Cost-effectiveness of emergency preparedness measures in response to infectious respiratory disease outbreaks: a systematic review and econometric analysis.

OBJECTIVES: Respiratory infectious disease outbreaks pose a threat for loss of life, economic instability and social disruption. We conducted a systematic review of published econometric analyses to assess the direct and indirect costs of infectious respiratory disease outbreaks that occurred between 2003 and 2019.
SETTING: Respiratory infectious disease outbreaks or public health preparedness measures or interventions responding to respiratory outbreaks in OECD countries (excluding South Korea and Japan) so as to assess studies relevant to the European context. The cost-effectiveness of interventions was assessed through a dominance ranking matrix approach. All cost data were adjusted to the 2017 Euro, with interventions compared with the null. We included data from 17 econometric studies. PRIMARY AND SECONDARY OUTCOME MEASURES: Direct and indirect costs for disease and preparedness and/or response or cost-benefit and cost-utility were measured.
RESULTS: Overall, the economic burden of infectious respiratory disease outbreaks was found to be significant to healthcare systems and society. Indirect costs were greater than direct costs mainly due to losses of productivity. With regard to non-pharmaceutical strategies, prehospitalisation screening and the use of protective masks were identified as both an effective strategy and cost-saving. Community contact reduction was effective but had ambiguous results for cost saving. School closure was an effective measure, but not cost-saving in the long term. Targeted antiviral prophylaxis was the most cost-saving and effective pharmaceutical intervention.
CONCLUSIONS: Our cost analysis results provide evidence to policymakers on the cost-effectiveness of pharmaceutical and non-pharmaceutical intervention strategies which may be applied to mitigate or respond to infectious respiratory disease outbreaks. © Author(s) (or their employer(s)) 2021. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

A systematic approach was followed, and the assessment of data quality indicated that the majority of studies included were of high quality.

The synthesis of the results was performed using the dominance ranking matrix approach, which allowed for a direct comparison of the cost-effectiveness of each intervention to the null.

Costs and resources varied between different countries, different regional settings and over time, making the cost component comparison of cost-effectiveness measures complex to interpret.

We only focused on EU and OECD analogous countries excluding Japan and South Korea, and hence our cost-effectiveness analyses are not applicable to other countries or settings.

Discrepancies in context and populations likely affect the implementation and efficacy of interventions.

This study was conducted prior to the COVID-19 pandemic.

Introduction

Emerging, re-emerging and endemic respiratory and influenza-like infectious diseases represent a threat for loss of life, economic instability and social disruption as they can rapidly spread within communities and across countries, affecting the whole globe. Annually, it is estimated that 5%–15% of the population will suffer from influenza-related respiratory tract infections, while 3–5 million people face severe illness due to influenza.1 In 2018, a total number of 109.5 million influenza virus episodes were identified among children under 5 years globally, with approximately 34 800 overall deaths. In Europe, seasonal influenza is estimated to lead to 4–50 million symptomatic cases and 15 000–70 000 deaths annually; however, this may differ between years, as the severe 2017/2018 influenza season led to an estimated 152 000 deaths in Europe alone.2 3

In order for robust national preparedness systems and response strategies to outbreaks to be established in the Europe, it is crucial for public health officers to receive recent data of the health impact and the economic burden of respiratory infectious disease outbreaks in contrast to emergency response and preparedness actions. This evidence will ensure well-informed decisions regarding, among others, the proper allocation of resources.4 5 To this extent, although there is substantial literature from previously published systematic reviews on the value of public health emergency preparedness, they either refer to an older timeframe6 or use mathematical models to predict the effectiveness and cost-effectiveness of measures.7 Hence, there is limited recent information on the economic evaluations of infectious respiratory disease outbreaks that provide an overview of the cost effectiveness of response measures.8

Within the above context, the aim of this systematic review of econometric analyses was to assess the economic impact of response and preparedness measures when contrasted with the cost of infectious respiratory disease outbreaks. We further synthesise the cost-effectiveness for each intervention using a dominance ranking matrix (DRM) approach.

Methods

Search strategy and selection criteria

A comprehensive systematic literature review of published econometric analyses was conducted between July and August 2019 using the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines9 and the Consolidated Health Economic Evaluation Reporting Standards (CHEERS)10 to identify peer-reviewed articles using two biomedical literature databases (PUBMED and EMBASE) and two economic literature databases (ECONLIT, IDEAS REPEC). The search strategy was designed for a broader study aiming to identify econometric studies on all types of infectious diseases, but due to the outbreak of COVID-19, and for the purposes of this specific article, we retained only those referring to respiratory infectious diseases. The complete search strategy and search terms are available in online supplemental appendix 1.

The inclusion criteria were as follows:

Exposure: respiratory infectious disease outbreaks or public health preparedness measures or interventions responding to respiratory outbreaks in OECD countries (excluding Asian countries South Korea and Japan due to the wide cultural differences with the EU context as this study was performed under contract for the European Center for Disease Prevention and Control (ECDC)).

