nursing
SITSO420
EBPProject-Article1.pdf
RESEARCH ARTICLE
The accuracy of healthcare worker versus self
collected (2-in-1) Oropharyngeal and Bilateral
Mid-Turbinate (OPMT) swabs and saliva
samples for SARS-CoV-2
Seow Yen TanID 1☯*, Hong Liang Tey2☯, Ernest Tian Hong Lim3☯, Song Tar Toh4☯, Yiong
Huak Chan 5 , Pei Ting Tan
6 , Sing Ai Lee
7 , Cheryl Xiaotong Tan
8 , Gerald Choon Huat Koh
9‡ ,
Thean Yen Tan 10‡
, Chuin Siau 11‡
1 Department of Infectious Diseases, Changi General Hospital, Singapore, Singapore, 2 Department of
Dermatology, National Skin Centre, Singapore, Singapore, 3 Emergency Department, Woodlands Health
Campus, Singapore, Singapore, 4 Department of Otorhinolaryngology- Head and Neck Surgery, Singapore
General Hospital, Singapore, Singapore, 5 Biostatistics Unit, Yong Loo Lin School of Medicine, Singapore,
Singapore, 6 Clinical Trials and Research Unit, Changi General Hospital, Singapore, Singapore, 7 Sheares
Healthcare Group Pte Ltd, Singapore, Singapore, 8 Temasek International Pte Ltd, Singapore, Singapore,
9 MOH Office for Healthcare Transformation, Singapore, Singapore, 10 Department of Laboratory Medicine,
Changi General Hospital, Singapore, Singapore, 11 Department of Respiratory & Critical Care Medicine,
Changi General Hospital, Singapore, Singapore
☯ These authors contributed equally to this work. ‡ These authors are joint senior authors on this work.
Abstract
Background
Self-sampling for SARS-CoV-2 would significantly raise testing capacity and reduce healthcare
worker (HCW) exposure to infectious droplets personal, and protective equipment (PPE) use.
Methods
We conducted a diagnostic accuracy study where subjects with a confirmed diagnosis of
COVID-19 (n = 401) and healthy volunteers (n = 100) were asked to self-swab from their
oropharynx and mid-turbinate (OPMT), and self-collect saliva. The results of these samples
were compared to an OPMT performed by a HCW in the same patient at the same session.
Results
In subjects confirmed to have COVID-19, the sensitivities of the HCW-swab, self-swab,
saliva, and combined self-swab plus saliva samples were 82.8%, 75.1%, 74.3% and 86.5%
respectively. All samples obtained from healthy volunteers were tested negative. Compared
to HCW-swab, the sensitivities of a self-swab sample and saliva sample were inferior by
8.7% (95%CI: 2.4% to 15.0%, p = 0.006) and 9.5% (95%CI: 3.1% to 15.8%, p = 0.003)
respectively. The combined detection rate of self-swab and saliva had a sensitivity of 2.7%
(95%CI: -2.6% to 8.0%, p = 0.321). The sensitivity of both the self-collection methods are
higher when the Ct value of the HCW swab is less than 30. The specificity of both the self-
swab and saliva testing was 100% (95% CI 96.4% to 100%).
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 1 / 11
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Tan SY, Tey HL, Lim ETH, Toh ST, Chan
YH, Tan PT, et al. (2020) The accuracy of
healthcare worker versus self collected (2-in-1)
Oropharyngeal and Bilateral Mid-Turbinate (OPMT)
swabs and saliva samples for SARS-CoV-2. PLoS
ONE 15(12): e0244417. https://doi.org/10.1371/
journal.pone.0244417
Editor: Dong-Yan Jin, University of Hong Kong,
HONG KONG
Received: September 10, 2020
Accepted: December 9, 2020
Published: December 16, 2020
Peer Review History: PLOS recognizes the
benefits of transparency in the peer review
process; therefore, we enable the publication of
all of the content of peer review and author
responses alongside final, published articles. The
editorial history of this article is available here:
https://doi.org/10.1371/journal.pone.0244417
Copyright: © 2020 Tan et al. This is an open access article distributed under the terms of the Creative
Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in
any medium, provided the original author and
source are credited.
