Chlamydia is a public health problem in the United States, with 481 reported diagnoses per 100,000 population in 2020.1 2
Chlamydial infections are often asymptomatic and can result in pelvic inflammatory disease (PID), with subsequent development of chronic pelvic pain (CPP), ectopic pregnancy (EP) and tubal factor infertility (TFI) in women.3 4
The recommendations are implemented as opportunistic screening: offering screening when eligible individuals present to health care settings for other reasons, or testing asymptomatic individuals who seek sexual health screens. Chlamydia screening is estimated to reduce risk of PID at 12 months post chlamydia test (pooled relative risk from clinical trials, 0.68; 95% confidence interval, 0.49–0.94; moderate evidence base).5 6–29
We developed and calibrated a chlamydia transmission model to multiple epidemiological datasets to estimate the impact of screening and PN on chlamydial infection outcomes during 2000 to 2015.2 30,31s
The purpose of the current study was to estimate the cost-effectiveness of chlamydia screening. Specifically, in this study, we conducted a cost-effectiveness analysis of chlamydia screening that included adverse health outcomes after chlamydial infection, and associated health care costs to evaluate the cost-effectiveness of the scale-up of screening and PN as implemented over 2000 to 2015. We also estimate the incremental benefits and cost-effectiveness of a hypothetical scenario in which all sexually active women aged 15 to 24 years were screened annually, compared with maintaining screening at 2015 levels, over the 2016 to 2019 period.
METHODS
Transmission Dynamic Model
Our pair formation model simulates chlamydia transmission in a heterosexual population in the United States.2
Model Calibration
The model was calibrated to multiple epidemiological data for years 2000 to 2015 to estimate posterior probability distributions for parameters governing the epidemiology of chlamydia (natural history, behavior, testing, and treatment). The model was calibrated to sex- and age-stratified chlamydia prevalence, chlamydia case reports and proportion of young people reporting having had sex.2 https://links.lww.com/OLQ/A917
Transmission Dynamic Model of Intervention Scenarios
The model population may get tested after attending health care with symptoms, or through asymptomatic screening, or after PN. Rates of testing of people with symptomatic infection were assumed to vary by sex and assumed to be constant over time and fixed across scenarios. Screening rates were allowed to vary over time and were parameterized by age and sex. The modeled screening frequency was highest in women younger than 25 years, although we allowed for lower screening frequency among women aged 25 to 39 years. Men aged 15 to 39 years were screened at a lower rate relative to the rate for women. We assumed that women and men older than 40 years were not screened but could receive testing and treatment due to symptoms or as part of PN. A scenario for estimated screening uptake over 2000 to 2015 (“Current policy”) was obtained from the calibrated transmission model. The calibrated model estimated annual screening rates (with 95% uncertainty interval [95% UI]) in 2000 to be 0–0.01 among women aged 15 to 18 years, 0–0.3 among women aged 19 to 24 years, and 0–0.01 among women aged 25 to 39 years. The respective 95% UI in 2015 were 0.3–0.7, 0.6–0.8, and 0.1–0.2 per year. The screening rates in men, relative to the women of the same age group, were 0.1–0.2 times as high for men younger than 25 years, and 0.01–0.03 times as high for men older than 25 years.
We modeled PN within the long-term partnerships. Probability of successful PN after diagnosis of the index case was stratified by sex and age of the index case, and parameters were constant over time. If a woman in a long-term partnership were the index case, the proportions of index case partners who receive PN and treatment if infected were estimated as 0.3–0.7, 0.6–0.9, 0.4–0.6, and 0.4–0.6 from the youngest to oldest age groups, respectively.2 2
Sequelae After Chlamydial Infection
We have previously conducted literature reviews to estimate probabilities, costs, utilities and durations of sequelae.32s,33s Ranges adopted for sequelae probabilities, costs and utilities are presented in Table 1 . We assumed women in chlamydia-infected health states experienced a constant rate of development of PID, and that women with repeat infections experienced a higher rate of PID based on observational evidence.34s,35s The sequelae development after PID was modeled as a probability tree, with risk that PID would develop into CPP, EP, and TFI. In men, we assumed that epididymitis was the only sequelae of untreated asymptomatic infection.
