The Estimated Lifetime Quality-Adjusted Life-Years Lost Due to Chlamydia, Gonorrhea, and Trichomoniasis in the United States in 2018

Abstract

Background

Comprehensive evaluation of the quality-adjusted life-years (QALYs) lost attributable to chlamydia, gonorrhea, andtrichomoniasis in the United States is lacking.

Methods

We adapted a previous probability-tree model to estimate the average number of lifetime QALYs lost due to genital chlamydia, gonorrhea, and trichomoniasis, per incident infection and at the population level, by sex and age group. We conducted multivariate sensitivity analyses to address uncertainty around key parameter values.

Results

The estimated total discounted lifetime QALYs lost for men and women, respectively, due to infections acquired in 2018, were 1541 (95% uncertainty interval [UI], 186–6358) and 111 872 (95% UI, 29 777–267 404) for chlamydia, 989 (95% UI, 127–3720) and 12 112 (95% UI, 2 410–33 895) for gonorrhea, and 386 (95% UI, 30–1851) and 4576 (95% UI, 13–30 355) for trichomoniasis. Total QALYs lost were highest among women aged 15–24 years with chlamydia. QALYs lost estimates were highly sensitive to disutilities (health losses) of infections and sequelae, and to duration of infections and chronic sequelae for chlamydia and gonorrhea in women.

Conclusions

The 3 sexually transmitted infections cause substantial health losses in the United States, particularly gonorrhea and chlamydia among women. The estimates of lifetime QALYs lost per infection help to prioritize prevention policies and inform cost-effectiveness analyses of sexually transmitted infection interventions.

Chlamydia, gonorrhea, and trichomoniasis cause substantial lifetime quality-adjusted life-years (QALYs) lost in the United States in 2018, particularly among women. The estimates of lifetime QALYs lost per infection help to prioritize prevention policies and inform cost-effectiveness analyses of STI interventions.

In the United States, there were 1 758 668 diagnoses of Chlamydia trachomatis infections and 583 405 diagnoses of Neisseria gonorrhoeae infections reported in 2018, making chlamydia and gonorrhea the 2 most common notifiable infections in that year [1]. Trichomonas vaginalis (trichomoniasis) is a common sexually transmitted protozoal infection. Trichomoniasis is not a notifiable disease in the United States, and over 80% of cases are estimated to be asymptomatic [2]. In 2016, there were 222 000 initial physician office visits for trichomoniasis [1]. Chlamydia and gonorrhea are important causes of pelvic inflammatory disease (PID), chronic pelvic pain (CPP), tubal factor infertility (TFI), and ectopic pregnancy (EP) in women and epididymitis (EDS) in men [3–5]. The risk of reproductive health outcomes following trichomoniasis is less understood, but it is associated with increased risk of PID [6]. In addition, chlamydia, gonorrhea, and trichomoniasis are associated with an increased risk of acquisition and transmission of human immunodeficiency virus (HIV) in observational studies [7–9].

Estimates of the quality of life impacts of chlamydia, gonorrhea, and trichomoniasis are needed to quantify the long-term health burden of these 3 sexually transmitted infections (STIs). The estimates can be used to inform cost-effectiveness analyses of STI prevention interventions. For example, gonorrhea vaccines are under development, and estimating the health burden caused by the infection is necessary to quantify the potential benefits from a vaccine [10].

Use of quality-adjusted life years (QALYs) enables comparison across a wide range of conditions and outcomes. Although health state utility values and QALYs lost associated with complications such as PID and/or TFI attributable to chlamydia and gonorrhea infections have been estimated in several primary studies [11–15], comprehensive evaluations of QALYs lost associated with infections and downstream health consequences secondary to infections have not been documented, particularly for trichomoniasis, and for all 3 infections in men. We estimated the QALYs lost per incident infection with C. trachomatis, N. gonorrhoeae, and T. vaginalis, and quantify the population-level QALYs lost attributable to each of the 3 STIs in 2018 by sex and age group.

