Him ORIGINAL CONTRIBUTION Policy Analysis of Cervical Cancer Screening Strategies in Low-Resource Settings Clinical Benefits and Cost-effectiveness Goldie, MD. MPH Context Cervical cancer isa leading cause of cancer-related death among women D in developing countries. In such low-resource settings, cytology-based screening isif- Ignette Denny, MMED(OAC), ficult to implement, and less complex strategies may offer additional options. BCOG(SA) Objective To assess the cost-ffectiveness of several cervical cancer screening strat- egies using population-specifc data Design and Setting Cost-effectiveness analysis using a mathematical model and a hypothetical cohort of previously unscreened 30-year-old black South African women, Screening tests included direct visual inspection (DVI) of the cervix, cytologic meth- ERVICALCANCERISALEADING — os, and testing for high-risk types of human papillomavirus (HPV) DNA. Strategies cause of cancer-related death — differed by number of clinical vst, screening frequency, and response to a positive among women in develop. test resut. Data sources included a South Alfican screening study, national surveys ingeountres with uptoso% and fee schedules, and published Itrature. offpatients presenting withadvanceddis- Main Outcome Measures Yea's of life saved (YLS), lifetime costs in US dollars, cease." In resource-poor areas, such as and incremental cost-effectiveness ratios (cost per YLS). Pollack, MD) Thomas C. Wright, MD sub-Saharan Africa, the growing num- Results When analyzing all strategies performed as a single lifetime screen at age ber of women infected with human im- 35 years compared with no screening, HPV testing followed by treatment of screen munodeliciency virus (HIV) may fur- postive women ata second vst, cost $39/YLS (27% cancer incidence reduction); DVI, ther compoun this problem because coupled with immediate treatment of screen-postve women at the ist vist was next they ha on inched otk at naman most effective (26% cancer incidence reduction) and was cost saving; cytology, fol- Haan EPL) lacrean the lowed by treatment of screen-postve women atasecond vist wasleast effective (19% et 7 amlection: !PS cancer incidence reduction) a a cost of $81/YLS. For any given screening frequency, causal agent of cervical cancer. when strategies were compared incrementally, HPV DNA testing generaly was more Cytology-based screening programs effective but also more costly than DVI, and always was more effective and less costly hhave decreased the incidence of inva- than cytology. When comparing all strategies simultaneously across screening fe- sive cervical cancer in many developed quencies, DVI was the nondominated strategy up toa frequency of every 3 years (in- counties."° However, cytologic screen- _crementalcost-effectiveness ratio, $460/YL5), and HPV testing every 3 years (incre- ing requires an established laboratory, mental costeffectiveness ratio, $11 500/YLS) was the most effective strategy. highly-trained eyotechnologiss,and up Conelusion Cervical cancer screening strategies that incorporate DVI or HPV DNA to 3 visits forscreening, evaluation of cy- testing and eliminate colposcopy may offer atiractive alteratives to cytology-based tologic abnormal results, and treat- screening programs in low-resource settings. ment, In low-resource settings, such a JAMA, 2001295:107-315, vw ama.com has proven difficult wo imple mentand sustain. Recently, ithas been colposcopy, allowing forsereeningand developed a comprehensive model ca suggested that less complex strategies treatment to be performed during the pable of assessing allernative screening may olferadditional options, Suchstra- same visit" These screening ap- strategies for cervical cancer im devel- cies replace the Papanicolaou sinear proaches have shown promise in large with simple visual screening methods, trials conducted in low-resource set- Aber Aflaons and Franc Dicer ied such as direct visual inspection (DVD) tings; however, determining whether UCeMgrtieatce go in which the cervix is viewed after the they will be clinically or cost-effective MO, MPH, Cente fr Rsk Arcs, Depatinet of i Heaith Poy and Management, Harvard Schoo of application ofan acetic acid solution, or requires formal evaluation using ana- ania and Managemen, Hana Sena of HPV DNA testing. Th also eliminate lytic modeling methods." Therefore, we MA02115-3524 (ema sldeahsph hava). (©2001 American Medical Association, A rights reserved. (Rept) JAMA, ane 27, 2001— 5.No.24 3107 Downloaded From: https:‘CERVICAL CANCER SCREENING STRATEGIES IN