0.05 to 0.1, and a duration of time affected ranging from 2 to 6 months.10–12,14 The resultant QALY loss using these alternatives is closer to our upper-bound scenario.
We produced a range of estimates of QALYs lost as there were considerable uncertainties involved, particularly in estimating the duration of time for which QoL is affected. The difference between our estimates and the wide confidence intervals reflect this uncertainty. First, there was a large loss to follow-up in terms of the case note review; less than 20% of the sample was declared free of warts at their final visit. In the past decade, a shift from clinic-based to home-based treatment has been seen in the United Kingdom and elsewhere, in line with European guidelines.26 The large proportion of people attending for a single visit, and the fact that only a small proportion of subjects had cleared their warts at the final visit of the EOC, appropriately reflect current treatment patterns and the frequent use of home treatment in this GUM clinic. While consistent with expected treatment patterns, these features of the data are problematic for estimating the duration of time for which QoL is affected. Those attending for only one visit essentially provide no data to inform the models and there is no information on what happens to individuals who still have genital warts after they leave the clinic. It is possible that this group do not return to the clinic because they no longer have symptoms, in which case our lower-bound estimate might be the most appropriate. However, the converse is also possible; patients’ warts may persist well after their last clinic visit. Other circumstances such as the inconvenience, cost or embarrassment of attending an STD clinic, or a change in relationship circumstances could feasibly discourage patients from remaining in active treatment.
Second, although we present 1 case as the upper bound, there will be a tendency toward underestimation in our estimates as we did not include any data relating to how long subjects’ warts had been present before first presenting at clinic. UK-specific data on treatment-seeking delay associated with genital warts is sparse. Median estimates of delay associated with STDs in general range from 7 to 30 days.27–30 As warts are not generally a painful condition, it is feasible that delay will be longer for warts than for other STDs. Further investigation of this issue would be needed to understand whether QoL is affected from the start of a genital warts episode, or if decreasing QoL is a trigger for treatment-seeking behavior.
Third, as the survival analysis is essentially a way of estimating time to clearance after first presentation at clinic, our results are based on the underlying premise that estimating the duration of genital warts is a valid proxy for estimating the time for which QoL is affected. This is not necessarily the case, as the level of impact on QoL may not be related solely to the presence or absence of symptoms; other factors such as age, sex, relationship status, perceptions of the diagnosis, and experience of treatment are likely to be related to the impact on QoL. There are no data on how QoL may be affected after successful treatment of an STI, but it is feasible that anxieties about recurrence and transmission contribute in some cases to a continued detriment to QoL even after clearance. The follow-up QoL data presented here are not conclusive on this issue. Furthermore, there is likely to be considerable variation in experiences of genital warts, as demonstrated by the range of responses to a disease-specific QoL questionnaire.8
Future research aimed at improving the accuracy of our estimates of duration would need to consider what happens before presenting at clinic as well as undertaking active follow-up for a time period sufficient to capture the end of most episodes, and some time after clearance. It would also be beneficial to carry out the analyses across multiple sites, to improve the generalizability of results.
Despite the challenges presented by the nature of the condition, its treatment and the population of interest, our study offers the most comprehensive estimate to date of the economic and QoL impact of genital warts, in the context of current clinical practice in the United Kingdom. Genital warts have the potential to have a significant impact on cost-effectiveness of HPV vaccination, particularly when comparing the bivalent and quadrivalent vaccinations. Information on the burden of genital warts should be considered in decisions about implementation of HPV vaccination.
1. Bosch X, de Sanjosé S. Chapter 1: Human papillomavirus and cervical cancer—burden and assessment of causality. J Natl Cancer Inst Monographs 2003; 31:3–13.
2. Brown DR, Schroeder DM, Bryan JT, et al. Detection of multiple human papillomavirus types in condylomata acuminata lesions from otherwise healthy and immunosuppressed patients. J Clin Microbiol 1999; 37:3316–3322.
3. Giannini SL, Hanon E, Moris P, et al. Enhanced humoral and memory B cellular immunity using HPV16/18 L1 VLP vaccine formulated with the MPL/aluminium salt combination (AS04) compared to aluminium salt only. Vaccine 2006; 24:5937–5949.
4. Health Protection Agency. STIs Annual Data 1997–2006— Data Tables. Avaliable at: http://www.hpa.org.uk/web/HPAweb&HPAwebStandard/HPAweb_C/1195733775264
. Accessed April 2008.
5. Cassell JA, Mercer CH, Sutcliffe L, et al. Trends in sexually transmitted infections in general practice 1990–2000: Population-based study using data from the UK general practice research database. BMJ 2006; 332:332–334.