METHODS

Analytic Overview

We used probability-tree models to represent clinical outcomes of chlamydia, gonorrhea, and trichomoniasis among men and women. We estimated health losses associated with the health outcomes along each distinct path of the probability tree and aggregated these into the total numbers of expected discounted lifetime QALYs lost per incident infection. Together with incidence estimates of chlamydia, gonorrhea, and trichomoniasis in 2018 we estimated the overall population-level discounted lifetime QALYs lost by age and sex. Model inputs and parameter values were derived from a variety of sources (Table 1 and Supplementary Table 7). We conducted multivariate sensitivity analyses accounting for uncertainties in input parameters. All analyses were undertaken in R version 3.5.2.

Table 1.

Model Input Parameters Describing Gonorrhea, Chlamydia, Trichomoniasis, and Sequelae Probabilities, Utilities, and Durations, and Proportions Tested and Treated in Men and Women

Parameter	Mean Estimate	95% Uncertainty Intervala
Probabilities related to infections
Proportion of infections that are symptomatic
ȃChlamydia, men	0.162	0.079 to 0.270
ȃChlamydia, women	0.254	0.172 to 0.345
ȃGonorrhea, men	0.588	0.296 to 0.840
ȃGonorrhea, women	0.318	0.154 to 0.507
ȃTrichomoniasis, men	0.088	0.027 to 0.184
ȃTrichomoniasis, women	0.195	0.106 to 0.304
Probability of treatment given symptomatic infectionb
ȃChlamydia, men	0.935	0.919 to 0.943
ȃChlamydia, women	0.892	0.842 to 0.922
ȃGonorrhea, men	0.740	0.720 to 0.753
ȃGonorrhea, women	0.745	0.716 to 0.762
ȃTrichomoniasis, men	0.374	0.078 to 0.624
ȃTrichomoniasis, women	0.832	0.530 to 0.945
Probability of treatment given asymptomatic infectionb
ȃChlamydia, men	0.140	0.075 to 0.227
ȃChlamydia, women	0.241	0.188 to 0.296
ȃGonorrhea, men	0.021	0.010 to 0.038
ȃGonorrhea, women	0.070	0.038 to 0.118
ȃTrichomoniasis, men	0	Fixed
ȃTrichomoniasis, women	0	Fixed
Probability of PID given symptomatic treated infectionc
ȃChlamydia	0.016	0.006 to 0.035
ȃGonorrhea	0.011	0.004 to 0.021
ȃTrichomoniasis	0.001	0.000 to 0.007
Probability of PID given asymptomatic treated infectionc
ȃChlamydia	0.027	0.008 to 0.065
ȃGonorrhea	0.027	0.008 to 0.065
ȃTrichomoniasis	0.002	0.000 to 0.019
Probability of PID given untreated infectionc
ȃChlamydia	0.161	0.069 to 0.284
ȃGonorrhea	0.038	0.014 to 0.078
ȃTrichomoniasis	0.002	0.000 to 0.026
Probability of CPP given PID	0.261	0.229 to 0.295
Probability of TFI given PID	0.168	0.114 to 0.229
Probability of EP given PID	0.070	0.048 to 0.098
Probability of urethritis given symptomatic infection	1	Fixed
Probability of urethritis given asymptomatic infection	0	Fixed
Probability of EDSd given untreated chlamydia and gonorrhea	0.043	0.001 to 0.160
Probability of EDSd given untreated trichomoniasis	0.003	0.000 to 0.029
Probability of EDSd given treated infection	0	Fixed
State-specific durations, y
Symptomatic treated infection
ȃChlamydia, mene	0.058	0.032 to 0.109
ȃChlamydia, womene	0.111	0.064 to 0.204
ȃGonorrhea, mene	0.036	0.023 to 0.061
ȃGonorrhea, womene	0.071	0.049 to 0.113
ȃTrichomoniasis, men	0.384	0.120 to 1.734
ȃTrichomoniasis, women	0.159	0.055 to 0.621
Asymptomatic treated infection
ȃChlamydia, men	0.342	0.288 to 0.398
ȃChlamydia, women	0.186	0.075 to 0.373
ȃGonorrhea, men	0.342	0.288 to 0.398
ȃGonorrhea, women	0.186	0.075 to 0.373
ȃTrichomoniasis, men	0.384	0.120 to 1.734
ȃTrichomoniasis, women	0.159	0.055 to 0.621
Untreated infectionf
ȃChlamydia, men	1.018	0.648 to 1.540
ȃChlamydia, women	1.090	0.931 to 1.274
ȃGonorrhea, men	0.129	0.076 to 0.206
ȃGonorrhea, women	0.258	0.154 to 0.411
ȃTrichomoniasis, men	0.155	0.088 to 0.255
ȃTrichomoniasis, women	0.738	0.434 to 1.191
PID	0.028	0.018 to 0.041
CPP	g	5 to 10 y
TFI	g	5 to 10 y
EP	0.078	0.051 to 0.113
EDS	0.019	0.013 to 0.028
State-specific utilities
Symptomatic chlamydia, womenh	0.994	0.749 to 1.000
Symptomatic gonorrhea, womenh	0.994	0.849 to 1.000
Symptomatic trichomoniasis, womenh	0.994	0.849 to 1.000
Symptomatic urethral infection, menh	0.961	0.839 to 0.991
Asymptomatic infection	1	Fixed
PIDi	0.756	0.645 to 0.856
CPPi	0.759	0.587 to 0.899
TFIi	0.905	0.814 to 0.970
EPi	0.731	0.570 to 0.870
EDSi	0.665	0.454 to 0.854