Comparator: (i) no intervention (cost of inaction) or current practice, (ii) cost of preparedness versus cost of response (for studies reporting cost and benefit of public health preparedness).

Outcome measures: direct and indirect costs for disease and preparedness and/or response or cost-benefit and cost-utility. Typical outcome measures of economic evaluations included: life years gained or cost per life-year gained with the intervention under investigation when incremental costs are combined, cost per quality-adjusted life year (QALY) gained, cases averted, monetary outcomes.

Perspective: all direct and indirect costs pertaining to all relevant perspectives (eg, individual, hospital, insurance and societal—including national and regional) and all direct and indirect costs pertaining to all relevant perspectives according to York Health Economics Consortium11 (health system perspective, including hospital, public health units; societal perspective; governmental perspective).

Study designs: all relevant analytical epidemiological designs which estimate cost either as full economic evaluation studies, including cost-minimisation, cost-effectiveness, cost-utility and cost-benefit studies; cost-outcome and economic modelling studies or partial economic evaluations.

Timeframe: from 2003 until August 2019, to reflect the timepoint from the 2003 SARS outbreak and onward12—this review refers to the pre-COVID-19 published evidence.

Studies that met the above inclusion criteria but did not report or perform any econometric analysis were excluded.

Data analysis and extraction

Studies identified from the searches were uploaded into a bibliographic database in which duplicate entries were removed. Initially, a pilot training screening process was used, where a random sample of 100 titles and abstracts were screened independently for eligibility by four reviewers (KN, KZ, RP, JLB) to enable consistency in screening and identify areas for amendments in the inclusion criteria. Following this, a random sample of 50% of titles and abstracts was screened independently by two reviewers. Since a high measure of inter-rater agreement was achieved (percentage agreement >88.7% and/or Cohen’s Kappa >0.646), the remaining titles and abstracts were screened for eligibility by one reviewer. Where insufficient information was available in the title and abstract to make a decision, the full-text article of the document was retrieved for further inspection. Full-text documents of potentially eligible studies were retrieved for the records marked for inclusion. All full-text documents were independently double-screened by two reviewers, and inter-rater agreement measures were calculated at 88.3%. Disagreements in every step of the process were subsequently discussed and agreed on. Documents that passed the inclusion criteria on the basis of the full-text screening were included in the current review.

Appraisal of methodological quality

For evaluating the methodological quality of the included studies, the Consensus on Health Economic Criteria (CHEC) checklist13 was used. This specific tool has been designed for the assessment of full economic evaluations and includes 19 items (questions) with answers of ‘Yes’ or ‘No’. For each positive answer on full economic evaluation studies, a single point was being assigned for the methodological quality, with a maximum score of 19. For the quality appraisal of partial economic evaluations, we used items from the CHEC checklist that were applicable—hence, the maximum score was 16. The quality appraisal process was completed by two reviewers, with a percentage of agreement in the three pilot studies, initially assessed by both, of 83.7%.

Comparative economic analysis approach

All cost data were adjusted to a common currency (Euro in 2017 (€2017)) and price year using the Campbell and Cochrane Economics Methods Group–Evidence for Policy and Practice Information and Coordinating Centre cost converter.14 We adjusted the original estimate of cost from the original price year to a target price year of the €2017, using a gross domestic product deflator index (GDPD), obtained from the International Monetary Fund World Economic Outlook Database GDPD index data set.15 Subsequently, we converted the price-year adjusted cost estimate from the original currency to €2017, using conversion rates based on purchasing power parities (PPP) for GDP (the 2017 implied conversion factor was US$1=€1.13, the €2017 conversion factor was €1=US$1.2, while with regard to British pounds, the conversion factor was £1=€0.88). PPP values adjust appropriately for differences in current price levels between countries, thus allowing comparisons based on a common set of average international prices; this is an advantage over pure exchange-rate conversions and GDP per capita approaches as PPPs eliminate differences in price levels between countries in the process of conversion. For studies that did not state the year of cost calculation, the costs were calculated 1 year before the publication year of each respective study.

Synthesis of cost-effectiveness

In order to synthesise the cost-effectiveness results, the DRM approach was used, which is a classification system developed for summarising and interpreting the results of economic evaluations in systematic reviews.16 The DRM is a three-by-three matrix with the following classification options:

Strong dominance for the intervention when the incremental cost-effectiveness measure shows the intervention compared with no intervention as: (i) more effective and less costly or (ii) as effective and less costly or (iii) more effective and equal cost.

Weak dominance for the intervention when the measure shows the intervention compared with no intervention as: (iv) effective and equally costly or (v) more effective and more costly or (vi) less effective and less costly.

Non-dominance for the intervention when the measure shows the intervention compared with no intervention as: (vii) less effective and more costly or (viii) less effective and equally as costly or (ix) as effective and more costly.

Within our DRM, only studies that compared interventions to no intervention were included in the matrix.