Data Availability Statement: All relevant data are
within the manuscript and its Supporting
Information files
Conclusion
Our study provides evidence that sensitivities of self-collected OPMT swab and saliva sam-
ples were inferior to a HCW swab, but they could still be useful testing tools in the appropri-
ate clinical settings.
Introduction
The current “gold standard” for testing for SARS-CoV-2 requires health care workers to collect
a nasopharyngeal (NP) sample from a patient. NP sampling is very uncomfortable for the
patient and requires deployment of trained personnel and use of personal protective equip-
ment (PPE) which are in limited supply.
A prior study has shown that a combination of oropharyngeal and anterior nares swabs is
equivalent in sensitivity to an NP swab in 190 ambulatory symptomatic patients [1]. In another
study on 236 ambulatory subjects, the performance of self-collected nasal and throat swabs is
at least equivalent to that of health worker collected swabs for the detection of SARS-CoV-2
and other respiratory viruses [2].
The international community is actively searching for an even less invasive means of sample
collection: saliva. In a recent study by Yale University on 29 subjects [3], it was suggested that a
large volume sample of saliva collected from COVID-19 inpatients can be more sensitive than
NP swabs for SARS-CoV-2 detection, and saliva samples had significantly higher COVID-19
viral titres than NP swabs (p = 0.001). Furthermore, the same study showed that sensitivity of
COVID-19 in saliva was more consistent throughout extended hospitalization compared to
NP swabs.
In addition, there are a number of studies done on saliva testing for COVID-19 which have
shown promising results, reporting 91.7%, and 100% positivity in saliva samples of COVID-19
patients [4, 5]. Iwasaki et al found an overall concordance rate of 97.4% for COVID-19 detec-
tion with a strong concordance between NP swabs and saliva sampling (κ = 0.874) among 66 COVID-19 negative and 10 COVID-19 positive subjects [6]. Furthermore, a study done by To
et al. showed that viral RNA could still be detected in saliva samples in a third of their twenty-
three patients 20 days or longer after symptoms onset despite the development of COVID-19
antibodies [7]. A meta-analysis conducted on 26 saliva studies also showed a positive detection
rate of 91%, comparable to the detection rate of 98% from nasopharyngeal swabs [8]. All these
studies had small sample sizes (all <30 COVID-positive subjects) and only one study also sam-
pled COVID-negative subjects.
It is still currently unknown whether a self-collected combined Oropharyngeal and Bilateral
Mid-Turbinate (OPMT) sample, or a self-collected saliva sample is equivalent to a swab done
by a health care worker (HCW). If the self-collection of samples is proven to be a reliable alter-
native to a HCW swab, it would reduce the reliance of trained personnel to collect samples and
enable a rapid increase in testing capacity. It would also reduce greatly the biosafety risk that is
posed to HCWs and help with PPE conservation efforts.
Materials and methods
Study design and trial oversight
This was a prospective study involving 401 subjects who were previously tested positive for
COVID-19 by RT-PCR, and 100 healthy volunteers. This study was approved by the
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 2 / 11
Funding: This study was funded by Sheares
Healthcare Group Pte Ltd. The funder provided
support in the form of salaries for authors SAL and
CXT, but did not have any additional role in the
study design, data collection and analysis, decision
to publish, or preparation of the manuscript. The
specific roles of these authors are articulated in the
‘author contributions’ section. Besides that, author
CXT is employed by Temasek International Pte Ltd,
and was acting on behalf of Sheares Healthcare
Group Pte Ltd for the study. Temasek International
Pte Ltd did not have any additional role in the study
design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing interests: Author CXT is employed by
Temasek International Pte Ltd, and was acting on
behalf of Sheares Healthcare Group Pte Ltd for the
study. This commercial affiliation does not alter our
adherence to PLOS ONE policies on sharing data
and materials.
SingHealth Centralised Institutional Review Board. Written informed consent was obtained
from the subjects.
