Costs and Effectiveness of Alternative Strategies for Cervical Cancer Screening in Military Beneficiaries

Maxwell, G. Larry MD; Carlson, Jay W. MC; Ochoa, Mark MC; Krivak, Tom MC; Rose, G. Scott MC; Myers, Evan R. MD, MPH

Obstetrics & Gynecology:
Original Research

OBJECTIVE: To estimate the potential effects, on costs and outcomes, of changes in sensitivity and specificity associated with new screening methods for cervical cancer in the military.

METHODS: A Markov model of the natural history of cervical cancer was created to simulate a cohort of 100,000 military beneficiaries aged 18–85. Probability estimates for various outcomes and the accuracy of screening tests were obtained from the literature. Cost estimates were obtained from military sources where available; otherwise, civilian costs were used. The outcomes and costs of conventional cytology, liquid‐based cytology, and liquid‐based cytology with human papillomavirus (HPV) triage were compared at 1‐, 2‐, and 3‐year screening frequencies.

RESULTS: Marginal reductions in the incidence of cervical cancer from increasing screening sensitivity are greater than reductions in cancer mortality at every screening interval. Incremental improvements in both cancer incidence and mortality are higher at less frequent screening intervals. Increases in the ratio of low‐ to high‐grade lesions result from increasing the sensitivity of the screening test or shortening the screening interval. Both liquid‐based cytology and liquid‐based cytology with HPV testing are cost effective (less than $50,000 per life‐year saved) when performed at 3‐year screening intervals. However, neither strategy is cost‐effective when performed more frequently than every 3 years.

CONCLUSION: Use of a more sensitive cervical cancer screening test increases costs. However, a more sensitive test performed less frequently may be more effective and less expensive than conventional cytology done annually. In the military setting, this has significant implications for both expense reduction and readiness enhancement.

In Brief

More sensitive cervical cancer screening performed less frequently may be more effective and less expensive than the conventional cytology done annually.

Author Information

Division of Gynecologic Oncology, Walter Reed Army Medical Center, Washington, District of Columbia; and Department of Obstetrics and Gynecology, Duke University Medical Center, Durham, North Carolina.

Address reprint requests to: Evan R. Myers, MD, MPH, Duke University Medical Center, Department of Obstetrics and Gynecology, Durham, NC 27710.

ERM has received unrestricted grant funding in the past from Digene Inc. for the evaluation of economic issues surrounding human papillomavirus testing. JWC authored the decision brief that led to the implementation of liquid‐based cytology and reflex human papillomavirus testing, utilizing ThinPrep and HC2, at all Army medical facilities worldwide.

Data reported here were presented at the 33rd Annual Meeting of the Society of Gynecologic Oncologists, Miami, Florida, March 2–7, 2002.

Received November 27, 2001. Received in revised form March 26, 2002. Accepted April 18, 2002.

Article Outline

Worldwide, cervical cancer is the second most common malignancy affecting women. In the United States, approximately 12,800 new cases were diagnosed and 4500 women died secondary to cervical cancer in the year 2000.1 Over the past 20 years, screening with the Papanicolaou smear has dramatically reduced the number of women who have developed cervical cancer.2 Data from the Surveillance, Epidemiology and End Results Registry has demonstrated a concurrent 46% reduction in mortality related to cervical cancer from 1973 to 1995.3

In settings where cervical cancer screening is available, a proportion of cases will arise from false‐negative screening tests. Several new technologies have recently been introduced in an effort to minimize the false‐negative rates associated with conventional cytology. Liquid‐based cytology is a method where the cervical cells are immersed in a liquid preservative before slide fixation. Using this method, artifact is decreased by removing obscuring blood, inflammatory cells, mucus, and debris. The improved quality of liquid‐based cytology results in potentially greater sensitivity when compared with the conventional Papanicolaou smear in the detection of preinvasive disease and cervical cancer.4