TABLE 1 -
Sequelae Probabilities and Costs
Variables
Dependency
Distribution
Mean (95% Range)
Source
Probability of symptomatic infection
Symptomatic infection in women
Infection in women
From the calibrated model
0.289 (0.281–0.297)
2
Symptomatic infection in men
Infection in men
From the calibrated model
0.262 (0.251–0.273)
2
Duration of sequelae (y)
Total duration of symptomatic infection in women
Among women with symptomatic infection
From the calibrated model
0.104 (0.100–0.108)
2
Total duration of symptomatic infection in men
Among men with symptomatic infection
From the calibrated model
0.112 (0.108–0.116)
2
Proportion of time with symptoms in women*
Duration of symptoms/total duration of infection among those who develop symptoms
Log-normal (2.63,0.021)/(Log-normal (3.19,0.46) + Log-normal (2.63,0.021))
0.37 (0.19–0.59)
54s
Proportion of time with symptoms in men*
Duration of symptoms/total duration infection among those who develop symptoms
Log-normal (1.87,0.036)/ (Log-normal (2.60,0.46) + Log-normal (1.87,0.036))
0.33 (0.16–0.55)
54s
Occurrence of sequelae
PID
Constant risk during infection, rate of PID per year of infection
Normal (0.154, 0.049)
0.154 (0.053–0.25)
33s
PID among women with a repeat infection
Allow for increased risk for PID among women with repeated CT
Uniform (1, 1.15)
RR 1.08 (1–1.15)
34s,35s
TFI
Probability given PID
Beta (27.75, 137.45)
0.17 (0.12–0.23)
55s–59s
Chronic pelvic pain
Probability given PID
Beta (155.11, 439.19)
0.26 (0.23–0.30)
55s,60s
Ectopic pregnancy
Probability given PID
Beta (31.58, 419.51)
0.07 (0.049–0.098)
55s,56s,58s,61s
Epididymitis
Probability given untreated infection in men
Beta (0.831, 18.491)
0.043 (0.00069–0.16)
62s
Duration of sequelae (years)
PID
Duration given PID
Beta (16.5, 571.1)
0.028 (0.018–0.041)
36s
EP
Duration given EP
Beta (15.5, 182.7)
0.078 (0.045–0.119)
36s
CPP
Duration given CPP
Uniform (5, 10)
7.5 (5.1–9.9)
36s
TFI
Duration given TFI
Uniform (5, 10)
7.5 (5.1–9.9)
36s
Epididymitis
Duration given epididymitis
Beta (16.8, 866.9)
0.019 (0.013–0.028)
Utilities
Baseline health utility
Age
NA
NA
63s
Symptomatic infection (women)
Utility given symptoms
Beta (4.791, 0.148)
0.97 (0.749–1)
36s,37s
Urethritis (symptomatic infection in men)
Utility given symptoms
Beta (22.08, 0.92)
0.961 (0.839–0.991)
36s,37s
PID
Utility given PID
Beta (42.2, 13.6)
0.756 (0.637–0.859)
CPP
Utility given CPP
Beta (22.1, 7.00)
0.759 (0.592–0.894)
36s,37s
EP
Utility given EP
Beta (23.5, 8.65)
0.731 (0.570–0.867)
36s,37s
TFI
Utility given TFI
Beta (47.7, 5.01)
0.905 (0.813–0.968)
36s,37s
Epididymitis
Utility given epididymitis
Beta (13.6, 6.84)
0.665 (0.453–0.847)
36s
Inpatient treatment (used for costs)
PID (inpatient)
Probability given PID
Uniform (0.098, 0.110)
0.104 (0.098–0.110)
64s
EP (inpatient)
Probability given EP
Uniform (0.129, 0.180)
0.155 (0.130–0.180)
64s
Epididymitis (inpatient)
Probability given EP
Uniform (0.0026, 0.0094)
0.006 (0.0028–0.0092)
65s
Costs
NAAT
Per test
Uniform (31.40, 95.61)
63.5 (33.00–94.00)
36s,53s,66s,67s
Non-NAAT†
RR to NAAT
Uniform (0.29, 0.61)
0.45 (0.30–0.60)
42s
Proportion of tests which are NAAT
Time-varying
Fixed
See Supp Figure S1
Azithromycin 1 g
Given treatment
Uniform (15.30, 43.71)
29.50 (16.00–43.00)
66s,68s–70s
Short clinic visit
Given diagnosis
Uniform (18.95, 61.05)
40.00 (20.00–60.00)
66s,67s,69s
Partner services costs‡
Per identified and infected partner
Uniform (738.92, 3022.08)
1880.50 (796.00–2965.00)
43s,44s
PID (inpatient treatment)
Given inpatient PID
Uniform (5558.40, 16,942.61)
11,205.50 (5843.00–16,658.00)
36s,53s,64s,66s–68s,70s–75s
PID (outpatient treatment)
Given outpatient PID
Uniform (304.42, 927.58)
616.00 (320.00–912.00)
36s,53s,64s,66s–68s,70s–75s