Participants
The first group consisted of patients who were confirmed to have COVID-19, and who were
cared for in either a hospital (Changi General Hospital), or a community care facility (Com-
munity Care Facility @ EXPO). Diagnosis of COVID-19 was confirmed via a positive RT-PCR
from a nasopharyngeal swab. The subjects in this group were recruited within 3 days of admis-
sion to the study site and they were recruited from 31 May 2020 to 10 June 2020. The patients
who were eligible were approached directly at the study site, and were invited to participate in
the study, and the study procedures were carried out on the same day. Recruitment was carried
out until the target sample size was achieved. At the time of the study, the majority of COVID-
19 cases belong to the migrant worker population, which primarily consisted of healthy young
male adults, mainly from Bangladesh and India. Hence, this group of subjects is not represen-
tative of the general population in Singapore.
Inclusion criteria applicable to this group include:
• Male and female patients, � 21 years-old
• Tested positive for COVID-19
• Admitted to study site within the previous 3 days
• Ability to provide informed consent
• Compliance with all aspects of study protocol, methods and provision of samples
• Ability to read and understand English
Exclusion criteria applicable to this group include:
• Nosebleeds in past 24 hours
• Previous nasal surgery in past 4 weeks
• Acute facial trauma within 8 weeks
• Unable to demonstrate understanding of study and instructions
• Experienced severe adverse reactions on prior nose and/or throat swabs
• Not willing to have all 3 samples collected
The second group comprised 100 healthy volunteers who were asymptomatic and well on
the day of the study, with no recent COVID-19 exposure. This was done on 18 and 19 July
2020. The study subjects were recruited via an open advertisement.
Inclusion Criteria for this group include:
• Males and females, � 21 years-old
• Ability to provide informed consent
• Capable of understanding and complying with the requirements of the study
• Ability to read and understand English
Exclusion Criteria applicable to this group were:
• Displaying symptoms of an acute respiratory infection
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 3 / 11
• Known close contact with an individual diagnosed with COVID-19 within the last 3 months
• Previously diagnosed with COVID-19
• Nosebleeds in past 24 hours
• Previous nasal surgery in past 4 weeks
• Acute facial trauma within 8 weeks
• Unable to demonstrate understanding of study and instructions
• Experienced severe adverse reactions on prior nose and/or throat swabs
• Not willing to have all 3 samples collected
Test procedures
Study subjects underwent three sequential test sample collection procedures within one study
visit in the following order:
1. Each subject self-collected a sample combining OP and bilateral MT swabs using a single
swab stick;
2. A trained healthcare worker then collected a combined OP and bilateral MT swab using
another single swab stick;
3. The subject then self-collected a saliva sample.
Study subjects were shown instructional videos for both the OPMT self-swab and saliva col-
lection prior to commencing the test procedures. Study team members were present on site to
observe and supervise the self-collection process. Posterior oropharyngeal saliva, commonly
described as deep throat saliva was collected for this study. Synthetic fibre swabs were used for
collection of the OP and MT samples by both subject and healthcare worker, and immediately
placed in universal transport medium (UTM), while saliva samples were collected using the
SAFER-Sample™ (by Lucence Diagnostics). All samples were double bagged and stored at air- conditioned room temperature in a chiller bag and transported to assigned laboratory on the
same day. Upon arrival in the laboratory, they were stored at 2˚C to 8˚C. All samples were pro-
cessed with 24 hours of sample collection.
Nucleic acid extraction was performed using PerkinElmer Nucleic Acid Extraction Kits
(KN0212) on the Pre-Nat II Automated Workstation (PerkinElmer1, United States), Extrac-
tion of swab samples followed the indicated protocol for oropharyngeal swabs, while extraction
of saliva samples followed a protocol consisting of pre-liquefaction with dithiothreitol (protocol
attached in S1 File). Reverse transcription polymerase chain reaction (RT-PCR) was performed
on the Quantstudio TM
5 Real Time PCR system (Thermo Fisher, United Kingdom) using the
PerkinElmer1 SARS-CoV-2 Real-time RT-PCR Assay. The targets were the ‘N’ gene and
‘ORF1ab’ gene. There is an internal control target that is present in every RT-PCR reaction. The
cycle threshold (Ct) values of the ‘N’ gene were used in the analysis involving Ct values.