The presence of human papillomavirus (HPV) is associated with an increased risk of having preinvasive disease or cervical cancer. Testing for high‐risk HPV subtypes has been proposed to complement cytology in cervical cancer screening algorithms. Combined cytology screening and HPV testing may be especially suitable for a more immediate triage of patients with atypical squamous cells of undetermined significance (ASCUS)5 or patients older than 35 with low‐grade squamous intraepithelial neoplasia.6 Patients with these types of abnormal cytologic findings can have testing for high‐risk HPV instead of repeat cytology. Patients who are screened with liquid‐based cytology may have the transport media remaining after cytologic screening used for the ancillary HPV.7 This “reflex” HPV testing can identify patients with a high‐risk HPV type who should be referred for immediate colposcopy instead of returning to the provider for serial cytologic testing. However, the roles of liquid‐based cytology and HPV testing remain controversial.8,9

The increasing use of liquid‐based cytology and reflex HPV triage of low‐grade cervical abnormalities has prompted several health maintenance organizations and insurance companies to initiate these newer techniques in their cervical cancer screening. The US Army has recently initiated a transition from conventional cytology to liquid‐based cytology and HPV triage using ThinPrep (Cytyc Corp., Boxborough, MA) and Hybrid Capture 2 (Digene Inc., Gaithersburg, MD) to provide more sensitive cervical cancer screening and reduce the number of patient encounters required for management of equivocal or borderline cervical cytology. Consistently, cost‐effectiveness analyses of cervical cancer screening programs have shown that the cost‐effectiveness ratio of screening, no matter what technology is used, increases exponentially as intervals between screening decrease, an effect that is increased as test sensitivity increases.10–12 More sensitive screening tests, by improving the negative predictive value of a normal result, may allow lengthening of the interval without jeopardizing safety; this issue is especially important in the military, where women are required to have normal test results before deployment.

The objective of this analysis was to estimate the cost‐effectiveness of conventional cytology, liquid‐based cytology, and liquid‐based cytology combined with HPV triage in a military health care setting.

Back to Top | Article Outline


We used a previously described Markov model that simulates the natural history of HPV infection and cervical carcinogenesis.13,14 A Markov model is a mathematical model involving a series of states (eg, Healthy, HPV Infected, Low‐Grade Squamous Intraepithelial Lesion, Stage I Cervical Cancer, Death From Cervical Cancer).15 During simulations, an individual or a population may move from one state to another based on predefined probabilities. For example, in a Markov simulation of a population with only two states, Alive and Dead, the probability of making the transition from Alive to Dead within a 1‐year time frame would be equivalent to the age‐specific mortality for that population. In this model, we simulate a cohort of women aged 18–85 and vary the probability of acquiring an HPV infection, persistence, progression, or regression of that infection through to invasive cervical cancer; death from other causes, death from cervical cancer; and hysterectomy for noncervical neoplasia‐related causes (Figure 1). The states and transition probabilities are shown in Table 1. We assume that all cervical cancer arises from HPV infection, and progresses from low‐grade through high‐grade squamous intraepithelial lesions (SILs) to invasive cancer of different stages. We assumed that the prevalence at age 18 of HPV infection is 10% and the prevalence of low‐grade SILs is 1%. Women infected with HPV can undergo regression, have no change, or undergo progression. Although most cases progress from HPV to low‐grade SILs, the model accounted for a proportion of HPV infections progressing directly to high‐grade SILs. The model has been extensively validated by comparing predicted age‐specific prevalence of HPV, low‐grade SILs, and high‐grade SILs with cross‐sectional rates.13,14 Women who died from causes unrelated to cervical cancer or had hysterectomies for reasons other than cervical cancer were removed from the pool of patients at risk for developing cervical cancer. A detailed description of the model is provided in previous publications.11,13,14 We simulated a cohort of 100,000 women with military health care benefits, assuming that screening begins at age 18, and observed the cohort through age 85. Costs and life expectancy were discounted at a 3% annual rate.

