INTRODUCTION

The coronavirus disease 2019 (COVID-19) pandemic has had profound effects on international tourism, as travel restrictions and bans have economically and socially impacted regions dependent on tourism [1]. In 2020, Phuket province, a major tourist destination in Thailand, reportedly lost around US$50 billion in tourism revenue [2]. However, reopening international tourism carries the risk of importing COVID-19 cases into local communities, potentially sparking outbreaks. Given the ongoing evolution of COVID-19 [3] and the critical need to revive the economies of tourism-reliant countries, it is essential to assess the effectiveness of previous pandemic travel protocols. This evaluation will help in developing future measures that facilitate safe travel while containing the spread of COVID-19.

The Phuket Sandbox Project was a tourism reintroduction initiative designed to balance the return of tourists with the prevention of COVID-19 importation into Thailand. Fully vaccinated travelers were permitted to enter the island of Phuket, subject to an island-confined quarantine following an initial quarantine period. The project achieved a high satisfaction rate, with 96.7% of travelers reporting a positive experience in Phuket [4]. The Phuket Project is considered successful in creating a positive tourist experience while enforcing mandatory screening, quarantine, and testing procedures. However, the effectiveness of the Sandbox Project in preventing the spread of COVID-19 has not yet been assessed. This project offered a unique opportunity to conduct a real-world cohort study to identify the risk of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection associated with international air travel under island-confined quarantine conditions.

The predominant variant of SAR-CoV-2 during the Phuket Sandbox Project was B.1.617.2, also known as the Delta variant. Vaccines have been shown to reduce the risk of infection from this and other variants in fully vaccinated individuals [5]. However, the effectiveness of vaccines varies among different brands and platforms [6], and their efficacy also tends to wane over time, depending on the brand and platform used [7]. Despite vaccination, breakthrough cases can still occur and pose a significant risk of infection to vulnerable populations due to their contagiousness [8]. Even with mandatory vaccination policies, the risk of importing cases during travel remains. Implementing screening and quarantine measures can help decrease the likelihood of imported infections and local transmission by enhancing detection rates among travelers [9]. Quarantines have been found to be more effective than travel bans at reducing the risk of imported cases [10]. Given the burden that border control policies place on local and tourist populations [11], quarantines offer a viable solution for facilitating tourism while containing the spread of the virus.

The primary objective of this study was to estimate the incidence of breakthrough SARS-CoV-2 infections among travelers participating in the Phuket Sandbox Program. The results could help validate the effectiveness of quarantine measures and inform future strategies for reintroducing tourism during a pandemic. Additionally, this study could redefine perceptions of island-confined quarantine, which has been traditionally viewed negatively, as a viable option for island tourist destinations. The second objective focused on assessing the impact of different vaccine brands and types on the risk of SARS-CoV-2 infection among fully vaccinated travelers. By comparing the risks associated with various vaccines, this analysis aimed to corroborate previous research findings and suggest potential enhancements to vaccine-related travel policies. The third objective was to compare the time to detection of the virus in asymptomatic versus symptomatic infected travelers. Understanding the differences in detection times between these 2 groups could lead to improvements in travel protocols and necessitate adjustments in the timing of preventive measures.

METHODS

Study Design and Setting

This retrospective cohort study analyzed data from 63 052 travelers who participated in the Phuket Sandbox Quarantine Project from July 1, 2021, to October 31, 2021. Of these travelers, 55 344 aged 18 years or older were fully vaccinated. The project included a failsafe mechanism to pause operations if certain conditions were met: more than 90 new infections per week, detection of the disease in all 3 Phuket districts and over 6 sub-districts, more than 3 outbreak clusters, widespread outbreaks with unidentifiable sources or connections, provincial hospital’s occupancy rates exceeding 80% for SARS-CoV-2 cases, or uncontrollable outbreaks of a mutant strain.

