In the modern era of radical attempts to limit health care expenditures, productivity and efficiency become central concerns, with a particular emphasis on relating costs to performance and outcomes. Thus, outcomes analysis is rapidly becoming the dominant paradigm by which health care critical pathways are judged. An essential part of outcomes analysis is a cost-effectiveness assessment of any given pathway.
Accurate costing of a specific preventive, diagnostic, or therapeutic intervention has been difficult and frequently hampered with insufficiencies in obtaining realistic numbers. Often hospital and surgeon charges for inpatient care are simply compared in an attempt to determine which treatment modality is more cost-effective. These attempts at cost analysis frequently do not incorporate the global societal resource utilization such as patient time, lost work hours, disability payments, or long-term complications that may be expensive to manage. In traditional and reimbursement-driven cost systems, it is also often not known what actually caused the costs of a certain product or service. To counteract these deficiencies in accounting and more accurately predict costs to a closed financial system, the idea of activity-based costing was developed. Activity-based costing traces all costs associated with a certain product or service to determine the true costs of this product or service distributed in a particular market or provided to a customer. This accounting technique is designed to eliminate cross-subsidies between products and services. Activity-based costing readily allows for isolation and identification of nonvalue-added activities. Thus, activity-based costing is well suited for establishing and evaluating performance measures.1
The loop electrosurgical excision procedure has been widely accepted in the management of cervical intraepithelial neoplasia (CIN), being used as a conization technique and as a replacement for ablative techniques. The loop electrosurgical excision procedure for CIN has been used either in a conventional setting, where colposcopy and directed-biopsy diagnosis preceded the procedure, or in a single-visit see-and-treat program.2,3 The see-and-treat management paradigm is appealing because it requires only one visit for evaluation and treatment, saves patients and physicians time, offers high patient satisfaction, and decreases the risk of loss to follow-up before the completion of evaluation and treatment of CIN.
We performed this study to compare direct and indirect costs associated with each of four management algorithms for patients with a screening Papanicolaou smear demonstrating changes consistent with high-grade squamous intraepithelial lesion (SIL). The algorithms used ranged from a conventional management scheme of colposcopy and directed biopsies, followed by cryotherapy or cold-knife conization, to a see-and-treat algorithm using the loop electrosurgical excision procedure. Using activity-based costing and incorporating direct and indirect costs associated with the management of complications and follow-up, we attempted to complete a more accurate assessment of resource utilization.
Materials and Methods
This cost-effectiveness study was performed adhering to the methodologic criteria set forth by a recent review on economic analyses in obstetrics and gynecology.4 Calculations in this study were based on the assumption that a total of 4000 theoretical patients were evaluated and treated for a screening Papanicolaou smear demonstrating high-grade SIL. One thousand individuals were placed in each of four management algorithms: a conventional algorithm of colposcopy and directed biopsies with endocervical curettage (ECC), followed by cryotherapy or cold-knife conization as indicated (conventional algorithm I); a conventional algorithm with the loop electrosurgical excision procedure substituting for cold-knife conization (conventional algorithm II); a conventional algorithm with the loop electrosurgical excision procedure substituting for cold-knife conization and cryotherapy (conventional algorithm III); or a single-visit see-and-treat algorithm using colposcopy and the loop electrosurgical excision procedure (see-and-treat algorithm IV). The study's null hypothesis postulated that there were no cost differences among these four management algorithms. All assumptions for the triaging of patients into different arms of the treatment algorithms were based on the peer-review literature as well as on data incorporated in major textbooks in obstetrics and gynecology and gynecologic oncology. Whenever there was a discrepancy between published data, a best-average assumption was made with a tendency to bias toward the null hypothesis.
