A total of 1300 records were identified through all methods. One hundred three studies met the inclusion criteria. Besides, we have also removed studies with different languages than English and those that describe only the disease's impact. We removed the studies by considering the Drummond checklist specifically if they have not mentioned the costs, the outcomes, and cost-effectiveness ratios and quality assessment tool we adopt. Fourteen studies were selected through cited reference searching and reference list screening, of which ten were excluded as they were not full economic evaluation studies. A total of 36 studies were finally included for the economic evaluation.
Description of the results
The studies were conducted mostly in developed countries. Among selected studies, only three studies, LMIC, South Africa, and China, were conducted in the developing countries, 33 studies were from the developed countries. The review indicated that 15 studies were conducted in the USA, followed by 3 in Singapore, 3 in the UK, 3 in China, 2 in the Netherlands, 1 in Australia, Canada, South Africa, Japan, France, Israel, and New Zealand. Two studies did not explicitly inform the country of the studies.
Among the studies, 14 of them were conducted on CUA, 15 on CEA, three on CBA, one on both CEA and CUA, two on both CBA and CEA, and one on both CBA and CUA. In terms of perspectives, 15 studies were conducted from a societal perspective, ten from a healthcare provider perspective, and eight from both a societal and a healthcare provider perspective. However, 3 of the studies did not clearly describe the perspective of the study. Most of the studies indicated the source of costs and the type of model used in the study. In table one below, we present a description of the reviewed studies.
Most of the studies have clearly described the costs used in the study. When using a societal perspective, there is a slight difference in the way productivity losses are estimated. For example, some studies have considered the future productivity loss due to death in the estimate , while other studies only considered the productivity loss at the time of infection, which might assume the future value of lost workers will be replaced [26-28]. The input cost for one study was not clearly described, and the link to other studies to see the estimate of the cost component did not work , and we failed in getting the information from the author. The cost input for one study was also based on the assumption that it should be interpreted with caution .
The cost-effectiveness of different interventions
Most of the studies have focused on pharmaceutical interventions related to antiviral treatment and vaccination, and some studies have conducted both pharmaceutical and non-pharmaceutical interventions. Generally, there are fewer studies conducted on non-pharmaceutical intervention.
In table two below, we present the cost-effectiveness of the reviewed studies.
In figure 3 below, we presented the cost-effectiveness of the interventions in 2019 USD. However, the interventions' findings are not consistent because of the settings of the studies and different assumptions, vaccinations, treatment of sick patients, and using personal protective equipment lie in a range of highly cost-effective interventions.
Cost-effectiveness of the reviewed studies
Stockpiling drugs and treatment of Antivirals
Seventeen studies were conducted on the stockpiling of drugs and treatment of patients with antivirals. Most of the studies used a do-nothing scenario as a comparator, and few have compared the interventions against the alternative course of actions or the addition of non-pharmaceutical interventions such as school closure. Stockpiling for the therapeutic use of the drugs to high-risk patients and post-exposure short term prophylaxis of all close contacts, including index patient treatment, is cost-saving. However, pre-exposure long term prophylaxis of entire patients is not cost-effective for stockpiling with less CFR.. Prophylaxis was economically beneficial in high-risk subpopulations and pandemics with a case fatality rate greater than 0.6% . It is cost-effective to stockpile drug when compared to no intervention at a pandemic reproductive rate of 1.8 and 20-40% population illness rate , and if the actual risk is less than 37% for 30 years but not cost-effective if less than 60% of the population would take the antiviral drugs or the attack rate is about 50% . All the studies reviewed had estimated that the stockpiling of antiviral agents for only treating patients had optimal economic benefits.
Testing all symptomatic patients and treating those only with positive test results is less cost-effective when compared to treating all symptomatic patients for a pandemic influenza disease . For stockpiling drugs and vaccine, antiviral treatment of those clinically infected is the most cost-effective, followed by population pre-pandemic vaccination, and then the combination of both antiviral treatment and population pre-pandemic vaccination when compared to a small stockpile of antiviral drugs for prophylaxis of case contacts and treatment of clinical cases and vaccine after six months .
A combination of interventions such as a reduction in household contacts and 60% of work/school contacts and stockpiling drugs is more cost-effective compared to the same intervention by adding school closure for 26 weeks. At 1% influenza mortality, moderate infectivity (Ro of 2.1 or greater), and 60% population compliance, a combination of adult and child social distancing, school closure, and antiviral treatment and prophylaxis is cost-effective. The addition of school closure is only cost-effective for pandemic with CFR of more than 1% and Ro higher than 1.6 .