Patient and public involvement

This study was performed under contract for the European Center for Disease Prevention and Control (ECDC). Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Results

The initial study search yielded 20 513 studies after removal of the duplicates and according to the specified selection criteria, only 66 were further assessed for eligibility via full text. Through the assessment of the full-texts, 52 studies were excluded for the following reasons: inadequate data on costs and/or cost-effectiveness (n=2), they were reviews (n=15), not referring to respiratory outbreaks (n=29), not referring to outbreaks of infectious diseases (n=2) and conference abstracts with no full text available (n=4). Additionally, three full-text papers were identified through the screening of the reference lists of the selected manuscripts, and hence, a total number of 17 econometric studies were considered in our analysis. The flowchart of the study selection process is presented in figure 1.

Figure 1

Flowchart.

Overall, 11 out of the 17 studies were of high methodological quality (>80%), 5 were categorised as of good quality (60%–80%) and only 1 was of medium quality (40%–60%) due to missing quality criteria not mentioned by the authors including the comparative intervention, sensitivity analysis, incremental costs and outcomes. Online supplemental appendix 2 presents the overall quality appraisal score, for studies related to cost of infectious disease outbreaks and for sources related to preparedness, preventive and response measures concerning infectious disease outbreaks. The quality appraisal of partial and full economic evaluation studies is in online supplemental appendices 3 and 4, respectively. It is important to note that for the studies where a partial economic evaluation was performed, we only performed calculations for the items of the quality appraisal tool that were applicable.

Comparative cost analysis of infectious respiratory disease outbreaks

Regarding infectious respiratory disease outbreaks, six studies were included.17–22 All studies referred to influenza as the disease, either relating to pandemic H1N1 or seasonal Influenza B. Geographically, the studies were performed in the USA,17 Spain,18 22 France,19 New Zealand and Australia.20 21 Five out of the six studies were observational in design (cross-sectional or retrospective) and used collected data;18–22 one study was based on a simulation model.17 Similarly, five out of the six studies assessed costs from a healthcare system perspective;17 18 20–22 however, societal (n=3),17–19 governmental (n=1)17 and payer (n=1)19 perspectives were also assessed. Discounting in costs was not necessary for any of the included studies as the implementation timeframe had a duration of less than 1 year, and sensitivity analyses were performed only in three studies.17 19 21 A detailed description of the characteristics of the included illness studies is presented in online supplemental appendix 5.

Table 1 presents an analytical overview of the direct and indirect costs associated with influenza outbreaks. Direct costs mainly refer to medical and healthcare costs related to the outbreaks, along with the costs of response measures. Indirect costs included the loss of income, the loss of business and the loss of productivity. The overall direct costs reported in the studies were calculated at the patient level where possible.

Table 1

Characteristics of cost of illness studies of influenza outbreaks*, expressed in Euros (base year 2017)

Study (Publication Year)	Setting, year	Perspective	Direct costs (€, 2017)	Indirect costs (€, 2017)
Prager et al (2017)17	USA, n/a	Healthcare system, governmental, societal	Seasonal (no vaccination): €5.92 billionSeasonal (vaccination): €9.96 billionPandemic (no vaccination): €81.18 billionPandemic (vaccination): €65.59 billion	Illness-related workdays losses(a) Vaccination and no vaccination in a seasonal scenario: vaccination contributes to more workday losses than no vaccination(b) Vaccination and no vaccination in a pandemic scenario: vaccination reduces workday losses by 22.2 million days compared with no vaccination
Morales-Suárez-Varela et al (2016)18	Spain, 2009–2010	Healthcare system, societal	Total direct cost/patientNon-pregnant women: €3 908.70Pregnant women: €2 227.10	Total indirect cost/patientNon-pregnant women: €107.18Pregnant women: €63.83
Silva et al (2014)19	France, 2010–2011	Payer, societal	Mean direct cost/patientAll ages—€53.430–4 years old—€73.915–14 years old—€52.7915–65 years old—€35.26≥65 years old—€44.13Total direct costsAll ages—€107 883 8350–4 years old—€18 908 2545–14 years old—€52 474 78115–65 years old—€21 590 741≥65 years old—€6 940 836	Mean daily allowance cost due to work leave/patientAll ages—€22.380–4 years old—€05–14 years old—€015–65 years old—€97.88≥65 years old—€0
Higgins et al (2011)20	Australia and New Zealand, 2009	Healthcare system	Total mean cost: €19 296 136Total ICU costs: €6 107 069Total non-ICU costs: €12 961 942Mean cost of ICU/patient: €61 368Mean cost of non-ICU/patient: €10 755Mean cost in ICU/per patient and per day: €4 767	Non-reported
Wilson et al (2009)21	New Zealand, 2009	Healthcare system	Total ICU costs: €40 807 660Median ICU cost/patient: €22 540Mean ICU cost/patient: €32 168Total hospital costs/patientMedian hospital cost: €39 696Mean hospital cost: 53 553Treatment costs in ICU per subgroup(a) Cost/patient with and without pre-existing comorbidity€16 100 and €28 980, respectively(b) Cost/patient with viral pneumonitis and with other influenza syndromes€22 212 and €12 880, respectively	Non-reported
Rodríguez-Rieiro et al (2009)22	Spain, 2009	Healthcare system	Total cost: €36 700 000Median cost per hospitalisation (concomitant chronic disease): €2 205Median cost per hospitalisation (without a medical condition): €1 172	Non-reported

The adjustment was performed from Canadian $, US$, Australian $, British pounds £ and converted to Euro (Germany has been selected as target currency in these cases). Currencies from European Union countries adjusted to their currency.

