Based on our decision analysis model, the total annual health care cost associated with chlamydia is greater when a screening program is implemented in both high and low burden settings. However, there is a significant decrease in chlamydia associated morbidity, which may offset the increased cost of screening given an appropriate willingness-to-pay (WTP). These results are consistent with several prior cost-effective models that show increased cost of chlamydia screening, often offset by quality adjusted life (QALY) units and monetary conversion, with a notable decrease in associated morbidity [
3,
5,
16,
17]. Current US guidelines recommend screening for chlamydia in early pregnancy despite a distinct lack of evidence as to outcomes [
18]. Since these recommendations were set knowing there is likely no net cost savings associated with chlamydia screening, the averted morbidity reported is of primary importance.
Main findings
The findings from the current investigation may offer guidance as to the future recommendations for chlamydia screening in both high and low burden settings. Despite increased cost expenditures, the cost thresh hold necessary to implement a program is low per individual, at $19.34 considering an 8% prevalence rate. The suggestion that this is a low cost per individual is pronounced when compared values for other STI screening recommendations such as that for Hepatitis B, which is estimated to cost $75.45 per individual when screening pregnant women in the US [
19]. It is important to mention that in this study, screening was shown to significantly decrease major chronic sequelae of chlamydia such as infertility. Rates decreased by almost 400%, which translates to large absolute numbers, particularly in high burden populations. This can be further extrapolated to account for QALYs saved for neonates when chlamydia-associated infertility and subsequent inability to carry a pregnancy is averted. Consequently, concerns regarding the cost associated with chlamydia screening for pregnant women, particularly as they apply to high burden settings, may be abated by monetary values revealed in this model when compared to other accepted programs.
Previous studies in our laboratory examined incidence and treatment outcomes of chlamydial conjunctivitis in the prescreening era offer support for our model’s results. Hammerschlag, et al. [
12] reported in a vertical transmission data study that prenatal screening and treatment of pregnant women was the most effective way to prevent neonatal chlamydial infections, especially as neonatal ocular prophylaxis has not been demonstrated to be effective in prevention of neonatal chlamydia conjunctivitis [
12,
20]. These results may also be relevant to other endpoints found in our model, and offers support for the importance of a screening program by reducing neonatal chlamydia-associated morbidity, as evidenced in the present study [
12].
Ong, et al [
5], in a study from Australia, found that antenatal screening of women aged 16–25 is likely to be cost-effective with significantly reduced morbidity. The findings from our current decision analysis model are in agreement with the aforementioned studies. Furthermore, Ong, et al [
5] agree with the current analysis that although there is net cost expenditure at predicted levels of chlamydia prevalence, infection rates rise cost savings quickly outpace expenditures [
5]. These results hold true despite using prevalence and vertical transmission data gathered through two different methodologies. Ong et al [
5] reported prevalence rates that were gathered from Australian family planning clinic records, and vertical transmission data was gathered from previously published studies in other populations [
5]. The prevalence rates used in the current study were based on screening of over 4,000 pregnant women in one medical center in central Brooklyn [
12]. Furthermore, vertical transmission data was based on screening 4357 pregnant women for cervical chlamydial infection, of whom 341 (8%) had positive cultures [
12]; 230 of their infants were for followed for 3 months and evaluated for development of neonatal chlamydial conjunctivitis, pneumonia and nasopharyngeal infection with serial cultures. Additionally, the rate of chlamydia infection among women less than 18 years of age was 14% [
12]. Despite values for these important variables coming from two different methodologies (Ong et al. [
5] and direct screening in the current study), the results remained grossly similar notwithstanding expected differences in exact numbers and ranges.
In the current investigation the increased cost associated with screening as determined by our model does not consider the long-term effect of decreasing prevalence of chlamydia in a population. Through comparison of results using the pre-screening prevalence of chlamydia (8%) to those of the post-screening modern era (6.7%), it appears as though a modern screening program results in a lower decrease in chlamydia-associated morbidity with a large increase in net cost, as compared to a non-screening scenario [
21]. Although these results are accurate for areas with a similar or otherwise low burden prevalence rate that have yet to implement a screening program, or implemented one despite a low prevalence rate, this does not hold true for areas with a high prevalence prescreening. In a situation with a high prevalence pre-screening that has, as a consequence of screening, decreased its prevalence rate, there is significant cost savings and decline in morbidity associated with a decrease in prevalence rates. As such, the estimates revealed in this model likely underestimate true cost savings and morbidity decline associated with the long-term implementation of a chlamydia screening program [
21].
Currently, there are no systemic review studies that have investigated the effect of chlamydia screening specifically in pregnant women. However, the United States Preventive Services Task Force (USPSTF) recommends screening for chlamydia in early pregnancy. Current guidelines also recommend screening in early pregnancy for other sexually transmitted infections (STIs) including HIV, gonorrhea and Hepatitis B [
18]. Similarly, little evidence exists for clinical outcomes associated with screening for these STIs, particularly in pregnancy. However given the low harm, high potential benefit and relatively low cost associated with screening for these STIs, as evidence by this model and other similar models, early screening for STIs during pregnancy has been determined to be an acceptable use of healthcare resources [
18].
