INFECTION WITH CHLAMYDIA TRACHOMATIS is the most prevalent bacterial sexually transmitted disease (STD) in the United States, with an estimated 4,000,000 cases acquired annually.1 Sequelae of chlamydial infection, including pelvic inflammatory disease (PID), ectopic pregnancy, tubal infertility, and chronic pelvic pain, have a profound impact on women's health. The annual cost of chlamydial infection and its sequelae was estimated at $2.2 billion in 1990.2 The Centers for Disease Control and Prevention (CDC) recently allocated expanded resources for the prevention and control of chlamydial disease.3
Only 10% to 30% of reported chlamydial infections in women are symptomatic, presenting primarily as mucopurulent cervicitis or salpingitis4,5; the remainder are usually detected on routine examination or at referral for contact with an infected sex partner. Asymptomatic infection results in prolonged, untreated disease; delayed treatment increases the risk of sequelae6 and transmission to sex partners. Although universal screening has been endorsed as the best strategy for detecting most asymptomatic disease,7–9 it is costly. Instead, the usual approach is to screen selectively or empirically treat sexually active women, using such criteria as young age, cervicitis (the definition of which may be highly variable), and sexual behavior risks.10–17 Although selective screening programs have been associated with local declines in chlamydia prevalence,18,19 their performance and cost-effectiveness across different clinical settings have not been fully assessed.
The current study had three objectives: to develop selective screening criteria for genital chlamydial infection in women from two large samples of family planning (FP) and STD clients, all of whom were universally screened for C. trachomatis; to assess the independent contribution of cervicitis to the performance of these selective screening criteria; and to evaluate the incremental cost-effectiveness of selective versus universal screening in the FP and STD populations.
Study Population and Design
A cross-sectional study design was used. Characteristics of the study population are shown in Table 1. The FP sample consisted of 11,141 women evaluated from October, 1989 to April, 1990 in 13 clinics throughout Washington State. The STD sample consisted of 19,884 women evaluated at 141 STD clinics in Washington, Oregon, Alaska, and Idaho (U.S. Department of Health and Human Services Region X) throughout 1993. All women undergoing pelvic examination were tested for chlamydia. Clients' age, race, and behavioral risks were recorded, including history of new sex partner in the preceding 30 days; two or more partners in the preceding 60 days; male partner with symptoms of urethritis (dysuria or discharge); and partner with other partners. Reported use of hormonal contraception or barrier contraceptives (defined as condom or diaphragm) was recorded, but temporal patterns of use were not. Current pregnancy and parity were recorded for FP but not STD clients.
Cervicitis was diagnosed if either mucopurulent cervical discharge or bleeding induced by passage of an endocervical swab was present. In STD clients, the presence of 10 or more polymorphonuclear leukocytes per high-power field on endocervical Gram's stain was also used, in addition to either mucopurulent discharge or bleeding induced by endocervical swab. PID was diagnosed clinically if abdominal, adnexal, or cervical motion tenderness were present.
Specimens for detection of C. trachomatis were obtained from the endocervical canal with a Dacron swab (Harwood Products Co., Guilford, ME) in FP clients and with either cytobrush or Dacron swab in STD clients. In FP clients, a direct fluorescent antibody (DFA) test was used, for which a positive test was defined by the presence of two or more elementary bodies.20 STD clients were tested using either a commercial enzyme immunoassay,21 a DNA probe test (Gen-Probe, Inc., San Diego, CA),22 or cell culture (Table 1). Enzyme immunoassay and DNA probe tests were performed according to manufacturers' specifications, and all cell cultures were performed in a single reference laboratory as previously described.23
Risk factors for chlamydial infection were analyzed by chi-square and multivariate logistic regression using SPSS (Chicago, IL). The effect of the chlamydial diagnostic method was evaluated by including test type as a categorical variable. Independent predictors of chlamydial infection were assessed using stepwise logistic regression, with variables entered into multivariate analysis on the basis of significance (P < 0.05) in univariate analysis and practical utility as screening criteria. FP and STD clients were examined individually in the multivariate analysis; for each group, one model including age younger than 20 years and individual risk factors, and another model including age younger than 20 years and a composite variable, “any behavioral risk factor,” were assessed. Selective screening criteria developed from the multivariate models were applied separately to the FP and STD clients and to each state. Criteria sensitivity (proportion of infected women identified by the criteria) and positive predictive value (proportion of women meeting the criteria who were infected) were calculated.
