key: cord-0788416-eomphiy6 authors: Booton, Ross. D.; Fu, Gengfeng; MacGregor, Louis; Li, Jianjun; Ong, Jason J.; Tucker, Joseph D.; Turner, Katy M.E.; Tang, Weiming; Vickerman, Peter; Mitchell, Kate M. title: Estimating the impact of disruptions due to COVID-19 on HIV transmission and control among men who have sex with men in China date: 2020-10-13 journal: medRxiv DOI: 10.1101/2020.10.08.20209072 sha: 5fb1cb021ae5cebc310a58a0532b4d9eaab40a90 doc_id: 788416 cord_uid: eomphiy6 INTRODUCTION: The COVID-19 pandemic is impacting HIV care globally, with gaps in HIV treatment expected to increase HIV transmission and HIV-related mortality. We estimated how COVID-19-related disruptions could impact HIV transmission and mortality among men who have sex with men (MSM) in four cities in China. METHODS: Regional data from China indicated that the number of MSM undergoing facility-based HIV testing reduced by 59% during the COVID-19 pandemic, alongside reductions in ART initiation (34%), numbers of sexual partners (62%) and consistency of condom use (25%). A deterministic mathematical model of HIV transmission and treatment among MSM in China was used to estimate the impact of these disruptions on the number of new HIV infections and HIV-related deaths. Disruption scenarios were assessed for their individual and combined impact over 1 and 5 years for a 3-, 4- or 6-month disruption period. RESULTS: Our China model predicted that new HIV infections and HIV-related deaths would be increased most by disruptions to viral suppression, with 25% reductions for a 3-month period increasing HIV infections by 5–14% over 1 year and deaths by 7–12%. Observed reductions in condom use increased HIV infections by 5–14% but had minimal impact (<1%) on deaths. Smaller impacts on infections and deaths (<3%) were seen for disruptions to facility testing and ART initiation, but reduced partner numbers resulted in 11–23% fewer infections and 0.4–1.0% fewer deaths. Longer disruption periods of 4 and 6 months amplified the impact of combined disruption scenarios. When all realistic disruptions were modelled simultaneously, an overall decrease in new HIV infections was always predicted over one year (3–17%), but not over 5 years (1% increase-4% decrease), while deaths mostly increased over one year (1–2%) and 5 years (1.2 increase - 0.3 decrease). CONCLUSIONS: The overall impact of COVID-19 on new HIV infections and HIV-related deaths is dependent on the nature, scale and length of the various disruptions. Resources should be directed to ensuring levels of viral suppression and condom use are maintained to mitigate any adverse effects of COVID-19 related disruption on HIV transmission and control among MSM in China. pandemic has presented further barriers to HIV control [5] . Quarantine projected that a 6-month interruption in ART supply across 50% of the population on 99 treatment could lead to a 60% increase in HIV-related deaths over a 1-year period [14] . 100 Another modelling study on low-and-middle-income countries (LMIC) projected that HIV-101 related deaths would increase by 10% over the next five years, with the greatest impact on 102 mortality estimated to be from ART interruptions [15] . However, neither of these studies 103 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint used observed COVID-19 impact data to inform their modelled disruptions, which is 104 essential for obtaining reliable projections for the true scale of COVID-19 disruptions. 105 106 To our knowledge no COVID-19 impact modelling has been published focussing on key 107 populations, who are the main groups affected by HIV [16] . In addition, no model projections 108 have to date incorporated observed data from the COVID-19 disruption. In this study, is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 13, 2020. (Table S6) . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. The percentage change in impact measures (new HIV infections and HIV-related deaths) did 220 not vary between each city, with greater within-city variation across the scenarios. Therefore, 221 all results are presented as the overall impact across four cities (Table S1, 222 Guangzhou/Shenzhen/Jinan/Qingdao), with results for each city in the supplement (Tables 223 S2-5) . 224 225 Realistic disruptions to facility-based HIV testing, ART initiation and condom use were each 227 estimated to lead to an increase in new HIV infections among MSM (Fig.1a , Tables S1-S5). 228 Disruptions to condom use (scenario-D) lasting 3 months were predicted to lead to the largest 229 overall relative increase in HIV infections, of 7.8% (95% CrI:4.5-13.8%) over one year 230 (Fig.1a) , with relative increases of 2.3% (1.7-2.9%) and 1.7% (1.2-2.4%) predicted over 1 231 year for realistic 3-month disruptions to facility-based HIV testing (scenario-A) and ART 232 initiations (scenario-B), respectively. Reductions in numbers of sexual partners (scenario-C) 233 were predicted to reduce HIV infections, by a median 16.2% (11.1-23.2%) over one year 234 among MSM following a 3-month disruption. 235 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. The effect of each disruption scenario on the relative percentage change in HIV infections 241 was always smaller over five years than after one year (but not the absolute difference in HIV 242 infections, which generally increased over 5 years), with a more rapid decrease in effect over 243 5 years seen for disruptions to partnership numbers, condom use and VS (Fig.1a) . (Figs.1a,b) . 