Skip to main content

Measures to contain the COVID-19 outbreak in migrant worker dormitories

  1. The vast majority of the COVID-19 cases in Singapore occurred in migrant worker dormitories, where the SARS-CoV-2 virus spread quickly due to their communal living arrangements. Earlier health and safe distancing measures turned out to be inadequate, given the highly infectious nature of the novel virus, and because pre-symptomatic and asymptomatic transmission was also taking place.
  2. In all, 54,505 out of the 58,320 who tested positive in Singapore for COVID-19 via a Polymerase Chain Reaction (PCR) test were migrant workers living in dormitories. This is out of a total of over 320,000 migrant workers who live in dormitories1. At the peak of the outbreak in April, more than 1,000 new cases a day were being detected in the dormitories.2
  3. We acted swiftly and decisively to stabilise the situation in the dormitories. Working closely with dormitory operators, employers, the medical community, NGOs and other community groups, we contained the outbreak and cleared the dormitories of the virus. By August, all migrant workers living in dormitories had been tested for COVID-19 at least once. Almost all of them have since also been cleared to resume work safely.
  4. Despite the scale of the outbreak in the dormitories, the morbidity and mortality rate among our migrant workers has been very low. There were 25 COVID-19-related ICU admissions amongst migrant workers living in dormitories and only two deaths due to COVID-19, including one of those who had been admitted to the ICU.

    Containment and testing strategy

    April – May 2020: Containing the outbreak
  5. Initially, our top priority was to contain the spread of COVID-19 in the dormitories. We placed all dormitories under isolation by mid-April, in line with the national Circuit Breaker measures. Throughout this phase, we stepped up our capacity to detect and isolate cases in the dormitories, and to care for the health of all our migrant workers.
  6. We stepped up testing to help us assess the situation in each dormitory. By the end of April, 1 in every 15 workers in dormitories had been PCR tested, far higher than the testing rates in other countries. Even so, due to the large numbers, we were not able to test all the migrant workers in the dormitories at once. We prioritised PCR testing for migrant workers in essential services so they could be cleared to continue working safely during the Circuit Breaker period.
  7. At the same time, any worker who reported sick or showed symptoms of acute respiratory infection during this period was isolated, and given medical care regardless of whether he had received a PCR test.
  8. To complement the testing and identify potential new cases quickly, thermometers and more than 25,000 oximeters were distributed. Every migrant worker living in the dormitories checked and reported his temperature and oximeter readings twice a day. Medical teams monitored these reports, and intervened early when they found abnormalities.
  9. Medical support was set up at the dormitories by the end of April to care for those who were unwell, and monitor the health of those who were well. These included:
    1. Medical posts staffed by teams of doctors, nurses and technicians at all 43 Purpose-Built Dormitories (PBDs).
    2. 8 medical posts, complemented by roving medical teams, that served all non-PBDs (e.g. Factory-Converted Dormitories and Construction Temporary Quarters).
    Migrant workers living in the community also had access to the nation-wide network of more than 900 Public Health Preparedness Clinics and Polyclinics.

    June – August 2020: Clearing the dormitories of COVID-19 infection
  10. By June, our focus shifted to clearing the dormitories of COVID-19, so the migrant workers could resume work safely. Testing was key.
  11. Our measures reflected the growing scientific understanding of COVID-19. We were discovering that a significant number of infected persons had no or mild symptoms, but could nevertheless spread COVID-19 to others. We also found that infected persons who had recovered could still continue shedding non-infectious viral fragments for several months.
  12. Our testing strategy therefore had to help us distinguish who had never been infected; who had had an old infection but had since recovered; and who was currently infected and still harbouring the virus, with or without symptoms. This would help us separate those who had never been infected or had had the virus earlier but were no longer infectious, from those who were currently infected or potentially harbouring the virus even without symptoms.
  13. Over time, more testing methods were becoming available, including serology tests. While PCR tests are used to diagnose current or new infections, serology tests identify those who had been infected in the past, by detecting the presence of COVID-19 antibodies in blood samples.
  14. A detailed comparison of various testing methods used is in Appendix I.
  15. In June, we took the extraordinary decision to systematically test all migrant workers living in dormitories, symptomatic or otherwise. In dormitories with a high incidence of infections, we applied a differentiated approach. All workers in these dormitories were screened with a combination of serology as well as PCR tests.
    1. If they tested serology-positive, it meant that they had been infected earlier. These workers underwent a seven-day period of isolation, at the end of which we could be confident that they were no longer infectious, and did not need to be tested further.
    2. Those who tested serology-negative were separately isolated for a longer period of 14 days, in case they were incubating the virus despite being asymptomatic. They were given a PCR test at the end of the isolation period to confirm that they were free from infection.
  16. This combination of PCR and serology tests enabled us avoid isolating or quarantining recovered workers for prolonged periods, so that they could return to their dormitories or workplaces sooner.
  17. By August, all migrant workers had been tested at least once for COVID-19. We were confident that the outbreak had been contained. By early November, more than 98 per cent of our migrant workers living in dormitories were cleared to resume work.