Cost-Effectiveness Analysis: General Approach
We performed an incremental cost-effectiveness analysis to compare the strategies of universal, selective, and no screening in hypothetical cohorts of 1 million FP and STD clients.24,25 Decision analysis (Figure 1) was used to assess potential outcomes of chlamydial infection, and was modeled on a computer spreadsheet. The main outcome measure was untreated chlamydial infection of the cervix in women. The analysis took a societal perspective, incorporating both direct medical costs and indirect costs resulting from lost productivity due to consequences of untreated chlamydial infections; intangible costs were not included. Costs were measured in 1993 U.S. dollars. Our base-case scenario assumed that either symptomatic or silent PID developed in 25% of untreated cases, that DFA was the diagnostic test used at a cost per test of $5 and sensitivity of 75%,26 and that chlamydia prevalence was 6.6%, the prevalence in our study population. Threshold analysis was used to determine the break-even population prevalence of C. trachomatis above which universal screening would generate cost savings relative to selective screening. The effect of using a diagnostic test with an estimated sensitivity of 95% and a unit cost of $10, such as the ligase chain reaction (LCR)27–29 assay, was also considered in threshold analysis. Sensitivity analysis was done for outcomes for which definitive, prospectively derived probabilities are not available.
Direct Costs.Table 2 displays the cost estimates used. Costs for clinic visits and for the treatment of uncomplicated chlamydial infection and outpatient PID were obtained from the Region X Office of Family Planning. Costs for inpatient treatment and for sequelae of untreated infection were updated from previously published estimates2 using the medical care component of the Consumer Price Index.30 The cost of a case of chronic pelvic pain was assigned that of inpatient treatment for PID, because recent data indicate that surgery (including lysis of adhesions, hysterectomy, salpingo-oophorectomy, and laparoscopy) is performed for 90% of women hospitalized for “chronic” PID (often defined by the presence of chronic pelvic pain).31 Costs for tubal infertility were discounted at a rate of 5% per year over 10 years, and those for ectopic pregnancy and chronic pelvic pain at a rate of 5% per year over 5 years. All other costs were assumed to occur within the first year after diagnosis of chlamydial infection.
Indirect Costs. Aggregated annual mean earnings of an individual by age and sex were used to calculate indirect costs associated with lost productivity.32 We estimated 10 nonspecified days of lost productivity for outpatient PID, 28 days for ectopic pregnancy, and 21 days each for inpatient PID, tubal infertility, and chronic pelvic pain.33 We estimated the average cost of a nonspecified day for women aged 15 to 40 years to be $80 by using the weighted annual mean earnings of labor and nonlabor (i.e., house work) for this age group ($29,068 divided by 365 days).32 For epididymitis in men aged 15 to 40 years, we estimated loss of 5 nonspecified days at $96 per day ($35,066 divided by 365 days). We did not include indirect costs of uncomplicated chlamydial infection because most of these infections in women are subclinical and do not incur productivity loss.