266 267 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint When comparing the impacts of two combined disruption scenarios (A+B+C+D and 269 A+B+C+D+E25) for disruptions lasting 3/4/6 months, we found impacts on both new HIV 270 infections and HIV-related deaths were approximately linear, with a 4-month disruption 271 leading to around 34% greater impact than a 3-month disruption, and a 6-month disruption 272 around two times the impact of a 3-month disruption, over both the 1-and 5-year time 273 horizons (Fig.2) . 274 275 The absolute numbers of predicted additional/prevented infections and deaths varied for each 277 city (Tables S2-S5 ), related to differing MSM population sizes, but the percentage changes in 278 infections and deaths did not vary substantially between cities (Fig.3, Tables S2-S5) . For 279 example, for scenario A+B+C+D, the overall predicted reduction in new HIV infections over 280 one year varied from 8.7% (-2.8-17.2%) in Qingdao to 9.1% (2.5-15.0%) in Jinan, with far 281 greater within-city than between-city uncertainty (Fig.3) . Over 1 or 5 years, with a 3-month disruption, the relationship between the magnitude of the 296 disruption (0,25,50,75,100%) and the projected impact was always linear (Fig.S4-5 , Table 297 S7-8), with higher values of % disruption leading to increases in new HIV infections and 298 HIV-related deaths for four disruption parameters (facility testing/ART initiation/condom 299 use/VS) but not for partnership disruption, (fewer infections and deaths). 300 301 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint For a 3-month disruption evaluated over a 1-year, theoretical disruption scenarios A-E50 302 (affecting 50% of MSM) increased new HIV-infections by 1.8%, 2.4%, -16.7% (decrease), 303 15.8% and 11.3% respectively, and increased HIV-related deaths by 0.1%, 2.5%, -0.4% 304 (decrease), 0.5% and 18.2% respectively (Fig.S4, Table S7 ). If disruption scenarios affected 305 100% of MSM (A-E100), then new HIV-infections were projected to increase to 3.7%, 4.9%, 306 -30.2% (decrease for 95% affecting MSM -the model requires >0 partnerships), 38.7% and 307 29.3% respectively. Scenarios A-E100 would also cause HIV related deaths to increase to 308 0.3%, 5.3%, -0.8% (decrease -for 95% disruption), 1% and 35.7% respectively. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. increases in HIV incidence of <1% from the suspension of HIV testing (compared to our 361 prediction of 1-2% for 50% reductions in facility testing, scenario-A50, all scenarios in Table 362 S7), <2% from no new ART initiations (2-4%, scenario-B50), 2-9% from the interruption of 363 condom availability (12-33%, scenario-D50), 4-89% from ART interruption (9-31% for 50% 364 reduction in VS, scenario-E50). Suspension of testing for 50% increased deaths by <1% 365 (<1%, scenario-A50), ART initiation <2% (2-3%, scenario-B50) and condom availability 0% 366 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint (0-1%, scenario-D50) with ART interruption causing an increase of 17-62% (14-25%, 367 scenario-E50). Our results align well with these estimates, considering the different 368 methodology/definitions of disruption/population (all adults/children, compared to solely 369 MSM) and underlying models and data (different settings/treatment/condom use). This is a 370 major strength of our study, the use of early data from China, to estimate data-driven (rather 371 than theoretical) magnitudes of COVID-19-related disruptions. Another major strength of this 372 study was performing our analysis on four separate cities from two distinct regions within 373 China. In addition, our analysis involves various scenarios and the effects of combining these 374 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint Joint United Nations Programme on HIV/AIDS (UNAIDS) Global epidemiology of HIV infection in men who have sex with men The prevalence of HIV among 461 MSM in China: A large-scale systematic analysis Challenges to HIV Care Psychological Health During the COVID-19 Pandemic Among People Living with 464 Maintaining HIV care during the COVID-19 pandemic Continuum and COVID-19 Outcomes Among People Living with HIV During the 469 COVID-19 Pandemic Characterizing the Impact of COVID-471 19 on Men Who Have Sex with Men Across the United States in April Evolving 474 ART crisis for people living with HIV in Indonesia Everything is a Mess": How COVID-19 is Impacting Engagement 477 with HIV Testing Services in Rural Southwestern Uganda Mental Health, HIV Prevention and HIV Treatment Impacts of COVID-19 and the 481 COVID-19 Response on a Global Sample of Cisgender Gay Men and Other Men Who 482 The Burden of COVID-484 19 in People Living with HIV: A Syndemic Perspective COVID-19 Pandemic Disrupts HIV Continuum of Care and 487 Prevention: Implications for Research and Practice Concerning Community Quick community survey on 490 the impact of COVID-19 outbreak for the healthcare of people living with HIV Potential 493 effects of disruption to HIV programmes in sub-Saharan Africa caused by COVID-19: 494 results from multiple mathematical models Potential impact of the COVID-19 pandemic on HIV, tuberculosis, and malaria in low-497 income and middle-income countries: a modelling study The COVID-19 emergency is impacting HIV care worldwide, as face-to-face consultations 401 and laboratory testing are reduced, drug and condom manufacture and transport are 402 interrupted, and lockdowns affect peoples' ability to access testing or collect medicines. Gaps . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted October 13, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)The copyright holder for this preprint this version posted October 13, 2020. . https://doi.org/10.1101/2020.10.08.20209072 doi: medRxiv preprint