APPROXIMATELY 4.5 MILLION chlamydial infections occur each year, making Chlamydia trachomatis the most common bacterial sexually transmitted disease (STD) in the United States.1 In women, undiagnosed and therefore untreated chlamydial infections can lead to pelvic inflammatory disease (PID) and its sequelae: chronic pelvic pain, ectopic pregnancy, and infertility. An estimated 500,000 cases of PID can be attributed to chlamydial infections,2 costing in excess of $1 billion each year.3 These costs represent a significant financial burden on the health sector and society in general.
To limit chlamydial infection and the development of its sequelae in an infected individual, to control the transmission of infection, and to contain the subsequent costs, partner notification programs have been implemented by public health departments as well as by private providers. In public health clinics (clinics focusing on reproductive health-STD, family planning clinics, and the like), disease intervention specialists (DIS) interview and counsel index patients infected with C. trachomatis. Identifying information for recent sex partners is collected and can be used actively to locate, notify, and offer therapy to the identified partners.
In this study, we used decision analysis to estimate the marginal cost-effectiveness of two health department partner notification strategies directed at decreasing the incidence of PID and its sequelae in women: strategy 1, locating and providing treatment for the female sex partners of male index patients; and strategy 2, locating and providing treatment for the male sex partners of female index patients. Strategy 1 offers early treatment for women who may be unaware of their infection, preventing the development of sequelae and limiting the extent of sequelae that may have already developed. Strategy 2 limits the duration of disease in male sex partners, decreasing the potential for reinfection of the female index patients who have received treatment, thus decreasing the chances of sequelae. Both strategies prevent second-generation transmission of chlamydial infection to sex partners not yet infected; however, this study focuses on prevention of PID and does not address prevention of second-generation transmission of chlamydial infection.
Taking a health care system perspective (one that includes the medical costs of inpatient and outpatient care for C. trachomatis infection and its sequelae), we developed a cost- and outcome-based decision model to assess the independent cost-effectiveness of strategies 1 and 2. We compared two hypothetical populations from a health care system perspective, those participating in each of the two partner notification strategies and those not participating in any type of partner notification. Computer-generated models (Figs. 1 and 2) were designed and analyzed using SMLTREE decision software (Version 2.9; James Hollenberg, Roslyn, NY).
Partner notification can be conducted in several different ways.4 We based our model on the program alternatives found to be cost-effective in previous studies.5–7 In our model, partner notification interviews are conducted by a health department DIS.5,6 When a sex partner presents for treatment, both male and female partners are presumptively treated with azithromycin.7 Outcome was measured by cases of PID prevented per 1,000 index patients, and costs included program expenditure as well as the benefit of averted medical costs. The probabilities and costs associated with each strategy were obtained from the published medical literature and expert opinion. Because probabilities and costs may differ by locality, we performed sensitivity and threshold analyses.
We calculated the number of cases of PID resulting from strategies 1 and 2, and no partner notification with a baseline ratio of one sex partner identified per index patient as well as the conditional probabilities that (1) a named partner will be brought in for treatment, (2) a named partner is infected, (3) an infected male partner is asymptomatic, (4) a symptomatic patient will seek treatment in the absence of intervention, (5) a female index patient will be reinfected by her untreated partner, and (6) an infected woman will contract PID if the chlamydial infections is not treated.
In Colorado Springs, Colorado, Zimmerman et al. found a chlamydial infection prevalence of 58% for female sex partners and 67% for male partners based on direct immunofluorescence assay (Table 1).8 Attempts by the local health department resulted in 55% of named female partners and 67% of named male partners receiving treatment.8 PID will develop in from 15% to 40% of women infected with C. trachomatis who do not receive treatment.11–13 We conservatively estimated that PID will develop in 20% of women with a first chlamydial infection if they do not receive treatment. The increased probability of PID associated with recurrent chlamydial infection is unknown. However, there is evidence of increased relative risk with two or more infections.14,15 We estimated that PID will develop in 30% of women with untreated recurrent chlamydial infections. In the absence of intervention, women may become reinfected if their infected male sex partner lacks clinical signs associated with chlamydial infection and does not seek treatment. We estimated the likelihood that a female would be reinfected in the absence of partner notification according to the following: Some symptomatic men may not recognize infection and will probably not seek care on their own. Combining a 43%8,16–18 chance of asymptomatic infection in men, an estimated 5% probability that symptomatic men will not receive treatment, and a 58% female partner chlamydial infection incidence yields a 27% chance of female reinfection in the absence of intervention.
