Andersen, Berit PhD*†; Gundgaard, Jens‡; Kretzschmar, Mirjam PhD§; Olsen, Jens‡; Welte, Robert PhD∥; Øster-gaard, Lars Dr Med Sci*
Chlamydia trachomatis infection is the most common treatable, sexually transmitted disease in industrialized countries.1 Some 50%–75% of C trachomatis-infected individuals are asymptomatic,2 and the ensuing nontreatment of such individuals causes the spread of the infection to continue. Untreated infections may result in pelvic inflammatory disease (PID) in women3,4 and may subsequently lead to infertility, ectopic pregnancy,5,6 or chronic pelvic pain (CPP).7 Newborn children of infected mothers are at risk of developing conjunctivitis or pneumonia.8 Untreated men are at risk of getting epididymitis.9 In general, the infection is easily and effectively treated with antibiotics.10
Opportunistic screening programs have been conducted11 and partner notification has been undertaken12 to control the spread of the disease, but its prevalence remains high in western countries despite these initiatives. In recent years, use of nucleic acid amplification tests (NAATs) and noninvasive samples for C trachomatis testing13–16 has paved the way for the implementation of outreach disease control strategies. Home sampling has been presented as a test principle,16 and 2 randomized trials have demonstrated that home sampling improves screening participation2 and outperforms current strategies in terms of number of partners tested.17 Home sampling would therefore seem to enjoy the potential of becoming an important future diagnostic avenue in efforts aimed at improving public health through a lowering of the number of urogenital complications and of the prevalence of this disease.
Any new diagnostic strategy should be evaluated in terms of its predicted costs, effectiveness (complications avoided), and influence on disease control. Several studies have evaluated the costs and effectiveness of introducing new diagnostic strategies for C trachomatis, but most of them have been based on static models.18 Urogenital chlamydia infections are, however, communicable, and improved detection and treatment will accordingly lower the risk of contracting a new infection among other members of a community, and more money will be saved over time. On the contrary, saved money will be lost in cases of reinfection. These opposing effects call for dynamic in contrast to static models to evaluate the effects of C trachomatis screening in order to obtain valid results of cost-effectiveness and to properly predict how a program affects disease control.19–23
This study deployed a dynamic simulation model to estimate the incremental effects in terms of avoided complications, disease control, and costs of implementing a population-based strategy involving home sampling for C trachomatis screening and partner notification as an alternative to the current strategy.
Materials and Methods
Current Danish Strategy
The current strategy is based on guidelines from the National Board of Health about diagnosis and treatment of C trachomatis infections.24 Testing is recommended if an individual shows up in a doctor’s office and presents symptoms or signs of chlamydia infection or if it is known that the individual has been exposed to chlamydia or another sexually transmitted disease. Women shall also be tested before transcervical procedures, and testing shall be considered if women in younger age groups have gynecological examinations performed for other reasons (opportunistic screening). There is no ongoing systematic screening program in Denmark. All Danish men and women, however, can have a free test for C trachomatis regardless of age, and if you want a test you can seek a general practitioner (GP) or a sexually transmitted disease clinic (STD-clinic) and discuss the relevance with health care professionals. Ninety percent of all samples for C trachomatis testing are obtained in general practices and the remaining 10% in STD-clinics and hospitals.25 If a patient has a positive test result, he or she will receive treatment from the health care provider requesting the test. Partner notification is performed by means of patient referral, i.e., the patient is asked to inform the partner about the potential exposure and inform him/her to seek at doctor for test and treatment.26
Alternative Home-Sampling Strategy
In the possible future alternative strategy, young individuals seen by a GP or a physician because of chlamydia symptoms or exposure to a sexually transmitted infection will have a free chlamydia test. The same applies for women undergoing transcervical procedures. Moreover, all 15- to 24-year-olds will yearly receive a direct mail invitation to be tested by means of a self-administered sample taken in the privacy of their own home, which should be sent to a central diagnostic laboratory (home sampling). Men will receive the actual test kit comprising a urine test tube, together with the invitation letter, and women will receive a postcard to be returned in order to have a test kit consisting of a vaginal swab mailed to them. The difference in contact procedure is based on an earlier study showing equal effect by means of detected infections for women, while men benefited from having the test kit provided, together with the invitation letter.2 The laboratory will mail the test result to the individual, and if it is positive, the infected individual must see a health care provider for antibiotic treatment and partner notification. The treating health care provider will hand out home sampling kits for the index patient to give to his or her partner(s) during the last 12 months. The partner(s) home sampling procedure will be the same as that described above.