    Overall case count and prevalence rate
  18. Our comprehensive testing strategy has provided us a more complete picture of the prevalence of COVID-19 in our migrant worker dormitories.

    PCR and serology test results of migrant workers
  19. As at 13 December 2020:
    1. 54,505 dormitory residents have tested positive using the PCR test.
    2. Another 98,289 have tested serology-positive though they did not have a positive PCR test.
    The data is summarised in Table 1.

    Table 1: Dormitory-dwelling migrant workers and their test results (as at 13 December 2020)
    A. PCR-positive* 54,505
    B. Serology-positive only 98,289
    Ratio of PCR-positive to serology-positive only (A:B) 1:1.8
    Prevalence rate, i.e. (A+B)/(total dormitory dwellers) 47% (of 323,000 dormitory dwellers)
    * Including some who have tested both PCR-positive and serology-positive.
  20. Among the migrant workers who tested PCR-positive or serology-positive, the vast majority were asymptomatic or had very mild symptoms. Only about 1 in 5 of migrant workers living in PBDs presented with symptoms, with the remaining 4 in 5 displaying very mild or no symptoms.3
  21. Including the serology test results, the prevalence rate of COVID-19 in the dormitories is currently 47 per cent. For every COVID-19 infection in the dormitories detected through PCR testing, another 1.8 cases were untested and undetected at the time, and were identified subsequently only through serology testing. This is not surprising as many migrant workers did not have any symptoms, and thus would not have sought treatment and received a PCR test in the process. Based on sample population, the ratio of PCR-positive to serology-positive among migrant workers in dormitories is comparable to the ratio of 1:1.8 for the whole of South Korea, and lower than 1:4 in Spain, and 1:15 in France4. Our low ratio reflects the extensive PCR testing we carried out in the dormitories.

    Reporting of cases
  22. Singapore reports COVID-19 cases in accordance with international practice. Every case who tested positive through a PCR test and was assessed to be an acute infection is reported and included in our case count. We follow WHO’s criterion that only positive results from confirmatory tests (i.e. PCR tests) are included in the case count. This ensures consistency in reporting cases across countries.
  23. A different approach is taken for reporting the results of serology tests. As these tests identify past cases of infection, they are used to aid epidemiological investigations and for retrospective assessment of the overall prevalence of infections within a population. Therefore, serology test outcomes are aggregated and presented as an estimate of prevalence in the population, separate from the daily case counts.
  24. Most countries only do serology testing on a sampling basis, to estimate the prevalence of infections in a population. But Singapore went further to do serology tests on our entire population of migrant workers living in the dormitories. This was a unique aspect of our efforts to clear the dormitories of COVID-19.
  25. As at 13 December 2020, we are still in the process of completing serology tests for 65,000 or so migrant workers living in dormitories who had not taken a serology test before. This will give us the full picture of the infection prevalence among our migrant workers.