Before departure, travelers were required to obtain a Thai Certificate of Entry (COE). The Thai government mandates that travelers provide a passport or visa, a vaccination certificate, health insurance with a minimum coverage of US$100 000, an airline ticket, a reservation at a hotel participating in the Phuket Sandbox Project, and proof of payment for the reversetranscription polymerase chain reaction (RT-PCR) testing required during the quarantine period to be eligible for a COE. Travelers had to submit a vaccination certificate confirming that they had completed the primary vaccination series against COVID-19 at least 14 days before departure. A complete primary series was defined as receiving at least 2 doses of CoronaVac, BBIBP-CorV (Sinopharm/COVILO), AZD1222/ChAdOx1 nCoV-19 (AstraZeneca/COVISHIELD), Gam-COVID-Vac (Sputnik V), BNT162b2 (Pfizer-BioNTech/COMIRNATYTM), or mRNA-1273 (Moderna), or at least 1 dose of Ad26.COV2.S (Johnson & Johnson). For 2-dose vaccine schedules, the second dose could be a different type than the first dose. The interval between doses should be at least 2 weeks for the CoronaVac vaccine, 3 weeks for the Sinopharm/COVILO, Pfizer-BioNTech/COMIRNATYTM, and Sputnik V vaccines, or 4 weeks for the AstraZeneca/COVISHIELD and Moderna vaccines. Fully vaccinated travelers were defined as individuals who had completed a primary series of COVID-19 vaccinations. Additionally, travelers were required to complete RT-PCR testing for COVID-19 within 72 hours of their flight departure and declare the aforementioned documents and information upon entry at Phuket International Airport (IATA: HKT).

Upon arrival, travelers underwent health control and immigration checks at the airport. They were tested for COVID-19 using an RT-PCR test upon arrival. Subsequently, travelers were transported from the airport to a participating hotel using project-certified transportation. While at the hotel, travelers awaited the results of the RT-PCR test conducted at the airport. If the result was negative, the traveler was allowed to move freely throughout Phuket. Conversely, a positive result led to hospital referral for treatment. Travelers who initially tested negative for COVID-19 were subjected to 2 additional tests during their quarantine period as part of the Phuket Sandbox Project. The second test was administered between the sixth and seventh days, and the third test between the twelfth and thirteenth days after arrival. Any individual testing positive during these subsequent tests was also referred to a hospital for treatment. In addition to the scheduled tests, travelers were required to use the “ThailandPlus” mobile application to assess whether further testing was necessary. While traveling on the island, they used this app to scan QR codes at public locations for check-in purposes. The app assigned a risk level to each traveler, which was updated based on their interactions with infected individuals or high-risk groups, utilizing Bluetooth contact tracing and global positioning system tracking. The application alerted travelers about their high-risk status and provided guidance on additional measures such as quarantine, symptom monitoring, or testing at a designated site. Travelers who tested negative for COVID-19 in their 12-day to 13-day RT-PCR test were released from quarantine and allowed to travel freely within Thailand.

Study Participants

Eligible participants included all air travelers arriving at IATA: HKT from July 1, 2021 to October 31, 2021. These individuals were required to have a COE and a medical certificate confirming a negative COVID-19 test result obtained within 72 hours prior to departure. Due to a policy mandating full COVID-19 vaccination for travelers aged 18 years and older, participants were categorized into 2 groups: all travelers and a subgroup of travelers aged 18 and older, excluding those under 18, for subsequent separate analyses. For those in the fully vaccinated group, a vaccination certificate was also required in addition to the aforementioned documents. Individuals with a negative COVID-19 test result were considered eligible as they were part of the population at risk of COVID-19 infection before the commencement of the study period. All individuals who met these eligibility criteria and participated in the Phuket Sandbox Program, which included testing and quarantine measures, were included in this study. Travelers lacking available COVID-19 test results were excluded.

Study Variables and Measurement

Traveler information, including age, sex, nationality, continent of departure, type of last dose vaccine, the number of days between the last dose of vaccine and departure, dates of RT-PCR testing before and during travel, and date of arrival, was obtained from the Phuket Immigration Checkpoint database. Dates of RT-PCR testing, testing results, and time of infection were collected from the Phuket Provincial Public Health Office Database. Age, sex, and nationality data were validated and supplemented with information from the Phuket Provincial Public Health Office Database. The type of last dose vaccine data was validated and supplemented using information from the databases of the Phuket Provincial Public Health Office and the Tourism Authority of Thailand, Phuket Office. The dates of RT-PCR testing and testing results were also validated and supplemented using information from the Tourism Authority of Thailand, Phuket Office Database. The continent of departure was determined by categorizing country of origin data into regional data using the World Health Organization region map. Testing results were provided serially, with the number of results depending on the traveler’s risk and whether a positive result occurred. The difference in time of infection between asymptomatic and symptomatic cases was determined by calculating the number of days between the date of arrival and the date of infection.