The cost data include all forms of cost derived from inclusion studies, such as overall/total cost, mean/average cost, income loss, labour cost, household cost, savings, cost per case, etc.

For studies without currency year indicated, the previous year of publication was selected for adjustment.

*Confirmed or extrapolated/hypothetical cases on which they base the economic evaluation.

ICU, intensive care unit.

The most recent study was a simulation study by Prager et al,17 in which multiple scenarios were assessed through simulation models for the US population so as to estimate the total economic burden of pandemic influenza outbreaks in the USA, taking into account both the scenario of an adequately vaccinated population and the opposite. The results indicated that medical expenditures for a pandemic influenza outbreak could reach 83.2 billion €2017 in the no vaccination scenario and 67.3 billion €2017 in the vaccination scenario. Notably, for indirect cost estimations, vaccination in a pandemic scenario would reduce workday losses by 22.2 million days, when compared with no vaccination.

Silva et al19 focused on an influenza outbreak in France between 2010 and 2011 and extrapolated the results to the entire country with a hypothetical approximate number of 2 million influenza cases (3.2% of the French population), for which they calculated an overall cost of 151 million €2017 for the French Health Insurance System. Direct costs per patient ranged between 35.26 €2017 and 73.91 €2017, with higher indirect costs of 97.88 €2017 per day due to absence from work, for those within the 15–65 age group.

Two studies assessed the cost of an influenza outbreak from an intensive care unit (ICU) and hospital perspective.20 21 One focused on ICU and hospital costs derived from an influenza pandemic in 2009 in New Zealand (among 1224 cases, of which 122 were admitted to ICUs), which surpassed 40.8 million €2017 at an average cost of 32 167 €2017 per patient, with significantly increased costs for patients with underlining comorbidities.21 The mean total hospitalisation cost (normal and ICU) per case surpassed 53 553 €2017. Similarly, in a study that included 762 H1N1 cases from both Australia and New Zealand, the mean cost per ICU patient was 61 368 €2017, with a per-day cost of 4767 €2017.20 On the contrary, the non-ICU patient had a mean cost of 10 755 €2017; however, overall non-ICU patient costs surpassed those of ICU patients (12.96 million €2017 vs 6.1 million €2017), leading to a total hospitalisation cost of 19.3 million €2017 for the 2009 influenza outbreak.

Similarly, Rodriquez-Rieiro et al22 studied the hospitalisation costs that occurred during the 2009 influenza pandemic in Spain, which reached 36.7 million €2017 for 11 449 hospitalisations—during which the appearance of comorbidities led to higher average costs per patient (2205 €2017 vs 1172 €2017, respectively). Specific populations in Spain were assessed by Morales-Suárez-Varela et al18 who estimated direct costs for medical visits, medication and diagnostic tests at €3908 €2017 for non-pregnant women and 2227 €2017 for pregnant women of reproductive age, with indirect costs estimated at 107 €2017 and 64 €2017, respectively.

Cost-effectiveness studies of measures in averting and/or responding to infectious respiratory disease outbreaks

We identified 11 studies23–33 referring to preparedness, preventative and response measures, to influenza outbreaks, presented in detail in online supplemental appendix 6. Two studies were observational (based in the Netherlands and the UK),23 24 and the remaining nine were simulation models (four US models, with one study each modelled for Canada, France, Australia, Israel and one referring to developed countries in general). All included studies either used a cost-effectiveness or a cost-utility economic evaluation approach. The studies’ timeframes ranged from 2004 to 2018. Regarding the perspective for direct and indirect costs, a healthcare system or society approach was consistently presented.

The preparedness, preventive and response measures described included three pharmaceutical interventions (vaccination as a response measure, general vaccination, antiviral drug therapy and stockpiling),31–33 four non-pharmaceutical interventions (screening at the point of contact, community contact reduction, volunteer isolation/quarantine, school closure and the use of personal protective measures)23–25 28 and four combined pharmaceutical and non-pharmaceutical interventions.26 27 29 30 Table 2 presents the details of the cost-effectiveness studies on preparedness and response measures for infectious respiratory disease outbreaks. Further details on the comparative analysis of health indexes gained when adverting or responding to respiratory outbreaks can be found in online supplemental appendix 7.