In algorithm I (Figure 1), patients with high-grade SIL on Papanicolaou smear underwent colposcopy with biopsies and ECC. Treatment with cryotherapy or cold-knife conization followed on a separate visit. Conization was performed for the classic indications: unsatisfactory colposcopy, lesion extending into the canal, positive ECC, marked discrepancy between Papanicolaou smear and colposcopy or biopsy findings, and suspected microinvasion.5,6 Complication rates were assumed to be as follows. For cold-knife conization, intraoperative or postoperative bleeding complications occurred in 12%7,8 and treatment failed in 5%9 of cases. Treatment failure referred to persistent or early recurrent preinvasive or invasive disease of the cervix. In 5% of patients having cold-knife conization, an evaluation for infection was assumed necessary.6 Significant bleeding is very unusual with cryotherapy.9,10 We assumed a need to evaluate for infection in 4%6,11 and treatment failure in 12% of patients.12,13 Follow-up was continued for 1 year after the index Papanicolaou smear because, with adequate follow-up, the majority of treatment failures are diagnosed in the first year.10,12 Follow-up visits were assumed at 3, 6, and 12 months after the initial therapeutic procedure, with 100% compliance to prescribed follow-up in all algorithms. The risk of new disease developing subsequently is reported to be similar to that of the high-risk population at large.14 The relative risk of developing invasive cancer in patients after conservative treatment of CIN remains five times greater than for the general population. This increase in risk appears independent of the conservative treatment modality or time elapsed since treatment and was thus assumed to be the same for each treatment algorithm.15Figure 2 outlines algorithm II, which followed the same management pathways and assumptions as algorithm I, except that the loop electrosurgical excision procedure was substituted for the cold-knife conization.
Complication rates for the loop electrosurgical excision procedure were set at 4% for bleeding,16,17 2% for infection,3,6,18 and 1% for incidental burns. A 5% treatment failure rate was assumed for the loop electrosurgical excision procedure.2,3,16,19,20Figure 3 demonstrates algorithm III, in which the loop electrosurgical excision procedure substituted for both cold-knife conization and cryotherapy, but still in a conventional setting where it was preceded by a colposcopic and histologic evaluation. In algorithm IV (Figure 4), the loop electrosurgical excision procedure was used in a see-and-treat single-visit setting.
Throughout all algorithms, we assumed that of the patients with high-grade SIL on a screening Papanicolaou smear, 5% would have colposcopy findings suspicious for invasive disease, of which 1% would indeed be confirmed to be cancer. Of the remaining 95% of patients, 10% would need a diagnostic conization, 85% would qualify for ablative treatment, and 5% would not require therapy. Patients with positive margins on a specimen obtained by cold-knife conization or the loop electrosurgical excision procedure were followed closely and not immediately retreated. For the purpose of this analysis, CIN was always treated conservatively, even if recurrent. Thus, hysterectomy was not incorporated into any of the management schemes. In the assumptions for this study, recurrent CIN was not treated with ablative therapy. Once the diagnosis of cancer was made, the cost of the patient's management was no longer included in the calculations. The treatment end point in all algorithms was successful management of cervical dysplasia. Because in this analysis, all costs occurred in the base year of the analysis, discounting to quantify the time value of money was not performed.
Cost assessment was performed for the four separate groups of 1000 theoretical patients with each group of patients managed by one of the defined algorithms (Figures 1–4). Patient time costs were derived from salary standards put forth by the University of California, Los Angeles (Table 1). The entry-level full-time employment (noncasual) was used for determining patient cost per hour ($8.25 per hour with benefits excluded). Physician costs were derived from Physician Marketplace Statistics 1997/98 put forth by the American Medical Association.21 The annual income averaged for gynecologists and family practitioners was $160,000 with a 56.5-hour work week and 3 weeks of vacation per year. Based on these numbers, an hourly rate for physician cost ($57.50) was derived. Patient and physician costs for appointments and procedures were based on the time necessary for completion of activities in our experience. We did not assume physician-patient encounters related to high-grade SIL other than those for evaluation, treatment, management of complications, and follow-up Papanicolaou smears (Table 1). Operative times and costs were based on cumulative data standards from the University of California, Los Angeles for the fiscal year 1994–1995. Disposable costs were based on University of California, Los Angeles costs and account for all procedure-related charges for hospital overhead, billing, and personnel. Disposable costs further included the initial costs for the equipment for cryotherapy ($1884) and the loop electrosurgical excision procedure ($2165). It was assumed that each instrument would have a lifetime sufficient for the management of all patients assigned per algorithm. Disposable costs also included all ongoing procedure-related costs such as electrosurgical loops or freezing media.
Statistical analysis was performed using Excel 6.0 software (Microsoft Corporation, Redmond, WA). Student t tests were performed with Bonferroni corrections for multiple comparisons to compare the average costs among the four groups.