Statistical Analysis

Continuous variables, including age and time to detection, were summarized using either the mean and standard deviation or the median and interquartile range, depending on their distribution. The independent samples t-test was employed to compare means between infected and non-infected travelers. Categorical variables were summarized using frequency and percentage, and comparisons of proportions across categories were conducted using the Pearson chi-squared test. The effects of different brands and types of COVID-19 vaccines on the outcome of SARS-CoV-2 infection were evaluated using multivariable Poisson regression with robust standard errors, adjusting for covariates such as sex, age, nationality, and continent of departure. The Kaplan-Meier failure curve was used to visualize the time to detect SARS-CoV-2 infection among infected travelers throughout the 14-day quarantine period. The logrank test was utilized to compare the failure curves between asymptomatic and symptomatic travelers.

Ethics Statements

This study was approved by the Hofstra University Human Subjects Committee (20220727-PH-HPHS-ROJ-2).

RESULTS

A total of 63 052 travelers were quarantined through the Phuket Sandbox Program, of which 55 344 were aged 18 years or older. The incidence of SARS-CoV-2 infection was approximately 3 cases per 1000 travelers (0.3%). The youngest traveler was 3 days old, while the oldest was 104 years old. The age group under 18 years old exhibited the highest incidence of SARS-CoV-2 infection, around 0.4%. The average age of infected travelers was significantly lower than that of those who were not infected. Most travelers were male and held non-Thai nationalities. However, the incidence of SARS-CoV-2 infection was similar across both sexes and between Thai and non-Thai nationals. Nearly half of the travelers, about 48.4%, came from the Eastern Mediterranean Region, the highest incidence of infection was observed among those from the Region of the Americas. The majority of travelers aged 18 and older, who were required to be fully vaccinated against COVID-19, received their last dose of BNT162b2 (55.5%). However, the highest incidence of infection (1.5%) was observed among travelers who received the Gam-COVID-Vac, despite this group being the smallest proportion of travelers. A higher incidence of infection was noted among travelers who received inactivated virus (0.4%) and non-replicating viral vector (0.5%) vaccines. The median time for the detection of SARS-CoV-2 infection among cases was on day 5. Infected cases were detected upon arrival in about 30.8% of instances. The majority of infected SARS-CoV-2 travelers were detected between days 1 and 6 after arrival (Table 1).

Table 2 presents the effects of various brands and types of COVID-19 vaccines on the risk of SARS-CoV-2 infection, with adjustments made for covariates to yield an adjusted risk ratio (aRR). In terms of the impact of vaccine brand, the first multivariable regression model used the BNT162b2 (Pfizer) vaccine as the reference. It found that travelers vaccinated with the Ad26.COV2.S (Johnson & Johnson), Gam-COVID-Vac (Sputnik V), and CoronaVac (Sinovac) vaccines experienced a significantly higher risk of SARS-CoV-2 infection. Among these, the Gam-COVID-Vac vaccine showed the highest risk ratio (aRR, 5.82). Regarding the influence of vaccine type on the risk of SARS-CoV-2 infection, the second multivariable model indicated that individuals who received a non-replicating viral vector vaccine faced a 2.35 times higher risk of infection compared to those who received RNA-based vaccines, which served as the reference (Table 2).

Of the total 186 cases, information about the presence of symptoms was available for 95.2% (177 cases); of these, approximately 82.5% were asymptomatic travelers. There seemed to be no significant difference in the proportion of asymptomatic and symptomatic infections across various ages, sexes, nationalities, and continents of departure. Furthermore, no significant differences were observed in the incidence of symptomatic cases among different vaccine brands or types. Additionally, when comparing the time to detection between asymptomatic and symptomatic cases, the number of days between arrival and detection of infection showed no significant differences between the 2 groups (Table 3).

The Kaplan-Meier failure curves illustrated the time to detection of SARS-CoV-2 infection among 177 infected travelers of all ages and 152 infected travelers aged 18 and older. For the entire group of 177 cases, the median detection time was 5 days. However, when analyzed by symptom presence, asymptomatic cases had a median detection time of 4 days, while symptomatic cases had a median detection time of 5 days. In the subgroup of travelers aged 18 and older, the median detection time was consistently 5 days, with both asymptomatic and symptomatic cases showing a median detection time of 4 days upon further stratification. The log-rank test indicated no significant difference in the distribution of detection times between asymptomatic and symptomatic cases (Figure 1).

