COVID-19 Policies and Sexually Transmitted Infections in 22 US States, January 2020–December 2021

Abstract

Objectives. To quantify the effects of 5 jurisdiction-level COVID-19 policies on chlamydia, gonorrhea, and syphilis rates in 22 US jurisdictions between January 2020 and December 2021.

Methods. We applied a mixed-effects, negative binomial, interrupted time series model to estimate the impact that each policy of interest had on reported cases.

Results. Mandatory stay-at-home periods were associated with a decrease in reported chlamydia (incident rate ratio [IRR] = 0.75; 95% confidence interval [CI] = 0.71, 0.80) and gonorrhea (IRR = 0.85; 95% CI = 0.78, 0.94) cases. We also observed decreased chlamydia case reporting when gatherings were restricted to 10 people (IRR = 0.88; 95% CI = 0.85, 0.92), masking was recommended (IRR = 0.90; 95% CI = 0.85, 0.97), or polymerase chain reaction testing was limited to symptomatic individuals (IRR = 0.72; 95% CI = 0.67, 0.77). Universal vaccine access corresponded to decreases in reported gonorrhea (IRR = 0.83; 95% CI = 0.75, 0.92) but increases in syphilis (IRR = 1.33; 95% CI = 1.04, 1.70) cases. We also observed effects by sex, race, and ethnicity.

Conclusions. COVID-19 policies had disparate effects on sexually transmitted infection rates that varied across demographic groups. Overall results were attenuated after the first lockdown period (March–May 2020), but demographic variations persisted. (Am J Public Health. 2025;115(5):789–798. https://doi.org/10.2105/AJPH.2024.307957)

COVID-19 pandemic policies had profound effects on the broader disease landscape. Global decreases in respiratory diseases were observed during their typical seasons in 2020 and 2021, potentially attributable to restrictions alleviating COVID-19 spread.¹ Simultaneously, stay-at-home orders restricted health care access.² Sexually transmitted infection (STI) rates were especially affected by the COVID-19 pandemic. Drastic decreases in US rates of chlamydia, gonorrhea, and syphilis were observed in March and April 2020, corresponding with stringent stay-at-home restrictions and STI clinic staff reassignments.^3,4

The driving factors behind these decreases and the subsequent resurgence of gonorrhea and syphilis are unclear but may relate to changes in sexual behaviors or decreases in testing and screening access.^5,6 One study among men who have sex with men noted that 25% experienced difficulties accessing STI testing services, and a majority reported decreases in sexual behavior in early 2020 that increased after stay-at-home requirements were lifted.⁷ A separate study found increases in sexual partnership during the initial lockdown and subsequent period.⁸ Another study found higher rates of chlamydia and gonorrhea in US counties with high COVID-19 case counts, suggesting that COVID-19 policy adherence affected STI rates; however, that study looked only at aggregate 2020 rates and did not directly examine policies.⁹

Literature on the impacts of COVID-19 on STI rates in the United States has focused on the initial shutdown period, limiting understanding on how policies may have affected later shifts in STI rates. Longer longitudinal studies also provide more opportunities to explore the effects of other COVID-19 policies, such as masking requirements, vaccination, and testing access. Furthermore, the majority of studies present nationwide results, which preclude examination of jurisdiction-specific policies. Understanding the implications of specific infectious disease policies can help inform decision-making in future disease outbreaks. We evaluated the impact of 5 COVID-19 policies (i.e., stay-at-home requirements, gathering restrictions, masking requirements, testing access, and vaccination access) on the rates of chlamydia, gonorrhea, and syphilis in the United States between January 2020 and December 2021.

METHODS

We compiled external data on 5 distinct COVID-19 policies and monthly counts of chlamydia, gonorrhea, and syphilis across multiple US jurisdictions. Details on data collection and the corresponding statistical analyses are enumerated.