Outcomes
The primary objective of the study was to evaluate the accuracy of self-collected (2-in-1) OPMT
swabs and self-collected saliva samples for SARS-CoV-2 versus that of HCW-collected (2-in-1)
MT and OP swabs. The secondary objective was to evaluate the correlation of PCR Ct values of
self-collected saliva samples and swabs with comparator healthcare worker-collected swabs.
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 4 / 11
Sample size
Firstly, we postulated that OPMT self-swabbing was as accurate as HCW-obtained swabs. Pos-
tulating a 100% accuracy, 400 subjects will be required to achieve a lower 95% confidence
interval 99.1% (which gives a less than 1% error rate). With the computed sample size of 400
subjects, a non-inferiority could be achieved with at most a 7% difference for OPMT self-swab-
bing compared to the HCW-obtained swabs. If the study included only subjects who were
diagnosed with COVID-19, all positive results would be regarded as true positives. Hence, to
address that gap in the form of specificity of self-swabs and saliva testing in the diagnosis of
COVID-19, a further study on 100 healthy subjects was conducted. The hypothesis was that
with 100% accuracy, the error rate for a false negative was 3.6%.
Statistical analysis
All analyses were performed using SPSS 25.0 with statistical significance set at p < 0.05.
The estimates for the positivity results of the 3 methods were presented as numbers and
percentages. The differences with 95% confidence interval (CI) between self-collection meth-
ods and HCW-obtained swabs to assess for non-inferiority was calculated. Sensitivity and
specificity of the two self-collection methods were compared with HCW-obtained swabs and
results were stratified by Ct values. Spearman’s test was used to assess the correlation of the
PCR Ct values across the 3 groups.
Results
A total of 401 COVID-19 positive and 100 COVID-19 negative subjects were recruited. Of the
401 COVID-19 positive subjects, 23 were recruited from Changi General Hospital, and 378
were recruited from the community care facility @ Expo. The symptomatic COVID-19 posi-
tive subjects that were recruited were well patients, whose clinical presentation was that of an
upper respiratory tract infection. None of the subjects required oxygen supplementation.
Only the demographic data of subjects from Changi General Hospital was known. The full
demographic data of the subjects that were admitted to the community care facility could not
be made available to us due to prevailing regulations of the study site during the period when
the study was conducted, hence we do not have the data of the age of the subjects that were
admitted to the community care facility. However, we were able to surmise that the age range
of patients admitted to the community care facility was 21 to 45, due to the admission criteria
to the facility, and the inclusion criteria for the study. A summary of the profile of recruited
subjects are listed in Tables 1 and 2 below.
All subjects went through the test procedures—500 participants (400 COVID-19 positive,
100 COVID-19 negative) were able to provide all 3 samples, and one subject was unable to
provide a saliva sample despite a prolonged attempt. All participants tolerated the test proce-
dures well and did not experience any adverse events.
In the group of subjects who were COVID-19 positive, twenty-seven (6.7%) patients were
tested negative across all 3 samples. This may be explained by the fact that they are recovering
and viral shedding may have ceased at point of testing. Forty-two (10.5%) subjects reported
�1 symptom of acute respiratory infection (ARI) (e.g. fever, cough, rhinorrhoea, sore throat,
malaise) on the day of study recruitment while 371 (92.5%) subjects reported being within 7
days from onset of COVID-19 illness.
The detection rates of the HCW swab, self-swab, saliva, and combined self-swab plus saliva
samples were 82.8%, 75.1%, 74.3% and 86.5% respectively (Table 3). Compared to HCW-
swabs, the detection rate was lower for self-swab by 8.7% (95% confidence interval, CI = 2.4%
to 15.0%, p = 0.006) and for saliva samples by 9.5% (95%CI = 3.1% to 15.8%, p = 0.003). When
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 5 / 11
the results of both the self-swab and saliva testing were combined, the detection rate was
higher by 2.7% (95%CI = -2.6% to 8.0%, p = 0.321) but this was not statistically significant.
The sensitivities of the self-swab, saliva and combined self-swab plus saliva testing, when
compared to the HCW swab were 83.6%, 80.6% and 92.3% respectively. Table 4 shows the con-
tingency tables comparing the HCW swab vs self-swab; HCW swab vs saliva, and HCW vs
combined self-swab plus saliva respectively.