Prenatal screening and treatment is more effective than other control methods, specifically, neonatal ocular prophylaxis. Hammerschlag, et al [
12] demonstrated that neonatal ocular prophylaxis with erythromycin or tetracycline ophthalmic ointments was ineffective for prevention of chlamydial ophthalmia in infants, as well as having no effect on respiratory infection [
12]. The Canadian Pediatric Society recently recommended that neonatal ocular prophylaxis for
C. trachomatis infection be stopped, and emphasized the importance of prenatal screening and treatment [
22,
23]; Currently, the American Academy of Pediatrics (AAP) is considering a program similar to that in Canada. Prenatal screening may be an important public health intervention in low resource areas, especially when, affordable, sensitive NAAT based screening tests become available. It should also be mentioned that some countries that currently do not screen actually have high prevalence of
C. trachomatis infection (>8%) in some of their populations, including the Netherlands (8%) [
24] and Ireland (5.6% overall, 9.1% in women 16–18 years of age) [
25].
Several advantages to expanding or initiating chlamydia screening programs outside of averted morbidity exist. It is possible, that with increased screening a higher burden of chlamydia in the population will be discovered, revealing the program to be increasingly cost-effective. Furthermore, since programs can be rolled out into already present antenatal care schedules, there is a decreased cost associated with administration, as well as decreased indirect costs to the patient. Sexual histories can be unreliable when determining the clinical risk of chlamydia infection, an opt-out approach to screening can avoid this uncertainty altogether [
5]. Finally, identification of an infected mother offers the opportunity to provide treatment for sexual partner to help prevent re-infection and chlamydia-associated morbidity in the partner.
Strengths and limitations
Strengths of this study included our ability to base our cost estimates and disease parameters from published studies without assumptions and with few calculated estimates. Those estimates that were calculated utilized linear regression between data points or simple algebra. Most epidemiological and cost data came from U.S. sources, adding to the accuracy of the results. Additionally, using data available in the literature, we were able to examine numerous endpoints associated with the health of both the pregnant woman and fetus/newborn. We were able to focus our study on a high burden setting and examine the effect of a screening program at two point prevalences: at implementation, and many years later after a screening program has been able to affect infection rate. Finally, it should be mentioned that given the accuracy of the data based in strong epidemiology, and robustness of our results in the setting of sensitivity analyses, this study is generalizable to other settings with a variety of chlamydia prevalence. The exception is that healthcare costs may differ between settings, which should be considered when applying these results to another location.
In addition to its strengths, several limitations to the study should be mentioned. We utilized a static decision analysis model that only accounted for 1 year of data. This was done to avoid complications associated with following the outcomes of pregnant women screened for chlamydia after the pregnancy has ended, and risks to the newborn and mother are greatly modified. It does, however, preclude a dynamic study of infection, which would account for changing prevalence and infection rates as the modeled screening program progresses. Additionally, as noted in previous studies, one aspect of a screening program that leads to greater cost-effectiveness is the ability to treat partners and subsequent decreased risk of reinfection. Due to a paucity of data regarding rates of partner treatment as well as asymptomatic chlamydial treatment, we were unable to incorporate this variable into the model. The risk of acquiring chlamydia-related sequelae is applied to all cases of chlamydia at the time of infection and does not account for the duration of the infection. Finally, there exists uncertainty regarding the estimated rates of chlamydia-related sequelae, which may lead to an overestimation of the cost-savings associated with a chlamydia screening program.
Interpretation
Increasingly accurate cost-benefit models are necessary for properly evaluating the development, implementation and maintenance of screening programs. To more accurately develop screening programs associated with chlamydia more epidemiological data should be gathered regarding three major variables that, if proper data existed, would have improved the robustness of this model’s results. These are the percent of chlamydia infection that goes untreated, rates of asymptomatic chlamydia and rates of partner treatment coverage as a result of screening. The addition of the variables to the model would result in more accurate results, and the model likely would have predicted increased cost savings and averted morbidity associated with screening than was presented. Future studies should focus on these epidemiological parameters.
Prior literature has reported that in low resource countries with high chlamydia prevalence, such as Papua New Guinea and many African countries where prevalence of infection in pregnant women may exceed 20%, increased expenditures may be of particular concern [
26]. Implementation of a screening program may prove particularly beneficial within a few years. Single dose treatment with azithromycin is the current standard of care [
27]. The development of sensitive and affordable point of care nucleic acid amplification tests (NAAT) will make testing feasible in these regions [
28]. This will help further offset costs associated with screening, and may even lead to net cost savings. Studies from Zambia show that an integrated approach to antenatal screening that builds on existing programs, such as those targeting HIV prevention, can help manage both STI infections in pregnancy, and help to overall increase antenatal attendance. This will help further offset costs associated with screening, especially when building a screening program based on pre-existing programs, and may even lead to net cost saving [
8].