Data Used in the Model
Data on Current Strategy for C trachomatis Testing.
Data on the number of tests and detected infections in Denmark were obtained from the Danish surveillance system for C trachomatis testing. In the year 2000, 24% of 15- to 24-year-old women and 4% of 15- to 24-year-old men were tested, and the prevalence of infection in the obtained samples was 25.7% for men and 8.3% for women (Table 1).25 Data on the prevalence of infection according to test indications (opportunistic screening/partner tracing/symptoms/transcervical procedures) were obtained from a Danish observational test indication study27 assessing a total of 11,423 samples (10,351 female samples and 1072 male samples).
Data on Home Sampling for Screening.
To study the effectiveness of a population-based screening program using home-obtained, mailed samples, we conducted a randomized study among 21- to 23-year-old men and women living in Aarhus County in October 1997.2 A total of 4000 women and 5000 men were given the opportunity of being tested by a home-obtained, mailed sample. We found an increased effect in terms of detected infections among both men and women. Among women who received a reply card with a test invitation, 24% were tested only by a home-obtained sample (infection prevalence, 8.0%). Among men receiving a test invitation, 25% were tested only by a home-obtained sample (infection prevalence, 5.9%). The participation and infection prevalence rates were considered to be the same among the15- to 24-year-olds and the 21- to 23-year-olds.
Data on Partner Testing.
We conducted a randomized multicenter study (4 Danish centers) where physicians treating C trachomatis–positive individuals were asked to include patients in the study.17 Patients accepting participation were randomized into 2 groups: Patients in the intervention group received a package containing home sampling kits for their partners during the last 12 months, while patients in the control group received the usual advice to contact partners and tell them to seek a health care provider for a test. A total of 414 women and 148 men accepted participation in this study. Among these, 74% of the female index patients and 47% of the male index patients in the intervention group had at least 1 partner tested for infection. In the control group, the corresponding numbers were 41% for women and 13% for men.
Parameter values pertaining to sexual behavior like sexual activity (high or low), type of partnership, and mix of partnerships were extracted from a Dutch survey from 1991.28 Key variables from the Dutch study were compared with available Danish data on sexual behavior,29 and no significant differences were found. A selection of behavioral data is seen in Table 2.
Natural History of the Chlamydia Infection
Data on transmission risk, incubation time, and recovery rates were obtained by a previous review of current literature and have been published previously.21 Key parameters are given in Table 2. Estimated risks of complications due to asymptomatic untreated C trachomatis infections were based on a review of the literature and are shown in Table 3.
Health Care Use and Costs
Unit cost estimates were calculated on the basis of assessment of the health care resources used for testing and treatment, diagnosis-related group (DRG) fees (based on the Danish case mix system of diagnosis-related groups30), GP reimbursement fees,31 and market prices for drugs (antibiotics). An expert panel provided estimates of the fraction of patients requiring in-hospital treatment versus outpatient treatment and estimates of the time lost due to C trachomatis complications. The productivity loss was evaluated on the basis of the average labor costs in the private sector and were adjusted for age, sex, and unemployment.32–34
Epidemiologic Simulation Model
The simulation model was developed by Kretzschmar and colleagues20,21 and estimates the yearly incidence and prevalence of C trachomatis in societies with different strategies for screening and partner notification. The model is an individual-based stochastic model based on Monte Carlo simulations in a population of 10,000 heterosexual men and women. The age range of the population is 15 to 64 and the sex ratio is 1:1, with a uniform age distribution. In the simulation model, 100 runs are performed for each of the chosen scenarios.