Data Collection

We sent requests for monthly aggregated count data between 2019 and 2022 for chlamydia, gonorrhea, and syphilis to 59 US jurisdictions with their own public health department. We also requested counts by sex, race, and ethnicity. We adjusted all received data per 100 000 population using the US Census Bureau’s American Community Survey 5-year data for the corresponding jurisdiction and demographic group.¹⁰

We extracted COVID-19 policies from the Oxford COVID-19 Government Response Tracker, which contains 24 distinct COVID-19 policies.¹¹ We limited our analysis to 5 policies that we hypothesized would affect STI rates (Table 1). We considered restrictions on gatherings and stay-at-home requirements “community health practices.” We considered masking requirements, testing access, and vaccination access “individual health practices.” We rounded the policy implementation date to the start of the month to align with monthly STI data. We hypothesized that stricter community health practices and masking requirements would be associated with fewer reported STI cases, as these policies may reduce high-risk sexual behavior frequency and clinic access. We hypothesized that increased testing and vaccination access would be associated with increased reported STI rates, given that infection status awareness and increased protection may result in a reversal of protective behaviors and increased clinic access.

TABLE 1—

COVID-19 Policies Included in the Analysis: 22 US States, January 2020–December 2021

Policy	Levels	Category
Gathering restrictions	No restriction (Ref) Restrictions on gatherings of > 1000 individuals Restrictions on gatherings of 101–1000 individuals Restrictions on gatherings of 11–100 individuals Restrictions on gatherings of ≤ 10 individuals	Community
Stay-at-home requirements	No restriction (Ref) Recommended Required with exceptions Required without exceptions	Community
Masking requirements	No restriction (Ref) Recommended Required in some shared spaces Required in occupied shared spaces Required in occupied or unoccupied shared spaces	Individual
PCR testing access	No policy (Ref) Access for specifica symptomatic individuals Access for any symptomatic individual Access for anyone via drive-through tests	Individual
Vaccine access	No vaccination (Ref) Available for 1 key groupb Available for 2 key groups Available for all key groups Available for all key groups plus select individuals Universal availability	Individual

Note. PCR = polymerase chain reaction.

^a Specific groups included essential workers, close contacts of individuals with confirmed COVID-19, and individuals who had recently traveled internationally.

^b We defined key groups as essential workers, immunocompromised individuals, and those aged 65 years and older.

Statistical Analysis

We constructed mixed-effects negative binomial interrupted time series models for each policy exposure and STI outcome combination. Interrupted time series models are used in health services research to model immediate and long-term effects of policies relative to a prepolicy period.¹² We chose a negative binomial model to account for the strictly positive nature of the outcome. We chose a mixed-effects model with a random effect for jurisdiction to account for repeated measures in a jurisdiction over time. Therefore, all models had the following structure:

$${\widehat{y}}_{s,d,i}={\beta }_0+{\beta }_1{X}_1+{\beta }_2{I}_{2,s}+{\beta }_3{X}_{3,s} +\left[{\beta }_4{I}_{4,s}+\dots +{\beta }_n{I}_{n,s}\right]+u_s,$$ (1)

where ${\widehat{y}}_{s,d,i}$ represents the estimated population-adjusted case counts for disease d in jurisdiction $s$ for month $i$; ${\beta }_0$ represents a fixed-effects intercept; ${\beta }_1{X}_{i,1}$ represents the number of months since January 1, 2020 (the study origin); ${\beta }_2{I}_{2,S}$ represents whether the month was before or after policy implementation (with ${\beta }_2$ interpreted as the immediate effect during the month when the policy was enacted); ${\beta }_3{X}_{3,S}$ represents the number of months after policy enactment (with ${\beta }_3$ interpreted as the long-term policy impact); ${\beta }_4{I}_4$ through ${\beta }_n{I}_{n,s}$ represent policy stringency as determined by the Oxford COVID-19 Government Response Tracker compared with a reference of no policy; and $u_s$ represents a random intercept per jurisdiction.¹¹ We ran separate models for each policy–disease combination, resulting in 15 models (3 diseases and 5 policies) for the main analysis.