Abbreviations: CPP, chronic pelvic pain; EDS, epididymitis; EP, ectopic pregnancy; PID, pelvic inflammatory disease; TFI, tubal factor infertility.

The distributions for the parameter uncertainty intervals, and the source studies are reported in Supplementary Table 7.

The method to estimate probability of symptomatic and asymptomatic infections that receive treatment is provided in Supplementary Section 2.

The probability of PID given infection for any of the 3 STIs was approximated by duration of the infection multiplied by the annual rate of PID given the infection with details in Supplementary Section 3. Durations of the 3 STIs are reported in Table 1 with details in Supplementary Section 2 and Supplementary Table 1.

We defined EDS to include epididymo-orchitis according to the STI Treatment Guidelines, 2021 [17].

For chlamydia and gonorrhea, the method to estimate durations of symptomatic treated infection is reported in Supplementary Section 2 and Supplementary Table 1.

Durations of untreated infections of chlamydia, gonorrhea, and trichomoniasis were estimated as the inverse of natural clearance rate, which are described in Supplementary Table 1.

Mean duration taken from uniform distribution between 5 years and 10 years (Supplementary Table 4a and 4b) according to previous analyses which assumed much shorter durations based on unspecified evidence [18]. We conducted 2 sensitivity analyses: (1) duration fixed at 5 years (Supplementary Table 4a) and (2) duration as lifetime (Supplementary Table 4c) using a lifetable approach described in the Li et al study [16].

Differences on distributions of symptomatic infections utilities between main analysis and sensitivity analyses are reported in Supplementary Section 4 in detail.

Differences on distributions of sequelae utilities between main analysis and sensitivity analyses are reported in Supplementary Section 4 in detail.

Model Structure and Probability of Key Clinical Outcomes

We used probability trees to model clinical outcomes of chlamydia, gonorrhea, and trichomoniasis, adapted from a model framework described in Li et al [16]. We assumed a similar set of secondary outcomes for all 3 STIs examined using the same probability-tree structures, but as relevant allowed model input values (ie, probabilities, durations, and utility) to vary across the 3 STIs. Separate trees were specified for men and women (Figure 1 and Figure 2).

Figure 1.

Probability tree for sequelae of gonorrhea, chlamydia, or trichomoniasis among women (Panel A), with complications of PID shown in Panel B. Abbreviations: CPP, chronic pelvic pain; EP, ectopic pregnancy; PID, pelvic inflammatory disease; TFI, tubal factor infertility.