Table 2

Characteristics of full economic evaluation studies on preparedness and response measures of influenza outbreaks, expressed in Euros (base year 2017)

Study (publication year)	Setting, Year	Population (n)	Interventions	Comparator	Economic evaluation outcomes
Non-pharmaceutical studies
Lankelma et al (2019)23	Netherlands, 2017–2018	Patients with acute RTI at the emergency department (1546 tests, 624 cases)	Point-of-care-testing for Influenza before hospital admission	2016–2017 influenza season	Net savings€388 317 (after subtraction with costs)More than 80% of the total savings are due to the shorter length of stay and decreased hospital admissions.The overall cost of intervention: €98 968Laboratory costs at €72 202Clinical aspects costs at €26 767
Orset (2018)25	France, 2014	200 participants, data extrapolated	7-day home confinement	No intervention	Costs associated with home confinement(a) Direct costsFor adults: €742/caseFor elderly: €1191/case(b) Indirect costsProductivity losses/caseFor adults: €550.For elderly: €125Costs of death/caseThe cost of death for children is estimated at €22–128, for adults at €63–361 and for elderly at €2667–15 389Loss of productivity due to influenza/caseProductivity loss in case of adult sickness: €88.70 (incl. absent from work+reduced productivity)Productivity loss in case of a sick child for the adult (mainly mother): €97.62
Sadique et al (2008)24	UK, 2005	Working parents with depending children	School closure	No intervention	Cost of school closure: between €280 million—€2.8 billion/weekCost of absenteeism: €1.4 billionAdjusting for informal care, the cost reduced between€552–€635 million per week.Adjusting for the elasticity of production the cost reduced to €970 327 320—€1.1 billion per week
Tracht et al (2012)28	USA, 2009–2010 influenza season	Simulation of the USA (302 million people:73 million children, 191 million adults and 38 million seniors)	Population use of face masks (N95) on the spread of a pandemic	No intervention	Net savingsIf masks are worn by 10% of the adult population: €418.75 billionIf masks are worn by 50% of the adult population: €501.9 billionEconomic burden, if no intervention: €728.28 billion (incl. direct and indirect costs)
Combined pharmaceutical and non-pharmaceutical strategies
Saunders-Hastings et al (2017)26	Canada, n/a	A simulation of Ottawa, Canada (1.2 million)	Vaccination+antiviral treatment Vaccination+antiviral treatment+antiviral prophylaxis Community contact reduction+personal protective measures+isolation Community-contact reduction+personal protective measures+isolation+antiviral treatment School closure+community contact reduction+personal protective measures+quarantine All interventions	No intervention	Cost/LYG vs no intervention €1700/LYG €1769/LYG €4394/LYG €4447/LYG €171 590/LYG €131 679/LYG Total economic burdenFor all scenarios, the economic burden ranges between €75 758 to €1 416 351
Halder et al (2011)27	Australia, 2009	A community in Western Australia (30 000)	Different combinations of durations of individual school closure, antiviral treatment, household antiviral prophylaxis, extended antiviral prophylaxis, 50% workplace closure, 50% community contact reduction	No intervention	Cost/case averted:Antiviral drug strategies+2 weeks school closure: €396 per case averted (cost-effective)Short-duration school closure: €820/case avertedISC, continuously+50% workplace. continuously: €6 204/case avertedIn case of 2 weeks for the above combination: €1891/case avertedISC, continuously: €2180/case avertedTotal cost, per 100 000 populationThe dual strategy of individual school closure for 2 weeks (ISC) along with the 50% community contact reduction (CCR): €3.39 millionThe dual strategy of continuous ISC along with the continuous—50% WP: €61.3 million.Productivity loss due to illness and interventions per 100 000 populationISC (cont.)+WP (cont.): €90.21 millionCombined antiviral treatment, household antiviral prophylaxis and extended antiviral prophylaxis: €4.63
Yarmand et al (2010)29	USA, 2009–2010 influenza season	North Carolina State University undergraduate students (23 087)	Vaccination	Self-isolation	High levels of interventionsSelf-isolation is incrementally more cost-effective than vaccinationThis has been presented for most of cost ratio values.Low levels of interventionsVaccination is incrementally more cost-effective than self-isolationThe results were robust, even in sensitivity analyses.
Sander et al (2009)30	USA, n/a	Residents of a 1.632-million-person city	HTAP25 with a stockpile for 25% of the population HTAP50 with a stockpile for 50% of the population HTAP with an unlimited stockpile School closure for 26 weeks Prevaccination 70% of the population with a low efficacy vaccine HTAP25+school closure HTAP50+school closure HTAP+school closure Prevaccination+school closure: prevaccinating 70% population with the low-efficacy vaccine, plus closing all schools for 26 weeks Treatment only: Treating all cases with antivirals FTAP25 for household contacts and 60% of work/school contacts, stockpile for 25% of the population FTAP50 for household contacts and 60% of work/school contacts, stockpile for 50% of population FTAP for household contacts and 60% of work/school contacts, stockpile unlimited FTAP25+school closure FTAP50+school closure FTAP+school closure	No intervention	Cost/capita and cost-effectiveness outcomes FTAP is cost-effective (54% reduction attack rate, €119 per capita) Prevaccination (48% reduction attack rate, €131 per capita) School closure in combination with each of the above is the least cost-effective (€2 524 per capita) ICUR of FTAP: €42 959ICUR of prevaccination and school closure: €43 106Cost-savingFTAP and prepandemic vaccination are cost-saving compared with no intervention
Pharmaceutical only strategies
Khazeni et al (2009)31	USA, n/a	A US metropolitan city (8.3 million)	Stockpiled strategy Expanded adjuvanted vaccination Expanded antiviral prophylaxis	No intervention	Intervention and treatment costs Stockpiled strategy: Total cost of €30.1 million and contribution to €288 million treatment costs Expanded adjuvanted vaccination: Total cost of €179 million and contribution to €166 million treatment costs Expanded antiviral prophylaxis: Total cost of €58.4 million and contribution to €266 million treatment costs No intervention: contribution to €462 million treatment costs Cost/QALY gained Stockpiled strategy compared with no intervention: €7894/QALY Expanded adjuvanted vaccination (at 80% effectiveness) relative to stockpiled strategy: €8600/QALY Expanded antiviral prophylaxis has a less favourable cost-effectiveness ratio than adjuvanted vaccination Expanded adjuvanted vaccination shown to be a cost-effective intervention because it contributes to 404 030 QALYs at $10 844 per QALY gained relative to stockpiled strategy.
Balicer et al (2005)32	Israel, n/a	Population of Israel (1 618 200 cases/patients)	Stockpiling with antiviral drugs Therapeutic use (all patients) Therapeutic use (high-risk patients) Pre-exposure long-term prophylaxis (all population) Pre-exposure long-term prophylaxis (high-risk population) Short-term postexposure prophylaxis for all close contacts	No intervention	CBATherapeutic use (incl. all and high-risk patients): 2.44–3.68Pre-exposure (incl. entire and high-risk population): 0.37–0.38Postexposure: 2.49Stockpiling with antiviral drugs for high-risk patients remain cost-saving strategy even if the annual probability of a pandemic remains >1 every 80 years.Overall costThe overall health-related costs: €56 234 057The overall cost to the economy: €535 245 986Workdays lost due to illness6 536 240 or 4 days/patient
Medema et al (2004)33	n/a,	Developed countries (1 billion people)	Egg-based vaccines with 17% population coverage Cell culture-based vaccines with 37% population coverage	No intervention	Cost per life-year gainedIn general, vaccination is cost-effective.Cell culture-based vaccines: €3376/LYG (cost-effective)Cost per interventionEgg-based: €2.6 billionCell culture-based: €5.87 billionNet savingsEgg-based: €8.5 billionCell culture-based: €5.87 billionSavings: €1.84 billion