Given that behaviors may have differed from March through May 2020 because of the overarching national lockdowns, we conducted a sensitivity analysis with June 2020 as the origin for 4 of the 5 policies. We excluded vaccination access from this sensitivity analysis, because vaccines were unavailable during the excluded period. For all analyses, we considered a P value less than .05 statistically significant. We performed all analyses in R version 4.3.0 using RStudio version 2023.06.0 + 421 (RStudio, Boston, MA).

RESULTS

Of 59 contacted jurisdictions, 50 (84.7%) replied to initial contact. Of respondents, 22 (44%) provided monthly aggregated count data. Jurisdictional representation was geographically diverse, with 7 from the South (32%), 6 from the Midwest (27%), 5 from the West (23%), and 4 from the Northeast (18%). Additional data on sex, race, and ethnicity were provided for 19 of 22 (86.4%) jurisdictions (Figure 1). We evaluated sex for men and women, as most jurisdictions provided binary data for this demographic. We evaluated race (i.e., White, Black/African American, Asian, American Indian/Alaska Native [AIAN], Native Hawaiian/Pacific Islander) and ethnicity (i.e., non-Hispanic/non-Latino, Hispanic/Latino) for the most consistent categories reported across jurisdiction.

FIGURE 1—

Jurisdictions Providing Data for the Study: 22 US States, January 2020–December 2021

Note. All jurisdictions shown in light blue provided monthly, aggregated data. Jurisdictions shown in dark blue provided additional monthly aggregated data by sex, race, and ethnicity.

We removed counts for additional race/ethnicity categories not listed (including “other”) to avoid comparing groups that were inconsistent between jurisdictions. We excluded jurisdictions without demographic data for a category from that subanalysis. We aggregated data by the jurisdictions according to their established policies and procedures. Appendix A (available as a supplement to the online version of this article at http://www.ajph.org) provides jurisdiction data notes.

COVID-19 Community Health Policy Impact

For chlamydia and gonorrhea, mandatory stay-at-home periods were associated with decreases in average disease rates (Figure 2). There were 25% fewer reported chlamydia cases (incident rate ratio [IRR] = 0.75; 95% confidence interval [CI] = 0.71, 0.80; P < .001) and 15% fewer gonorrhea cases (IRR = 0.85; 95% CI = 0.78, 0.94; P = .001) on average compared with periods without restrictions. This was consistent in the majority of demographic subgroups to varying degrees (Appendix B, available as a supplement to the online version of this article at http://www.ajph.org). Hispanic individuals had the greatest average decrease for both chlamydia (IRR = 0.61; 95% CI = 0.55, 0.69; P < .001) and gonorrhea (IRR = 0.71; 95% CI = 0.61, 0.81; P < .001). Mandatory stay-at-home periods were not associated with syphilis rates.

FIGURE 2—

Estimated Effects of COVID-19 Community Health Policies on Chlamydia, Gonorrhea, and Syphilis Rates per 100 000 Persons: 22 US States, January 2020–December 2021

Note. AIAN = American Indian/Alaska Native; CI = confidence interval; IRR = incident rate ratio; NHPI = Native Hawaiian/Pacific Islander. Red points with IRR > 1 indicate a significantly increased rate of the sexually transmitted infection of interest during the month of policy enactment compared with a period of no policy restrictions, and green points with IRR < 1 indicate a significantly decreased rate. Black points indicate no significant difference. Missing points indicate that not enough jurisdictions provided data for that category to create a model, and missing CIs denote especially tight bands that were not present given the figure scale.

Recommended stay-at-home periods had an effect on gonorrhea but not chlamydia. There were 15% more gonorrhea cases on average (IRR = 1.15; 95% CI = 1.09, 1.21; P < .001) during recommended stay-at-home periods than during no policy periods, with the largest increase among women (IRR = 1.24; 95% CI = 1.16, 1.32; P < .001). Recommended stay-at-home periods had no statistical effect on overall syphilis rates. There were changes among demographic groups, with Asians experiencing the greatest average decrease (IRR = 0.53; 95% CI = 0.28, 0.98; P = .043).