Figure 2.

Probability tree for sequelae of gonorrhea, chlamydia, or trichomoniasis among men. Abbreviation: EDS, epididymitis.

For women, each STI was categorized as symptomatic or asymptomatic with different durations of infections, which determined the probability of developing PID and subsequent complications (Figure 1) [19]. Although treatment for infections is not explicitly distinguished as a separate branch in Figure 1, the estimated durations of infections accounted for the fraction of cases treated. Possible complications following PID were CPP, EP, and TFI (Figure 1B ). The probability that an infection would be symptomatic was derived from estimates from Kreisel et al and Lewis et al [20, 21]. Proportions of treated symptomatic or asymptomatic infections, and durations of symptomatic and asymptomatic infections (distinct for treated or untreated) were estimated based on the parameters in these studies [20, 21] (Supplementary Material Section 2 and Supplementary Table 1). Probabilities of PID and sequelae (CPP, EP, and TFI) were estimated based on synthesizing evidence (Supplementary Material Section 3).

For men, infections were categorized as symptomatic or asymptomatic urethral infections (Figure 2). We assumed that urethritis was present in all symptomatic infections for the duration from symptom onset to testing due to symptoms (Supplementary Table 1), and that urethritis was not present in asymptomatic infections [22]. There is sparse evidence on the relationship between duration of infection and sequelae development for men, and EDS was modeled using a probability that was independent of duration (Supplementary Material Section 3).

Lifetime QALYs Lost per Incident Infection

Utilities for clinical outcomes related to the 3 STIs and sequelae were estimated based on weights from the Global Burden of Disease Study (GBD) disability weights study [23] and a modeling study from the Institute of Medicine (IOM) using the Health Utility Index [15]. Disability weights represent the magnitude of health loss associated with specific health outcomes are measured on a scale from 0 to 1, where 0 equals a state of full health and 1 equals death. Compared to the IOM study, disability weights from the GBD implied smaller health losses associated with each outcome, and the GBD estimates can be seen as more conservative estimates of QALYs lost due to the STIs. In our model, utility values for an infection with any of the 3 STIs was specified as a distribution with mean value equal to estimated utility (1 minus the disability weight from the GBD). The lower bound of the 95% uncertainty interval (UI) was set as the infection-specific IOM estimate with the upper bound as 1. For sequelae of infection, distributions around the utility values were defined with mean values taken as the average of estimates reported by GBD and IOM, with lower and upper bounds defined by the IOM and GBD estimates, respectively. Details are provided in Table 1 and Supplementary Material Section 4.

Utility weights for infections and sequelae were multiplied by background utilities reflecting chronic comorbidities by age, based on nationally representative EQ-5D index scores (Supplementary Table 4c) [24]. CPP and TFI were treated as chronic sequelae occurring 5 years after infection, and durations were assumed to vary between 5 to 10 years and were incorporated into QALYs lost using a life table approach [16] (Supplementary Table 4a and 4b). Age-specific mortality rates for women were derived from National Center for Health Statistics data [25]. QALYs lost incurred in years after the incidence of infection were discounted to the year of infection (2018), using a discount rate of 3% per year [26]. Based on the probability tree models in Figure 1 and Figure 2, discounted lifetime QALYs lost per incident infection were estimated by summing the expected losses for each unique sequela or sequelae combination (the product of the probability, duration, and disutility).

QALYs Lost at Population Level and Outcome Measures

Incidence estimates for the United States in 2018 were obtained from Kreisel et al [20]. (chlamydia and gonorrhea) and from the Lewis et al (trichomoniasis) [21]. Consistent with these studies [20, 21], we stratified the population in our model by sex and age groups defined as 15–24 and 25–39 years for chlamydia and gonorrhea, and 15–24, 25–39, and 40–59 years for trichomoniasis. Estimates of incidence by disease, sex, and age group are reported in Supplementary Table 5. The population size in the United States in 2018 by sex and age group was estimated using the American Community Survey full resident population estimates [20, 21, 27] (Supplementary Table 6). We computed the QALYs lost at the population level as the product of population size, incidence rates per population, and discounted lifetime QALYs lost per incident infection for each STI. Results of QALYs lost are summarized in terms of aggregated population-level counts, as well as per 1000 person-years (calculated as the total numbers of QALYs lost in the United States due to infection acquired in 2018 divided by the population size and multiplied by 1000). From our modeled results, we extracted the composition of lifetime QALYs lost due to each sequela by sex and STI.