CBA, cost-benefit ratio; CCR, community contact reduction; FTAP, full targeted antiviral prophylaxis; HTAP, household targeted antiviral prophylaxis; ICER, incremental cost-effectiveness ratio; ICUR, incremental cost-utility ratio; ISC, individual school closure; LYG, life-year gained; QALY, quality-adjusted life year; VSL, value of statistical life; WP, workplace closure.

With regard to studies that compared multiple interventions, a simulation model of pandemic influenza in the USA studied the cost-effectiveness of stockpile strategy and concluded that expanded adjuvanted vaccination seemed to be the most cost-effective strategy, averting 68% of infections and deaths and gaining 404 303 QALYs at $10 844 (€9600 €2017) per QALY gained relative to the stockpiling strategy.30 Saunders-Hastings et al,26 using a simulated population of 1.2 million people (reflective of Ottawa, Canada), performed a cost-effectiveness analysis of six interventions including vaccination, school closure, antiviral prophylaxis and other measures. The authors concluded that vaccination was the most cost-effective intervention when compared with other interventions while the least cost-effective intervention was school closure in conjunction with community-contact reduction, personal protective measures, voluntary isolation and quarantine. In particular, the cost per life-year saved was estimated to be $2581 (1700 €2017) for combined vaccination and antiviral treatment, while an estimated cost of $260 472/life-year saved (€171 590 €2017) was noted for school closure in conjunction with other interventions. Finally, Halder et al27 aimed to evaluate the most cost-effective strategies suitable for a future pandemic with H1N1 2009 characteristics in Australia. The results showed that the strategy with the lowest cost was the dual strategy of individual school closure for 2 weeks along with antiviral drug strategies, with a total cost of approximately AU$632 (376.31 €2017) per case averted. The strategy with the highest cost was the dual strategy of school closure along with the continuous—50% workplace closure, with a cost of $103 million (61.3 million €2017), per 100 000 population.

Comparative cost-effectiveness analysis

A DRM approach is presented in figure 2. These interventions include both pharmaceutical measures and non-pharmaceutical measures. The interventions were compared with the ‘no intervention’ scenario, with the exception of one study29 in which the comparators were vaccination versus self-isolation, which was subsequently excluded from the DRM.

Figure 2

Dominance ranking matrix for pharmaceutical and non-pharmaceutical strategies. *+: the intervention is less cost saving than the comparator; 0: the intervention is equally cost saving with the comparator; −: the intervention is more cost saving than the comparator. **+: The intervention is more effective than the comparator; 0: the intervention is equally effective with the comparator; −: the intervention is less effective than the comparator.