Using the Ct values (‘N’ gene) of HCW swabs as reference, 3 categories of Ct values (i.e. <25,
25–30 and >30) were studied. It was observed that the sensitivity of self-swab (Table 5) and
saliva testing (Table 6) performed better at the lower Ct values, suggesting that the sensitivity of
self-collection methods approaches to that of HCW swab, when the viral load was higher.
There was a good correlation of PCR Ct values between self-swab and HCW swab (r = 0.825,
p<0.001) but moderate correlation between saliva samples and HCW swab (r = 0.528, p<0.001).
The self-swab has a better agreement with the HCW swab. Using Wilcoxon Signed Rank Test,
the difference in CT values between self-swab and HCW swab is statistically significant, where
p = 0.026. Similarly for the saliva and HCW swab, where p<0.001. Figs 1 and 2 show the scatter-
plot of the correlation between the Ct values of the HCW swab and the self-swab as well as the
saliva respectively. Table 7 shows the distribution of the Ct values of the 3 tests.
One hundred healthy volunteers were recruited, and all of them were able to provide the 3
required samples. All the samples obtained from the healthy volunteers were tested negative
for SARS-CoV2. This implies that the specificity of the self-swab and saliva sampling was
100% (95% CI 96.4% to 100%) with an error rate of 3.6% for having a false negative.
Table 2. Profile of COVID-19 negative subjects (N = 100).
Gender N (%)
Female 51 (51.0)
Male 49 (49.0)
Age (years)
Mean (SD) 38.24 (10.16)
Range 22–70
https://doi.org/10.1371/journal.pone.0244417.t002
Table 1. Profile of COVID-19 positive subjects (N = 401).
N (%)
Age, years
Min–max 21–54
Mean (SD)� 37.26 (6.4)
Male 401 (100.0)
Presence of symptoms on study day
No 359 (89.5)
Yes 42 (10.5)
Duration between illness onset to study day, days
Range 1–25
Mean (SD) 5.65 (2.1)
Days from first positive swab to study day, days
Range 1–20
Mean (SD) 5.48 (1.8)
� Calculation based on the known age of 24 subjects.
https://doi.org/10.1371/journal.pone.0244417.t001
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 6 / 11
Discussion
This study shows that the sensitivity of a self-swab or saliva sample on its own is lower than
HCW swab. However, the sensitivity of a combined self-swab and saliva collection is equiv-
alent to that of a HCW swab. Another significant finding is that the self-swab and saliva
samples have a higher sensitivity when the viral load is higher, and this generally occurs
during the early stages of COVID-19. The sensitivity of both self-swab and saliva testing
drops significantly when the Ct values of the HCW swab is more than 30. A study from Sin-
gapore [9] reported that viral cultures were negative from samples with Ct values > = 30
(i.e. when viral load is low), and the SARS-CoV-2 virus often cannot be isolated or cultured
after day 11 of illness [10]. Thus, the results of this study support the use of self-testing
methods as a replacement for a HCW swab in the early phase of COVID-19 illness when
viral loads are high, and the sensitivities of the self-swab and saliva are similar to that of the
HCW swab.
The strength of our study is the large number of subjects confirmed to have COVID-
19. Besides that, the study also included a high proportion of asymptomatic individuals
who were picked up because of Singapore’s proactive mass screening policy. The combi-
nation of self-swab and saliva sampling performed well in these asymptomatic subjects,
implying that the strategy of combined self-testing, has the ability diagnose COVID-19 in
asymptomatic individuals with a sensitivity equivalent to that of a swab by a HCW. The
study results from the healthy volunteers indicate a low false positive rate with self-collec-
tion methods.
These findings, indicate that self-collection methods may be a useful tool for COVID-19
surveillance in the asymptomatic individuals, and in situations where testing capacity needs to
be scaled up rapidly, without a need for large increase of manpower, and without increased
infectious exposure to the swabbing staff. Testing strategies can be tailored based on the target
population and the intended use of the various tests on its own or in combination.
Table 3. Detection rates of various modalities in all subjects.