Sensitivity Analysis

We performed multivariate sensitivity analyses accounting for uncertainty in (1) incidence in each subpopulation (sex and age group), (2) probabilities that infections are symptomatic, (3) probabilities of infected persons receiving treatment by each subpopulation and symptom status, (4) durations of infections by subpopulation and symptom status, (5) sequelae probabilities, (6) durations of sequelae, and (7) utility values associated with each health outcome. Input parameters were drawn from the distributions estimated by studies in Supplementary Table 7.

We conducted additional sensitivity analyses on the discount rate and our quality-of-life assumptions. First, we present undiscounted estimates in addition to the base case estimates that used a 3% discount rate. Second, we conducted 2 sensitivity analyses around the choice of utility values and sequelae durations to represent extreme cases, including (1) using utility estimates from the GBD Disability Weights study with duration of chronic sequelae fixed at 5 years (Supplementary Table 4a) for a lower extreme on QALYs; and (2) using utility estimates from the IOM with lifetime duration of chronic sequelae (Supplementary Table 4c) for an upper extreme.

We sampled parameters 1000 times independently from the parameters’ distributions reported in Table 1. We ran the probability-tree models with the sampled parameter sets for the main and sensitivity analyses separately. We calculated mean and 95% uncertainty intervals values for each study outcome.

RESULTS

Discounted Lifetime QALYs Lost per Incident Infection

The estimated discounted lifetime QALYs lost per 1000 incident infections among men were 0.94 (95% UI, 0.12–3.85) due to chlamydia (for ages 15–39 years), 1.43 (95% UI, 0.19–5.48) due to gonorrhea (ages 15–39 years), and 0.12 (95% UI, 0.01–0.52) due to trichomoniasis (ages 15–59 years) (Figure 3 and Supplementary Table 8). The relatively small numbers of QALYs lost in men relate to relatively short durations, from several days to weeks, of sequelae for men. On a per-infection basis, the average number of QALYs lost for gonorrhea was 1.5 times that of chlamydia and 11.9 times that of trichomoniasis.

Figure 3.

Discounted lifetime QALYs lost per incident infection of chlamydia, gonorrhea, and trichomoniasis by sex and age group in the United States in 2018. Abbreviations: QALY, quality-adjusted life-year; UI, uncertainty interval.

The discounted lifetime QALYs lost per 1000 incident infections of chlamydia and gonorrhea were higher for women than men, with estimated QALYs for women of 46.98 (95% UI, 12.87–113.27) for chlamydia and 14.22 (95% UI, 3.26–39.14) for gonorrhea (ages 15–39 years) (Figure 3). For trichomoniasis, QALYs lost per 1000 incident infections in women (ages 15–59 years) was much smaller (1.35; 95% UI, 0.003–10.67; Supplementary Table 8). The average number of discounted lifetime QALYs lost per incident infection was 3.3-fold higher for chlamydia than for gonorrhea in women, and 34.8-fold higher for chlamydia than for trichomoniasis. QALYs lost per infection were higher for chlamydia than gonorrhea due to a higher probability of developing PID and chronic sequelae of CPP and TFI. Trichomoniasis had the lowest QALYs lost per incident infection given risk of PID during trichomoniasis is thought to be substantially lower than for chlamydia [9, 28–30].