LOW-RESOURCE SETTINGS Figure 1. Overview of the Moael ‘feather nthe model incorporate cenval dave dala, human papllonavs (APO) nection sau, ‘dhuman inmanodetidenyvius (1) ecto status. Fach math women can pores repesin tht Cerviealdseae: those lowest rica disease progression have ne delete ol -sk HPV DNA and have HIV infection those thighs i of disease progression have detectable high-k types of HPV DNA an {einiater sages of HV infection ach month, worn who ate HVsnfecea may progress inthe WV de ‘se Nel shown ae women who nay de Yom cequted mnmunodefeny syndrome, cera cance or tet uses. SIL indeates squarous aegis ‘oping countries and used the model 1o- gether with country-specific data to con- duct a policy analysis comparing the clinical benefits and costeffectiveness of different cervical eancer screening strat- ‘egies in black South African women, METHODS: We developed a state transition com- puter-based model to stmmulate the nati ral history of HPV-induced cervical neo- plasia and cervical cancer screening, diagnosis, and treatment in acohort of previously unscreened 30-year-old black South African women, Model out- ‘comes included life expectancy and lif lime costs, Comparative performance of different screening strategies was measured by the incremental cost effectiveness ratio, defined as the ad- ditional cost ofaspecifiesereening stral- egy, divided by its additional clinical benefit, compared with the next least expensive strategy.” We adopted as0- ictal perspective (ie, all costs and ben- efits are included regardless to whom they accrue). Time preference was in- corporated by discounting costs and benefits 3% annually. The implica- Lions of alternative assumptions regard ing the natural history of cervical neo- plasia, effectiveness of screening and treatment, prevalence of HIV, and di- 3108 JAMA June 27,2001 Vol (Repintes) reet medical and time costs were eval ated in sensitivity analyses, Natural History Model Health states in the model incorporate cervical disease (normal, low-grade, and high-grade squamous intraepithelial le- sions [SILs], and local, regional, and dis- tant invasive cervical cancer), HPV DNA status (high-risk types of HPV DNA de- tected or not detected), and HIV di case (HIV-uninfected and HIV-infected with CD4 lymphocyte cell counts higher than 500 X10*/L, between 201-500 X10°/L, and 200 X10"/L and lower) (EIGURE 1). Transition probabilities, ex- Luapolated from published literature, are used to move women through different health states over tine, The time hor zon of the analysis incorporates a wom- ans entire lifetime and is divided into equal monthly increments (ie, Markov cycles) during which women transition from one health state to another. Tran- sition probabilities governing the natu ral history of cervical disease are condi- Ldonalon HPV DNA statsin that women with detectable high-risk HPV DNA are at greater risk of SIL progression than all other women. However, even among these omen, the risk of invasive ean- cer is not homogeneous, and we as sumed a proportion of them will have persistent HPV infection and beat great- fest risk." The HPY status also de- pends on the presence of HIV coinfec- lion, with HIV-infected women having ‘higher prevalence of HPV than those who were not HIV infected." In each cycle, women who are HIV-infected ean progress in their HIV disease. Women face competing mortality risks from ac quired immunodeficiency syndrome (AIDS), cervical eancer, and other causes. Screening Model and Screening Strategies Screening strategies were distin- {guished by the number of clinical vis- Its, the use of 1 or 2 screening tests, screening frequency, and ages to target for sercening. Screening tests included. DVI, cervical eytology using a conve ional Papanicolaou smear, and HPV DNA testing using a high-risk HPV probe (Hybrid Capture I; Digene Corp, Gaithersburg, Md). Three-visit strate- files, the standard ofcare in most devel- ‘oped countries, nelude an initial screen- ing evaluation, a second visit for a diagnostic workup incorporating col- poscopy and biopsy in women with post tive test results, and a third visit for treat- ment of women with confirmed SIL, Two-visil strategies incorporate a visit at which screening takes place fol- lowed by a second visit at which women ‘with positive test results undergo treat- ment, without evaluation by colpos- copy. One-visit strategies, (ie, sereenand treat on the same day) incorporate im- mediate treatment in all screen- positive women, In strategies incorpo- rating a combination of 2 tests, women ‘with