Tables 2–5 summarize the activity-based costing results for the four different management algorithms for patients with a screening Papanicolaou smear showing high-grade SIL. Conventional evaluation of patients with colposcopy and biopsies followed by cryotherapy or cold-knife conization (Table 2, algorithm I) was the most costly management scheme. A total of $899,405 would be spent on 1000 patients with high-grade SIL. Substituting cold-knife conization with the loop electrosurgical excision procedure in an otherwise unchanged management pathway (Table 3, algorithm II) significantly (P < .001) decreased the cost by 32%, to $610,484. Substituting both cold-knife conization and cryotherapy with the loop electrosurgical excision procedure in an otherwise unchanged management pathway (Table 4, algorithm III) still significantly (P < .001) reduced the cost but only by 25%, to $674,830. The most cost-effective management (Table 5, algorithm IV) was the see-and-treat single-visit pathway at $531,281, with a 41% cost reduction compared with algorithm I (P < .001).
The average cost per patient who would qualify to be managed with ablative therapy in each of the treatment algorithms would be as follows: algorithm I, $751 per patient; algorithm II, $611 per patient; algorithm III, $691 per patient; and algorithm IV, $535 per patient. The average cost per patient who required a diagnostic conization was $1888 in algorithm I, $671 in algorithms II and III, and $467 in algorithm IV. Once a diagnosis of cancer was established, further management was excluded from the cost analysis. Thus, patients requiring conization appear to consume less resources in algorithms III and IV than patients qualifying for ablation. Including initial colposcopy and close follow-up Papanicolaou smears, the average cost per patient who did not require treatment was $402 in algorithms I–III and $504 in algorithm IV.
To correct for potential inaccuracy in the pathologic interpretation of cervical cytology smears and for errors in colposcopic assessment, we performed sensitivity calculations for unnecessary treatment in the see-and-treat algorithm IV. Algorithm I would be more cost-effective than algorithm IV only if fewer than 170 patients required treatment. Similarly, only if fewer than 425 or fewer than 319 patients ultimately required treatment would algorithms II and III, respectively, be more cost-effective than the see-and-treat algorithm IV. Because of the single-visit setup, the see-and-treat algorithm would continue to be the most cost-effective even if all patients qualified for ablative therapy or if there were no differences in the therapeutic efficacy of the different treatment modalities.
Endocervical curettage was not included in the routine follow-up in management algorithms I–IV. We therefore calculated additional costs if ECCs were included either routinely in all high-grade SIL patients at the first follow-up Papanicolaou smear or only for those patients who had a diagnostic and therapeutic conization performed using the cold knife or loop electrosurgical excision procedure. Assuming that the performance of an ECC at the first follow-up Papanicolaou smear would add $38.75 in total cost per patient, the routine use of ECCs in the follow-up of these patients would add $38,362 for each of the four treatment algorithms. If only patients who had undergone conizations were to receive an ECC at the first follow-up, $5618.75 would be added to the costs of each treatment algorithm. Because the underlying distribution of disease was assumed to be the same for each algorithm, there was no difference in the costs added by the performance of ECCs in each of the management algorithms.
Using activity-based costing, we have demonstrated that a technique that appears to use more resources actually ends up using less, even though more procedures are performed. Under the assumption that all four algorithms are equally effective at evaluating and treating high-grade SIL, the pathway associated with the least cost is the one that should be used. In our present study, this was demonstrated to be the see-and-treat algorithm. The second best strategy from a financial perspective was the conventional algorithm in which the loop electrosurgical excision procedure substituted for cold-knife conization, but not for cryotherapy. Clearly, the protocols that first evaluate with colposcopy and biopsies, followed by the loop electrosurgical excision procedure substituting for both cold-knife conization and cryotherapy, are the least cost-effective algorithms using the loop electrosurgical excision technique.
Because not only financial concerns but also patient selection and success of the regimen dictate cost-effectiveness, several issues arise. There is a concern that treatment of SIL may impair fertility. This may occur in four ways: induction of cervical stenosis, alteration in cervical mucus, ascending infection with subsequent tubal damage, and cervical incompetence. The peer-review literature provides little information about the differential effects of the treatment modalities studied here on fertility. Cold-knife conization of the cervix has been associated with an increased risk of preterm delivery. There are no data that demonstrate that cryotherapy or the loop electrosurgical excision procedure has a negative impact on fertility. However, studies addressing the issue are scant.22 Thus, it is beyond the scope of the present cost analysis to address the potential long-term effects of the different management modalities on fertility and cervical competence.