We took the perspective of the military health system in this analysis, and considered only the direct medical costs associated with screening, diagnosis, and treatment of low‐grade SILs, high‐grade SILs, and invasive cervical cancer. Although a more realistic model would have ignored costs incurred after discharge from the military (similar to a civilian model of a health maintenance organization including only expected length of enrollment), it is possible that such a model would have concluded that screening for cervical neoplasia in the military is not justified at all because most cases of cancer would occur after discharge. Because dependents are included in the military system, because a growing number of women seek care within the Veterans Affairs medical system, and because the majority of other published cost‐effectiveness analyses use a lifetime perspective, we have taken a similar approach. Nonmedical costs (such as time away from active duty) were not included in the assessment of cost‐effectiveness because we did not have accurate estimates of the actual time lost for any of the outcomes. We also did not attempt to adjust for quality of life, because we are unaware of any validated data of women's preferences for all of the states associated with cervical cancer prevention, ranging from routine screening to diagnosis of HPV to invasive cancer.

The underlying model is based on the natural history of HPV infection and cervical carcinogenesis. The probability that a screening test will detect any histological abnormality of cervical intraepithelial neoplasia (CIN) I or greater is equal to the sensitivity of the screening test. The sensitivity of the conventional Papanicolaou smear, based on unbiased studies where test negatives underwent colposcopy, is between 50% and 55% where ASCUS or greater is considered an abnormal Papanicolaou result and CIN I or greater is considered an abnormal histological result.13,16 The probability that a screening test will be falsely positive when no histological abnormality is present (women who are normal, or who have detectable HPV deoxyribonucleic acid [DNA] without histological changes) is the specificity. The probability that an abnormal cytologic test will give a specific result (ASCUS, Low‐Grade SILs, High‐Grade SILs, Cancer) is also dependent on the underlying histology; these probabilities were derived from the College of American Pathologists Q‐Probes study.17

In patients with ASCUS, we estimated the proportion of HPV‐positive women with a given histological abnormality who would have a high‐risk HPV type (assumed to be 75% for abnormal cytology but with normal histology in patients younger than 35 and 60% for those women 35 or older, 85% for women with CIN I, 90% for women with CIN II/III, and 95% for women with cancer).18,19 The overall probability of a positive test for HPV is then the above probability multiplied by the sensitivity of the test for detection of HPV DNA, assumed to be 98%. Strategies for ASCUS were either repeat cytology in 6 months with colposcopy for women with persistent ASCUS or higher, or HPV testing with immediate colposcopy referral for women with high‐risk HPV types. Because the incidence of HPV in patients with low‐grade SILs is approximately 83%,18 patients with cytologic findings of low‐grade SILs were referred for immediate colposcopy instead of HPV testing. The model assumed that 20% of patients with low‐grade intraepithelial biopsy results would be treated with cryotherapy, with the rest observed conservatively, and all high‐grade intraepithelial neoplasia biopsy results would be treated with the loop excisional electrocautery procedure. Once treatment was implemented, the patient would return to the scheduled screening interval. All treatments were assumed to be 95% effective.

Estimates for the test characteristics of conventional cytology were derived from a systematic review of the literature.13,16 In the base case, we estimated that conventional cytology had a sensitivity of 51% and a specificity of 97%, using a cytologic threshold of ASCUS for abnormality and a histological threshold of CIN I or higher, the values found in the studies of conventional cytology where verification bias was minimized by performing colposcopy on all or a random portion of women with negative cytology. We varied test sensitivity from 51% to 85%. Decreases in specificity associated with increased sensitivity were modeled using the summary odds ratio, a measure that incorporates both test sensitivity and specificity. For a cytologic threshold of ASCUS and a histological threshold of CIN I or higher, the summary odds ratio for conventional cytology was 1.863. Using this value, the specificity associated with a given sensitivity can be calculated using the formula 1.863 = ln {Sensitivity/[1 – Sensitivity]} + ln {Specificity/[1 – Specificity]}.13