an abnormal result on either test are considered screen positive. The sen- sitivity and specificity of the combina- lion of DVI and cytology were 96% and 82%, respectively; for DVLand HPV, they ‘were 07% and 75%, respectively, and for HPV and cytology, they were both 86%. In strategies incorporating 2 sequential tests, women with an abnormal result on the first test undergo a second test, and only those with positive results on both tests receive treatment. The sensitivity and specificity of DVI followed by tology were 50% and 90%, and 65% and (©2001 American Medical Association, All rights reserved.98% for HPV followed by cytology, re- spectively. Sereening could be con- ducted 1, 2, oF 3 times per lifetime or at any specified interval Thefollowing assumptions were made: ()allwomensuitable for outpatient reat- ment receive eryotherapy; (2) among, ‘women undergoing cryotherapy, 10%re- ceive nabenelitand another 1O%develop recurrent disease within 1 year"; (3) Sho women receiving cryotherapy have minor symptoms (eg, discharge, bleed- ing, and minor infection requiringaclinic visit and oral antibiotics), and 1% have a more serious complication (eg, infee- tion or bleeding) requiring 1 to2 days of hospitalization; (4) all sereen-positive ‘women undergo DVI prior tocryotherspy thusallowing thosewitha4-quadrantle- sion or suspicious cancer to be referred toaphysician; (5) 10% ofscreen-positive women havea 4-quadrantlesion orsus- picious cancer and are referred to a physician—those with cancer undergo slagingand teatment consistentwith eur- rent practicein South Africa,"and those ‘without eancer undergo a loop electro- surgical excision procedure (50%), con- ization (25%), orno treatment (25%); (6) 2%ofall women with high-gradedisease have undetected microinvasivecanceror adenocarcinoma in situ; (7) one third of ‘women with microinvasive cancer orad- ‘enocarcinoma in situare undetected, one third are relerted toa physician, and one third are treated with cryosurgery; (8) ‘women with microinvasive cancer, inap- propriately reated with eryotherapy,2e- ceive no clinical benelit from the proce- dure; (9) women with false-negative re- sults are only detected if they develop symptoms from invasive cancer; (10) ‘women with true-negative results expe- rience thecostsand complications elated. to the screening process but are eligible tobe screened again; and (11) thereisa 15% loss to follow-up with each visit. ss values (and plau- sible ranges) used for the base case. Clinical parameters were obtained from, four South African screening study, as ‘well as extrapolated from a systematic literature review, with preference given, CERVICAL CANCER SCREENING STRATEGIES IN LOW-RESOURCE SETTINGS ‘Table 1: Model Varabler bace-Cave Valuer ana Ranges Used Senay Anaser Variable ace Case, Ye Pegge: TTT HEV DNA detocteat 20 6.33 Towegade SCP 32 Tae High-grade SEE 1238 Teal vaste one 00e028 Fgnalinvasve caneer 010030 Doro O30 ean Tranaion Probables forthe Overall Population (HPV Statue Unknown) Progression ate ermal owgrade SL 30y 15.45 TowgadetonghgaceSE— By Ty Toso Low-grade SIL to normal eoindy 20.00 Fras es ‘ean Transition Probabiliee Conditional on High-Risk HPV DNA Pasitviys sm Prograssion rata with delectable persetent Tigre HEV CNA, ermal low-grade Sl. 35y 20. [owgyade to nghrrace ST Tay S520 Figh grade SE to rvasie career Zay Toso Fageason rate ih delectable prsterk Tighrisk HEV ON SL tonal {wade SIL to normal 2oindy 10-80 igh grade SE to normal Tonty 520 Frogiosson rata vith ro, out ‘iran detect HPV DNA Normal low-grade SI toy 0820 Towgyade to nghrgrace ST Tay Oso To cancer 7 zs thro, Ena “rain detctacle HPY DNA Tow-gade SIL To narmal eoindy 30.00 Figh-gade SE to normal aondy 15:60 ‘Screening Test Performance Cenycaleyclogy"*™* Senay 60 20.00 Spectiaty i aoa Dic vec napecton™™ Senay 2 Speci % PPV probe astay eincar-coleclag Sena 2 Speci PPV probe astay Gal caleciem ‘Senay 8 55.85 Speciity 3S Tea Sereenng testa saquanee Seay ca 7.00 e000 Winox compieaions jor compteatons 7 (©2001 American Medical Association, A rights reserved. (Reprints) JAMA, Jne 27,2001 Vol 295, No. 288108‘CERVICAL CANCER SCREENING STRATEGIES IN LOW-RESOURCE SETTINGS data from the National Cancer Insti- hhute's Surveillance, Epidemiology, and "I Model Variables Sane-Case Values an Ranges Used i Senstily Analyses” (Cont Vari Base Cane, End Results Survey were used in the ‘Gancer Stage Distribution == sensilivity analysis. Estimates for com- oe a peting mortality fom ager, sexe, and Sasa = Tace-ajuned al atscs were obiained = Irom life tables developed