Population-Level Discounted QALYs Lost

The population-level discounted lifetime QALYs lost for all the 3 STIs acquired in 2018 was 131 477 (95% UI, 34 325–324 487), with 128 561 (95% UI, 33 177–315 663) among women and 2916 (95% UI, 380–11 504) among men. The population-level discounted lifetime QALYs lost for infections acquired in 2018 were 1541 (95% UI, 186–6358) due to chlamydia among men aged 15–39 years, 989 (95% UI, 127–3720) due to gonorrhea among men aged 15–39 years, and 386 (95% UI, 30–1851) due to trichomoniasis among men aged 15–59 years (Supplementary Table 9). The population-level discounted lifetime QALYs lost by age groups analyzed were highest for chlamydia among men aged 15–24 and 25–39 years, followed by gonorrhea among men aged 25–39 years (Figure 4), primarily driven by patterns of incidence across infections and ages (highest for chlamydia in men aged 15–24 years) (Supplementary Table 5), and a larger number of QALYs lost per incident infection for gonorrhea compared to chlamydia or trichomoniasis for men (Figure 3 and Supplementary Table 8).

Figure 4.

Population-level number of discounted lifetime QALYs lost associated with infection of chlamydia, gonorrhea, and trichomoniasis by sex and age group in the United States in 2018. Abbreviations: QALY, quality-adjusted life-year; UI, uncertainty interval.

The population-level discounted lifetime QALYs lost for infections acquired in 2018 were 111 872 (95% UI, 29 777–267 404) due to chlamydia among women aged 15–39 years, 12 112 (95% UI, 2 410–33 895) due to gonorrhea among women aged 15–39 years, and 4576 (95% UI, 13–30 355) due to trichomoniasis among women aged 15–59 years (Figure 4 and Supplementary Table 9). The population-level discounted lifetime QALYs lost were highest for chlamydia among women aged 15–24 years and, followed by chlamydia among women aged 25–39 years (Figure 4). Both the incidence rate (Supplementary Table 5) and QALYs lost per infection were much higher among women aged 15–24 years with chlamydia than gonorrhea and trichomoniasis (Figure 3 and Supplementary Table 8).

Lifetime QALYs lost per 1000 person-years reflect variation in incidence of chlamydia, gonorrhea, and trichomoniasis by sex across age groups. For both men and women, the QALYs lost per 1000 person-years were highest among individuals aged 15–24 years with chlamydia among all the 3 infections across population subgroups (Figure 5), primarily due to the highest estimated incidence for chlamydia in ages 15–24 years (Supplementary Table 5).

Figure 5.

Population-level number of discounted lifetime QALYs lost per 1000 population associated with infection of chlamydia, gonorrhea, and trichomoniasis by sex and age group in the United States in 2018. Abbreviations: QALY, quality-adjusted life-year; UI, uncertainty interval.

Composition of Lifetime Discounted QALYs Lost

Among men the main contributor to the total QALYs lost was urethritis, which contributed 74% (95% UI, 25%–99%) of QALYs lost for chlamydia, 85% (95% UI, 44%–99%) for gonorrhea, and 92% (95% UI, 22%–100%) for trichomoniasis (Figure 6). Symptomatic urethral infections were more common for gonorrhea than for chlamydia (Table 1).

Figure 6.

Decomposition of total discounted lifetime QALYs lost by infections and sequelae of chlamydia, gonorrhea, and trichomoniasis by sex in the United States in 2018. The QALYs lost decomposed in this figure are the discounted lifetime QALYs lost per infection reported in Figure 3. The first bar for each of the 3 STIs for women represents the QALYs lost from the symptomatic infection itself. Bars 2 through 5 represent the QALYs lost due to each sequela (arising from both symptomatic and asymptomatic infection), and the sum of the proportions of QALYs lost from infection and sequelae for each infection is 100%. Means are indentified as historgrams and 95% UIs are indentified as error bars. Abbreviations: CPP, chronic pelvic pain; CT, chlamydia; EDS, epididymitis; EP, ectopic pregnancy; GC, gonorrhea; PID, pelvic inflammatory disease; TFI, tubal factor infertility; TV, trichomoniasis; UTS, urethritis.