The use of IV antiviral treatment for hospitalised patients with influenza-like illness is cost-effective for the smaller cost of antiviral treatment . Using Oseltamivir for Influenza was cost-effective from a healthcare perspective and cost-saving from a societal perspective . The empirical treatment of all sick patients with Oseltamivir costs less money and have higher effectiveness (dominant) when compared to the post rapid influenza diagnostic test treatment with Oseltamivir, and is cost-effective when compared with no intervention . For high-risk groups, antiviral treatment, although less effective, seems more feasible and cost-effective than prophylaxis and should be chosen, mainly if limited drug availability .
For a variety of pandemic influenza scenarios (attack rate 20% or more, probability of preterm birth for women with Influenza 12% or more, mortality for a preterm neonate 2% or more, and the probability of influenza-attributable hospitalisation 4.8%), the use of antiviral medications for post-exposure prophylaxis among pregnant women in a pandemic influenza scenario is cost-effective compared to no intervention for pandemic influenza, but not in a seasonal influenza setting. As the probability of preterm death increases and CAR, the use of prophylaxis will be cost-saving . PEP with Oseltamivir is probable to be a cost-effective strategy for family contacts when influenza-like illness contact attack rates are 8% or higher .
Most of the interventions related to stockpiling and treatment are cost-effective, given that there are higher infectivity and severity of pandemic influenza. The probability of pandemic in a given year and contact rate is also a determinant factor whether the treatment/stockpiling is cost-effective. In general, the cost-effectiveness of pandemic treatment/ stockpiling is determined by the probability of pandemic in a given year, contact rates, pre-existing immunity, age-specific mortality, changes in attack rates, case fatality rate, discount rate, and AV drug efficacy.
Stockpiling vaccines and vaccination
Among the reviewed studies, nine of them were conducted mainly on vaccination strategies. For influenza pandemic, early availability of vaccine before the peak time of pandemic determines its economic value. Vaccinating high-risk groups is highly cost-effective, followed by extending the vaccination to schools and then low-risk groups . Vaccinating low-risk group is also cost-effective when compared to no intervention. When compared to seasonal prophylaxis, vaccination of the total population is preferred. Vaccinating the population at risk is very cost-effective, followed by a priority population. The least cost-effective scenario is vaccinating the total population for a pandemic with a 25% attack rate . Under a wide range of scenarios, vaccination for pH1N1 for children and working-age adults is cost-effective compared to other preventive health interventions. The vaccine availability delays had a substantial impact on the cost-effectiveness of vaccination strategies .
At a lower cost ($21 per vaccine), there is a net saving to society if persons in all age groups are vaccinated, but at $62 per vaccine and gross attack rates of 25%, there will be a net loss if persons not at high risk for complications are vaccinated. At a lower vaccination cost, vaccinating all populations would be cost-effective. However, at higher vaccine costs, vaccinating high-risk populations would be cost-effective . Vaccination initiated before the outbreak is also cost-saving , and pre-vaccination of 70% of the population with low efficacy vaccine is cost-effective .
Maternal influenza vaccination using either the single or two-dose strategy is a cost-effective approach when influenza prevalence greater than or equal to 7.5% and influenza-attributable mortality is greater than or equal to 1.05% (consistent with epidemic strains). This will be cost-saving as the prevalence of Influenza become more than 30% . Assuming each primary infection causes 1.5 secondary infections, vaccinating 40% of the population would be cost-saving. Vaccination is even more cost-saving if more extended incubation periods, lower rates of infectiousness, or increased implementation of non-pharmaceutical interventions delay time to the peak of the pandemic. Vaccination saves fewer lives and is less cost-effective if the epidemic peaks earlier . Mass vaccination was cost-effective and depended on the timing of vaccination and vaccine effectiveness . Vaccination and stockpiling pneumococcal vaccine for the time of influenza pandemic are also very cost-effective [28, 46].
For severe pandemics or pandemics in which pre-pandemic influenza vaccine is unavailable, stockpiling of 23-valent pneumococcal polysaccharide vaccine (PPSV23) for a pneumococcal disease can be a cost-effective strategy for reducing the health and economic burden associated with secondary pneumococcal infections in a high-risk population. However, for a mildly severe pandemic in which pre-pandemic influenza vaccine is available, stockpiling of PPSV23 may not be cost-effective . The cost-effectiveness of vaccine stockpile is determined by the vaccine efficacy and strain match and the costs of the vaccine. For higher vaccine efficacy, higher strain match, and lower vaccine cost, stockpiling vaccines are cost-effective .