HCW Swab Self-Swab Saliva Self-Swab + Saliva
Count 336 301 297 347
Percentage 83.8% 75.1% 74.3% 86.5%
95% CI 79.8% - 87.3% 70.1% - 79.2% 69.7% - 78.5% 82.8% - 89.7%
https://doi.org/10.1371/journal.pone.0244417.t003
Table 4. Comparison between HCW swabs and the self-swab/saliva.
HCW Swab
Not detected Detected p value�
Self-swab
Not detected 45 55 <0.001
Detected 20 281 (83.6)
Saliva
Not detected 37 65 <0.001
Detected 27 270
Self-swab plus saliva
Not detected 27 26 0.207
Detected 37 310
� p value was obtained from McNemar test
https://doi.org/10.1371/journal.pone.0244417.t004
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 7 / 11
The way the study findings were presented are unlike most studies involving saliva testing
for COVID-19. This is probably due to the fact that our study is carried out on subjects who
are already known to have COVID-19, unlike most studies which are done in testing centres
where the potential subjects’ results are still unknown. This also meant that the sampling was
done later in the subjects’ trajectory of illness, as they were first tested positive for COVID-19,
then enrolled into the study. The later sampling possibly had a negative impact on the sensitiv-
ity of the saliva [11].
Another key study limitation, is that the demographics of the COVID-19-positive popula-
tion was skewed, consisting solely of male migrant workers, the worst affected group of the
pandemic in Singapore, at the time this study was conducted. Hence the results from this
study might not be applicable to the general population, without the inclusion of paediatric
and elderly population segments. The migrant worker population in this study, which consist
of generally young and healthy males, is also not representative of the demographics of
Singapore.
The addition of the stabilising solution to the deep throat saliva sample, could have also
decreased the yield of the saliva testing. Studies utilising saliva test kits that do not require the
addition of stabilising fluid generally report equivalent sensitivities of the saliva test when com-
pared to a HCW swab [3, 12]. Hence the use of stabilising solution is a key consideration in
future design of saliva test kits.
The study team members observed that, despite clear instructions, many subjects still
needed guidance with the self-collection methods. For the self-swab, the most commonly
encountered scenario was that, the subjects needed guidance in breaking the swab stick. The
saliva collection presented a greater challenge to the subjects. The flow of saliva from the fun-
nel into the collection container was not smooth, and the additional step of adding the stabilis-
ing fluid required prompting. These necessitated the presence of a trained staff to troubleshoot
and ensure that the correct steps are carried out. We believe that these observations are useful
in the re-design of collection containers to enhance results and end users’ acceptability. Both
the self-swab and saliva collection require dexterity and this would limit its applicability in seg-
ments of the population who are not able to do so.
We caution against widespread, unsupervised implementation of self-collection methods.
The reliability and effectiveness of self-collection methods may also be dependent on social
and economic drivers, hence potentially influencing the test performance. For example, indi-
viduals who face a potential loss of income or unemployment if tested positive or travellers
having a test done at immigration clearance may deliberately do a suboptimal self-test to influ-
ence the test outcome.
Table 5. Sensitivity of self-swab, stratified by Ct values of HCW swab.
HCW Swab Ct Number of subjects Sensitivity
<25 60 100% (94.0–100)
25–30 81 96.3% (89.6–99.2)
>30 195 73.3% (66.5–79.4)
https://doi.org/10.1371/journal.pone.0244417.t005
Table 6. Sensitivity of saliva, stratified by Ct values of HCW swab.
HCW Swab Ct Number of subjects Sensitivity
<25 60 96.7% (88.5–99.6)
25–30 81 92.6% (84.6–97.2)
>30 194 70.6% (63.7–76.9)
https://doi.org/10.1371/journal.pone.0244417.t006
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 8 / 11
Hence, it is important to have designated personnel to supervise the self-collection process,
ensuring that the correct test procedures are carried out. These personnel need not be a HCW
and the supervision process will have a lower exposure risk (supervisor can be >1m away from
subject), compared to the HCW-swabbing process where a HCW is <1m away and face-to-
face with the subject.