Estimates for the test characteristics of liquid‐based cytology relative to conventional cytology are problematic given the available literature. For example, one of the largest studies showed an absolute increase in sensitivity of 15% for liquid cytology compared with conventional cytology, but the slides were interpreted by different pathologists with different levels of training, which precluded definitive estimates of relative test performance.4 For the base case, we assumed that liquid‐based cytology would increase sensitivity by 60% (82% versus 51%) for the same histological and cytologic thresholds relative to conventional cytology. Data on the specificity of liquid‐based cytology are limited. Because the highest quality studies of cervical cancer screening have consistently found that specificity decreases with increased sensitivity, we assumed that there would be a decrease in specificity associated with improvements in sensitivity with liquid‐based cytology. For the base case, we assumed that relative specificity of liquid‐based cytology would be 92%, compared with 97% for conventional cytology. We examined the effects of uncertainty about relative performance of liquid cytology by varying relative sensitivity and specificity simultaneously with conventional Papanicolaou sensitivity and specificity.

These estimates of test characteristics for the base case were validated by comparing model predictions to cross‐sectional data from the Atypical Squamous Cells of Undetermined Significance/Low‐Grade Squamous Intraepithelial Lesions Triage Study.19 In this trial, 56.1% (95% confidence interval [CI] 54.1, 58.1%) of women triaged on the basis of a positive HPV test were referred to colposcopy, whereas 58.6% (95% CI 56.5, 60.6%) were referred to colposcopy on the basis of a liquid Papanicolaou test of ASCUS or higher. Using the parameters above for test characteristics, and matching the prevalence of underlying histological abnormalities to the Atypical Squamous Cells of Undetermined Significance/Low‐Grade Squamous Intraepithelial Lesions Triage Study, the model predicted referral rates of 55.5% for HPV triage and 58.9% for triage based on a repeat liquid‐based cytology Papanicolaou test of ASCUS or higher.

Costs for cervical cancer screening were derived from military estimates. Because costs in the base case of a previously reported civilian model included estimates of the total cost of obtaining a cytologic smear (including equipment, physician and staff time, etc), we derived total Army costs by subtracting the difference between total civilian costs and the Army processing/reading costs (component cost) for conventional cytology. This enabled us to estimate the total Army costs for conventional cytology to be $28 for women younger than 65 and $40 for women 65 or older. Sensitivity analysis varied the price of the follow‐up visit from $25 to $75. The cost of liquid‐based cytology was expressed as a marginal difference between liquid‐based cytology and conventional cytology of $5 (varied up to $30 in sensitivity analysis). The estimated cost of HPV testing, which included the costs of storage, processing, and interpretation, was $65 (varied from $50 to $100 in the sensitivity analysis). The costs for both screening and treatment of abnormal cytology are shown in Table 2.

Because only component charges were available for estimating Army costs of diagnosis and treatment, civilian‐based costs for preinvasive disease, derived from econometric analysis of a variety of sources, were used in the analysis (Table 1). We used mean costs per episode of care for a given level of abnormality, so that the costs of complications, unexpected visits, and so on were incorporated into our estimates. Costs were in US dollars, updated to year 2000 using the medical care component of the consumer price index. Given recent evidence that adjuvant chemotherapy improves outcomes in invasive cervical cancer,20–23 we also included potential cost of chemotherapy in sensitivity analysis.

Ten strategies were compared: no screening, conventional cytology, liquid‐based cytology, and liquid‐based cytology with HPV testing. Each screening measure was then compared at 1‐, 2‐, and 3‐year intervals. Total office visits for screening, diagnosis and treatment of invasive disease, total abnormal Papanicolaou smears, cervical cancer incidence, stage distribution, and mortality were estimated using the model. The cost‐effectiveness of the three different screening tests was then compared at different screening intervals using incremental cost per life‐year saved, calculated by dividing the difference in predicted costs between two strategies by the difference in average life expectancy predicted for the strategies.