This review showed that the cost-effectiveness of vaccination strategies is sensitive to the risk of death, CAR (the overall size of the epidemic), cost of vaccines, model assumptions, QALY loss per case, hospitalisation rates and costs, and case-fatality ratios, the number of vaccine doses needed, costs of vaccination, illness rates, and timing of vaccine delivery.
Most of the reviewed studies related to school closure are conducted in combination with other studies. Closing school is a non-pharmaceutical response that is usually implemented at the early phase of a pandemic. Individual-based school closure with a lower threshold to trigger school closure is more cost-effective than district level or system level school closure . Likewise, closing schools for a lower reproductive number, less than 1.6, is not considered economically efficient for society . As most of the pandemic interventions are implemented in combination, the extent to which adding schools closure to pharmaceutical interventions is also reviewed. Combination of antiviral prophylaxis for those with contact history and school closure is considered more efficient when compared to full targeted antiviral prophylaxis. Besides, a combination of pre-vaccination with a low-efficacy vaccine before the outbreak of a pandemic and school closure is more cost-effective than full targeted antiviral prophylaxis . When compared to do nothing scenario, a combination of interventions such as adult and child social distancing, school closure, AV treatment, and prophylaxis is cost-effective for a pandemic with a reproductive number of 2.1 and a case fatality rate of 1%. The findings in those studies are sensitive to the infectivity, case fatality rate, level of population compliance, and antiviral effectiveness .
Tracking Exposed Persons, Testing, and Quarantines
A combination of healthcare testing, contact tracing, isolation centre, and mass symptom screen compared to only healthcare testing was found to be cost-effective with a reproductive number of 1.5 or with 0.1% prevalence of the disease to prevent COVID-19. The findings were sensitive to reproductive number, efficacies of contact tracing, isolation centres, and Mass screening to detect infections; cost of isolation of cases and quarantine centres . At school for re-opening, weekly test with a test sensitivity of 80% (when compared to test sensitivity of 70% is cost-effective) with reproduction number of 2.5 and case fatality rate of 0.05% . PCR for any symptomatic patients plus one-time PCR for the entire population is cost-effective when compared to PCR for any COVID-19-consistent symptoms with self-isolation if positive with an effective reproduction number of 2.0. The findings are sensitive to Varying rates of presentation to hospital care and ICU survival, input parameters on infections and deaths . PCR testing with 90% sensitivity, and 100% specificity with a probability of Influenza-like illness being Influenza of 10%, 20%, 30% are cost-effective when compared to no interventions. Point of care test with 50% sensitivity and 95% specificity with the same probability is also cost-effective. The findings are sensitive to the probability of ILI being Influenza .
Personal protective equipment
Personal protective equipment in preventing COVID-19 cases was also assessed for some of the studies. For example, a study which compared N95 mask and face mask estimated the additional cost to be ranging from US$ 490 to USD 1230 per cases prevented with a clinical attack rate of 4.6. The findings were sensitive to the CRI attack rate and intervention effectiveness . To protect healthcare workers, using personal protective equipment when compared with no intervention was 4,448 USD per HCW life saved for HCW infection as a percent of total infection of 9.6% and case fatality ratio of 1.38% . For spontaneous and induced labour at a delivery service, universal screening is the preferred strategy given the high cost of universal PPE in the base case. However, for a planned cesarean delivery, universal PPE is cost-saving compared with universal screening. Universal screening is more cost-effective than PPE to prevent the transmission of COVID-19 to health workers for spontaneous and induced labour . For a hospital protection measure response against respiratory infections, protection measures targeting only infected patients (hospitals full PPE for a suspected case, contact tracing, AV treatment for all cases for Pandemic (H1N1) 2009) yielded low incremental cost per death averted of $23,000 (US$) for pandemic (H1N1) 2009. Enforced protection in high-risk areas and full protection throughout the hospital averted deaths but came at a higher incremental cost .
Other non-pharmaceutical interventions
For interventions like border closure, the net present value is high if the case fatality rate is high. Border closure was relatively cost-effective when seen from the provider perspective and cost-saving from the societal perspective. The findings were sensitive to the number of deaths assumed, the value put on lost productivity, and the value of a monetised .