Gathering size restrictions had varying impacts (Appendix C, available as a supplement to the online version of this article at http://www.ajph.org). When gatherings were restricted to 10 people, chlamydia decreased by 12% on average (IRR = 0.88; 95% CI = 0.85, 0.92; P < .001) compared with no policy periods, with Hispanic individuals experiencing the greatest decrease (IRR = 0.85; 95% CI = 0.79, 0.92; P < .001). When gatherings were restricted to 1000 people, men (IRR = 0.86; 95% CI = 0.75, 0.98; P = .025) and Hispanic individuals (IRR = 0.73; 95% CI = 0.56, 0.96; P = .024) experienced case decreases.

Although gonorrhea cases overall were not significantly affected during these periods, Black individuals (IRR = 1.11; 95% CI = 1.01, 1.21; P = .025) and women (IRR = 1.10; 95% CI = 1.02, 1.18; P = .016) had significantly more average cases than in periods without restrictions. This effect persisted when gathering sizes were relaxed, with additional demographic groups (i.e., men, non-Hispanic, Hispanic) experiencing a significant increase in cases. Syphilis rates were not significantly affected by gathering restrictions.

COVID-19 Individual Health Policy Impact

When masking was recommended, there were 10% fewer average chlamydia cases (IRR = 0.90; 95% CI = 0.85, 0.97; P = .003) than in periods without masking policies (Figure 3). This trend was consistent across demographic groups, with the largest effect among AIAN individuals (IRR = 0.54; 95% CI = 0.35, 0.83; P = .005). When masking was always required in shared spaces, chlamydia cases were higher among women (IRR = 1.07; 95% CI = 1.01, 1.14; P = .024) and non-Hispanic individuals (IRR = 1.12; 95% CI = 1.04, 1.21; P = .003) than in no policy periods (Appendix D, available as a supplement to the online version of this article at http://www.ajph.org). Overall gonorrhea cases increased by 32% on average (IRR = 1.32; 95% CI = 1.23, 1.43; P < .001) during these periods, an effect shared by many demographic groups. We observed no significant effect for syphilis.

FIGURE 3—

Estimated Effects of COVID-19 Masking Policies on Chlamydia, Gonorrhea, and Syphilis Rates per 100 000 Persons: 22 US states, January 2020–December 2021

Note. AIAN = American Indian/Alaska Native; CI = confidence interval; IRR = incident rate ratio; NHPI = Native Hawaiian/Pacific Islander. Red points with IRR > 1 indicate a significantly increased rate of the sexually transmitted infection of interest during the month of policy enactment compared with a period of no policy restrictions, and green points with IRR < 1 indicate a significantly decreased rate. Black points indicate no significant difference. Missing points indicate that not enough jurisdictions provided data for that category to create a model, and missing CIs denote especially tight bands that are not present given the figure scale.

The impact of COVID-19 polymerase chain reaction (PCR) testing access on STI rates was consistent (Appendix E, available as a supplement to the online version of this article at http://www.ajph.org). Observed STI rates decreased as PCR testing was restricted to symptomatic individuals in limited populations (e.g., those who had recently traveled internationally, essential workers). There were 20% fewer gonorrhea cases (IRR = 0.80; 95% CI = 0.72, 0.90; P < .001), 28% fewer chlamydia cases (IRR = 0.72; 95% CI = 0.67, 0.78; P < .001), and 29% fewer syphilis cases (IRR = 0.71; 95% CI = 0.55, 0.91; P = .006) on average compared with periods when PCR testing access was not measured, an effect paralleled across most demographic subgroups. When PCR tests were readily available to any individual, effects diverged.

We found fewer average chlamydia cases (IRR = 0.87; 95% CI = 0.80, 0.95; P = .002), more average gonorrhea cases (IRR = 1.19; 95% CI = 1.04, 1.37; P = .011), and no significant change in syphilis cases (IRR = 0.83; 95% CI = 0.63, 1.09; P = .182). Women (IRR = 1.25; 95% CI = 1.07, 1.46; P = .006) and Hispanic individuals (IRR = 1.33; 95% CI = 1.09, 1.62; P = .006) had a significant increase in gonorrhea cases when testing was more readily available than in no policy periods.