The model incorporated the direct program costs associated with partner notification, and the direct medical costs associated with therapy and treatment of C. trachomatis infection and its sequelae (Table 2). Indirect costs, such as lost production, and intangible costs, such as pain and suffering, were not estimated. All cost data were adjusted to 1994 U.S. dollars, using the medical care component of the Consumer Price Index.19 A 5% discount rate was applied to future costs and benefits. Katz et al reviewed records of 338 male index patients with gonorrhea, nongonococcal urethritis, or known chlamydial infections seen in an Indianapolis STD clinic, and estimated that a DIS interview takes 22.5 minutes at an average cost of $24 per index patient with one named sex partner.5 Locating costs, including time (average, 30.1 minutes) and expenses, were estimated by Katz et al to be $36 per named partner.5 A single 1-g dose of azithromycin costs the public health department $9.50 per patient treated (public health price).
We used a weighted cost of $4,891 per case of PID as estimated in a cost-effectiveness analysis of azithromycin versus doxycycline for treatment of chlamydial infection by Haddix et al.7 Haddix et al calculated the weighted cost of PID by combining the probability of sequelae (hospitalization for PID, chronic pelvic pain, ectopic pregnancy, and infertility) and the associated costs as estimated by Washington and Katz who reviewed discharge data from California acute care hospitals, charges in a San Francisco hospital, and cost data from San Francisco practices and the American Medical Association.3 Fourteen percent of PID cases will result in hospitalization within 12 months of chlamydial infection, for an average cost of $1,961 per episode.3,20 Women with untreated PID have an 18% chance of experiencing chronic pelvic pain necessitating surgery 2 years after the chlamydial infection, at an average cost of $12,392. An ectopic pregnancy will occur in approximately 6% of untreated women in the fifth year after the chlamydial infection, at an average cost of $11,938, and approximately 20% of untreated women will have tubal factor infertility. Twenty-five percent of these women will seek infertility treatment 10 years after the chlamydial infection at an average cost of $4,956.3,21,22
In a hypothetical cohort of 1,000 male index patients not receiving partner notification activities, 116 cases of PID would occur in their female sex partners, costing $567,000 in sequelae costs. Alternatively, implementation of partner notification activities in this cohort (strategy 1) prevents 64 cases of PID, 12 cases of chronic pelvic pain, 4 ectopic pregnancies, and 13 cases of tubal factor infertility. Strategy 1 compared with no partner notification results in a cost savings of $247 per index patient and a health care savings of $3,900 per case of PID prevented (Table 3). In the hypothetical cohort of 1,000 female index patients not receiving partner notification, 30 cases of PID would occur due to reinfection, costing $147,000 in sequelae costs. Strategy 2 prevents 20 cases of PID, 4 cases of chronic pelvic pain, 1 ectopic pregnancy, and 5 cases of tubal factor infertility. Compared with no partner notification, strategy 2 (which prevents reinfection in the female index patient) results in a cost savings of $33 per index patient and a health care savings of $1,600 per case of PID prevented (Table 3).
Sensitivity analyses reveal that strategy 1 is cost-effective in preventing PID across a wide range of parameters. Break-even values are the parameter values where the cost of partner notification strategy is equal to that of no partner notification; Table 4 gives break-even values for each parameter. Strategy 1 provides a cost savings to the health care system as long as at least 11% of the named female partners receive treatment, and when the chlamydial prevalence in female partners is greater than 12%. Increases in program costs do not affect the cost-saving ability of strategy 1 until costs rise above $300 per index patient. Decreases in the probability of PID do not affect the cost savings of strategy 1 unless PID develops in fewer than 4% of untreated women.
Sensitivity analyses reveal that strategy 2 is cost-effective in baseline analyses. Yet, this strategy is sensitive to changes in several parameters (Table 4). If fewer than 43% of the named male partners present for treatment, strategy 2 loses its cost-saving ability. Strategy 2 provides a cost savings to the health care system as long as the chance of a female index patient becoming reinfected is greater than 18%, or PID develops in more than 20% of women with untreated reinfection. Increases in program costs do not affect the cost-effectiveness of strategy 2 unless costs increase above $92 per patient. Decreases in the cost of a case of PID do not affect the cost-effectiveness of strategy 2 until the cost of a case of PID and other sequelae decreases below $3,300.
We expanded the analysis of strategy 1 to reflect the cost effectiveness of this strategy when an index patient names more than one sex partner (Table 5). We estimated that an increase in the number of partners named in a partner notification interview will increase the cost of the interview by 2% per additional partner. If the number of female sex partners per male index patient increases from one to five, the cost-saving capability of strategy 1 increases by more than fivefold (a 109% increase in savings per each partner for two or more partners). Because strategy 2 models reinfection rather than initial infection, this method of expansion cannot be applied.