Back to Top | Article Outline


The estimated screening and cancer‐related outcomes of each cervical cancer screening strategy, over the lifetime of the military cohort, are shown in Table 2. Each 1‐year decrease in screening interval results in an almost twofold increase in the number of office visits required. The largest reduction in both cancer incidence and mortality occurs with any cervical cancer screening method, even at 3‐year intervals. Marginal reductions in cancer incidence from increasing screening sensitivity are greater than marginal reductions in cancer mortality at every screening interval. Marginal reductions in both cancer incidence and mortality from increasing screening sensitivity are greater at less frequent screening intervals. There are three main findings: 1) more sensitive tests result in fewer cancer cases and deaths, but with more office visits within a given screening interval; 2) use of HPV triage consistently results in fewer visits, fewer cancer cases, and fewer cancer deaths than liquid‐based cytology alone at every screening interval; and 3) liquid‐based cytology plus HPV triage results in fewer office visits, fewer cancer cases, and fewer cancer deaths than conventional cytology performed at the next most frequent screening interval (eg, liquid‐based cytology plus HPV every 3 years is superior to conventional Papanicolaou every 2 years). Increasing test sensitivity and/or decreasing screening interval results in proportionally greater increases in detection of CIN I relative to CIN II/III, as shown in Table 3.

Table 4 translates these findings into cost‐effectiveness as measured by cost per year of life saved. Conventional cytology and liquid‐based cytology are both cost‐effective methods of screening when performed at 3‐year intervals. None of these strategies cost less than the commonly used standard of $50,000 per life‐year saved at screening intervals more frequent than every 3 years. The use of HPV testing for triage of ASCUS smears is both less expensive and slightly more effective than liquid‐based cytology alone at every interval.

The costs of diagnosis and treatment of preinvasive disease or cancer, including possible chemotherapy costs, did not change the relative rankings or substantially alter the cost‐effectiveness ratios. Several variables did have an impact. If the marginal cost of liquid‐based cytology compared with conventional cytology was more than $25, then the cost‐effectiveness ratios of all liquid‐based cytology–based strategies were greater than $50,000, even at a triennial screening interval. However, the overall costs of using HPV testing remained lower than liquid‐based cytology alone even if the cost of HPV testing was $100.

Table 5 shows the effects of conventional cytology test characteristics on the cost‐effectiveness of liquid‐based cytology and HPV triage. As conventional cytology sensitivity increases, specificity decreases. The marginal gains in life expectancy from more sensitive screening decrease, whereas costs associated with decreased specificity rise. At high levels of sensitivity (98% for liquid cytology), only one option, liquid‐based cytology plus HPV testing every 3 years, costs less than $50,000 per year of life saved.

Back to Top | Article Outline


Our analysis found that increasing test sensitivity resulted in decreases in predicted cancer incidence and mortality; these reductions in morbidity and mortality were accompanied by increases in the number of office visits, which translated into significantly increased costs. As screening is performed more frequently, costs increase more rapidly than reductions in mortality, resulting in rapid increases in cost‐effectiveness ratios. However, one important finding is that more sensitive tests may be both more effective and less expensive than conventional cytology at more frequent intervals (eg, based on our estimates, liquid‐based cytology and HPV triage every 2 years is more effective and less expensive than conventional cytology every year).

These findings are qualitatively similar to the findings of other model‐based explorations of cervical cancer screening, which have consistently found trade‐offs between test sensitivity and specificity and screening frequency.10–12,24 Cost‐effectiveness ratios increase dramatically when screening is performed at more frequent intervals, especially if test sensitivity is increased. At frequent screening intervals, the marginal costs of individual cancer screening tests (ie, conventional cytology, liquid‐based cytology, and HPV testing) exceed the $50,000 per life‐year saved level considered by many health economists to be the threshold for cost‐effectiveness.25 Two factors play a role in this observation—the costs of the tests themselves and the increased likelihood of detecting transient lower grade cytologic abnormalities that, though highly unlikely to lead to cancer, generate substantial follow‐up costs. These effects are exacerbated if greater test sensitivity is accompanied by decreased specificity.