Initially, model simulations were performed for the current strategy in Denmark until a steady-state situation was reached (baseline). At that point, we introduced home sampling for screening and partner notification, and the simulations were run for a 10-year period. Only the effect of partner notification related to test of current partner was taken into account.
To differentiate the effect of home-based systematic screening and intensified partner notification on the future prevalence of infection, we separated the effect of these 2 parameters. Thus, we also simulated the effect of continuing the ongoing Danish opportunistic screening program supplemented with improved home-based partner notification and the effect of implementing population-based screening without improving partner notification.
Based on simulations from the epidemiologic model, we estimated the number of major outcomes averted (MOA) that could be gained each year as a consequence of implementing the new strategy. Subsequently, the averted costs and the increased test and treatment costs were estimated in order to calculate the costs per MOA. A societal perspective was adopted, and all future costs and effects were discounted at 3% per year. A mean of 10 years of simulation of the present Danish strategy was used as base case.
Complications, here defined as MOA, were symptomatic PID, CPP, infertility, ectopic pregnancy, and neonatal pneumonia, which are the most painful or severe C trachomatis–related complications. In the model, C trachomatis cases were divided into symptomatic and asymptomatic cases. Symptomatic cases were defined as individuals seeing a GP for a C trachomatis test because of symptoms and subsequently being diagnosed and treated for the infection. We assumed that their treatment would show no complications, but asymptomatic cases were considered to be at risk of complications. The possible consequences for these individuals are given in a flowchart (Fig. 1), and the estimated probability of each complication is featured in a table (Table 3).
Direct and indirect costs were computed in the national currency (DKK) at 2002 prices, and the total cost was converted into US dollars ($1 = DKK 8).
The costs were broken down into direct intervention costs, averted direct costs, and indirect costs. Direct intervention costs are changes in costs due to the increase in testing (e.g., additional tests, launch and operation of the program, fixed costs such as rent, etc.) and the use of antibiotics for treatment. Averted direct costs are costs avoided owing to the drop in the number of complications as a consequence of the program. Indirect costs include the value of productivity lost due to absenteeism from work in connection with chlamydia complications. The human capital approach was applied to measure the productivity loss of paid work, while unpaid work (e.g., household work, informal care) was not considered.
The costs of C trachomatis complications were discounted through estimates of the period (in years) between infection and appearance of complications. It was assumed that PID, CPP, neonatal conjunctivitis, neonatal pneumonia, and epididymitis would occur in the year of C trachomatis infection. The probability of having an ectopic pregnancy or being identified as infertile would depend on the active interest in having a child, which was approximated by fertility rates and rates for women with at least 1 child. We discounted the period between infection and the year of attempted conception plus an estimated 2 years for diagnosis and subsequent in vitro fertilization (IVF).35
Intangible costs like anxiety and discomfort in connection with C trachomatis infections or complications were not estimated.
Antibiotic Efficacy and Positive Predictive Value
The epidemiologic model estimates the number of asymptomatic, infected, effectively treated individuals found by screening or partner notification for each year. However, to estimate the direct intervention costs more precisely, we estimated the total number of treated individuals, including false-positive individuals and ineffectively treated persons. For infections detected as part of current strategy, we estimated that 100% would receive treatment for the infection because the testing doctor is responsible for contacting the tested individual in cases of untreated, detected infections. For infections detected as part of the alternative home sampling strategy, we estimated that 95% would receive treatment corresponding to results from a earlier screening study (unpublished data).
To test the robustness of the model, we performed univariate sensitivity analysis on the following parameters: discount rate, level of partner notification, expansion of age group to include all 15- to 29-year-old men and women, risk of PID, and costs of screening test.
Each of the 4 strategies was implemented among 15- to 24-year-olds (Fig. 2). Continuation of the current strategy yielded an infection prevalence close to 4%, which is roughly equivalent to the present level; implementation of the total alternative home sampling strategy for screening and partner notification caused the prevalence among 15- to 29-year-olds to decline from 4.2% to 1.0% over a 10-year period; systematic screening without improved partner notification produced a 10-year prevalence of 2.9%, but the prevalence dropped to 1.8% if partner notification was improved and the ongoing opportunistic screening program was continued.