QALYs lost were higher in women than in men for all 3 STIs and across all age groups. The duration of chronic sequelae in women was estimated to be substantially longer than the duration of sequelae in men. For chlamydia in women, the composition of total QALYs lost was estimated to be 75% (95% UI, 48%–91%) due to CPP and 20% (95% UI, 6%–40%) due to TFI (Figure 6). The composition of total QALYs lost among women with gonorrhea was similar to that for chlamydia (Figure 6). The composition of total QALYs lost among women with trichomoniasis was estimated to be 78% (95% UI, 1%–100%) from symptomatic trichomoniasis and 17% (95% UI, 0%–83%) from CPP, primarily due to the assumed lower risk of PID for trichomoniasis [9, 28–30]. We report QALYs lost per case of PID in Supplementary Table 13.

Results of Sensitivity Analyses

QALYs lost per incident infection were highly sensitive to the choice of utility values and durations. Comparing estimates based on IOM utilities versus estimates based on GBD utilities, the discounted lifetime QALYs lost per incident infection quadrupled for chlamydia, gonorrhea, and trichomoniasis for men (Supplementary Tables 10 and 11).

Utilities and durations for chronic sequelae (CPP and TFI) were important determinants of overall QALYs lost for women. The discounted lifetime QALYs lost due to CPP when using utilities from the IOM study and assuming lifetime duration was almost 30 times higher than the estimate using utilities from the GBD study and assuming duration of 5 years (Supplementary Table 4c). This 30-fold difference for CPP resulted in a similar difference for the overall discounted lifetime QALYs lost per infection for women based on the 2 sets of assumptions (Supplementary Tables 10 and 11), as CPP contributes the largest proportion of QALYs lost among infection and sequelae for women (Figure 6).

Discounting did not affect estimates of QALYs lost for men, given sequalae are short term and occur in the same year of infection (Supplementary Tables 8 and 12). For women, undiscounted lifetime QALYs lost per incident infection were nearly 30% higher than those when using a discount rate of 3% per year for all 3 infections (Supplementary Tables 8 and 12). Although the lifetime QALYs lost per incident infection and population-level total QALYs lost for women were sensitive to discounting, the composition of total QALYs lost for women changed minimally in relation to the choice over discounting.

DISCUSSION

We quantified the discounted lifetime QALYs lost in the United States due to chlamydia, gonorrhea, and trichomoniasis acquired in 2018. On both a per-infection basis and in terms of total aggregate QALYs lost nationwide, chlamydia imposed the greatest burden, followed by gonorrhea, then trichomoniasis. Published estimates of the QALYs lost in the United States due to syphilis [31] suggest that syphilis has a higher QALY lost per infection (0.09; 95% UI, 0.03–0.19) than the 3 STIs we examined, and that syphilis is the second only to chlamydia in terms of the aggregate QALYs lost nationwide (13 349; 95% UI, 5071–31 360).

We accounted for a wide spectrum of long-term sequelae caused by the infections. Our estimates of QALYs lost in women revealed the substantial burden of CPP, which has been described as a neglected reproductive health morbidity [32]. While there has been research on acute PID and TFI for women with chlamydia and gonorrhea [33–35], our results suggest that CPP causes the most health losses among women, contributing more than 70% (74.5%; 95% UI, 45.1%–90.6%) of the total QALYs lost. This underscores the importance of accounting for CPP when considering the benefits of chlamydia and gonorrhea interventions for women, and the need for further research on the burden of CPP and potential strategies to reduce this burden [36].

The estimated lifetime QALYs lost attributable to the 3 STIs among men were substantially lower than those for women. However, transmission dynamics for STIs imply that sexual health among men will affect the health of women and their children during pregnancy and childbirth [37]. Thus, the estimates of QALYs lost in relation to the primary infection do not account for the full value of interventions that aim to avert these infections. Several studies have demonstrated that chlamydia screening in high-risk men and combined screening in men and women were cost-saving [38, 39]. In addition, interventions promoting partner notification and treatment of the 3 STIs contribute to reduce repeat infections in women and incidence and long-term QALYs lost in the entire population [40–42].