Although these basic findings should be valid for any health system, they have additional implications for the military. In a military population, the employer provides the resources for the health care of active duty soldiers in meeting the mission goals of providing a physically fit force. Our findings suggest that the use of liquid‐based cytology plus HPV triage helps to meet this objective in two ways. First, testing for HPV consistently results in fewer visits than liquid‐based cytology alone, resulting in both lower direct medical costs and less time away from the military unit. One study in a civilian population found that the average time spent away from work due to screening visits was 1.76 hours.26 If time losses are similar for the military, the economic argument for the use of HPV testing is even stronger. Second, these findings suggest that, even at relatively small gains in sensitivity over conventional cytology, liquid‐based cytology plus HPV testing is both less expensive and more effective at less frequent intervals. The Army Medical Department currently has a regulation requiring annual screening in its active duty female population. If studies currently underway verify these predictions about relative sensitivity, use of these technologies means that the interval between screening tests can be safely lengthened, which would result in both decreased appointments and improved health because the negative predictive value of a single normal test result is greater.

As with all models, uncertainty about the values of both probabilities and costs affects the conclusions of the analysis; however, varying these over a wide range did not result in substantial changes in the main findings (Table 6). We did not incorporate quality of life measures primarily because of a lack of validated measures for health states relevant to cervical cancer screening. We did not measure nonmedical costs, such as time lost from work or transportation costs, but as discussed above, it is likely that these would only strengthen the findings on the trade‐offs of costs, sensitivity, specificity, and screening frequency. We also assumed perfect compliance with follow‐up of abnormal smears. However, preliminary results from a cost‐effectiveness analysis of a screening trial where compliance was directly measured found that cost‐effectiveness ratios were actually improved with less than perfect compliance because the majority of missed visits were related to either false‐positive tests or tests that detected lower grade abnormalities (personal communication, S. Kulasingam, 2001).

In conclusion, we found that liquid‐based cytology with HPV triage is both less expensive and more effective than liquid‐based cytology alone at all screening intervals. Replacing conventional cytology with these technologies does not meet traditional cost‐effectiveness thresholds at annual screening intervals; however, using these technologies at less frequent intervals is consistently both less expensive and more effective. The Army's conversion from the conventional Papanicolaou smear to liquid‐based cytology with reflex HPV triage is the first such broad implementation in a large government organization. Concurrent with this implementation are initiatives to improve compliance with screening and treatment algorithms. If additional data from this initiative and other ongoing studies confirm the predictions of this analysis, increasing the screening interval may be possible, with improvements in both the health of individual soldiers and the overall readiness of the Army.