Partner notification of both male and female partners helps identify disease in the individual and is beneficial from an individual health perspective. From a health care system perspective, partner notification reduces the incidence of C. trachomatis infection, PID, and other sequelae in the population, and in many public health settings results in a net savings to society. At baseline, both strategy 1 and strategy 2 are cost-effective, saving $3,900 per case of PID prevented and $1,700 per case of PID prevented, respectively. The cost effectiveness of strategy 1 is robust across a wide range of probabilities and cost values, where most of the break-even values fall outside the probability and cost ranges likely to be seen in a public health setting. The cost-effectiveness of strategy 2 is robust across a wide range of probability and cost values, yet it appears sensitive to changes in the probability of reinfection or the probability of PID after recurrent infection. As the probability of reinfection or the probability of PID after recurrent chlamydial infection increases, the cost-saving ability of strategy 2 increases. For example, if fewer than the estimated 95% of all symptomatic men who seek care receive treatment in the absence of intervention, or if female partner prevalence is higher than that used here, the probability of reinfection in the female index patient increases, increasing the cost-saving ability of strategy 2.
We believe the estimates used in this analysis are conservative, the true values being much higher. A study in Baltimore, Maryland analyzing paired specimens from sex partners indicated 67% increases in male chlamydial prevalence and approximately 4% increases in female prevalence when diagnosis is made by polymerase chain reaction rather than by cell culture.10 The true chlamydial prevalence in partner populations may be substantially underestimated in clinics using direct fluorescent antibody tests or cell culture, and many infected individuals may not be receiving treatment. This ongoing research implies that many infected individuals may be missed when assays with low sensitivities are used. The true risk of female reinfection is difficult to estimate because of the subtle differences between reinfection, new infection, and persistent infection. Studies are in progress to assess these rates. In addition, 50% to 70% of all PID is subclinical,20,23,24 leading us to believe that a 20% chance of PID after an untreated initial chlamydial infection and a 30% chance of PID after recurrent chlamydial infection are conservative estimates owing to the probable high level of unrecognized disease. A 25% increase in the probability of PID would increase the cost-saving potential of strategy 1 by 32%, and that of strategy 2 by 75%.
We used conservative estimates of the benefits associated with PID prevention. We did not include indirect and intangible benefits of partner notification, such as the prevented pain and suffering associated with PID, ectopic pregnancy, and infertility. In addition, 25% to 50% of infants born to infected mothers acquire infant conjunctivitis, and 8% to 16% acquire chlamydial pneumonia.25,26 Prevention of C. trachomatis infection and timely eradication of infection in pregnant women prevents the development of these infections in their newborns. We did not model the effect of prevented second-generation infection. Inclusion of one or all of these additional benefits would increase the cost-effectiveness of both strategies 1 and 2.
One advantage of a such model is that it can illustrate where critical data are lacking and what research is needed to help inform policy. We found few studies investigating the clinical features of C. trachomatis infection and health-seeking behavior with regard to STDs. Specifically, refined estimates of PID resulting from untreated chlamydial infection as well as the probability of reinfection in the absence of intervention are needed. Studies exploring the accuracy of information provided by index patients in DIS-conducted interviews would also improve our ability accurately to estimate the benefits of health department partner notification. Future studies comparing the cost-effectiveness of partner notification and other prevention programs, such as selective screening, would allow for development of an optimally cost-effective set of prevention strategies. In addition, because of lack of data from the United States, we used European studies for several key estimates. These studies may not be generalizable to U.S. settings because of cultural differences in sexual behavior. More U.S.-based studies are needed to provide the data necessary for accurate modeling.
Using economics in medical decision making aids policy makers in most effectively obtaining health value for each dollar spent. The results of our study suggest that health department partner notification of the female sex partner of infected male patients should be implemented in settings where female partner prevalence is at least 12%. If male sex partners do not receive treatment, the chlamydial infection can be passed back to the female index patient, reintroducing the potential for sequelae and failing to break the disease transmission cycle. Health department partner notification of the male sex partners of infected female patients should be implemented when the chlamydial infection prevalence in male partners is at least 29%.
Partner notification has the potential to effectively control the spread of C. trachomatis infection, prevent the development of sequelae, and save health care costs in the future. As such, partner notification of both men and women deserves consideration in public health programs designed to prevent and control the transmission of C. trachomatis infection and its sequelae.
1. Washington AE, Johnson RE, Sanders LL. Chlamydia trachomatis
infections in the united states: what are they costing us? JAMA 1987; 257:2070–2072.
2. Quinn, T. Update on Chlamydia trachomatis
infections. Infections in Medicine 1994; 11:210–211.