Estimates for disability weights (utilities) differ substantially across previous studies [43] including the 2 studies used in our analysis, GBD reporting community-based measures and IOM reporting measures from expert perspective [15, 23]. However, there is consistency in the rankings of disutilities across sequelae between the GBD and the IOM study (eg, larger disability weights for CPP than PID and TFI in both studies). Community-based utilities are usually higher than expert-based utilities, and we estimated QALYs lost by averaging utilities from the 2 sources as well as considering extreme cases in sensitivity analyses. Future studies that apply our estimates should account explicitly for large uncertainty around the health state utility values.

Estimates of lifetime discounted QALYs lost per incidence of gonorrhea were similar to the estimates in our previous study [16]. For example, using utility estimates from the IOM study with lifetime duration of chronic sequelae, our estimated discounted lifetime QALYs lost per incidence of gonorrhea for women aged 15–39 years was 0.087 (95% UI, 0.026–0.202) (Supplementary Table 11), similar to that reported in our previous study (0.093; 95% UI, 0.022–0.220) [16]. When using utilities estimates synthesized from the GBD 2019 study and the IOM study, our estimates of QALYs lost per infection (Figure 3) and per case of PID due to the 3 infections (Supplementary Table 13) were smaller than estimates in previous studies [12, 16].

Substantial uncertainty around the probability of developing PID from trichomoniasis limited the estimate's reliability. We varied this probability from 0 to 0.79 (relative to the probability of PID due to chlamydia), in line with high variability across previous studies [5, 22]. The large uncertainty in estimates of QALYs lost from trichomoniasis reflects the data gaps in the natural history of trichomoniasis. We relied on estimates of incidence, probabilities of symptoms, and natural clearance rate from 2 modeling studies [20, 21], which used similar methodology across the infections, and represent the most up-to-date evidence of the epidemiology of these infections. Reliance on one source per STI for infection epidemiology is a limitation and reflects data gaps in the field.

We did not include QALYs lost associated with STI-associated HIV infections, or pregnancy-related adverse outcomes [44] due to the STIs, given relatively limited evidence base on the proportion of these outcomes that are attributable to the 3 STIs compared to the direct consequences emphasized in our study. Rigorous epidemiological studies are needed to improve our understanding of the role of specific STIs and the many confounding factors present. The restriction of the analysis to a small number of selected sequelae reflects current evidence on the burden of STIs and results in conservative estimates. Like any mathematical modelling study, our estimation relied on a number of simplifying assumptions. For example, we assumed that the probability of PID for women with gonorrhea was the same as for chlamydia, and we did not consider the increasing risk of PID among women with repeated chlamydial infection. Gonorrhea may be associated with higher probability of PID than chlamydia and repeated infections may increase the risk of PID and sequelae [5, 45], and our estimates of QALYs lost due to chlamydia and gonorrhea for women could be underestimated.

Our study provides comprehensive estimates of the lifetime QALYs lost per incident infection and overall population health burden of chlamydia, gonorrhea, and trichomoniasis acquired in 2018. Estimates of lifetime QALYs lost for incident STIs in the United States are valuable for informing cost-effectiveness analyses of STIs prevention interventions and prioritizing prevention policies. Our detailed documentation of methodology allows future studies to adjust relevant parameters and select appropriate estimates for their research question. We hope our findings from sensitivity analyses promote more research into these important but poorly estimated parameters in the future.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Disclaimer. The findings and conclusions in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Financial support. This work was supported by the Centers for Disease Control and Prevention, National Center for HIV, Viral Hepatitis, STD, and TB Prevention Epidemiologic and Economic Modeling Agreement (grant number 5 NU38PS004651). Funding to pay the Open Access publication charges for this article was provided by the Centers for Disease Control and Prevention, National Center for HIV, Viral Hepatitis, STD, and TB Prevention Epidemiologic and Economic Modeling Agreement (grant number 5 NU38PS004651).