Conditional Probabilities. The conditional probabilities for outcomes were derived from review of the medical literature using Medline from 1985 to September, 1994 (Table 2). Estimates for the development of symptomatic PID arising from untreated cervical chlamydial infection range from 10% to 40%.4,34,35 However, “silent” PID—subclinical upper genital tract inflammation—is estimated to account for 50% to 75% of all PID, and has adverse reproductive health outcomes.36–38 We estimated that 60% of all PID was silent; this translates to 10% and 15% for the probabilities of symptomatic and silent PID, respectively, arising from untreated chlamydial infection in the base-case scenario. Women were assumed to be at risk for ectopic pregnancy, tubal infertility, and chronic pelvic pain after either symptomatic or silent PID, although costs for the treatment of PID were assigned only to symptomatic women.31 The probability of ectopic pregnancy occurring after chlamydial PID has been estimated at 5% to 10%,38,39 whereas that of tubal infertility ranges from 8% to 40%, depending on the severity and cumulative number of PID episodes.40–44 We assumed that 25% of women with tubal infertility would seek medical evaluation.45 The probability of chronic pelvic pain used was 18% based on one prospective study of laparoscopically confirmed PID.46
The prevalence of pregnancy (10% for FP clients and 3% for STD clients) was obtained from participating clinics. Maternal cervical infection with C. trachomatis is estimated to infect 60% to 70% of infants, with neonatal conjunctivitis resulting in 25% to 50% and pneumonia in 10% to 20%.47,48 Cure rates for uncomplicated chlamydial infection with doxycycline are 95%, with estimates of compliance ranging from 70% to 100%.2 Complications of treatment, defined as either vaginal candidiasis or gastrointestinal side effects necessitating reevaluation, occur in 5%.49 For the probability of chlamydial transmission to a male sex partner, we used 0.33.50 Although a study using polymerase chain reaction (PCR) found concordant infection in 75% of male sex partners of infected women,51 the applicability of this finding to risk of chlamydial transmission is not yet clear. Symptomatic urethritis prompting clinical evaluation develops in 40% of infected men,52 with epididymitis occurring in 1%.2
Characteristics of Study Subjects
As shown in Table 1, C. trachomatis prevalence among the 11,141 FP clients and 19,884 STD clients was 6.6% in both groups (95% confidence interval [CI], 6.2–7.0%). Above 20 years of age, C. trachomatis prevalence declined sharply (data not shown); therefore, we chose 20 years or younger as the cut-off for assessment of young age as a risk. Compared to FP clients, STD clients were older, more racially diverse, more likely to report all behavioral risks and barrier contraceptive use, and were diagnosed more frequently with cervicitis.
The results of the univariate analysis of risk factors for chlamydial infection are shown in Table 3. The strongest demographic predictor in both populations was young age (≤20 years). Neither race nor use of nonbarrier or no contraception was significantly associated with chlamydial infection in either group, nor was there a significant association between chlamydial infection and clinical diagnosis of PID in FP clients. Pregnancy and nulliparity were associated with slightly elevated risks among FP clients. All behavioral risks were significantly associated with infection in both the FP and STD groups, the strongest being report of a symptomatic male partner. Cervicitis was associated with a threefold increased risk of infection in both FP and STD clients.
In the multivariate model that considered risks without cervicitis, complete information was available for 10,798 (97.0%) FP and 18,936 (95.2%) STD clients. The results are shown in Table 4. Young age and behavioral risks associated with sex partners were the only independent predictors of chlamydial infection in both populations. Age 20 years or younger remained the strongest predictor of infection (odds ratio [OR] 2.3 in FP, 3.6 in STD). Reporting a symptomatic male sex partner more than doubled the risk of infection in both groups. When cervicitis was included in the model, data for analysis were complete on 10,504 FP (94%) and 16,822 STD clients (84.6%). The independent risk of infection associated with cervicitis was higher in FP clients (OR = 3.7) than in STD clients (OR = 2.7); however, this difference was not statistically significant. The ORs for the remaining variables did not change significantly (data not shown). When behavioral risks were combined into a single measure, the ORs for the remaining variables did not change significantly (data not shown). Variables that were not significantly associated with chlamydial infection included use of nonbarrier or no contraception and type of diagnostic test.
Development and Performance of Selective Screening Criteria
The prevalence of chlamydial infection in FP clients whose only risk was age 20 years or younger was 7.3%; chlamydia prevalence in women 20 years of age or younger increased to 12.1% if any behavioral risk was reported. The presence of two or more behavioral risks, regardless of age, was associated with chlamydia prevalences ranging from 13% to 33%. Nevertheless, of all FP clients with C. trachomatis, 26.1% had none of these risk factors. In the STD clients, the prevalence of infection in women whose only risk was age 20 years or younger was 7.5%; in the presence of any behavioral risk, prevalence among women aged 20 years or younger rose to 14.3%. However, in contrast to the FP population, only 6.3% of infected STD clients were older than 21 years of age or denied behavioral risk.
On the basis of these findings, the selective screening criteria chosen were age 20 years or younger and the presence of at least one behavioral risk. Use of nonbarrier or no contraceptive was not included because information describing clients' recent use of contraceptive method was not recorded, and because this variable was not independently associated with infection. As shown in Table 5, use of these criteria would require screening 52.7% of the FP clients, and 73.9% of the infections would have been detected. Adding cervicitis to the criteria increased the sensitivity from 73.9% to 78.2%. In the STD clients, 76.8% would be screened, and 93.7% of all infections detected; adding cervicitis to the criteria increased the sensitivity to 95.0%. The sensitivities of the selective screening criteria did not vary significantly by state, or according to the type of diagnostic test (data not shown).