CPP
Given CPP
Uniform (625.05, 1902.95)
1,264,00 (657.00–1871.00)
64s,66s,68s
EP (inpatient treatment)
Given inpatient EP
Uniform (6275.68, 19,128.32)
12,702 (6597.00–18,807.00)
64s,66s,68s
EP (outpatient treatment)
Given outpatient EP
Uniform (1694.24, 5164.76)
3429.50 (1781.00–5078.00)
64s,66s,68s
TFI
Given TFI
Uniform (3256.29, 9924.71)
6590.50 (3423.00–9758.00)
64s,66s,68s
Epididymitis (inpatient treatment)
Given inpatient Epididymitis
Uniform (3872.74, 11,803.26)
7838.00 (4071.00–11,605.00)
65s,66s,68s,70s,72s,73s,75s
Epididymitis (outpatient treatment)
Given outpatient Epididymitis
Uniform (221.63, 676.37)
449.00 (233.00–665.00)
65s,66s,68s,70s,72s,73s,75s
*We calculate the proportion of infection spent with symptoms and apply the proportion to the transmission model estimate of duration in the symptomatic compartment. Kreisel et al (2021)54s estimated 14.83 days (5.55–32.96) until symptom development, and additional 6.47 (6.04–6.94) days until testing and treatment for men, and 26.78 (9.95–59.81) days and 13.90 (13.34–14.49) days in women, respectively. These provides the framework to estimate total duration of infection given symptomatic infection (time to symptom development + time to testing and treatment), and proportion of this spent with symptoms.
† We defined the costs so that non-NAAT are less expensive than NAAT. This was based on CMS Clinical Laboratory Fee Schedule where relative costs between non-NAATs (chlamydia antibody, and chlamydia DNA direct probe) have remained constant compared with NAATs between 2011 and 2019.
‡ In the chlamydia transmission model, successful PN and treatment occurs in pairs where both partners are infected, and the index case is identified by chlamydia testing, and we matched our costs to this. In Silverman et al (2019)43s costs per partner treated after interview and costs per partner treated after DIS interviewed ranged between $702 and 2869, which was used as the basis to define a broad range of costs associated with partner services.
Utilities Associated With Health States
For both sexes symptomatic infections were associated with lower quality of life (reduced utility) in the acute phase, while asymptomatic infection did not cause reduced utility unless the infection progressed to sequelae. We divided symptomatic infection into pre-symptomatic and symptomatic periods, with disutility assigned only to the latter. We assumed all sequelae incurred health utility losses.
CPP and TFI had the longest sequelae duration. For CPP and TFI, we modeled a lag of 5 years from infection to the sequelae development. A study from the Institute of Medicine (IOM) estimated the duration of the conditions to be between 5 years to lifetime, based on expert opinion.36s We examined a shorter duration of sequelae (5–10 years) in the main analysis and included a longer duration in sensitivity analyses.
We combined information on utility estimates from the IOM Health Utilities Index (HUI)36s and Global Burden of Disease (GBD) 2019 Disability Weights37s (Supplemental Tables S2 and S3, https://links.lww.com/OLQ/A917 38s IOM HUI estimates have a higher disutility compared with GBD across all the health outcomes included in this study. In the utility estimates used in the main analysis, the mean values were defined based on the average of GBD and IOM estimates, and lower and upper ranges were based on IOM and GBD, respectively.38s We performed sensitivity analyses on the impact of the utility estimates on the cost-effectiveness results. Further details on utilities are in the Supplemental Material (https://links.lww.com/OLQ/A917 39s and the results reflect the discounted, expected lifetime QALYs lost due to chlamydial infection.