DISCUSSION

The overall infection rate of SARS-CoV-2 among air travelers was 0.3%, or 3 infections per 1000 travelers. While this rate may seem relatively low, it is important to note that it was achieved under stringent conditions: mandatory full COVID-19 vaccination, pre-departure RT-PCR testing, on-arrival testing, and quarantine with intermittent testing. Furthermore, passengers under the age of 18 had the highest incidence of SARS-CoV-2 infection. This higher incidence in younger travelers could be attributed to the ongoing evaluation of COVID-19 vaccine efficacy and safety for children and adolescents during the Phuket Sandbox Project, which led to lower vaccination rates in this age group.

After evaluating the relative impact of different COVID-19 vaccine brands and types on the risk of SARS-CoV-2 infection, and adjusting for covariates such as age, sex, ethnicity, and continent of departure, it was found that travelers vaccinated with Gam-COVID-Vac, Ad26.COV2.S, and CoronaVac faced a significantly higher risk of SARS-CoV-2 infection compared to those vaccinated with BNT162b2. Additionally, travelers who received non-replicating viral vector vaccines were at 2.35 times the risk of infection compared to those who received the mRNA vaccine. There was no significant difference in the time to detection between asymptomatic and symptomatic travelers.

The policy mandating full vaccination for entry was anticipated to significantly reduce the incidence of infection. However, the observed incidence of SARS-CoV-2 infection indicates that while vaccines may lower the risk of contracting the virus [5], they do not entirely prevent it. Consequently, vaccines alone may not suffice to prevent the importation of cases through travel. The low incidence observed aligns with prior evidence [12]; nonetheless, it is important to note that this study implemented specific preventative measures. Without such measures, air travel might experience a higher rate of infection. Given these findings, there remains a strong rationale for maintaining SARS-CoV-2 surveillance systems to monitor tourists and balance the risk of local transmission against the economic benefits of the tourism industry.

The type and brand of vaccine received by travelers before entering Thailand influenced their differential risk of infection. Previous studies have shown that Gam-COVID-Vac, Ad26.COV2.S, and CoronaVac exhibit lower efficacies [13,14]; this is in contrast to the BNT162b2 vaccine, which has demonstrated high efficacy among other SARS-CoV-2 vaccines [14]. This finding aligns with the increased risk associated with non-replicating viral vector-type vaccines, such as Ad26.COV2.S and Gam-COVID-Vac [14]. Generally, non-replicating viral vector vaccines are less effective at preventing infection compared to mRNA vaccines [15]. However, the efficacy of these vaccine brands and types may have shifted since the conclusion of the study period, given the continuously evolving nature of COVID-19 [3].

Symptomatic and asymptomatic infections showed no difference in the time to detection among infected travelers. Previous knowledge indicates that the viral loads in symptomatic and asymptomatic infections are not significantly different, and factors such as age and sex may influence the manifestation of symptoms [16]. Therefore, it was consistent to find that symptomatic and asymptomatic individuals developed infections at the same rate. This observation supports earlier findings that the timing of infection does not affect the ratio of symptomatic to asymptomatic infections [17]. There was a high incidence of asymptomatic cases in the study population compared to previous studies [18]; asymptomatic individuals may act as a reservoir for the virus [16], potentially leading to the spread of SARS-CoV-2 into local communities without intervention. The mandatory quarantine system, coupled with intermittent testing, appears to have effectively prevented the spread of the infection, as the project did not reach the thresholds that would necessitate pausing or canceling travel. This supports earlier evidence that pre-departure testing and quarantine interspersed with frequent testing can reduce the importation of cases [19]. We anticipate that the future implementation of an island quarantine system could preserve the benefits of mandatory quarantine while mitigating its negative impacts on local and tourist populations [2,11].