by the World ‘Year Survival Invasive Cervical Gance Health Organization for South Af- Lozi cance 73 Spartans mora oo ie rin" Separate all-cause morality tes Fes forthe HIV-infected and uninfected sec- Ostiriree ores a for he LIV infected and uninected s HIN-Related Monthy Mortality, per 100" venaitivity urelyses. For these analyses, cosets 500 a0 an sensitivity analyses For hese aac ‘we used 1989-1091 lifetables for black Females to represent the sector without HIV infection’ and the most conserva tive survival rates from previously vall- ‘Sites cevea rimepineianecsincs EN iingrawsest neaiersaveuroutnaianes dated natural history models for the sec- Psa fend Sera ans ccnona atu tor with HIV infection. >= tostudies with larger sample sizes, well- Second, aseries of equations were used Costs defined control groups, and longer fol- tosplitout these mean population prob- Micracosting techniques were used to Iow-up.'!°""" To establish a plausible abilities into HPV-specific probabili- estimate direct medical and nonmedi- range for screening test characteris- ties, based on the relative risk of SILin cal costs (TABLE 2)-!? Country lies, we gave priority ostudiesin which women with HPV (compared with the specific data expressed in 1999 South potential verification bias wasavoided."* relative risk of those without HPV) and Africa Rand were converted to US dol- estimates from studies particularly el. HPV prevalence.' "0" Finally, the — larsatan exchange rate of R6.2/S1.” We evant (0 our target population were not _ model was calibrated such that the pro- used the Representative Association of ‘within this plausible range, the range ected lifetime risk of cervical eancerin Medical Schemes recommended scale ‘was widened to accommodate them. unscreened women was consistent with of benefits for the public sector asa sur- published data rogate for most direet medical costs.” Natural History ‘Based on the mean prevalence rates of ‘The cost of the HPV test was based on We estimated the probabilities of SIL HPV (22%% overall) and HIV (8% ove progression and regression using 2 all) reported.in the South Africa seree methods. First, ransition probabilities ing study and the probability of HPV in indirect costs of similar laboratory tests. the quoted price of the test kit for de- veloping countries and manpower and conditional on HPV that were available HIV-infected women compared with We also included the costs of treat from the literature were extrapo- those who were not infected with HIV ment complications and physician r lated. *#6!8!9289424269 For probabili- (le, the rate ratio), weestimated that 78% ferrals of women ineligible for eryo- lies conditional on HIV, data froma pre- ofall detectable HPV was attributable to therapy. Resource use associated with viously published analysis of cervical women who were not HIV infected” each stage of cervical cancer was based, cancer screening in HIV-infected women We used a rate ralio of 3.8 for the base on clinical protocols from 2 regional were used.””* The probability of high- the probability of HPV in HIV hospitalsand the literature. °"" Costs agrade SIL progression to invasive can- infected vs uninfected women), and es- of hospitalization and radiation therapy ccerwas estimated using simulation tech- tablished plausible rangeof 3.510 5.6, were estimated from a cost-compari- niques as described previously."? A _ based on published studies, including 2 son study conducted at the University second method was used to estimale conducted in Africa." Similar of Cape Town,” and we assumed che- transition probabilities conditional on methods were used to estimate the pro-_ motherapy would not be available. For HPV that were not directly available portion of SIL and invasive cancer al- HIV-infected women, data from the from the literature. First, clinical stud-tributable to HIV-infected women and World Bank were used to estimate the ies lacking HPV status of individual those who were not cost of treating opportunistic infec- women were used tod Iransition probabilities fora hypothett- Mortality Rates Estimates of time spent traveling, calpopulationofwomen,someofwhom Stage-specific survival rates for cervical waiting, and receiving health care and. would be expected to acquire and clear cancer were based on data from a ter- the cost of transportation were based HPV, and some of whom would be ex-_tiary referral center in Cape Town and on 2 national surveys.