Costs
We conducted the analysis from a health sector perspective, considering direct medical costs without regard to payor. Costs were expressed in 2020 US dollars using the medical care component of the consumer price index.40s We included costs of testing and treatment, clinic visits, PN and treatment of sequelae. Sequelae costs were stratified by inpatient and outpatient costs. We modeled the increased use of NAATs for chlamydia diagnosis over the period of 2000 to 2015 (Supplemental Fig. S1, https://links.lww.com/OLQ/A917 41s and they were less costly than NAAT. We incorporated the price difference in the model by using Clinical Laboratory Schedule estimates to estimate the relative costs of non-NAAT versus NAAT.42s To estimate costs of PN, we relied on studies that provided estimates of costs per diagnosed partner of an index case, which corresponds closely to how PN was implemented in the model.43s,44s
Provider-initiated PN is less common for chlamydia than for gonorrhea and syphilis,45s and we include this in the model by allowing for both patient self-initiated PN and provider-initiated PN with costs incurred only from the latter.
Provider-initiated PN, when included in a given scenario, was assumed to account for 10% to 40% of all PN in that scenario.45s We assumed patient-initiated PN occurred in all scenarios, and only provider-initiated PN was removed in the “no PN” scenario examined. The model counts the number of partners successfully treated, and we estimated a cost range to reflect this. We assumed costs occur in the year of the incident infection except for CPP and TFI, which occur 5 years after the infection, implemented as one-off costs. Costs and QALYs lost were discounted using an annual discount rate of 3%46s to the beginning of the analysis period (2000 or 2016, depending on the scenario in question) for years after the first year of the analysis period.
Multivariate Uncertainty Reflected in the Mechanistic Models
We sampled 60,000 draws from the calibrated transmission posterior parameter estimates, and PID development, sequelae probabilities, duration, utilities and costs. These capture uncertainty in the natural history of infection, PID development, sexual behaviors, and chlamydia screening from 2000 to 2015.
Analysis
We performed 2 sets of analyses (Table 2 ). The first estimated the impact of chlamydia screening and PN in the United States as implemented during 2000 to 2015, compared with a counterfactual without screening and PN over the same period. This scenario reflects the impact of current policy including increasing screening coverage and changing testing from culture to NAAT over the 16-year period.
TABLE 2 -
Scenarios of Chlamydia Screening and PN in the United States Included in the Study
Scenario Name
Screening
PN
1. 2000–2015
Current policy
Screening coverage as estimated in the model*
PN uptake as estimated in the model†
No screening and no PN
No screening, but people with symptomatic infection can be tested and treated
No PN‡
2.1. 2016–2019
Guidelines
Annual screening of all sexually active women aged 15–24 y§
PN uptake estimated in the model†
Current policy
Screening coverage as estimated in the model in 2015*
PN uptake as estimated in the model†
2.2. 2016–2019 + 5 y (2020–2024)
Guidelines
For 2016–2019: 100% screening coverage (annual) of all sexually-active women aged 15–24 y§
PN uptake as estimated in the model†
Current policy
Screening coverage as of 2015 estimated in the model*
PN uptake as estimated in the model†
In scenario 1, the reference strategy is “no screening & no PN”. In scenarios 2.1 and 2.2, the reference strategy is “current policy”.
*Screening coverage for 2000–2015 was as estimated in the calibrated model.
2 For scenarios 2.1 and 2.2, screening uptake in year 2016 and beyond was assumed to be the same as the calibrated model estimate for 2015.
† PN uptake for 2000–2015 was as estimated in the calibrated model.
2 For scenarios 2.1 and 2.2, PN uptake in year 2016 and beyond was assumed to be the same as the calibrated model estimate for 2015.
‡ In the “no PN” scenario, physician-initiated PN is set to 0, but patient-initiated PN is retained.
§ In the annual screening coverage scenario, all sexually-active women aged 15–24 years are screened annually, and screening coverage of women aged 25 years and over is the same as in the “current policy” scenario.
The second analysis estimated the incremental impact of increasing screening from the levels of screening achieved by 2015 to full realization of the guidelines (annual screening of all sexually active women aged 15–24 years). We assumed scaled up interventions would be implemented over 2016 to 2019, and we modeled outcomes over an additional 5-year follow-up period to capture some of the sustained effects of the intervention, relating to reduced prevalence and averted secondary infections. The scenario reflects the additional benefits achievable, and the resources needed if we wanted to increase screening from the levels achieved by 2015.