Costs may be considered from an overall societal perspective and from the narrower perspective of the health care sector. Assuming the latter, only direct costs are relevant. The direct costs peaked in the first year of screening ($14,747) but decreased during the 10-year period and became slightly negative in year 10 (Table 4). The initial rise in direct costs sprang from the need for extra human resources to initiate the program, and the lower costs at the end of the 10-year period are explained by the discovery of few new infections in year 10 causing a decline in patients’ consumption of primary health care resources and prescription antibiotics (data not shown). The incremental direct cost per MOA (incremental cost-effectiveness ratio, ICER) was $4096 ($14,747/3.6) in year 1. The program carried a direct cost till year 9 (ICER = $22.6 [$488/21.5]), but became slightly cost saving in year 10 (calculations based on Table 4). After 10 years of screening, the ICER was $292 ($47,226/161.5) (no discounting), and the total alternative program was not cost saving compared with the current Danish strategy (calculations based on Table 4).
A societal cost perspective requires inclusion of both direct and indirect costs. The ICER was $3186 in year 1, $164 in year 2, but in year 3 the indirect costs exceeded the direct costs and the program effectively became cost saving from a societal point of view (Table 4). Looking at the accumulated direct and indirect costs (Fig. 3), we see that the indirect costs offset the direct costs between years 3 and 4, and the program is subsequently cost-saving to society.
The MOA were mainly PID (83%) and CPP (12%) (Table 4).
The univariate sensitivity analysis (Table 5) showed that the estimated costs were very sensitive to the degree of partner referral. If improved partner referral was not included in the program, the ICER (including all costs) changed from domination to $1535 within 10 years. The estimated risk of PID for asymptomatic untreated women was also a very important parameter, but if this risk was estimated at 10% instead of 20%, the alternative strategy was still dominant. Expansion of the age group to include all 15- to 29-year-olds only influenced the prevalence of infection after 10 years to a very limited degree, but such an expansion was very expensive. Other parameters used for the sensitivity analysis had a limited influence on the results.
We estimated that the prevalence of C trachomatis infection would decline from 4.2% to 1.0% after 10 years if home sampling for screening and partner notification were implemented in a Danish setting. Inclusion of both indirect and direct costs yielded an ICER (costs/MOA) of $3186 in the first year of screening, but the accumulated indirect costs offset the accumulated direct costs in year 4, after which the program was cost-saving to society. When direct costs of the program were considered alone, it remained a cost to society.
This study enjoys the advantage that the alternative strategy was evaluated on the basis of data on participation rates and infection prevalence drawn from major randomized clinical trials. Furthermore, the use of a simulation model to predict the effect of the alternative strategy is a major advantage since C trachomatis is a communicable disease and detection and treatment of an infected individual therefore affects the C trachomatis incidence and prevalence. Results from our study might not be fully or directly applicable to societies with other health care systems or registration of inhabitants, and important parameters regarding, for example, participation rates in screening and partner notification programs might also be very different in other areas. I these cases, however, we believe that this study represents an applicable future method to evaluate the same or other C trachomatis prevention programs in other communities.
The main limitations of this study spring from the limitations of the available data used for modeling and the general limitations pertaining to mathematical models.36 However, the model applied in the hypothetical population produced an infection prevalence corresponding to the a priori prevalence estimation in our region and this supports the usefulness of the model on our society. In an earlier Danish population-based screening study,2 the infection prevalence of participating 21- to 23-year-old men and women was 5.7% to 8.0%. In the same study, we found that a strategy aimed at including more women did, indeed, raise the participation rate, but the infection prevalence in the submitted samples declined. Furthermore, analysis of the reasons for not participating indicated that the infection prevalence could be lower among nonparticipants than among participants. Thus, an infection prevalence of 4% in the whole population of 15- to 29-year-old men and women can reasonably be expected. However, limitations definitely apply to some of the data used in the model. Data on sexual behavior used in our study are based on the best available surveys from The Netherlands28 and Denmark.29 In order to obtain detailed data on sexual behavior, however, major nationwide studies in all age groups should be conducted, as has been the case in the United Kingdom,37–39 but due to the intimate subject, one should still be aware of the possible information bias related to the evidence gathered.