Vaccination access had a contradictory impact relative to other policies (Appendix F, available as a supplement to the online version of this article at http://www.ajph.org). Although vaccine access did not have an overall effect on chlamydia cases, men (IRR = 1.17; 95% CI = 10.6, 1.28; P = .001), White individuals (IRR = 1.21; 95% CI = 1.08, 1.36; P = .001), non-Hispanic individuals (IRR = 1.20; 95% CI = 1.09, 1.32; P < .001), and Hispanic individuals (IRR = 1.28; 95% CI = 1.10, 1.48; P = .001) experienced increases when vaccines became universally available. There were 33% more syphilis cases on average (IRR = 1.33; 95% CI = 1.04, 1.70; P = .022) than in periods when COVID-19 vaccines were unavailable. There were 13% fewer gonorrhea cases on average (IRR = 0.83; 95% CI = 0.75, 0.92; P = .001) when vaccines were universally available, with women experiencing a significant decrease (IRR = 0.83; 95% CI = 0.73, 0.93; P = .002).

Sensitivity Analysis

When we excluded March to May 2020 from the analysis, the overall impact of each policy on STI rates was no longer statistically significant. Substantial differences remained for select demographic groups. There were 16% fewer average chlamydia cases in men (IRR = 0.84; 95% CI = 0.74, 0.97; P = .017) and 24% fewer in Hispanic individuals (IRR = 0.76; 95% CI = 0.59, 0.98; P = .037) when gatherings were restricted to 1000 persons or fewer than in periods without restrictions. No significant associations were found for other demographic subgroups or STIs.

There were 13% fewer chlamydia cases in women (IRR = 0.87; 95% CI = 0.77, 0.98), 22% fewer in Hispanic individuals (IRR = 0.79; 95% CI = 0.62, 0.99; P = .043), and 32% fewer in Asian individuals (IRR = 0.68; 95% CI = 0.47, 1.00; P = .048) on average when stay-at-home requirements were mandatory than in periods without stay-at-home policies. There were 24% fewer gonorrhea cases in men (IRR = 0.76; 95% CI = 0.62, 0.93; P = .006), 35% fewer in Hispanic individuals (IRR = 0.65; 95% CI = 0.49, 0.86; P = .003), 34% fewer in White individuals (IRR = 0.66; 95% CI = 0.48, 0.92; P = .013), and 23% fewer in Black individuals (IRR = 0.77; 95% CI = 0.59, 1.00; P = .048) on average when stay-at-home requirements were mandatory. All other stay-at-home associations were not statistically significant.

On average, chlamydia cases among Hispanic individuals were 48% lower (IRR = 0.52; 95% CI = 0.31, 0.88; P = .015) during periods when masking was recommended than during periods without masking requirements. There were 30% more chlamydia cases in Asian individuals (IRR = 1.30; 95% CI = 1.08, 1.58; P = .006) on average when masking was sometimes required in shared public spaces. There were 37% fewer gonorrhea cases in men (IRR = 0.63; 95% CI = 0.47, 0.84; P = .002), 34% fewer in women (IRR = 0.66; 95% CI = 0.56, 0.95; P = .025), 43% fewer in non-Hispanic individuals (IRR = 0.57; 95% CI = 0.40, 0.82; P = .002), 59% fewer in Hispanic individuals (IRR = 0.41; 95% CI = 0.28, 0.61; P < .001), 35% fewer in White individuals (IRR = 0.65; 95% CI = 0.46, 0.92; P = .016), 57% fewer in Black individuals (IRR = 0.43; 95% CI = 0.22, 0.83; P = .012), and 96% fewer in AIAN individuals (IRR = 0.06; 95% CI = 0.01, 0.71; P = .025) on average when masks were recommended. During these same periods, there were 73% more cases of syphilis in White individuals (IRR = 1.73; 95% CI = 1.03, 2.92; P = .040) on average. All other masking associations were not statistically significant.

There were 33% fewer gonorrhea cases in Hispanic individuals (IRR = 0.67; 95% CI = 0.48, 0.92; P = .015) and 74% fewer in Asian individuals (IRR = 0.26; 95% CI = 0.08, 0.89; P = .032) on average when PCR testing was only available to symptomatic persons than in periods without PCR testing access policies. All other PCR testing associations were not statistically significant.