We estimated the cumulative sequelae averted in the different scenarios. For the cost-effectiveness analysis, we estimated cumulative discounted costs and lifetime QALYs gained relating to infections acquired over the time horizon defined for each analysis. We computed the incremental cost-effectiveness ratio (ICER) using the ratio of the mean incremental cumulative costs and mean incremental QALYs gained. We constructed cost-effectiveness acceptability curves (CEACs) to show the percentage of model simulations in which the cost per QALY gained by a given screening strategy was below a given threshold (cost per QALY gained × QALYs gained − costs).
In addition, we conducted 3 sensitivity analyses. Two of the 3 analyses focused on the choice of utility estimates and duration of long-term sequelae. We estimated the ICER and CEAC when assuming higher disutility estimates for all outcomes and longer duration of CPP and TFI (based on IOM estimates), and again when assuming lower disutility estimates for all outcomes (based on GBD estimates) and the same duration of CPP and TFI as in the main analysis (Supplemental Table S2 and S3, https://links.lww.com/OLQ/A917
RESULTS
Current Policy Compared With No Screening and No PN
From 2000 through 2015, screening and PN, as implemented in the United States, were estimated to have averted a mean of 1.3 million (95% UI, 490,000–2.3 million) cases of PID, 430,000 (95% UI, 160,000–760,000) cases of CPP, 300,000 (95% UI, 104,000–570,000) cases of TFI, and 140,000 (95% UI, 47,000–260,000) cases of EP in women, and 230,000 (95% UI, 2900–880,000) cases of epididymitis in men, compared with a scenario in which there were no screening and no provider-initiated PN during this period (Fig. 1 A).
Figure 1: Cumulative number of sequelae outcomes averted by chlamydia screening. Shown as mean and 95% UI. (A) Sequelae averted by chlamydia screening and PN, 2000–2015*. (B) Sequelae averted by screening as indicated by the guidelines, 2016–2019**. (C) Sequelae averted by screening as indicated by the guidelines, 2016–2019 with the additional 5 years of follow up (2020–2024)**. Epid, epididymitis. Notes: *No screening and no physician-initiated PN compared with current policy (screening coverage and PN as estimated in the calibrated model
2 ). **Guidelines: all sexually active women aged 15 to 24 years are screened annually; for years 2020–2024 screening coverage as estimated in the model in 2015. Compared with current policy (screening uptake in year 2016 and beyond was assumed to be the same as the calibrated model estimate for 2015).
We estimated that chlamydia screening and PN (“current policy”) cost $9700 (Table 3 ) per QALY gained compared with no screening and no PN. We estimated the current policy to have cost an additional $5.2 billion (95% UI, $–270,000 million to 11 billion) over the 16-year period (2000–2015), and there were 530,000 QALYs gained (95% UI, 160,000–1.2 million) as a result. In our main analysis, the current policy cost less than $50,000 per QALY gained in over 90% of simulations (Fig. 2 ). Yearly average estimates across key outcomes are presented in Supplemental Table S7 (https://links.lww.com/OLQ/A917
TABLE 3 -
Incremental Costs and Benefits and Cost-Effectiveness Ratio for the Current Policy Compared With No Screening and No PN
Scenario
Cumulative Costs in '000s (Discounted) ($)
Cumulative QALYs Lost in '000s (Discounted)
Incremental Costs, in ‘000s ($)
Incremental QALYs Gained, in ‘000s
ICER ($/QALY Gained)
1. 2000–2015
No screening and no PN*
15,969,000
1908.90
NA
NA
NA
Current policy†
21,157,000
1374.10
5,188,000
534.80
9700.8
95% Uncertainty
NA
NA
−270,670; 11,369,000
157.35; 1176.8
Cost-saving; 49,478
2.1. 2016–2019
Current Policy†
6,919,100
349.96
NA
NA
NA
Guidelines‡
8,042,900
311.93
1,123,800
38.030
29,550
95% Uncertainty
NA
NA
416,440; 1,938,200
11.383; 82.157
7832.8; 117,210
2.2. 2016–2019 + 5 y
Current policy†
14,541,000
738.34
NA
NA
NA
Guidelines‡
15,494,000
681.35
953,000
56.990
16,722
95% Uncertainty
NA
NA
202,170; 1,785,400
16.844; 124.58
2511.4; 73,877
Costs are in $2000 US dollars. All results are shown to 5 significant digits.