This study’s strengths included the large size of the travel population, the utilization of real-world data, and the diversity of vaccines administered to travelers. However, despite meticulous efforts to collect all routine data in accordance with the practical protocol of the Phuket Sandbox Project, a significant limitation was the omission of non-routine variables not specified in the protocol. These omitted variables, such as travelers’ underlying diseases, could potentially act as confounders in evaluating the risk of SARS-CoV-2 infection in relation to the efficacy of various vaccine brands and types. Nevertheless, the presence of common underlying diseases—such as cardiovascular disease and diabetes—underscores the critical need for COVID-19 vaccination to prevent infection across different vaccine types. This commonality may help mitigate the confounding effect and facilitate more accurate comparisons among the vaccines. Moreover, the data provided clear insights into how the risk of SARS-CoV-2 infection has evolved with the implementation of vaccines and the practice of island-confined quarantine, as well as the similarities in detection times between asymptomatic and symptomatic cases. We recommend that researchers and policymakers consider incorporating island quarantine into future strategies for reintroducing travel to local communities during new waves of SARS-CoV-2 or other global pandemics. We anticipate that the implementation of combined preventative measures, both established and novel, will enable safe travel, ultimately reducing the risk of local transmission while balancing the benefits of reopening the travel and tourism sectors.

Notes

Conflict of Interest

The authors have no conflicts of interest associated with the material presented in this paper.

Funding

None.

Author Contributions

Conceptualization: Sirijantradilok T, Rojanaworarit C, Kallayanasit W, Yodkhunnathum P, Khamsakhon S, Suerungruang S. Data curation: Sirijantradilok T, Photisan N. Formal analysis: Rojanaworarit C, Photisan N. Funding acquisition: None. Methodology: Sirijantradilok T, Rojanaworarit C, Andrade I, Photisan N. Project administration: Sirijantradilok T, Rojanaworarit C. Writing – original draft: Andrade I, Sirijantradilok T, Rojanaworarit C. Writing – review & editing: Sirijantradilok T, Rojanaworarit C, Andrade I, Kallayanasit W, Yodkhunnathum P, Khamsakhon S, Suerungruang S, Photisan N.

Acknowledgements

Figure. 1.

Cumulative probability of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection among cases detected during quarantine. (A) Cumulative probability of SARS-CoV-2 infection among all infected travelers. (B) Cumulative probability of SARS-CoV-2 infection stratified by presentation of symptoms. (C) Cumulative probability of SARS-CoV-2 infection among infected travelers aged ≥18. (D) Cumulative probability of SARS-CoV-2 infection among infected travelers aged ≥18 and stratified by presentation of symptoms. Values are presented as median detection time, day (95% confidence interval).