Conclusion
This study demonstrates that while self-collection methods have a sensitivity of approximately
75%, it is inferior to the rate obtained by the health care worker administered swab (83.8%). The
sensitivity of the self-collection methods is, however, higher and correlates better when Ct values
of the HCW swabs are less than 30. The combined results of the saliva and self-swab test achieve a
sensitivity equivalent to that of a health care worker administered swab. The specificity of the self-
collection methods is 100%. Together with high specificity, we postulate that self-collection meth-
ods have their roles in diagnosis in early disease, where the viral load, and infectivity is high.
Fig 1. Correlation of Ct values of HCW swab and self-swab.
https://doi.org/10.1371/journal.pone.0244417.g001
Fig 2. Correlation of Ct values of HCW swab and saliva.
https://doi.org/10.1371/journal.pone.0244417.g002
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 9 / 11
Supporting information
S1 File. Provisional protocol for saliva sample collected in Lucence SAFER kit.
(PDF)
S2 File. Study protocol.
(PDF)
S3 File. Table with Ct values of N gene.
(XLSX)
Acknowledgments
We thank all clinical, nursing and allied health staff who provided care for the patients at
Changi General Hospital, and Community Care Facility @ EXPO; staff in the Changi General
Hospital Clinical Trials & Research Unit for coordinating patient recruitment, logistics man-
agement and assistance.
Author Contributions
Conceptualization: Seow Yen Tan, Hong Liang Tey, Ernest Tian Hong Lim, Song Tar Toh,
Yiong Huak Chan, Sing Ai Lee, Cheryl Xiaotong Tan, Gerald Choon Huat Koh, Thean Yen
Tan, Chuin Siau.
Data curation: Seow Yen Tan, Hong Liang Tey, Ernest Tian Hong Lim, Yiong Huak Chan.
Formal analysis: Seow Yen Tan, Hong Liang Tey, Song Tar Toh, Yiong Huak Chan, Pei Ting
Tan, Gerald Choon Huat Koh.
Funding acquisition: Sing Ai Lee, Cheryl Xiaotong Tan.
Investigation: Seow Yen Tan, Hong Liang Tey, Ernest Tian Hong Lim, Song Tar Toh, Gerald
Choon Huat Koh, Thean Yen Tan, Chuin Siau.
Methodology: Seow Yen Tan, Hong Liang Tey, Song Tar Toh, Yiong Huak Chan, Gerald
Choon Huat Koh, Thean Yen Tan, Chuin Siau.
Project administration: Seow Yen Tan, Hong Liang Tey, Ernest Tian Hong Lim, Song Tar
Toh, Pei Ting Tan, Chuin Siau.
Resources: Ernest Tian Hong Lim, Song Tar Toh, Sing Ai Lee, Cheryl Xiaotong Tan, Chuin
Siau.
Supervision: Hong Liang Tey, Ernest Tian Hong Lim, Song Tar Toh, Gerald Choon Huat
Koh, Thean Yen Tan, Chuin Siau.
Validation: Seow Yen Tan, Gerald Choon Huat Koh.
Visualization: Sing Ai Lee, Cheryl Xiaotong Tan, Chuin Siau.
Table 7. Distribution of Ct values of the HCW swab, self-swab and saliva.
Test Median (IQR�) of Ct value
HCW Swab 31.59 (26.77, 35.62)
Self-swab 31.65 (26.65, 35.94)
Saliva 33.10 (28.25, 36.23)
� IQR = Interquartile Range
https://doi.org/10.1371/journal.pone.0244417.t007
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 10 / 11
Writing – original draft: Seow Yen Tan, Hong Liang Tey, Ernest Tian Hong Lim, Song Tar
Toh, Yiong Huak Chan.
Writing – review & editing: Seow Yen Tan, Hong Liang Tey, Ernest Tian Hong Lim, Song
Tar Toh, Yiong Huak Chan, Pei Ting Tan, Sing Ai Lee, Cheryl Xiaotong Tan, Gerald
Choon Huat Koh, Thean Yen Tan, Chuin Siau.