*No screening and no physician-initiated PN for 2000–2015.
† Screening coverage and PN for 2000–2015 were as estimated in the calibrated model.
2 For scenarios 2.1 and 2.2, screening uptake in year 2016 and beyond was assumed to be the same as the calibrated model estimate for 2015.
‡ In screening by guidelines: all sexually-active women aged 15–24 years are screened annually, and screening coverage of women 25 years and older is the same as in the “current policy” scenario. For scenario 2.2: for years 2020–2024 screening coverage as estimated in the model in 2015.
Figure 2: CEAC for the 3 scenarios performed. Figure show the percentage of simulations in which the cost per QALY gained was less than the given threshold. The lines present the 3 main scenarios: CEAC for 2000–2015 current policy compared with no screening & no PN*; CEAC for screening at guidelines level compared with current policy (2016–2019)**; CEAC for screening at guidelines level compared with current policy with additional 5 years of follow-up (2016–2019 + 5 years)**. Notes: *No screening and no physician-initiated PN compared with current policy (screening coverage and PN as estimated in the calibrated model
2 ). **Guidelines: all sexually-active women aged 15 to 24 years are screened annually; for years 2020–2024 screening coverage as estimated in the model in 2015. Compared with current policy (screening uptake in year 2016 and beyond was assumed to be the same as the calibrated model estimate for 2015).
In the sensitivity analyses for the current policy, using the lower disutility values from the GBD study was estimated to result in an ICER of $12,000 per QALY gained (Supplemental Table S4, https://links.lww.com/OLQ/A917 https://links.lww.com/OLQ/A917 https://links.lww.com/OLQ/A917
Screening at the Level Indicated by the Guidelines Compared With the Current Policy
From 2016 through 2019, the full realization of chlamydia screening guidelines would have averted an estimated mean of 94,000 (95% UI, 36,000–160,000) cases of PID, 31,000 (95% UI, 12,000–53,000) cases of CPP, 21,000 (95% UI, 7500–39,000) cases of TFI, 9600 (95% UI, 3400–18,000) cases of EP in women, and 17,000 (95% UI, 220–65,000) cases of epididymitis in men, compared with a scenario in which chlamydia screening coverage was maintained at 2015 levels (Fig. 1 B). The number of sequelae averted was higher when we included the 5-year follow-up in the analysis: 1.25 times as high for women, and 1.15 times as high for men.
From 2016 through 2019, the full realization of chlamydia screening guidelines was estimated to cost $30,000 per QALY gained, compared with a scenario in which chlamydia screening coverage was maintained at 2015 levels (Table 3 ). Screening at guidelines level cost an additional $1.1 billion (95% UI $420 million to 1.9 billion) over the 4-year period and was associated with 38,000 (95% UI 11,000–82,000) QALYs gained. When we included a 5-year follow-up period, the ICER was lower at $17,000. There were at least 60% of the simulations where the costs per QALY gained were estimated to be at or below $50,000 when using 2016 to 2019 timeframe (Fig. 2 ). Yearly average estimates across key outcomes are presented in Supplemental Table S7 (https://links.lww.com/OLQ/A917
In the sensitivity analyses, when we assumed higher disutility estimates for all outcomes and longer duration of CPP and TF, the cost-effectiveness ratios per QALY gained were $8100 and $4600 without and with the 5-year follow-up period included, respectively. When assuming lower disutility estimates for all outcomes, the cost-effectiveness ratios were $37,000 and $21,000, respectively. When discounting to the year of infection the associated ICER values were $29,000, and $16,000, respectively (Supplemental Tables S4, S5, S6, https://links.lww.com/OLQ/A917
DISCUSSION
In this study, we examined the effectiveness of chlamydia screening and PN practices as implemented in the United States from 2000 through 2015. We found that screening and PN for chlamydia have improved the population-level health of Americans by averting a substantive number of adverse reproductive health outcomes. We estimated that chlamydia screening and PN has improved population health and provided good value for money, and further gains are possible if screening coverage can be increased.