Pharmaceutical measures

Vaccination as a response measure

With the application of our inclusion and exclusion criteria, four studies assessed vaccination as a response measure in the context of an outbreak and included a cost analysis. Overall, as highlighted in the majority of the studies, vaccination as a response measure was noted to have a more significant clinical effect than comparators and was more cost-saving in most cases. According to Sander et al,30 the most clinically effective intervention was expanded adjuvant vaccination which contributed to 404 030 QALYs. Similarly, Khazeni calculated that with expanded adjuvanted vaccination, 45 941 deaths would be averted.31 Additionally, Saunders-Hastings et al26 concluded that the most cost-effective approach for controlling a pandemic was vaccination in combination with antiviral therapy and prophylaxis. However, a review of the results showed that much of the cost-effectiveness of pharmaceutical interventions were driven by vigorous vaccination campaigns, while the contribution of antiviral drugs’ was not of significance. Finally, Madema et al33 through a simulation model of an influenza pandemic among developing countries calculated the costs and assessed the effectiveness of two types of vaccines, an egg-based and a cell culture-based, in comparison with no intervention. Overall, vaccination was more cost-effective than no intervention; however, vaccination with cell culture-based vaccines was the most cost-effective strategy with a cost of 3779 €2017 per life-year gained. General vaccination was also assessed by Sander et al,30 who noted it to be both more cost-saving and effective than the unmitigated pandemic scenario, although when comparing prevaccination with low-efficacy vaccines with full targeted antiviral prophylaxis, it was less effective and more costly.

Antiviral drugs

Antiviral drug strategies were assessed in five studies, where it was noted that they were both more effective and cost-saving than the no intervention scenario, primarily when used as targeted prophylaxis. According to Halder et al,27 antiviral drug strategies such as antiviral treatment and antiviral treatment in combination with household confinement and extended prophylaxis can result in reduced attack rates of 7.6% and 3.5% in comparison to the unmitigated attack rate of 13%. The costs of these strategies are also lower than the cost of no intervention.

Moreover, therapeutic treatment and postexposure prophylaxis for exposed individuals (targeted prophylaxis) were shown to be the most cost-saving.32 Consistent with the above, antiviral therapy in combination with a layered non-pharmaceutical approach seemed to reduce the overall economic costs the most and was identified as more effective when compared with no intervention.26 Furthermore, it was noted that expanded antiviral prophylaxis could help delay a pandemic when additional strategies are implemented and would also lead to averting 32 745 deaths in the USA.31 Finally, Sander et al30 used a stochastic simulation model of pandemic influenza in the USA, aiming to evaluate the potential economic impact of 16 different mitigation interventions from a societal perspective. Conclusively, targeted antiviral prophylaxis was both the most cost-saving and effective intervention with a cost of $127 per capita (€118.73 €2017), with the scenario of implementation of expanded antiviral prophylaxis leading to a total of 282 329 QALYs gained.

Stockpile strategy

The stockpile strategy was assessed in three of the studies included in this systematic review. Based on the findings, stockpiling antiviral prophylaxis in the context of a pandemic was noted to be both cost-saving for the society and avert loss of life compared with no intervention.30 Moreover, prepandemic stockpiling of antiviral drugs would be more effective and cost-saving than no intervention if antiviral drugs were administered either solely as a treatment or as short-term prophylaxis for exposed individuals.32 Finally, stockpiling was also found more effective than a no intervention scenario (averting 29 761 deaths in the USA), although when compared with other interventions, expanded vaccination and prophylaxis were found to be more effective.31

Non-pharmaceutical measures

Pre hospitalisation screening

Lankelma et al23 assessed the cost-effectiveness of screening patients with acute respiratory tract infection for influenza before hospital admission. Overall costs of screening were estimated at 98 968 €2017 for 1546 tests and 624 cases and reported net savings of 388 317 €2017 for the healthcare system. Point-of-care testing for influenza before hospital admission was identified as a cost-effective intervention.23

Community contact reduction

Community contact reduction was assessed in two studies, where it was either implemented solely or in combination with other pharmaceutical and non-pharmaceutical measures. Home confinement was noted as cost-effective as a preventive measure in the context of influenza epidemics, if the proportion of compliance is adequate and infected individuals ask for medical assistance, regardless of the severity level of the pandemic.26 Isolation of infected individuals was found to be among the most effective interventions, whereas combined with community contact reduction, personal protective measures and antiviral treatment, self-isolation had the lowest cost.27