6. Kjaer SK, Tran TN, Sparen P, et al. The burden of genital warts: A study of nearly 70,000 women from the general female population in the 4 Nordic Countries. J Infect Dis 2007; 196:1447–1454.
7. Lacey CJ, Lowndes CM, Shah KV. Burden and management of non-cancerous HPV-related conditions: HPV 6/11 disease. Vaccine 2006; 24(suppl 3):S3/35–41.
8. Woodhall S, Ramsey T, Cai C, et al. Estimation of the impact of genital warts on health-related quality of life. Sex Transm Infect 2008; 84:161–166.
9. Jit M, Choi YH, Edmunds WJ. Economic evaluation of human papillomavirus vaccination in the United Kingdom. BMJ 2008; 337:a769.
10. Elbasha EH, Dasbach EJ, Insinga RP. Model for assessing human papillomavirus vaccination strategies. Emerg Infect Dis 2007; 13:28–41.
11. Kulasingam SL, Benard S, Barnabas RV, et al. Adding a quadrivalent human papillomavirus vaccine to the UK cervical cancer screening programme: A cost-effectiveness analysis. Cost Eff Resour Alloc 2008; 6:4.
12. Brisson M, Van de Velde N, De Wals P, et al. The potential cost-effectiveness of prophylactic human papillomavirus vaccines in Canada. Vaccine 2007; 25:5399–5408.
13. Dasbach EJ, Insinga RP, Elbasha EH. The epidemiological and economic impact of a quadrivalent human papillomavirus vaccine (6/11/16/18) in the UK. BJOG 2008; 115:947–956.
14. Chesson HW, Ekwueme DU, Saraiya M, et al. Cost-effectiveness of human papillomavirus vaccination in the United States. Emerg Infect Dis 2008; 14:244–251.
15. Kim JJ, Goldie SJ. Health and economic implications of HPV vaccination in the United States. N Engl J Med 2008; 359:821–832.
16. Insinga RP, Dasbach EJ, Myers ER. The health and economic burden of genital warts in a set of private health plans in the United States. Clin Infect Dis 2003; 36:1397–1403.
17. Chesson HW, Blandford JM, Gift TL, et al. The estimated direct medical cost of sexually transmitted diseases among Am youth, 2000. Perspect Sex Reprod Health 2004; 36:11–19.
18. Alam M, Stiller M. Direct medical costs for surgical and medical treatment of condylomata acuminata. Arch Dermatol 2001; 137:337–341.
19. Langley PC, White DJ, Drake SM. The costs of treating external genital warts in England and Wales: A treatment pattern analysis. Int J STD AIDS 2004; 15:501–508.
20. Brown RE, Breugelmans JG, Theodoratou D, et al. Costs of detection and treatment of cervical cancer, cervical dysplasia and genital warts in the UK. Curr Med Res Opin 2006; 22:663–670.
21. Myers ER, Green S, Lipkus I. Patient preferences for health states related to HPV infection: Visual analogue scales versus time trade off elicitation. In: Proceedings of the 21st International Papillomavirus Conference; February 20–26, 2004; Mexico.
22. Stratton K, Durch J, Lawrence R, eds. Vaccines for the 21st Century: A Tool for Decision Making. Washington, DC: National Academy Press; 2000.
24. Curtis L. Unit Costs of Health and Social Care 2007. PSSRU: Kent. 2007.
25. Wilson JD, Brown CB, Walker PP. Factors involved in clearance of genital warts. Int J STD AIDS 2001; 12:789–792.
26. Von Krogh G, Lacey CJN, Gross G, et al. European course on HPV-associated pathology: Guidelines for primary care physicians for the diagnosis and management of anogenital warts. Sex Transm Infect 2000; 76:162–168.
27. Mercer CH, Sutcliffe L, Johnson AM, et al. How much do delayed healthcare seeking, delayed care provision, and diversion from primary care contribute to the transmission of STIs? Sex Transm Infect 2007; 83:400–405.
28. Pitts MK, Woolliscroft J, Cannon S, et al. Factors influencing delay in treatment seeking by first-time attenders at a genitourinary clinic. Int J STD AIDS 2000; 11:375–378.
29. Gott CM, Rogstad KE, Riley V, et al. Delay in symptom presentation among a sample of older GUM clinic attenders. Int J STD AIDS 1999; 10:43–46.
30. Cassell JA, Brook MG, Mercer CH, et al. Treating sexually transmitted infections in primary care: A missed opportunity? Sex Transm Infect 2003; 79:134–136.
31. Br National Formulary 2007 BNF 54. Br Medical Association, London.