References 1. LeBlanc JJ, Heinstein C, MacDonald J, Pettipas J, Hatchette TF, Patriquin G. A combined oropharyn-
geal/nares swab is a suitable alternative to nasopharyngeal swabs for the detection of SARS-CoV-2. J
Clin Virol. 2020; 128:104442. https://doi.org/10.1016/j.jcv.2020.104442 PMID: 32540034
2. Wehrhahn MC, Robson J, Brown S, Bursle E, Byrne S, New D, et al. Self-collection: An appropriate
alternative during the SARS-CoV-2 pandemic. J Clin Virol. 2020; 128:104417. https://doi.org/10.1016/j.
jcv.2020.104417 PMID: 32403007
3. Wyllie AL, Fourmier J, Casanovas-Massana A, Campbell M, Tokuyama M, Vijayakumar P, et al. Saliva
or Nasopharyngeal Swab Specimens for Detection of SARS-CoV-2. N Engl J Med 2020; 383:1283–
1286. https://doi.org/10.1056/NEJMc2016359 PMID: 32857487
4. To KK, Tsang OT, Yip CC, Chan K, Wu T, Chan JM, et al. Consistent detection of 2019 novel coronavi-
rus in saliva. Clinical Infectious Diseases 2020; 71(15):841–3. https://doi.org/10.1093/cid/ciaa149
PMID: 32047895
5. Azzi L, Carcano G, Gianfagna F, Grossi P, Gasperina DD, Genoni A, et al. Saliva is a reliable tool to
detect SARS-CoV-2. J Infect. 2020; 81(1):e45–e50. https://doi.org/10.1016/j.jinf.2020.04.005 PMID:
32298676
6. Iwasaki S, Fujisawa S, Nakakubo S, Kamada K, Yamashita Y, Fukumoto T, et al. Comparison of
SARS-CoV-2 detection in nasopharyngeal swab and saliva [published online ahead of print, 2020 Jun
4]. J Infect. 2020; 81(2):e145–e147. https://doi.org/10.1016/j.jinf.2020.05.071 PMID: 32504740
7. To KK, Tsang OT, Leung WS, Tam AR, Wu T, Lung DC, et al. Temporal profiles of viral load in posterior
oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an
observational cohort study. Lancet Infect Dis. 2020; 20(5):565–574. https://doi.org/10.1016/S1473-
3099(20)30196-1 PMID: 32213337
8. Czumbel LM, Kiss S, Farkas N, Mandel I, Hegyi A, Nagy Á, et al. Saliva as a Candidate for COVID-19
Diagnostic Testing: A Meta-Analysis. Front Med (Lausanne). 2020 Aug 4; 7:465. https://doi.org/10.
3389/fmed.2020.00465 PMID: 32903849; PMCID: PMC7438940.
9. Young BE, Ong SWX, Ng LFP, Anderson DE, Chia WN, Chia PY et al. Immunological and Viral Corre-
lates of COVID-19 Disease Severity: A Prospective Cohort Study of the First 100 Patients in Singapore
(4/15/2020). Available at SSRN: https://ssrn.com/abstract=3576846 or https://doi.org/10.2139/ssrn.
3576846
10. Position Statement from the National Centre for Infectious Diseases and the Chapter of Infectious Dis-
ease Physicians, Academy of Medicine, Singapore– 23 May 2020
11. Jamal AJ, Mozafarihashjin M, Coomes E, Powis J, Li AX, Paterson A, et al. Sensitivity of nasopharyn-
geal swabs and saliva for the detection of severe acute respiratory syndrome coronavirus 2 (SARS-
CoV-2). Clin Infect Dis. 2020 Jun 25:ciaa848. https://doi.org/10.1093/cid/ciaa848 Epub ahead of print.
PMID: 32584972; PMCID: PMC7337630.
12. Rao M, Rashid FA, Sabri FSAH, Jamil NN, Zain R, Hashim et al. Comparing nasopharyngeal swab and
early morning saliva for the identification of SARS-CoV-2. Clin Infect Dis. 2020 Aug 6:ciaa1156. https://
doi.org/10.1093/cid/ciaa1156 Epub ahead of print. PMID: 32761244
PLOS ONE Accuracy of self-testing for COVID-19
PLOS ONE | https://doi.org/10.1371/journal.pone.0244417 December 16, 2020 11 / 11
