A one-way deterministic sensitivity analysis was conducted. First, the percentage reductions in the number of travelers for each extra day of quarantine (14.49% for inbound travelers and 9.66% for returning travelers for base case analysis) were increased and decreased 0.5-fold. Second, the length of quarantine to enter countries in the ROW was fixed at 0 days or 14 days while the length of quarantine to enter Singapore varied across 0–14 days.
Third, we fixed the COVID-19 prevalence rates in Singapore and ROW at their maximal and minimal values over the period from January to the middle of June 2022. The base case analysis used the average prevalence rates over this period. Higher prevalence rates may require more stringent testing and quarantine policies to achieve a higher NMB and vice versa. Fourth, we increased the R0 of Omicron from 8.2 (26 (link)) to a high estimate of 10 (27 (link)) and an assumed value of 15 to model the possibility of a new variant. Fifth, we reduced the vaccination coverage from 76% in ROW and 92% in SG, which was the coverage for having received the full regimen plus a booster shot or full regimen only as of June 2022, to 30% in ROW and 78% in SG, which was the coverage of having received the full regimen plus a booster shot as of June 2022 (19 , 20 ). When vaccination programs become non-mandatory and out-of-pocket payment is required for vaccination, the vaccination coverage may be lower than the current status. Sixth, the assumed efficacy of vaccines among inbound travelers was lowered from the efficacy of mRNA vaccine (base case analysis) to the efficacy of the inactivated vaccine (28 (link)). Seventh, we increased the PCR processing time to 2 days. An increase in processing time is likely if the daily PCR testing capacity is overwhelmed by a large number of travelers.
Eighth, we incorporated the tourism multiplier effect, as receipts in the tourism sector will likely have spill-over effect on upstream sectors (29 (link)). This spill-over effect is quantified by a multiplier derived using Leontief's matrix based on Singapore's Input-Output Table (21 ) and historical tourism receipt components during 2016–2020 (18 ). Ninth, we reduced the productivity loss due to quarantine to 0%, which may be plausible since returning travelers may still work remotely while in quarantine or use their annual leave. Tenth, the healthcare expenditure levels in destination countries relative to Singapore were varied from a lower bound of 0.04 (PCHE of Indonesia relative to Singapore) to an upper bound of 1.90 (PCHE of Australia relative to Singapore) among the 7 major destination countries for Singaporean travelers (24 , 25 ). Eleventh, we doubled the estimated medical costs of COVID-19 cases. Twelfth, the QALY losses due to morbidity for each type of symptomatic case were matched to high estimates from literature (30 (link)) on other respiratory diseases. Thirteenth, the CET of Singapore was varied from $39199 (0.4 times of GDP per capita, to proxy a supply-side CET) to $293,394 (3 times of GDP per capita) (21 , 31 (link), 32 (link)), allowing for lower and higher economic impacts of health loss due to COVID-19 morbidity and mortality.
Third, we fixed the COVID-19 prevalence rates in Singapore and ROW at their maximal and minimal values over the period from January to the middle of June 2022. The base case analysis used the average prevalence rates over this period. Higher prevalence rates may require more stringent testing and quarantine policies to achieve a higher NMB and vice versa. Fourth, we increased the R0 of Omicron from 8.2 (26 (link)) to a high estimate of 10 (27 (link)) and an assumed value of 15 to model the possibility of a new variant. Fifth, we reduced the vaccination coverage from 76% in ROW and 92% in SG, which was the coverage for having received the full regimen plus a booster shot or full regimen only as of June 2022, to 30% in ROW and 78% in SG, which was the coverage of having received the full regimen plus a booster shot as of June 2022 (19 , 20 ). When vaccination programs become non-mandatory and out-of-pocket payment is required for vaccination, the vaccination coverage may be lower than the current status. Sixth, the assumed efficacy of vaccines among inbound travelers was lowered from the efficacy of mRNA vaccine (base case analysis) to the efficacy of the inactivated vaccine (28 (link)). Seventh, we increased the PCR processing time to 2 days. An increase in processing time is likely if the daily PCR testing capacity is overwhelmed by a large number of travelers.
Eighth, we incorporated the tourism multiplier effect, as receipts in the tourism sector will likely have spill-over effect on upstream sectors (29 (link)). This spill-over effect is quantified by a multiplier derived using Leontief's matrix based on Singapore's Input-Output Table (21 ) and historical tourism receipt components during 2016–2020 (18 ). Ninth, we reduced the productivity loss due to quarantine to 0%, which may be plausible since returning travelers may still work remotely while in quarantine or use their annual leave. Tenth, the healthcare expenditure levels in destination countries relative to Singapore were varied from a lower bound of 0.04 (PCHE of Indonesia relative to Singapore) to an upper bound of 1.90 (PCHE of Australia relative to Singapore) among the 7 major destination countries for Singaporean travelers (24 , 25 ). Eleventh, we doubled the estimated medical costs of COVID-19 cases. Twelfth, the QALY losses due to morbidity for each type of symptomatic case were matched to high estimates from literature (30 (link)) on other respiratory diseases. Thirteenth, the CET of Singapore was varied from $39199 (0.4 times of GDP per capita, to proxy a supply-side CET) to $293,394 (3 times of GDP per capita) (21 , 31 (link), 32 (link)), allowing for lower and higher economic impacts of health loss due to COVID-19 morbidity and mortality.
Free full text: Click here