It is controversial whether the four algorithms studied here are equally effective at treating cervical precancers and thus preventing cervical cancer. We do not account for different patient populations and patient compliance, or lack thereof, which would have a substantially different impact within each of the different management algorithms for cervical dysplasia. Previous studies have reported that as many as 27–70% of women with biopsy-proved high-grade dysplasias will be lost to follow-up during the initial evaluation and thus will not undergo definitive treatment.23,24 In addition, patients with true cervical precancerous lesions tend to confront the greatest socioeconomic and cultural barriers to adequate screening, treatment, and follow-up, such as language, logistics, or costs.25–27 Providing about 90% of patients with adequate treatment during the first visit,2,3 the see-and-treat approach has the clear advantage of circumventing some of the barriers to appropriate follow-up. Because progression of high-grade SIL to invasive cancer has been reported to occur in 10–25% of patients,28 the cost savings of the see-and-treat approach have probably been grossly underestimated by our activity-based costing analysis.
Another aspect of this study in which the value of see-and-treat may be underestimated is the reliance on the skills of the colposcopist in conventional algorithms I-III. Numerous publications have consistently revealed a small, but important, number of patients with microinvasive disease found upon histologic evaluation of a cold-knife conization, loop electrosurgical excision procedure, or hysterectomy specimen that had been missed by colposcopy.2,6,16 Using a historic control population, Benedet et al29 found that the change from conization to colposcopically directed biopsies followed by ablative treatment was associated with a reduction in the detection of microinvasive disease from 3.5% to 1%. Similarly, the detection rate of unsuspected carcinoma after the loop electrosurgical excision procedure is reported to approach 1%.9,17 Timely discovery and treatment of microinvasive carcinoma confer a substantial improvement in curative potential and cost savings.
The most frequent argument against the see-and-treat management of patients with cervical dysplasia is the concern about a high rate of procedures performed for which the specimen is without evidence of dysplasia, particularly in patients who present with minimally abnormal Papanicolaou smears. We therefore restricted our analysis to a theoretical group of patients with high-grade SIL on the Papanicolaou smear. Using the four treatment algorithms, our study showed that only if 57.5–83% of patients were treated unnecessarily would see-and-treat be less cost-effective than conventional evaluation by colposcopy and biopsies, followed by cold-knife conization, loop electrosurgical excision procedure, or cryotherapy as indicated in algorithms I–III. However, two large series reporting on see-and-treat for high-grade SIL in a composite group of more than 1000 patients found that specimens from the loop electrosurgical excision procedure were negative for dysplasia in only 2–6% of cases.2,3 Thus, strong consideration should be given to accepting see-and-treat as the new standard for management of patients with high-grade SIL. Concerns regarding overtreatment and side effects appear to be outweighed by the global cost savings and the capability to minimize rates of missed microinvasive cancer and follow-up failure.
1. Miller JA. Activity/product costing. In: Miller JA, ed. Implementing activity-based management in daily operations. New York: John Wiley & Sons, Inc, 1996:49–68.
2. Bigrigg MA, Codling BW, Pearson P, Read MD, Swingler GR. Colposcopic diagnosis and treatment of cervical dysplasia at a single clinic visit. Experience of low-voltage diathermy loop in 1000 patients. Lancet 1990;336:229–31.
3. Keijser KGG, Kenemans P, van der Zanden P, Shijf CPT, Vooijs GP, Rolland R. Diathermy loop excision in the management of cervical intraepithelial neoplasia: Diagnosis and treatment in one procedure. Am J Obstet Gynecol 1992;166:1281–7.
4. Smith WJ, Blackmore CC. Economic analyses in obstetrics and gynecology: A methodologic evaluation of the literature. Obstet Gynecol 1998;91:472–8.
5. Hill EC. Premalignant and malignant disorders of the uterine cervix. In: Pernoll ML, ed. Current obstetric and gynecologic diagnosis and treatment. A Lange medical book. 7th ed. Norwalk, Connecticut: Appleton & Lange, 1991:937–54.
6. Townsend DE, Morrow CP. Premalignant and related disorders of the lower genital tract. In: Morrow CP, Curtin JP, eds. Synopsis of gynecologic oncology. 5th ed. New York: Churchill Livingstone, 1998:17–60.
7. Bjerre B, Eliasson G, Linell F, Söderberg H, Sjöberg NO. Conization as only treatment of carcinoma in situ of the uterine cervix. Am J Obstet Gynecol 1976;125:143–52.
8. Luesley DM, McCrum A, Terry PB, Wade-Evans T, Nicholson HO, Mylotte MJ, et al. Complications of cone biopsy related to the dimensions of the cone and the influence of prior colposcopic assessment. Br J Obstet Gynaecol 1985;92:158–64.