The results of the incremental cost-effectiveness analysis are shown in Table 6. The total cost resulting from no screening was comparable in the two populations. Compared with no screening, selective screening in either setting saved approximately $1,000 for every case prevented. However, selective screening in STD clinics cost less per case prevented and prevented more cases than in the FP setting, reflecting the superior performance of the selective screening criteria in STD clients. Proceeding from selective screening to universal screening in the STD setting incurred a small net expenditure ($53 per case prevented). In contrast, in the FP setting, universal screening was more cost-effective than selective screening, saving an additional $667 per case prevented.
The break-even prevalence was defined as the population prevalence of C. trachomatis above which universal screening generated cost savings relative to selective screening. In the base-case scenario, the break-even prevalence for FP clinics was 3.1% (95% CI, 2.7–3.5%); that for STD clinics was 6.9% (95% CI, 6.5–7.3%). The higher break-even prevalence in STD clients reflects the higher sensitivity of the selective screening criteria in this group. Universal screening was the more cost-effective strategy in the FP population except at very low prevalence (≤3%). Substituting a more sensitive, more costly diagnostic test (such as LCR) for DFA in the base-case scenario had minimal effect, increasing the break-even prevalences to 3.3% for FP and 7.3% for STD. These break-even prevalences remained constant through the range of compliance with medical therapy assessed.
Figure 2 displays the results of the sensitivity analysis that assessed different probabilities of symptomatic or silent PID. The break-even prevalence was most sensitive to variation in the range of PID probabilities below 15%; increasing the probability of PID above this level had little effect on the break-even prevalence in either group.
In this study of women attending FP and STD clinics who were universally screened for chlamydia, age 20 years or younger, behavioral risks regarding sex partners, and cervicitis were independent risk factors for genital infection with C. trachomatis. Simple selective screening criteria composed of age 20 years or younger and any behavioral risk considerably reduced the number of women screened in both groups. However, whereas only 6% of chlamydial infections would be undetected in STD clients screened according to these criteria, 26% would be missed in FP clients. The addition of cervicitis to these criteria did not substantially change their sensitivity, increasing the percentage of total disease detected by only 1% to 4%. These results indicate that women who are more likely to have chlamydial infection can be identified, and a decision to test an individual patient for C. trachomatis made before pelvic examination is performed. Diagnostic testing may be indicated when signs of cervicitis are present,1,4 but cervicitis is not a necessary criterion for screening.
The difference in the performance of selective screening criteria in FP clients compared with STD clients strongly affected their relative cost-effectiveness. Among FP clients, in whom infection was less commonly associated with reported risk behaviors, universal screening was more cost-effective than selective screening at all but very low disease prevalences. In STD clients, because of the superior performance of the selective screening criteria, a higher prevalence was required (≥7%) for universal screening to be the more cost-effective strategy.
The diagnostic tests used in this study are now known to have sensitivities for the detection of cervical chlamydial infection substantially lower than those of amplified DNA tests such as LCR and PCR,28,29 and the prevalence of disease defined in our study was probably underestimated. However, break-even prevalences for selective versus universal screening were minimally altered by the use of LCR (assuming a 95% sensitivity and a $10 unit cost), suggesting that the benefit of increased detection of C. trachomatis offset the higher cost. Amplified DNA tests may detect infection in people whose risk profiles differ from those of our subjects, which could modify selective screening criteria themselves. Nevertheless, until large-scale studies of LCR and PCR assays can be undertaken, our study is applicable to tests that are likely to be used in many clinical settings for some time.
The risk of salpingitis associated with chlamydial infection is uncertain, but we believe that 25% (the figure used for “all PID” in our base-case scenario) is a conservative estimate when both symptomatic and silent salpingitis are considered. In the cost-effectiveness analysis, we varied the probability of salpingitis from 1% to 40% and demonstrated that the break-even prevalence was minimally affected by increasing the risk beyond 15% (Figure 2). Therefore, although better data are needed on the true risk of PID and fallopian tube damage after untreated infection, refined estimates are unlikely to affect substantially cost-effectiveness analyses of screening strategies.