*""? Time was pected to develop persistent HPV. published literat coverall mean lions and providing palliative care.” 215581 although valued using survey data in which 3190 JAMA June 27,2001 Vol 5, No 24 Rented) (©2001 American Medical Association. All rights reservedhouseholds were divided into quintiles based on mean income. Since our tar- {get population mainly Fell into the bo tom 2 quintiles, a mean hourly wage rate was estimated for all women using the mean monthly income for these lower 2 categories (R307 or $64) and assuming 20 working days per month. This approach was used because even though unemployment is between 40% and 60% for such women, nearly 40% of this sector of the population fs either a legal female headed household or a de facto female-headed household, and a sub- stantial proportion of the women hold jobs in the informal sector.*" RESULTS Cost-effectiveness of Cervical Cancer Screening at Different Intervals The discounted lifetime costsand life e pectancy of strategies performed at di Ferent screening intervals, beginning at 135 years of age, are shown in FIGURE 2. Cervical cancer screening increased dis- ‘counted life expectancy by 0.84 to 3.50 months, depending on the screening strategy. The cost-elfectiveness of mov- ing from one screening strategy to an- other more costly alternative is repre- sented by the difference in cost divided by the difference in life expectancy as- sociated with the 2 strategies Strategies Iying on the efficiency curve dominate those lying to the right of the curve be cause they are more ellective and either cost less oF have a more attractive cost effectiveness ratio than the next best strategy. A costeffectiveness ratio is shown for each nondominated strategy aand is the reciprocal of the slope of the line connecting the 2 screening strate- ‘es under comparison; thisslope will be steeper when the net gain in life expec- taney per US dollar is greater.” Asingle lifetime screen, usinga Lvisit strategy with DVI, was moreelfectiveand less costly than no screening (Figure 2) Screening with a I-visit DVL twice in a lifesime (at ages 35 and 40 years) pro- vided additional life expectancy ben- cfitsfor S70/YLS, compared with sereen- ing only once in a lifetime. Performing CERVICAL CANCER SCREENING STRATEGIES IN LOW-RESOURCE SETTINGS this strategy 3 times per lifetime (atages _YLS, compared with performing DVI 35, 40, and 45 years) rather than twice every 3 years. All 3-visit cytology str Increased mean lifeexpectancy by only egies (ie,sereening.at the first visit, per 3 days and was dominated by the DVI forming colposcopy at the second visi. strategy performed every Syears. Atmost and treating biopsy-confirmed SIL at a screening frequencies, a 2-visit HPV third visit) were dominated because they DNA testingstrategy wasdominated by cost more, although they were less elfec- amore frequent L-visitstrategy with DVL tive, than either Lvisit DVL oF 2-visit Testing for HPV every 3 years was the HPV DNA testing strategies. At all only nondominated HPV strategy, with screening frequencies, the conven- a cost-effectiveness ratio of $11500/ tional 3-vist cytology strategy resulted. ‘Table 2 Selected Corts nee Cave Valier tnd Ranges Used Senstity Anayse=™ age-Case Public ‘Sector Gost Ranget South atican| aS Wario Rand uss "uss Sacenng ea ‘Cerda eogy conventional Papariola tes 920 7.00-12.00 Tat 080-300 rect visual napecton FPVDNA hybrid assay 500-2005 Fel care coste™ ‘Screening vist + Ghoteapy vat ¥ Extended cine vat ¥ Diagpeste wort and weatren™ ‘Caposcopy + Biosys) + Gost or raoDgE RATES ¥ i ¥ GostoTRtoiogis mays t t + ¥ corzation (eo te) + ‘Selected patent tne cos mean Targa ‘Screening vist 1 (8:20 min + ‘Coposcopy vt 207 ¥ ‘Greeny va 201 + Treat with racket h + Tresimantwihradaton thrapy, 7770 rin ¥ ine, 112 (60-248) min ¥ relocate, 168 00-525) ih ¥ Wait 45 B6-90) rin ¥ Transporation be fr average Sve a ¥ Sct total weatrart ons (ect modi and to costs) “Chyatherary in patents sutable + For eutpatnt treatment Thea vasve cane canoar Tea Fgjonainvasve covial career 256-1074 Ditantinvacive canal cancer ‘Coma rivralated ath care Tresment of epportnate octions 1884 200 + Patave care aT 9 ¥ (©2001 American Medical Association, A rights reserved. (Reprintee) JAMA, Jne 27, 2001 Vt 205, No. 283141‘CERVICAL CANCER SCREENING STRATEGIES IN LOW-RESOURCE SETTINGS in lower reduction in cervical cancer imescreen usingaself-collected sample ogy were dominated by a 1-visit HPV incidence compared with a I-visit DVL for HPV would cost only $26/YLS com- DNA testing strategy because they cost (oF 2-visit HPV strategy’ pared with no screening. more, although they were less effe Allscreening tests