Data on the nature of the infection were extracted from the literature, but considerable limitations apply to the evidence on the nature of untreated asymptomatic C trachomatis infections. We estimated the risk of developing PID among untreated asymptomatic women on the basis of randomized studies showing that screening for C trachomatis reduces the incidence of symptomatic PID 15%3 and 30%4 within 1 year, and we chose to set the risk in the simulation model to 20%. However, a recent study followed untreated C trachomatis infections among 30 women for 1 year, and none of these women developed symptomatic PID during follow-up.40 A similar situation may apply to the risk of infertility after symptomatic PID, which has earlier been estimated in 2 Swedish studies41,42; but one recent study could not confirm an increased risk of infertility among women with C trachomatis endometritis.43 Thus, the evidence also falls short of adequacy as far as the nature of the infection is concerned, and more studies connecting C trachomatis exposure and PID and long-term C trachomatis complications are needed.
Despite the limitations pertaining to the use of a simulation model and the limited quality of the modeling data, we find that the advantages of using a dynamic model outweigh the limitations and that the dynamic model is the best available method for evaluating decision alternatives within this field. Static models are often used,18,44,45 but they have an even greater limitation because they cannot appropriately cater to the dynamic nature of C trachomatis infections, i.e., they ignore that C trachomatis is an infection.
Improved partner notification was instrumental in triggering a decline in prevalence and had a significant effect on the cost-effectiveness of the program. The percentages used in our model are based on the assumption that the treating GPs and physicians accept to provide infected patients with the partner packages. Implementing changes in general practice may be difficult, but the use of thorough implementation strategies has proved useful.46 Difficulties in implementing partner notification may cause the test ratio of current partners of infected individuals to be overestimated. This possible overestimation may be outweighed by the fact that the model only included tests of current partners, even though we intended to include partners during the past 12 months in the national program. A Swedish study47 showed that earlier partners had a higher risk of infection than current partners. Inclusion of earlier partners would hence heavily affect the outcome of the program, but they were not included in the modeling.
From a society point of view, our study indicates that the alternative program should run for a number of years after its implementation because the first years will be most expensive and the gains will appear later. However, the study also indicates that the program should probably not run for a total of 10 years without modifications because the infection prevalence is then predicted to have declined to only 1%. At this infection prevalence, the positive predictive value will be unacceptably low.48 Studies have shown that a positive test result causes major concerns49 and the false-positive diagnoses should therefore be reduced to a minimum. Continuous evaluation of the program is hence very important, and the more the infection prevalence declines, the more relevant it becomes to seek to identify subgroups of individuals having a higher risk of infection, e.g., by use of selective self-reportable screening criteria.50
This study suggests that implementing home sampling for population-based screening and partner notification improves the control of C trachomatis infections and that the infection prevalence will fall significantly. Furthermore, the implementation of such a program may be cost-saving to society after 4 years. Home sampling for screening and partner notification should be therefore be considered as an alternative to the current strategy.
1. Gerbase AL, Rowley TR, Mertens EM. Global epidemiology of sexually transmitted diseases. Lancet 1998; 351(suppl):2–4.
2. Andersen B, Olesen F, Møller JK, et al. Population-based strategies for outreach screening of urogenital Chlamydia trachomatis
infections: a randomised controlled trial. J Infect Dis 2002; 185:252–258.
3. Scholes D, Stergachis A, Heidrich FE, et al. Prevention of pelvic inflammatory disease by screening for cervical chlamydial infection. N Engl J Med 1996; 334:1362–1366.
4. Ostergaard L, Andersen B, Moller JK, et al. Home sampling versus conventional swab sampling for screening of Chlamydia trachomatis
in women: a cluster-randomized 1-year follow-up study. Clin Infect Dis 2000; 31:951–957.
5. Nilsson U, Hellberg D, Shoubnikova M, et al. Sexual behavior risk factors associated with bacterial vaginosis and Chlamydia trachomatis
infection. Sex Transm Dis 1997; 24:241–246.