DISCUSSION

To our knowledge, this study is the first to quantify the consequences of 5 COVID-19 policies (i.e., stay-at-home requirements, gathering restrictions, masking requirements, testing access, and vaccination access) on chlamydia, gonorrhea, and syphilis rates across the United States. In 22 geographically diverse jurisdictions, chlamydia decreased when stay-at-home requirements were mandatory, gathering restrictions were most restrictive, masking was recommended, and PCR tests were universally available. Gonorrhea cases decreased when stay-at-home requirements were mandatory and vaccines were more available, and they rose when gathering restrictions were loosened and testing access increased. Syphilis was largely unaffected by community health policies, whereas universal vaccination access corresponded with case increases. The magnitude and significance of these effects varied across demographics and period. Although the overall association between COVID-19 policies and STI rates were attenuated after the first US lockdown, relationships in select demographic groups persisted. These heterogeneous results suggest the complex impact diverse public health policies can have on STI rates across demographic groups.

We selected the 5 evaluated COVID-19 policies to capture various facets of health behavior. Stay-at-home requirements and gathering restrictions were contagion mitigation policies that, although effective, were highly divisive.^13,14 Thus, these policies are likely to be reserved for crises and may demonstrate how the most stringent emergency policies for 1 disease affect another. However, masking policies are less burdensome and may offer a middle ground for contagion mitigation that balances easy implementation with effectiveness. Yet the political partisanship of masking may inhibit their utility in future infectious disease outbreaks. The restrictive nature of these policies led to our hypothesis that increases in their stringency would be associated with a decrease in STI rates.

In addition, testing and vaccine policies may influence human behavior. Previous work has shown that patients diagnosed with influenza traveled less frequently during their illness than during illness-free periods, suggesting that knowledge of disease state may influence behaviors.¹⁵ Thus, we hypothesized that widespread knowledge of COVID-19 status (as measured through testing access) would similarly affect sexual behaviors and thus STI rates. We hypothesized that vaccination availability would have a similar effect, as patients vaccinated against COVID-19 may be more likely to reengage in sexual behaviors after lockdown. Alternatively, taking a COVID-19 test and receiving a vaccination are health-seeking behaviors that decrease transmission, and individuals exhibiting these behaviors may also be less likely to engage in high-risk behaviors that could increase their STI risk.¹⁶

Researchers have speculated that the STI decrease during the US national lockdown was attributable to reduced access to sexual health services.^{3–5,7,17,18} Others have argued that it may be owing to sexual behavior changes.^1,19,20 Our results support the clinic access hypothesis, as we observed significantly fewer chlamydia and gonorrhea cases when stay-at-home orders were mandated and when gatherings were limited to 10 or fewer people (chlamydia only). This effect disappeared when we excluded March to May 2020, even though similar lockdown policies were reintroduced in select jurisdictions afterward. This supports the idea that closing clinics and reducing testing services—not the lockdowns—contributed to the decreases in testing. However, the policies affected almost every demographic group individually after this period. These results align with the hypothesis that increased stringency in stay-at-home orders and gathering restrictions would be associated with lower STI rates. Future work should examine each group individually to identify unique determinants that affected these results and have long-term influences on sexual health service access.

Periods with the most stringent masking policies were not significantly associated with STI rates. However, this shifted as masking policies loosened. Whereas gonorrhea cases increased when masks were sometimes required, chlamydia cases decreased when masking was recommended. This effect differed among women and non-Hispanic individuals, who experienced significant increases in cases when masking was sometimes required. These increases may also be connected to clinic reopenings and resumption of regular testing, and chlamydia increases among women and non-Hispanic individuals may suggest that they resumed normal visits earlier than did others. These findings were partially aligned with our hypothesis, and variations may be attributable to adherence, which we did not measure in this study.