Numerous other studies have also consistently identified young age, behavioral risks (usually a recent history of a new sex partner or more than one partner), and cervicitis as strong predictors of chlamydial infection in women.10–17 Although selective screening criteria based on these and other variables have been reliable in some settings,10,11 few criteria have maintained high sensitivity across different settings. For example, the criteria derived by Handsfield et al10 were sensitive (90%) and reduced the percentage of FP clients screened by about 35%, but were either insensitive or failed to significantly reduce the number of women screened when applied in other FP settings.7,12,53 Performance of selective screening criteria has also been variable in STD clinics.13,14
Our findings are in agreement with previous cost-effectiveness analyses, which concluded that universal screening for C. trachomatis is cost-effective in populations of women with disease prevalence above 2% to 6%.8,9,54–56 However, our study has several unique strengths. Our sample size approximates that of all prior studies combined. We assessed the independent contribution of cervicitis to disease detection, and showed that diagnosing cervicitis is not crucial to screening decisions. In the cost-effectiveness analysis, we accounted for asymptomatic PID and included indirect costs only for outcomes that developed in women with untreated chlamydial infection. We performed sensitivity analysis for critical probabilities (namely, PID and its sequelae) and used conservative probability estimates in the basecase scenario. Our decision analysis can be easily applied to other geographic settings by modifications that incorporate local variations in demographics, costs, disease prevalence, and choice of diagnostic tests.
Our study has several limitations. First, the limited value of cervicitis as an independent predictor of chlamydial infection in this study may reflect variability in its diagnosis or in the relative proportions of incident versus prevalent and symptomatic versus asymptomatic infections in these two settings. Second, limitations are inherent in any cost-effectiveness analysis when definitive data do not exist on the probability of some or all outcomes. The natural history of upper genital tract infection and the incidence of long-term sequelae may be modified by antibiotic therapy, or may vary in women with symptomatic versus asymptomatic PID. Although representative of many FP and STD clinics, the direct costs used in our model may underestimate costs in other settings, such as private physicians' offices. Third, lack of data on duration and consistency of current contraceptive use limited our ability to demonstrate an association with chlamydial infection. Fourth, we did not include several less common but costly outcomes in which C. trachomatis may play a causative role, such as low birth weight,57 postpartum endometritis, and premature labor58,59; more precise data on the risks of these outcomes will be required for consideration in future analyses. Finally, as for most cost-effectiveness analyses, the actual costs incurred or saved are imprecise; the comparative costs of the various strategies are more important than absolute costs.
The CDC recently issued recommendations for selective screening of chlamydial infections in women.1 When applied to our sample of FP clients, the CDC criteria had a sensitivity of 89%, compared with 74% for our criteria, but required screening 72% of women rather than 53%. In our STD clients, the CDC criteria had a slightly lower sensitivity than our own (91% vs. 94%), and would have required screening 70% of women rather than 77%. The CDC criteria are distinguished from ours primarily by their inclusion of nonbarrier or no contraceptive use as an indication for screening in women older than 20 years of age. In our study, the absence of an association between chlamydial infection and lack of barrier contraceptive use was unexpected; however, our subjects provided information regarding only current contraceptive use, and none on duration or consistency of use. In addition, we found substantial variability in reported barrier use among FP clients across the four states, which may reflect a general difficulty in reliably obtaining or recording this information in nonresearch clinical settings. Accordingly, contraceptive method may not be a practical selective screening criterion in some settings. Nevertheless, reliable contraceptive history collected from individual patients may prove useful in a decision to screen. Comprehensive analysis of the performance of the CDC's screening criteria in our study population will be published separately.
The results of our study imply a need to modify the approach to C. trachomatis screening in many FP and STD programs. Many STD programs screen universally, whereas most FP programs screen selectively, but our findings support the reverse strategy: at chlamydia prevalences common in most FP clinics (i.e., >3%) universal screening is preferred, whereas selective screening appears cost-effective at the intermediate prevalences (i.e., <7%) currently seen in many STD clinics. However, the minimal cost associated with proceeding to universal screening in STD clinics should be taken into account, particularly for clinics in which prevalence falls between 6.5% and 7.3% (the 95% CI limits on the break-even prevalence in our STD population); such clinics may elect to continue universal screening. For most STD clinics, where the chlamydia prevalence is greater than 7.3%, universal screening is the recommended strategy. The performance of selective screening criteria may change as chlamydia prevalence falls,19 and should be reevaluated periodically.