    Phase Three
  26. Having tested all migrant workers living in the dormitories, we will continue with Rostered Routine Testing (RRT) of every worker who may still be susceptible to infection, in order to detect and contain new infections rapidly. All such workers living in dormitories, and those who work in the construction, marine and process (CMP) sectors, have been undergoing RRT once every 14 days.
  27. After several RRT cycles, the number of new infections have remained very low. Since October, no new cases were detected in the dormitories on many days.
  28. Having brought the outbreak in the dormitories under control, and instituted Safe Living and Safe Working measures at all dormitories and worksites, we will progressively ease the restrictions on migrant workers.
    1. With the transition into Phase Three, we are preparing to return migrant workers to the community in a controlled manner, with strict measures in place.
    2. We will start a pilot scheme in the first quarter of 2021 to allow migrant workers in some dormitories to access the community once a month, subject to compliance with RRT, wearing of contact-tracing devices and safe living measures.
  29. We will continue to keep our migrant workers safe, and to ensure that we detect and contain any new cases or clusters quickly.
    1. Meanwhile, we are monitoring the earliest cohort of migrant workers who have recovered from COVID-19 and are currently exempt from RRT. We are studying how their antibodies change over time. We will resume RRT for these workers if we detect their antibodies starting to fade, or if there is evidence of re-infection among them.
    2. We will continue our multi-layered strategy of aggressive routine testing using both PCR and antigen rapid testing, accompanied by isolation strategies.
    3. By end-December, we will complete distributing contact-tracing devices to more than 450,000 workers living in dormitories, or working in the construction, marine and process sectors. These devices will improve our ability to isolate and ringfence potential cases once they are detected.
    4. We are also building new dormitories with improved safety standards to minimise the risk of a resurgence of COVID-19 among migrant workers, and prevent new public health threats.
  30. The Government will continue to work with our partners and the community to ensure a safe living and working environment for our migrant workers, as we move towards Phase Three.

Appendix I: Comparison of PCR, Serology and Antigen Rapid Tests

Aspects Polymerase Chain Reaction (PCR) Antigen Rapid Test Serology
Scientific basis Detection of SARS-CoV-2 viral sequences by nucleic acid amplification tests in respiratory tract specimens. Detection of SARS-CoV-2 viral proteins (antigens) in respiratory tract specimens. Detection of antibodies produced by the human body in response to infection with the SARS-CoV-2.
Aim of test Diagnosis of SARS-CoV-2 infection Diagnosis of SARS-CoV-2 infection Check for previous SARS-CoV-2 infection as part of epidemiological investigations
Sample type Nasopharyngeal (NP), Oropharyngeal (OP), Midturbinate (MT) Nasopharyngeal (NP), Oropharyngeal (OP), Midturbinate (MT) Venous blood, fingerprick, for point-of-care testing (POCT)
Turnaround time
  • 4-6 hours per run (lab)
  • ~1 hour per run in POCT PCR or PCR that does not require separate extraction step (Cepheid)
  • 15-30 minutes per run
  • All kits are POCT, no analyser machines required for most kits
  • Lab-based: 40-120 minutes per test; 100-200 tests per hour
  • POCT: ~30 minutes per test
Clinical performance
(Sensitivity / Specificity)
>99.5%/100% Variable sensitivity but generally higher sensitivity for individuals with high viral load.

 

 

WHO criteria for antigen-detecting rapid diagnostic tests: >80%/97%
Performance of serologic assays varies widely in different testing groups (such as disease severity, age), timing of testing and the target viral protein.

 

 

In general, lab-based tests using venous blood has higher sensitivity/specificity than POCT.
Examples of use cases and role in overall testing strategy
  • Symptomatic individuals
  • Stay-Home Notice (SHN) exit swab
  • Quarantine Order (QO) entry and exit swab
  • Rostered routine testing
  • Screening for pre-event testing, rostered routine testing
  • Differentiate between acute and old infections in cases that test positive for COVID-19
Limitations
  • Unable to differentiate between acute and old infections due to persistent shedding of viral fragments among recovered individuals.
  • Potentially high false negative rate in individuals with low viral load.
  • Higher false positive rate than PCR tests.
  • Unable to rule out acute/early infection if serology-negative.

FOOTNOTE

  1. Includes all Purpose-Built Dormitories (PBDs) and non-PBDs (e.g. Factory-Converted Dormitories, Construction Temporary Quarters, and temporary living quarters).
  2. Our highest daily COVID-19 case count in the dormitories was recorded on 20 April 2020, with a total of 1,397 new cases detected among dormitory residents that day.
  3. Based on a study of migrant workers living in PBDs who had tested positive by PCR or serology tests as of 25 July 2020.
  4. Lai et al. Population-based seroprevalence surveys of anti-SARS-CoV-2 antibody: An up-to-date review. Int Journal of Infectious Diseases 101 (2020) 314-322.