Back to Top | Article Outline


1. American Cancer Society. Statistics for 2002. Atlanta: American Cancer Society Inc., 2002. Available at: Accessed 2002 Aug 21.
2. Gustafsson L, Ponten J, Zack M, Adami HO. International incidence rates of invasive cervical cancer after introduction of cytologic screening. Cancer Causes Control 1997; 8:755–63.
3. Ries L, Kosary C, Hankey B, Miller BA, Edwards BK. SEER cancer statistics review, 1973–1995. Bethesda, MD: National Cancer Institute, 1998.
4. Hutchinson ML, Zahniser DJ, Sherman ME, Herrero R, Alfaro M, Bratti MC, et al. Utility of liquid-based cytology for cervical carcinoma screening. Cancer 1999;87:45–55.
5. Manos MM, Kinney WK, Hurley LB, Sherman ME, Shieh-Ngai J, Kurman RJ, et al. Identifying women with cervical neoplasia: Using human papillomavirus DNA testing for equivocal Papanicolaou results. JAMA 1999; 281:1605–10.
6. Wright TC, Lorincz A, Ferris DG, Richart RM, Ferenczy A, Mielzynska I, et al. Reflex human papillomavirus deoxyribonucleic acid testing in women with abnormal Papanicolaou smears. Am J Obstet Gynecol 1998;178:962–6.
7. Sherman ME, Chiffman MH, Lorincz AT, Herrero R, Hutchinson ML, Bratti C, et al. Cervical specimens collected in liquid buffer are suitable for both cytologic screening and ancillary human papillomavirus testing. Cancer 1997;81:89–97.
8. Hartmann KE, Nanda K, Hall S, Myers E. Technologic advances for evaluation of cervical cytology: Is newer better? Obstet Gynecol Surv 2001;56:765–74.
9. Kaufman RH. Is there a role for human papillomavirus testing in clinical practice? Obstet Gynecol 2001;98:724–5.
10. Eddy DM: Screening for cervical cancer. Ann Intern Med 1990;113:214–26.
11. Myers ER, McCrory DC, Subramanian S, McCall N, Nanda K, Bastian L, et al. Toward a better cervical screening test: Criteria for a cost-effective test for cervical neoplasia screening. Obstet Gynecol 2000;96:645–52.
12. Goldie SJ, Kuhn L, Denny L, Pollack A, Wright TC. Policy analysis of cervical cancer screening strategies in low-resource settings: Clinical benefits and cost-effectiveness. JAMA 2001;285:3107–15.
13. McCrory D, Matcher DB, Bastian L, Datta S, Hasselblad V, Hickey JD, et al. Evaluation of cervical cytology. Evidence report/technology assessment no. 5. AHCPR publication no. 99-E010. Rockville, MD: Agency for Health Care Policy and Research, 1999.
14. Myers ER, McCrory DC, Nanda K, Badstian L, Matchar DB. Mathematical model for the natural history of human papillomavirus infection and cervical carcinogenesis. Am J Epidemiol 2000;151:1158–71.
15. Sonnenberg FA, Beck JR. Markov models in medical decision making: A practical guide. Med Decis Making 1993;13:322–38.
16. Nanda K, McCrory DC, Myers ER, Bastian LA, Hasselblad V, Hickey JD, et al. Accuracy of the Papanicolaou test in screening for and follow-up of cervical cytological abnormalities: A systematic review. Ann Intern Med 2000; 132:810–9.
17. Jones BA, Novis DA. Cervical biopsy-cytology correlation: A College of American Pathologists Q-Probes study of 22,439 correlations in 348 laboratories. Arch Pathol Lab Med 1996;120:523–31.
18. Solomon D, Schulman M, Tarn R. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: Baseline results from a randomized trial. J Natl Cancer Inst 2001;93:293–9.
19. The Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group. Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: Baseline data from a randomized trial. J Natl Cancer Inst 2000;92:397–402.
20. Keys HM, Bundy BN, Stehman FB, Muderspach LI, Chafe WE, Suggs CL, et al. Cisplatin, radiation and adjuvant hysterectomy compared with radiation and adjuvant hysterectomy for bulky stage IB cervical carcinoma. N Engl J Med 1999;340:1154–61.
21. Rose PG, Bundy BN, Watkins EB, Thigpen T, Deppe G, Maiman MA, et al. Concurrent cisplatin-based radiotherapy and chemotherapy for locally advanced cervical cancer. N Engl J Med 1999;340:1144–53.
22. Morris M, Eifel PJ, Lu J, Grigsby PW, Levenbeck C, Stevens RE, et al. Pelvic radiation with concurrent chemotherapy compared with pelvic and para-aortic radiation for high-risk cervical cancer. N Engl J Med 1999;340:1137–43.
23. Peters WA, Lu PY, Barrett RJ, Stock RJ, Monk BJ, Berek BJ, et al. Concurrent chemotherapy and pelvic radiation therapy compared with pelvic radiation therapy alone as adjuvant therapy after radical surgery in high-risk early-stage cancer of the cervix. J Clin Oncol 2000;18:1606–13.
24. Brown AD, Garber AM. Cost-effectiveness of 3 methods to enhance the sensitivity of Papanicolaou testing. JAMA 1999;281:347–53.
25. Garber AM, Phelps CE. Economic foundations of cost-effectiveness analysis. J Health Econ 1997;16:1–31.
26. Shireman TI, Tsevat J, Goldie SJ. Time costs associated with cervical cancer screening. Int J Technol Assess Health Care 2001;17:146–52.
© 2002 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.