1. Centers for Disease Control and Prevention. Recommendations for the prevention and management of Chlamydia trachomatis
infections, 1993. MMWR Morb Mortal Wkly Rep 1993; 42(No. RR-12):1–9.
2. Washington AE, Johnson RE, Sanders LL. Chlamydia trachomatis
infections in the United States: What are they costing us? JAMA 1987; 257:2070–2072.
3. American Social Health Association. Update on fiscal year 1994: Funding for the STD Program of the CDC and NIH. Sex Transm Dis 1994; 21:55–58.
4. Stamm WE, Holmes KK. Chlamydia trachomatis
infections of the adult. In: Holmes KK, Mårdh P-A, Sparling PF, et al, eds. Sexually transmitted diseases. 2nd ed. New York: McGraw-Hill; 1990:181–193.
5. Stamm WE. Toward control of sexually transmitted chlamydial infections. Ann Intern Med 1993; 119:432–434.
6. Hillis SD, Joesoef R, Marchbanks PA, Wasserheit JN, Cates W, Westrom L. Delayed care of pelvic inflammatory disease as a risk factor for impaired fertility. Am J Obstet Gynecol 1993; 168:1503–1509.
7. Weinstock HS, Bolan GA, Kohn R, Balladares C, Back A, Oliva G. Chlamydia trachomatis
infection in women: A need for universal screening in high prevalence populations? Am J Epidemiol 1992; 135:41–47.
8. Trachtenberg AI, Washington AE, Halldorson S. A cost-based decision analysis for chlamydia screening in California family planning clinics. Obstet Gynecol 1988; 71:101–108.
9. Humphreys JT, Henneberry JF, Rickard RS, Beebe JL. Costbenefit analysis of selective screening criteria for Chlamydia trachomatis
infection in women attending Colorado family planning clinics. Sex Transm Dis 1991; 19:47–53.
10. Handsfield HH, Jasman LL, Roberts PL, Hanson VW, Kothenbeutel RL, Stamm WE. Criteria for selective screening for Chlamydia trachomatis
infection in women attending family planning clinics. JAMA 1986; 255:1730–1734.
11. Addiss DG, Vaughn ML, Golubjatnikov R, Pfister J, Kurtycz DF, Davis JP. Chlamydia trachomatis
infection in women attending urban Midwestern family planing and community health clinics: Risk factors, selective screening, and evaluation of non-culture techniques. Sex Transm Dis 1990; 17:138–146.
12. Winter L, Goldy AS, Baer C. Prevalence and epidemilogic correlates of Chlamydia trachomatis
in rural and urban populations. Sex Transm Dis 1990; 17:30–36.
13. Magder LS, Harrison RH, Ehret JM, Anderson TS, Judson FN. Factors related to genital Chlamydia trachomatis
and its diagnosis by culture in a sexually transmitted disease clinic. Am J Epidemiol 1988; 128:298–308.
14. Kent GP, Harrison HR, Berman SM, Keenlyside RA. Screening for Chlamydia trachomatis
infection in a sexually transmitted disease clinic: Comparison of diagnostic tests with clinical and historical risk factors. Sex Transm Dis 1988; 15:51–57.
15. Winter L, Goldy AS, Baer C. Prevalence and epidemiologic correlates of Chlamydia trachomatis
in rural and urban populations. Sex Transm Dis 1990; 17:30–36.
16. Sellors JW, Pickard L, Gafni A, et al. Effectiveness and efficiency of selective vs. universal screening for chlamydial infection in sexually active young women. Arch Intern Med 1992; 152:1837–1844.
17. Magder LS, Harrison RH, Ehret JM, Anderson TS, Judson FN. Factors related to genital Chlamydia trachomatis
and its diagnosis by culture in a sexually transmitted disease clinic. Am J Epidemiol 1988; 128:298–308.
18. Britten TF, DeLisle S, Fine D. STDs and family planning clinics: A regional program for Chlamydia
control that works. Am J Gynecol Health 1992; 6(3):24–31.
19. Addiss DG, Vaughn ML, Ludka D, Pfister J, Davis JP. Decreased prevalence of Chlamydia trachomatis
infection associated with a selective screening program in family planning clinics in Wisconsin. Sex Transm Dis 1993; 20:28–35.