6. Westrom L, Bengtsson LP, Mardh PA. Incidence, trends, and risks of ectopic pregnancy in a population of women. BMJ Clin Res Ed 1981; 282:15–18.
7. Westrom L. Effect of acute pelvic inflammatory disease on fertility. Am J Obstet Gynecol 1975; 121:707–713.
8. Schachter J, Grossman M, Sweet RL, et al. Prospective study of perinatal transmission of Chlamydia trachomatis.
JAMA 1986; 255:3374–3377.
9. Eley A, Oxley KM, Spencer RC, et al. Detection of Chlamydia trachomatis
by the polymerase chain reaction in young patients with acute epididymitis. Eur J Clin Microbiol Infect Dis 1992; 11:620–623.
10. Hillis SD, Coles FB, Litchfield B, et al. Doxycycline and azithromycin for prevention of chlamydial persistence or recurrence one month after treatment in women: a use-effectiveness study in public health settings. Sex Transm Dis 1998; 25:5–11.
11. Santer M, Warner P, Wyke S, et al. Opportunistic screening for chlamydia infection in general practice: can we reach young women? J Med Screen 2000; 7:175–176.
12. Mathews C, Coetzee N, Zwarenstein M, et al. Strategies for Partner Notification for Sexually Transmitted Diseases (Cochrane Review). In: The Cochrane Library, Issue 1. Oxford: Update Software, 2002.
13. Jaschek G, Gaydos CA, Welsh LE, et al. Direct detection of Chlamydia trachomatis
in urine specimens from symptomatic and asymptomatic men by using a rapid polymerase chain reaction assay. J Clin Microbiol 1993; 31:1209–1212.
14. Lee HH, 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.
15. Schachter J, McCormack WM, Chernesky MA, et al. Vaginal swabs are appropriate specimens for diagnosis of genital tract infection with Chlamydia trachomatis.
J Clin Microbiol 2003; 41:3784–3789.
16. Østergaard L, Møller JK, Andersen B, et al. Diagnosis of urogenital Chlamydia trachomatis
infection in women based on mailed samples obtained at home: multipractice comparative study. BMJ 1996; 313:1186–1189.
17. Østergaard L, Andersen B, Møller JK, et al. Managing partners of people diagnosed with Chlamydia trachomatis
: a comparison of two partner testing methods. Sex Transm Dis 2003; 79:358–362.
18. Honey E, Augood C, Templeton A, et al. Cost-effectiveness of screening for Chlamydia trachomatis
: a review of published studies. Sex Transm Infect 2002; 78:406–412.
19. Roberts T, Robinson S, Barton P, et al. The correct approach to modelling and evaluating chlamydia screening [letter]. Sex Transm Infect 2004; 80:324.
20. Kretzschmar M, Welte R, van den HA, et al. Comparative model-based analysis of screening programs for Chlamydia trachomatis
infections. Am J Epidemiol 2001; 153:90–101.
21. Kretzschmar M, van Duynhoven YT, Severijnen AJ. Modeling prevention strategies for gonorrhea and chlamydia using stochastic network simulations. Am J Epidemiol 1996; 144:306–317.
22. Welte R, Kretzschmar M, Leidl R, et al. Cost-effectiveness of screening programs for Chlamydia trachomatis
: a population-based dynamic approach. Sex Transm Dis 2000; 27:518–529.
23. Welte R, Postma MJ, Leidl R, et al. Costs and effects of chlamydial screening: dynamic versus static modelling. Sex Transm Dis 2005; 32:474–483.
24. Sundhedsstyrelsen Vejledning for Diagnose og Behandling af Seksuelt Overførbare Sygdomme [in Danish]. 1999. Copenhagen, Denmark: Sundhedsstyrelsen.
25. Serum Institut. Klamydia 2000. EPI-NYT
, week 38 [in Danish].
26. Andersen B, Østergaard L, Nygård B, et al. Urogenital Chlamydia trachomatis
infections in general practice: diagnosis, treatment, follow-up and contact tracing. Fam Pract 1998; 15:223–228.
27. Møller JK, Andersen B, Olesen F, et al. Reasons for Chlamydia trachomatis
testing and the associated age-specific prevalences. Scand J Lab Clin Invest 2003; 63:339–346.