Similarly, although COVID-19 vaccine access was not associated overall with chlamydia, we observed an increase uniquely among men, Hispanic individuals, and White individuals when the vaccine became universally available. This may suggest that these individuals were last to return to regular clinic visits, especially because universal COVID-19 vaccine access was not granted until spring 2021. It could also indicate a resurgence in sexual behavior among these groups, an effect observed directly after national lockdowns both in and outside the United States.^5,21,22 However, testing and vaccination access are critical components of policy adherence that we did not measure, which could explain the lack of an overall relationship between COVID-19 vaccine access and STI rates.

There is limited earlier work measuring associations between COVID-19 policies and STI rates. One study among men who have sex with men in the Netherlands found a decrease in STIs during the nationwide lockdown that was not significant during subsequent policy shifts.²¹ There were fewer chlamydia and gonorrhea cases among men in this study when stay-at-home and gathering restrictions were implemented after the main lockdown period. Because earlier work evaluated only men who have sex with men, future work should stratify analyses by sexual behavior to determine what role this may play. Another study found no association between adherence to nonpharmaceutical COVID-19 policies and a history of STI among adolescent and young adult Black men, although temporality of the STI relative to policy enactment was not evaluated.²³ By contrast, we found a marked association between policy implementation and STI rates among both Black/African American individuals and men.

We hypothesize that our results demonstrate challenges in testing access that differentially affected demographic groups. The decrease in STI rates during the complete shutdown period correlates with data in other work that showed that STI testing decreased during the same period.⁶ The dramatic increase in congenital syphilis shown in other work suggests that sexual behaviors and STI transmission still occurred during this time window.²⁴ Given this pattern, we expected that syphilis rates in our study would also change dramatically; however, syphilis rates did not fluctuate widely. We hypothesize that this discrepancy is attributable to testing and treatment changes that remained after lockdown orders were rescinded. The racial and ethnic differences in STI rates after lockdown suggest that there may have been additional inequitable access to testing and STI services in particular groups. Thus, our results emphasize the crucial need to identify areas for technical and process innovations that can enhance equitable access to testing.

Limitations

Several study limitations are of note. First, the ecological nature of this study precluded our ability to establish causality and may have obfuscated individual trends. Future work could directly evaluate COVID-19 policy adherence and STI rates over time, especially beyond the initial lockdown. This could also confirm the precise temporality of cases relative to policy implication, as we assumed policy implementation on the first of every month.

Second, data were not available for certain demographic groups, and future work could repeat this analysis for those groups, including men who have sex with men. Similarly, although the public health jurisdictions that provided data were geographically diverse, they were not comprehensive. We hypothesize that this did not affect results, assuming jurisdictions are missing at random.

Third, given the heterogeneity in individual jurisdictions, we did not adjust models for jurisdiction-level variables, thereby potentially exaggerating the observed effects. Similarly, we did not incorporate substate policies specific to a certain county or city into our analysis, as data were not available at a lower level of geography. We assumed that state policies superseded local policies, but future work could examine the effect of more localized policies on STI rates (especially for jurisdictions where the city and state policies differed).

Finally, our chosen model, although straightforward in interpretation, assumed a linear relationship and serial correlation that did not diminish over time. Future work could explore alternative model structures (e.g., an autoregressive integrated moving average) to assess the relationship between COVID-19 policy stringency and STI rates.

Conclusions

COVID-19 stay-at-home orders, gathering restrictions, masking requirements, PCR testing access, and vaccine access had differential effects across demographic groups and STI rates. Chlamydia decreased with more stringent community health policies, whereas gonorrhea decreased for more stringent policies before increasing when policies were loosened. Syphilis was unaffected except when vaccines became universally available, which corresponded to case increases. Aggregated impacts became statistically insignificant when policy effects were limited to June 2020 onward but persisted among demographic groups. Taken together, these results suggest that more stringent community health COVID-19 policies had a greater influence on STI rates than did less stringent individual health policies. These relationships could be mediated by policy adherence or testing access, which we did not capture and which should be examined in future work. Before enacting new policies for a disease, policymakers and health experts should consider how they may fit into the broader disease landscape, including both overall and group-specific consequences.