The basis of the analysis was a calibrated transmission model of chlamydia in a heterosexual population, which was developed to evaluate the uncertainties present both in the natural history of infection and in changes in screening coverage and reporting.2
The pair formation modeling structure captures risk of reinfection from the main partner. Models, which omit risk of reinfection, can overestimate the impact of screening.47s The modeled PN can underestimate the full impact achievable by the intervention: the model included PN in the pairs, which reflect long-term partnerships. We did not model PN for short-term partnerships. For short-term partnerships, PN may have less benefit for the index case, if the sexual partnership has ended, but it may have a larger impact for the wider sexual network. For the long-term sequelae, CPP and TFI, we included one-off costs, which reflect costs of diagnosis and short-term treatment of the conditions.36s In focusing on the diagnostic and short-term costs, we excluded potential long-term costs, such as the cost of management of chronic pain related to CPP. We also excluded costs associated with lost productivity, lost wages, and other societal costs.48s,49s
There are also reasons why the health impact and cost-effectiveness of screening may be less favorable than we estimated. The parameterization of the natural history of chlamydia and associated reproductive health outcomes is highly uncertain and, if sequelae occurrence owing to chlamydial infection was lower than estimated, we would overestimate of the impact of screening. The modeled screening coverage is uniform within the age groups, which can overestimate the reach of the opportunistic screening program. The modeled guidelines scenario assumes that all sexually active women younger than 25 years uptake annual screening of chlamydia, a target which has proven to be difficult to reach in practice.50s,51s We may have underestimated the costs of increasing screening in the United States between 2000 and 2015 by only including the increased testing costs and clinic visits but not considering the infrastructure, and personnel training costs that would have been needed. For example, reaching populations at higher risk for chlamydia and achieving uniform screening coverage among young people may require additional efforts not considered. Whether chlamydial infection induces meaningful, lasting immunity is unclear.8
This study highlights the achievements in chlamydia prevention in the United States between 2000 and 2015, such as improving diagnostic techniques and increasing screening coverage.
Our analysis shows the value in continued investment in chlamydia prevention strategies and the potential benefits of increased screening coverage. Improving the testing and treatment cascade and investing in more frequent testing of individuals with a higher risk of infection may improve cost-effectiveness of chlamydia prevention further. Other approaches, such as point of care testing52s and expedited partner therapy,53s can aid in increasing the impact and efficiency of chlamydia control strategies.
REFERENCES
1. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance, 2020. Available at:
https://www.cdc.gov/std/statistics/2020/default.htm . Accessed October 8, 2021.
2. Rönn MM, Tuite AR, Menzies NA, et al. The impact of screening and partner notification on chlamydia prevalence and numbers of infections averted in the United States, 2000–2015: Evaluation of epidemiologic trends using a pair-formation transmission model. Am J Epidemiol 2019; 188:545–554.
3. Workowski KA, Bachmann LH, Chan PA, et al. Sexually transmitted infections treatment guidelines, 2021. MMWR Recomm Rep 2021; 70:1–187.
4. LeFevre ML; U.S. Preventive Services Task Force. Screening for chlamydia and gonorrhea: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2014; 161:902–910.
5. Low N, Redmond S, Uusküla A, et al. Screening for genital chlamydia infection. Cochrane Database Syst Rev 2016; 9:CD010866.
6. Jackson LJ, Auguste P, Low N, et al. Valuing the health states associated with chlamydia trachomatis infections and their sequelae: A systematic review of economic evaluations and primary studies. Value Health 2014; 17:116–130.
7. Roberts TE, Robinson S, Barton P, et al; Chlamydia Screening Studies (ClaSS) Group. Screening for
Chlamydia trachomatis : A systematic review of the economic evaluations and modelling. Sex Transm Infect 2006; 82:193–200; discussion 201.
8. Rönn MM, Wolf EE, Chesson H, et al. The use of mathematical models of chlamydia transmission to address public health policy questions: A systematic review. Sex Transm Dis 2017; 44:278–283.
9. 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.
10. Welte R, Postma M, Leidl R, et al. Costs and effects of chlamydial screening: Dynamic versus static modeling. Sex Transm Dis 2005; 32:474–483.