REFERENCES

• 1. Movono A, Scheyvens R, Auckram S. Silver linings around dark clouds: tourism, Covid-19 and a return to traditional values, villages and the vanua. Asia Pac Viewp 2022;10.1111/apv.12340. https://doi.org/10.1111/apv.12340ArticlePMC
• 2. Kuhakan J, Thepgumpanat P. Thailand’s “paradise island” Phuket reopens to tourists. Reuters; 2021 Jul 1 [cited 2024 Feb 28]. Available from: https://www.reuters.com/article/idUSKCN2E739D/
• 3. Koelle K, Martin MA, Antia R, Lopman B, Dean NE. The changing epidemiology of SARS-CoV-2. Science 2022;375(6585):1116-1121. https://doi.org/10.1126/science.abm4915ArticlePubMedPMC
• 4. Thaicharoen S, Meunrat S, Leng-Ee W, Koyadun S, Ronnasiri N, Iamsirithaworn S, et al. How Thailand’s tourism industry coped with COVID-19 pandemics: a lesson from the pilot Phuket Tourism Sandbox project. J Travel Med 2023;30(5):taac151. https://doi.org/10.1093/jtm/taac151ArticlePubMed
• 5. Fan YJ, Chan KH, Hung IF. Safety and efficacy of COVID-19 vaccines: a systematic review and meta-analysis of different vaccines at phase 3. Vaccines (Basel) 2021;9(9):989. https://doi.org/10.3390/vaccines9090989ArticlePubMedPMC
• 6. Tregoning JS, Flight KE, Higham SL, Wang Z, Pierce BF. Progress of the COVID-19 vaccine effort: viruses, vaccines and variants versus efficacy, effectiveness and escape. Nat Rev Immunol 2021;21(10):626-636. https://doi.org/10.1038/s41577-021-00592-1ArticlePubMedPMC
• 7. Nordström P, Ballin M, Nordström A. Risk of infection, hospitalisation, and death up to 9 months after a second dose of COVID-19 vaccine: a retrospective, total population cohort study in Sweden. Lancet 2022;399(10327):814-823. https://doi.org/10.1016/S0140-6736(22)00089-7ArticlePubMedPMC
• 8. Bergwerk M, Gonen T, Lustig Y, Amit S, Lipsitch M, Cohen C, et al. Covid-19 breakthrough infections in vaccinated health care workers. N Engl J Med 2021;385(16):1474-1484. https://doi.org/10.1056/NEJMoa2109072ArticlePubMed
• 9. Grépin KA, Aston J, Burns J. Effectiveness of international border control measures during the COVID-19 pandemic: a narrative synthesis of published systematic reviews. Philos Trans A Math Phys Eng Sci 2023;381(2257):20230134. https://doi.org/10.1098/rsta.2023.0134ArticlePubMedPMC
• 10. Zhong L. A dynamic pandemic model evaluating reopening strategies amid COVID-19. PLoS One 2021;16(3):e0248302. https://doi.org/10.1371/journal.pone.0248302ArticlePubMedPMC
• 11. Peeling RW, Heymann DL, Teo YY, Garcia PJ. Diagnostics for COVID-19: moving from pandemic response to control. Lancet 2022;399(10326):757-768. https://doi.org/10.1016/S0140-6736(21)02346-1ArticlePubMed
• 12. Rosca EC, Heneghan C, Spencer EA, Brassey J, Plüddemann A, Onakpoya IJ, et al. Transmission of SARS-CoV-2 associated with aircraft travel: a systematic review. J Travel Med 2021;28(7):taab133. https://doi.org/10.1093/jtm/taab133ArticlePubMed
• 13. Jin L, Li Z, Zhang X, Li J, Zhu F. CoronaVac: a review of efficacy, safety, and immunogenicity of the inactivated vaccine against SARS-CoV-2. Hum Vaccin Immunother 2022;18(6):2096970. https://doi.org/10.1080/21645515.2022.2096970ArticlePubMedPMC
• 14. Rahman MM, Masum MH, Wajed S, Talukder A. A comprehensive review on COVID-19 vaccines: development, effectiveness, adverse effects, distribution and challenges. Virusdisease 2022;33(1):1-22. https://doi.org/10.1007/s13337-022-00755-1ArticlePubMedPMC
• 15. Pordanjani SR, Pordanjani AR, Askarpour H, Arjmand M, Babakhanian M, Amiri M, et al. A comprehensive review on various aspects of SARS-CoV-2 (COVID-19) vaccines. Int J Prev Med 2022;13: 151. https://doi.org/10.4103/ijpvm.ijpvm_513_21ArticlePubMedPMC
• 16. Zuin M, Gentili V, Cervellati C, Rizzo R, Zuliani G. Viral load difference between symptomatic and asymptomatic COVID-19 patients: systematic review and meta-analysis. Infect Dis Rep 2021;13(3):645-653. https://doi.org/10.3390/idr13030061ArticlePubMedPMC
• 17. Shi N, Huang J, Ai J, Wang Q, Cui T, Yang L, et al. Transmissibility and pathogenicity of the severe acute respiratory syndrome coronavirus 2: a systematic review and meta-analysis of secondary attack rate and asymptomatic infection. J Infect Public Health 2022;15(3):297-306. https://doi.org/10.1016/j.jiph.2022.01.015ArticlePMC
• 18. Ma Q, Liu J, Liu Q, Kang L, Liu R, Jing W, et al. Global percentage of asymptomatic SARS-CoV-2 infections among the tested population and individuals with confirmed COVID-19 diagnosis: a systematic review and meta-analysis. JAMA Netw Open 2021;4(12):e2137257. https://doi.org/10.1001/jamanetworkopen.2021.37257ArticlePubMedPMC
• 19. Dickens BL, Koo JR, Lim JT, Park M, Sun H, Sun Y, et al. Determining quarantine length and testing frequency for international border opening during the COVID-19 pandemic. J Travel Med 2021;28(7):taab088. https://doi.org/10.1093/jtm/taab088ArticlePubMedPMC

Figure & Data

References

Citations

Citations to this article as recorded by