28. van Zessen G, Sandfort TGM. Seksualiteit in Nederland: Seksuaal gedrag, risico en preventie van AIDS [in Dutch]. Amsterdam, The Nederlands: Swets & Zeitlinger, 1991.
29. Ung 99: En Seksuel Profil (bind 1–3) [in Danish]. 1999.
30. Danish National Board of Health. Takstsystemet 2002: Vejledning 2001 [in Danish]. 2001.
31. Sygesikringens Forhandlingsudvalg og Praktiserende Lægers Organisation Overenskomst for Almen Praksis [in Danish]. 1999.
32. Statbank Denmark. Statistics Denmark. Available at: http://www.dst.dk
. Accessed 2003.
33. Løn-og indkomststatistik. 2001:10 [in Danish]. Available at: http://www.dst.dk
. Accessed 2003.
34. Statistics Denmark. Statistiske efterretninger, arbejdsmarked. 2001:23 [in Danish]. 2003.
35. Danmarks Statistik Befolkningens bevægelser 2000 [in Danish]. 2001.
36. Garnett GP. An introduction to mathematical models in sexually transmitted disease epidemiology. Sex Transm Infect 2002; 78:7–12.
37. Wellings K, Nanchahal K, Macdowall W, et al. Sexual behaviour in Britain: early heterosexual experience. Lancet 2001; 358:1843–1850.
38. Johnson AM, Mercer CH, Erens B, et al. Sexual behaviour in Britain: partnerships, practices, and HIV risk behaviours. Lancet 2001; 358:1835–1842.
39. Fenton KA, Korovessis C, Johnson AM, et al. Sexual behaviour in Britain: reported sexually transmitted infections and prevalent genital Chlamydia trachomatis
infection. Lancet 2001; 358:1851–1854.
40. Morre SA, van den Brule AJ, Rozendaal L, et al. The natural course of asymptomatic Chlamydia trachomatis
infections: 45% clearance and no development of clinical PID after one-year follow-up. Int J STD AIDS 2002; 13(suppl 2):12–18.
41. Westrom L, Joesoef R, Reynolds G, et al. Pelvic inflammatory disease and fertility: a cohort study of 1,844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results. Sex Transm Dis 1992; 19:185–192.
42. Westrom L. Influence of sexually transmitted diseases on sterility and ectopic pregnancy. Acta Eur Fertil 1985; 16:21–24.
43. Haggerty CL, Ness RB, Amortegui A, et al. Endometritis does not predict reproductive morbidity after pelvic inflammatory disease. Am J Obstet Gynecol 2003; 188:141–148.
44. Genc M, Mardh A. A cost-effectiveness analysis of screening and treatment for Chlamydia trachomatis
infection in asymptomatic women. Ann Intern Med 1996; 124:1–7.
45. van Valkengoed IG, Postma MJ, Morre SA, et al. Cost-effectiveness analysis of a population based screening programme for asymptomatic Chlamydia trachomatis
infections in women by means of home obtained urine specimens. Sex Transm Infect 2001; 77:276–282.
46. Shafer MA, Tebb KP, Pantell RH, et al. Effect of a clinical practice improvement intervention on chlamydial screening among adolescent girls. JAMA 2002; 288:2846–2852.
47. Ramstedt K, Forssman L, Johannisson G. Contact tracing in the control of genital Chlamydia trachomatis
infection. Int J STD AIDS 1991; 2:116–118.
48. Østergaard L. Detection of Chlamydia trachomatis
in a low prevalence population [letter]. Eur J Clin Microbiol Infect Dis 1995; 14:471–472.
49. Duncan B, Hart G, Scoular A, et al. Qualitative analysis of psychosocial impact of diagnosis of Chlamydia trachomatis
: implications for screening. BMJ 2001; 322:195–199.
50. Andersen B, van Valkengoed IG, Olesen F, et al. Value of self-reportable screening criteria to identify asymptomatic individuals in the general population for urogential Chlamydia trachomatis
infection screening. Clin Infect Dis 2003; 36:837–844.