may not be equally tive. All strategies that incorporated 2 available in low-resource settings, and Cost-effectiveness of Once sereening {esis either in combination oF certain sercening tests may be selected ina Lifetime Screening sequentially, either eost more and wer for programmatic reasons. For example, insomeregions,itmaybedifficultto pro- relative efficiency of different sereen- strong dominance) or had a less attrac- While Figure 2 provides insight into the less effective (le, thereby eliminated by vide the ongoing training necessary for ing strategies, the majority had cost- ive eost-elfectiveness ratio than other performing DVL Thereareanumberof effectiveness ratios beyond the thresh- more elective strategies (ie, thereby ‘ways in which HPV DNA testing could _ old ofthe poorest countries." Therefore, eliminated by weak dominance). These beincorporated intothescreeningstrat- weconductedananalysis comparing the single-lifetime strategies, when tar- egy in low-resource settings. In addi- screening strategies when performed _geted to women younger than age 30 tion to the 2-visit clinician-collected only once ina lifetime, at35 years ofage years or older than age 45 years, wer strategy used inthe base case,onecould (TABLE 3, upper half). Compared with never as cost-effective as targeting, consider using HPV testing of self no screening, a L-visit DVI strategy re- women in their mid-to-late 30s collectedsamples, Another feasiblestrat- duced the incidence of cancer by 26%, Because these screening strategies may egy is rapid processing of clinician- increased life expectancy by 0.84 notbe equally available in low-resource collected samples in the cli screeningand teatmentatthesame vist. mean total lifetime costs. L-visitstrat- also were calculated by comparing the allowing months, and was associated with lower settings, their costelfectiveness ratios Assuming these3 strategies wereequally egy using aclinician-collected sample or incremental costs and benefits of each a Levisit strategy with clini HPV reduced cancer incidence by 32% strategy with no screening (Table 3, ian-collected samples was more effec- and, compared with the DVI strategy, in- lower half). With the exception of a live and less costly than either of the creased life expectancy by 0.19 months _single-lifetime DVI, the cost-efective- acessib 2-visitstrateges. However, intheabsence for $LIS/YLS. All 2-vist strategies and ness ratios ranged from $14 to $147/ ofa L-visit strategy, a 2-vist single life- the 3-visit strategy using cervical eytol-_YLS, each compared with no screening. Figure 2. Decourted istrne Conan Berets of crening st Dferentntenaiewin Sensitivity Analysis {Vist DVL 2-Visit HPV DNA Testing of Cincian-Collected Samples, and 3-VisitCenical Results wer Ghelosy ral history of SIL, sensitivity, and cost of the screening tests, cost of cancer care, ‘most sensitive to the natu- 10. ae Te and HPV prevalence. However when we sone varied these parameters one ata time (i we Lay sensitivity analysis) overa plas re ovven sy crssoiem570 aiblerange the rank ordering ofsereen- » We. ing strategies didnot change i oomnn eens We condicted a series of sensitivity Hwee [ER | /vacimare analyses in which 2 and 3 variables were B or sean varied simultaneously (te, 2-and 3-way 4 ote fori sensitivity analysis). The choice b : Vi collected HPV strategy oF a single- 8 J sine Visit DVL performed once in lifetime ep Zs Tressaws ] depended on 3 things test cost, tstsen- seco] grr” | stuvity, and the costeffectiveness Ttcor| _Ubreshold (e, willingness to pay fora seco YLS). Givena cost-effectivenessthresh- ES SS OAS ETT Ee Te He old of S/VLS,a Lvist DVI was nearly ‘tec Tne Cots U8 8 always preferred toa 2-visit HPV DNA ‘inate: dee vaalnpecon, VS, year offend and HPV, honanpaplonavius ategasty. _‘esling the cost-elfectiveness thresh Ingonthe cen cave sorte hoe ng ote igh of cae beau te se moe and_ old i SSO/VLS, a I-visit DVI was only ‘her cot les or hoves more atacive con-efechenex than te nxt et stateey Compsreswth preferred toa 2-visit HPV testing ithe a urening te ncdence of cncer nae educed by 26% nthe tet Ov sategy bys nthe dat Preferred to a2-visit HPV testing ith Syste: ye nthe Ov every ya statey. oy 61% he Dviewey Sys Sey araby sensitivity of DVI exceeded 58%, and 5% inte HPV evry 3 yeas ststgy The cortices rato ete apes oe sope fe nell other variables were held constant. courage ewes Sep es onpaien AL this costeffectiveness threshold, 3112, JAMA june 27, 2001¥ol 5, No 24 Rented) (©2001 American Medical Association. All rights reserved