3. Washington AE, Katz P. Cost and payment source for pelvic inflammatory disease. JAMA 1991; 266:2565–2569.
4. Millson ME, Rasooly I, Scott EAF, et al. Partner notification for sexually transmitted diseases: Proposed practice guidelines. Can J Public Health 1994; 85:S53–S55.
5. Katz BP, Danos TS, Quinn TS, Caine V, Jones RB. Efficiency and cost-effectiveness of field follow-up for patients with Chlamydia trachomatis
in a sexually transmitted disease clinic. Sex Transm Dis 1988; 15:11–16.
6. Alary M, Joly JR, Poulin C. Gonorrhea and chlamydial infection: Comparison of contact tracing performed by physicians or by a specialized service. Can J Public Health 1991; 82:132–134.
7. Haddix A, Hillis SD, Kassler W. The cost-effectiveness of single-dose therapy for Chlamydia trachomatis
infections in women. Sex Transm Dis 1995; 22:274–280.
8. Zimmerman HL, Potterat JJ, Dukes RL, et al. Epidemiologic differences between chlamydia and gonorrhea. Am J Public Health 1990; 80:1338–1342.
9. Thelin I, Wennstrom AM, Mardh P-A. Contact-tracing in patients with genital chlamydial infection. British Journal of Venereal Disease 1980; 56:259–262.
10. 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.
11. Jones RB, Mammel JB, Shepard M, Risher RR. Recovery of Chlamydia trachomatis
from the endometrium of women at risk for chlamydial infection. Am J Obstet Gynecol 1986; 155:35–39.
12. Rees E. The treatment of pelvic inflammatory disease. Am J Obstet Gynecol 1980; 138:1042.
13. Stamm WE, Guinan ME, Johnson C, Starcher T, Holmes KK, McCormack WM. Effect of treatment regimes for Neisseria gonorrhea
on simultaneous infection with Chlamydia trachomatis.
N Engl J Med 1984; 310:545–549.
14. Hillis SD, Owens LM, Marchbanks PA, Amsterdam LE, MacKenzie WR. Recurrent chlamydial infections increase the risk of ectopic pregnancy. In: Orfila J, et al, eds. Am J Obstet Gynecol 1997; 176:103–107.
15. Hillis SD, Harms L, Marchbanks P, Amsterdam LE, MacKenzie WR. Recurrent chlamydia infections as a risk factor for ectopic pregnancy. In: Orfila J, et al, eds. Chlamydial Infections. Proceedings of the Eighth International Symposium on Human Chlamydial Infections; June 19–24, 1994, Gouvieux-Chantilly, France; Bologna, Italy: Societa Editrice Esculapio; 1994: pp 611–613.
16. Stamm WE, Koutsky LA, Benedetti JK, Jourden JL, Brunham RC, Holmes KK. Chlamydia trachomatis
urethral infection in men: Prevalence, risk factors and clinical manifestations. Ann Intern Med 1984; 100:47–51.
17. Stamm WE, Cole B. Asymptomatic Chlamydia trachomatis
urethritis in men. Sex Transm Dis 1986; 13:163–165.
18. Fish ANJ, Fairweather DVJ, Oriel JD, Ridgeway GI. Chlamydia trachomatis
infection in a gynecology clinic population: Identification of high-risk groups and the value of contact tracing. Eur J Obstet Gynecol Reprod Biol 1989; 31:67–74.
19. Statistical Abstract of the United States, 1994. Table No. 163: Consumer Price Indexes of Medical Care Prices: 1970–1994; Table No. 756: Consumer Price Indexes by Major Groups: 1960 to 1994; U.S. Bureau of Labor Statistics, CPI Detailed Report, January, 1995. Monthly Labor Review and Handbook of Labor Statistics (periodic).
20. Cates W, Joesoef MR, Goldman MB. A typical pelvic inflammatory disease: Can we identify the clinical predictors? Am J Obstet Gynecol 1993; 169:341–346.
21. Fuchs VR, Perrault L. Expenditures for reproduction-related health care. JAMA 1986; 225:76–81.
22. Forest JD, Gold RB, Kenney AM. The Need, Availability and Financing of Reproductive Health Services. New York: Alan Guttmacher Institute, 1989.
23. 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.
24. Cates W, Wasserheit JN. Genital chlamydial infections: Epidemiology and reproductive sequelae. Am J Obstet Gynecol 1991; 164:1771–1781.
25. Schacter J, Grossman M, Sweet RL, Holt J, Jordan C, Bishop E. Prospective study of perinatal transmission of Chlamydia trachomatis.
JAMA 1986; 255:3374–3377.
26. Hammerschlag MR, Anerlea M, Semine DZ, McComb D, McCormick WM. Prospective study of maternal and fetal infection with Chlamydia trachomatis.
Pediatrics 1979; 64:142–148.