9. Hatch KD, Hacker NF. Intraepithelial disease of the cervix, vagina, and vulva. In: Berek JS, Adashi EY, Hillard PA, eds. Novak's gynecology. 12th ed. Baltimore: Williams & Wilkins, 1996:447–86.
10. Townsend DE. Cryosurgery. Surg Clin North Am 1978;58:97–109.
11. Creasman WT, Weed JC, Curry SL, Johnston WW, Parker RT. Efficacy of cryosurgical treatment of severe cervical intraepithelial neoplasia. Obstet Gynecol 1973;41:501–6.
12. Ostergard DR. Cryosurgical treatment of cervical intraepithelial neoplasia. Obstet Gynecol 1980;56:231–3.
13. Hatch KD, Shingleton HM, Austin JM, Soong SJ, Bradley DH. Cryosurgery of cervical intraepithelial neoplasia. Obstet Gynecol 1981;57:692–8.
14. Richart RM, Townsend DE, Crisp W, DePetrillo A, Ferenczy A, Johnson G, et al. An analysis of “long-term” follow-up results in patients with cervical intraepithelial neoplasia treated by cryotherapy. Am J Obstet Gynecol 1980;137:823–6.
15. Soutter WP, de Barros Lopes A, Fletcher A, Monaghan JM, Duncan ID, Paraskevaidis E, et al. Invasive cervical cancer after conservative therapy for cervical intraepithelial neoplasia. Lancet 1997;349:978–80.
16. Prendiville W, Cullimore J, Norman S. Large loop excision of the transformation zone (LLETZ). A new method of management for women with cervical intraepithelial neoplasia. Br J Obstet Gynaecol 1989;96:1054–60.
17. Luesley DM, Cullimore J, Redman CW, Lawton FG, Emens JM, Rollason TP, et al. Loop diathermy excision of the cervical transformation zone in patients with abnormal cervical smears. BMJ 1990;300:1690–3.
18. Mor-Yosef S, Lopes A, Pearson S, Monaghan JM. Loop diathermy cone biopsy. Obstet Gynecol 1990;75:884–6.
19. Gunasekera PC, Phipps JH, Lewis BV. Large loop excision of the transformation zone (LLETZ) compared to carbon dioxide laser in the treatment of CIN: A superior method of treatment. Br J Obstet Gynaecol 1990;97:995–8.
20. Murdoch JB, Morgan PR, Lopes A, Monaghan JM. Histological incomplete excision of CIN after large loop excision of the transformation zone (LLETZ) merits careful follow-up, not retreatment. Br J Obstet Gynaecol 1992;99:990–3.
21. Gonzalez ML, Zhang P. American Medical Association/Center for Health Policy Research: Physician marketplace statistics 1997/98. Chicago: American Medical Association, 1998.
22. Montz FJ. Impact of therapy for cervical intraepithelial neoplasia on fertility. Am J Obstet Gynecol 1996;175:1129–36.
23. Santos C, Galdos R, Alvarez M, Velarde C, Barriga O, Dyer R, et al. One-session management of cervical intraepithelial neoplasia: A solution for developing countries. Gynecol Oncol 1996;61:11–5.
24. Spitzer M, Chernys AE, Seltzer VL. The use of large loop excision of the transformation zone in an inner-city population. Obstet Gynecol 1993;82:731–5.
25. National Cancer Institute. Healthy people 2000. DHHS publication no. (PHS) 91-50212. Bethesda, Maryland: United States Department of Health and Human Services, 1990.
26. Orbell S, Crombie I, Robertson A, Johnston G, Kenicer M. Assessing the effectiveness of a screening campaign: Who is missed by 80% cervical screening coverage? J R Soc Med 1995;88:389–94.
27. The National Cancer Institute Cancer Screening Consortium for Underserved Women. Breast and cervical cancer screening among underserved women. Baseline survey results from six states. Arch Fam Med 1995;4:617–24.
28. Mclndoe WA, McLean MR, Jones RW, Mullins PR. The invasive potential of carcinoma in situ of the cervix. Obstet Gynecol 1984;64:451–8.
29. Benedet JL, Anderson GH, Boyes DA. Colposcopic accuracy in the diagnosis of microinvasive and occult invasive carcinoma of the cervix. Obstet Gynecol 1985;65:557–62.