11. Andersen B, Gundgaard J, Kretzschmar M, et al. Prediction of costs, effectiveness, and disease control of a population-based program using home sampling for diagnosis of urogenital
Chlamydia trachomatis infections. Sex Transm Dis 2006; 33:407–415.
12. Townshend JRP, Turner HS. Analysing the effectiveness of chlamydia screening. J Oper Res Soc 2000; 51:812–824.
13. Adams EJ, Turner KM, Edmunds WJ. The cost effectiveness of opportunistic chlamydia screening in England. Sex Transm Infect 2007; 83:267–274; discussion 274-275.
14. Gillespie P, O'Neill C, Adams E, et al. The cost and cost-effectiveness of opportunistic screening for
Chlamydia trachomatis in Ireland. Sex Transm Infect 2012; 88:222–228.
15. de Vries R, van Bergen JE, de Jong-van den Berg LT, et al. Systematic screening for
Chlamydia trachomatis : Estimating cost-effectiveness using dynamic modeling and Dutch data. Value Health 2006; 9:1–11.
16. Low N, McCarthy A, Macleod J, et al. Epidemiological, social, diagnostic and economic evaluation of population screening for genital chlamydial infection. Health Technol Assess 2007; 11:iii–iv, ix–xii, 1–165.
17. Roberts TE, Robinson S, Barton PM, et al. Cost effectiveness of home based population screening for
Chlamydia trachomatis in the UK: Economic evaluation of chlamydia screening studies (ClaSS) project. BMJ 2007; 335:291.
18. Fisman DN, Spain CV, Salmon ME, et al. The Philadelphia high-school STD screening program: Key insights from dynamic transmission modeling. Sex Transm Dis 2008; 35(11 Suppl):S61–S65.
19. Gift TL, Gaydos CA, Kent CK, et al. The program cost and cost-effectiveness of screening men for chlamydia to prevent pelvic inflammatory disease in women. Sex Transm Dis 2008; 35(11 Suppl):S66–S75.
20. Tuite AR, Jayaraman GC, Allen VG, et al. Estimation of the burden of disease and costs of genital
Chlamydia trachomatis infection in Canada. Sex Transm Dis 2012; 39:260–267.
21. de Wit GA, Over EA, Schmid BV, et al. Chlamydia screening is not cost-effective at low participation rates: Evidence from a repeated register-based implementation study in the Netherlands. Sex Transm Infect 2015; 91:423–429.
22. Vriend HJ, Lugnér AK, Xiridou M, et al. Sexually transmitted infections screening at HIV treatment centers for MSM can be cost-effective. AIDS 2013; 27:2281–2290.
23. Looker KJ, Wallace LA, Turner KM. Impact and cost-effectiveness of chlamydia testing in Scotland: A mathematical modelling study. Theor Biol Med Model 2015; 12:2.
24. Tuite A, Fisman D. Agent-Based Modelling of Chlamydia Trachomatis Transmission in a Canadian Subpopulation. Winnipeg, Manitoba: National Collaboration Centre for Infectious Diseases, 2014.
25. Armbruster B, Brandeau ML. Contact tracing to control infectious disease: When enough is enough. Health Care Manag Sci 2007; 10:341–355.
26. Owusu-Edusei K Jr., Chesson HW, Gift TL, et al. Cost-effectiveness of chlamydia vaccination programs for young women. Emerg Infect Dis 2015; 21:960–968.
27. Clarke J, White KAJ, Turner K. Approximating optimal controls for networks when there are combinations of population-level and targeted measures available: Chlamydia infection as a case-study. Bull Math Biol 2013; 75:1747–1777.
28. Teng Y, Kong N, Tu W. Optimizing strategies for population-based chlamydia infection screening among young women: An age-structured system dynamics approach. BMC Public Health 2015; 15:639.
29. de Vries R, van Bergen JE, de Jong-van den Berg LT, et al. Cost-utility of repeated screening for
Chlamydia trachomatis . Value Health 2008; 11:272–274.
30. Owusu-Edusei K, Bohm MK, Kent CK. Diagnostic methodologies for chlamydia screening in females aged 15 to 25 years from private insurance claims data in the United States, 2001 to 2005. Sex Transm Dis 2009; 36:419–421.
For further references, please see “Supplemental References,” https://links.lww.com/OLQ/A918