20. Stamm WE, Harrison HR, Alexander ER, Cles LD, Spence MR, Quinn TC. Diagnosis of Chlamydia trachomatis
infection by direct immunofluorescence staining of genital secretions: A multicenter trial. Ann Intern Med 1984;101:683–41.
21. Clark A, Stamm WE, Gaydos C, et al. Multicenter evaluation of the AntigEnz chlamydia enzyme immunoassay for diagnosis of Chlamydia trachomatis
genital infection. J Clin Microbiol 1992; 30:2762–2764.
22. Kluytmans JA, Niesters HG, Mouton JW, et al. Performance of a nonisotopic DNA probe for detection of Chlamydia trachomatis
in urogenital specimens. J Clin Microbiol 1991; 29:2685–2689.
23. Stamm WE, Tamm MR, Koester M, Cles LD. Detection of Chlamydia trachomatis
inclusions in McCoy cell cultures with fluorescein-conjugated monoclonal antibodies. J Clin Microbiol 1983; 17:666–668.
24. Eisenberg JM. A guide to the economic analysis of clinical practices. JAMA 1989; 262:2879–2886.
25. Udvarhelyi IS, Colditz GA, Rai A, Epstein AM. Cost-effectiveness and cost-benefit analyses in the medical literature: Are the methods being used correctly? Ann Intern Med 1992; 116:238–244.
26. Stamm WE, Mårdh P-A. Chlamydia trachomatis.
In: Holmes KK, Mårdh P-A, Sparling PF, et al, eds. Sexually transmitted diseases. 2nd ed. New York: McGraw-Hill; 1990:917–925.
27. Dille BJ, Butzen CC, Birkenmeyer LG. Amplification of Chlamydia trachomatis
DNA by ligase chain reaction. J Clin Microbiol 1993; 31:729–731.
28. Chernesky MA, Lee H, Schachter J, et al. Diagnosis of Chlamydia trachomatis
urethral infection in symptomatic and asymptomatic men by testing first-void urine in a ligase chain reaction assay. J Infect Dis 1994; 170:1308–1311.
29. Lee H, Chernesky MA, Schachter J, et al. Diagnosis of Chlamydia trachomatis
genitourinary infection in women by ligase chain reaction assay of urine. Lancet 1995; 345:213–216.
30. Statistical Abstracts of the U.S., 1993. Table No. 756: Consumer Price by Major Groups: 1960 to 1992; Table No. 163: Consumer Price Indexes of Medical Prices: 1970 to 1992. Washington, DC: Department of Commerce, Bureau of the Census, 1993.
31. Rolfs RT, Galaid EI, Zaidi AA. Pelvic inflammatory disease: Trends in hospitalizations and office visits, 1979 through 1988. Am J Obstet Gynecol 1992; 166:983–990.
32. Centers for Disease Control and Prevention. A Practical Guide to Prevention Effectiveness: Decision and Economic Analyses. Atlanta, GA: Epidemiology Program Office; 1994.
33. Washington AE, Arno PS, Brooks MA. The economic cost of pelvic inflammatory disease. JAMA 1986; 255:1735–1738.
34. Stamm WE, Guinan ME, Johnson C, Starcher T, Holmes KK, McCormack WM. Effect of treatment regimens for N. gonorrhoeae
on simultaneous infection with C. trachomatis.
N Engl J Med 1984; 310:545–549.
35. Paavonen J, Kiviat N, Brunham RC, et al. Prevalence and manifestations of endometritis among women with cervicitis. Am J Obstet Gynecol 1985; 152:280–286.
36. Cumining DC, Honore LH, Scott JZ, Williams KE. Microscopic evidence of silent inflammation in grossly normal fallopian tubes with ectopic pregnancy. Int J Fertil 1988; 33:324–328.
37. Cates W, Joesoef MR, Goldman MB. Atypical pelvic inflammatory disease: Can we identify the clinical predictors? Am J Obstet Gynecol 1993; 169:341–346.
38. Cates W, Wasserheit JN. Genital chlamydial infections: Epidemiology and reproductive sequelae. Am J Obstet Gynecol 1991; 164:1771–1781.
39. Westrom L, Bengtsson LPH, Mårdh P-A. Incidence, trends, and risks of ectopic pregnancy in a population of women. Br Med J 1981; 282:15–18.
40. Westrom L, Joesoef R, Reynolds G, Hagdu A, Thompson SE. Pelvic inflammatory disease and fertility: A cohort study of 1844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results. Sex Transm Dis 1992; 19:185–192.
41. Brunham RC, Maclean IW, Binns B, Peeling RW. Chlamydia trachomatis:
Its role in tubal infertility. J Infect Dis 1985; 152:1275–1282.
42. Cates W, Rolfs RT, Aral SO. Sexually transmitted diseases, pelvic inflammatory disease, and infertility: An epidemiologic update. Epidemiol Rev 1990; 12:199–220.
43. Sellors JW, Mahoney JB, Chernesky MA, Rath DJ. Tubal factor infertility: An association with prior chlamydial infection and asymptomatic salpingitis. Fertil Steril 1988; 49:451–457.
44. Svensson L, Mårdh PA, Westrom L. Infertility after acute salpingitis with special reference to Chlamydia trachomatis.
Fertil Steril 1983; 40:322–329.
45. Hirsch MB, Mosher WD. Characteristics of infertile women in the United States and their use of infertility services. Fertil Steril 1987; 47:618–625.
46. Westrom L. Incidence, prevalence, and trends of acute pelvic inflammatory disease and its consequences in industrialized countries. Am J Obstet Gynecol 1980; 138:880–892.
47. Schachter J, Grossman M, Sweet RL, Holt J, Jordan C, Bishop E. Prospective study of perinatal transmission of Chlamydia trachomatis.
JAMA 1986; 255:3374–3377.
48. Hammerschlag MR, Anderlea M, Semine DZ, McComb D, McCormack WM. Prospective study of maternal and infertile infection with Chlamydia trachomatis.
Pediatrics 1979; 64:142–148.
49. Tartaglione TA, Russo ME. Pharmacology of drugs used in venereology. In: Holmes KK, Mårdh P-A, Sparling PF, et al, eds. Sexually transmitted diseases. 2nd ed. New York: McGraw-Hill; 1990:993–1019.
50. Katz BP, Caine VA, Jones RB. Estimation of transmission probabilities for chlamydial infection. In: Bowie WR, Caldwell HD, Jones RB, et al, eds. Chlamydia infections. New York: Cambridge University Press; 1990:567–570.
51. Viscidi RP, Bobo L, Hook EW, Quinn TC. Transmission of Chlamydia trachomatis
among sex partners assessed by polymerase chain reaction. J Infect Dis 1993; 168:488–492.
52. Stamm WE, Cole B. Asymptomatic Chlamydia trachomatis
urethritis in men. Sex Transm Dis 1986; 13:163–165.
53. Stergachis A, Scholes D, Heidrich FE, Sherer DM, Holmes KK, Stamm WE. Selective screening for Chlamydia trachomatis
infection in a primary care population of women. Am J Epidemiol 1993; 138:143–153.
54. Nettleman MD, Jones RB, Roberts SD, et al. Cost-effectiveness of culturing for Chlamydia trachomatis:
A study in a clinic for sexually transmitted diseases. Ann Intern Med 1986; 105:189–196.
55. Phillips RS, Aronson MD, Taylor WC, Safran C. Should tests for Chlamydia trachomatis
cervical infection be done during routine gynecologic visits? Ann Intern Med 1987; 107:188–194.
56. Genc M, Mårdh P-A. A cost-effectiveness analysis of screening and treatment for Chlamydia trachomatis
infection in asymptomatic women. Ann Intern Med 1996; 124:1–7.
57. Wager GP, Martin DH, Koutsky L, et al. Puerperal infectious morbidity: Relationship to route of delivery and to antepartum Chlamydia trachomatis
infection. Am J Obstet Gynecol 1980; 138:1028–1033.
58. Martius J, Krohn MA, Hillier SL, Stamm WE, Holmes KK, Eschenbach DE. Relationship of vaginal Lactobacillus
species, cervical Chlamydia trachomatis,
and bacterial vaginosis to preterm birth. Obstet Gynecol 1988; 71:89–95.
59. Gravett MG, Nelson HP, DeRouen T, Critchlow C, Eschenbach DA, Holmes KK. Independent associations of bacterial vaginosis and Chlamydia trachomatis
infection with adverse pregnancy outcome. JAMA 1986; 256:1899–1903.