School closure

The effectiveness and the economic burden of school closure were evaluated in four studies, highlighting that the duration of school closure and potentially combined strategies significantly affect its impact. Sadique et al24 estimated the economic burden of school closure in the UK from a societal perspective and showed that the estimated costs of school closure were high, at 0.28–1.68 billion €2017 per week and the authors concluded that school closure was likely to significantly add an extra economic burden on the health system through staff absenteeism, even if school closure may delay infectious disease transmission. Similarly, Sander et al,30 who studied school closure as an additional intervention to full targeted antiviral prophylaxis or prevaccination found that while school closure further improves health outcomes (gaining 51 QALYs), it was the least cost-effective measure as it increased the total cost to society by $2700 per capita (€2524 €2017). Additionally, school closure produced only a small reduction in attack rate, whether implemented in combination with other interventions or alone.26 Finally, exclusive school closure for 2 weeks along with the continuous 50% workplace closure, antiviral treatment, household antiviral prophylaxis and extended antiviral prophylaxis, had the lowest illness attack rate (2.4%) and one of the lowest costs. On the contrary, school closure as a sole intervention to counterbalance infectious respiratory diseases was not a cost-effective measure.27

Personal protective measures

Personal protective measures such as face masks and hand hygiene were assessed in two of the included studies, noting that they could contribute to the control of a pandemic, dependant though on the exposed and susceptible individuals’ compliance rate, the setting and the overall burden of the respiratory pandemic.26 28 Tracht et al aimed to assess the cost-effectiveness of facemasks (N95 grade) in reducing the spread of pandemic (H1N1) 2009, using a simulation model of the US population and identified an economic burden of 728.28 billion €2017 (incl. direct and indirect costs). Notably, if masks are worn by 10% and 50% of the adult population of the US net savings were calculated at 418.75 billion €2017and 501.9 billion €2017, respectively. Hence, the use of face masks were identified as a cost-effective preventive measure depending on the population’s level of compliance.

Discussion

The aim of this systematic literature review of econometric analysis studies was to assess the economics of preparedness when contrasted with the cost of infectious respiratory disease outbreaks primarily within the context of European and OECD countries (excluding Japan and South Korea). Overall, the economic burden of infectious disease outbreaks is costly to healthcare systems, or to governments and society reflecting the medical costs for response activities including both the treatment of the confirmed cases and the surveillance and elimination of the disease’s transmission, as well as indirect costs which were also substantial.

In general, the majority of direct costs seemed to primarily reflect cost of additional personnel hours, which are mandatory for the management of the infected cases, for the organisation of response planning and contact tracing, for providing educational training and materials as well as laboratory costs. With regard to indirect costs, these could in many cases be greater than the direct costs, especially when school closures and/or workplace closures are enacted across a population, which in turn impact productivity and increase the economic burden.

While all the identified pharmaceutical and non-pharmaceutical interventions lead to a health benefit for the individual or the society, the cost benefit of such interventions differs. With regard to the potential non-pharmaceutical strategies, we identified that the use of personal protective measures, such as a facemask, is both cost-saving and effective, as also is prehospitalisation screening among suspect cases. On the other hand, all studies that assessed the impact of school closure noted that although it is an effective measure in reducing transmission, it is not cost-saving as it leads to increased economic burden. Moreover, when school closure was used as a sole intervention, then the use of limited duration school closure was significantly more cost-effective compared with continuous school closure.24 Community contact reduction was identified to have a positive health impact but had ambiguous results with regard to its potential cost saving as one study26 noted that it is a cost-saving intervention, while the other27 noted that social distancing strategies, such as reduced workplace attendance, were not a cost-saving measure primarily due to productivity losses, especially during longer periods of closure. Productivity losses primarily were noted to arise from pandemic related deaths and illness coupled with those losses due to interventions such as workplace closure and child-care of an ill child. It is important to note that non-pharmaceutical strategies were mostly applied complementary with a pharmaceutical measure or in combination with other non-pharmaceutical strategies in order to enhance their effectiveness. However, their cost-effectiveness highly depended on the duration, the level of compliance from the population and the type and burden of the infectious disease. It should moreover be noted that cost-effectiveness of measures will vary depending on the epidemiology of the disease in question.

With regard to pharmaceutical interventions, vaccination as a rapid response measure for infected and suspected individuals was noted to have a more significant clinical effect than comparators and was more cost-saving in most cases. As for antiviral treatment, the majority of the findings noted that it is a cost-effective strategy, especially when combined with other pharmaceutical and non-pharmaceutical interventions or when used as targeted prophylaxis for exposed individuals. Targeted antiviral prophylaxis was the most cost-saving and effective intervention, while stockpiling was cost saving in most cases and averted loss of life when compared with no intervention.

The current number of economic evaluation or cost-effectiveness studies of influenza outbreak preparedness measures is small, with an increase shown since the 2009 influenza pandemic; however, it is important to note that these studies refer to the evidence published before the COVID-19 pandemic. There are only a limited number of related reviews, however of different scope focusing primarily on policy recommendations34 or used dynamic transmission models in the included economic assessments of pandemic influenza preparedness measures based on significantly older studies.6 Additionally, most of the existing review studies either evaluate the overall economic burden of the disease or the cost-effectiveness of different pharmaceutical and non-pharmaceutical interventions without necessarily them reflecting the economics of outbreaks of infectious respiratory diseases.

Placing the above into context and following the assessment of the methodological approaches used across studies, it is essential to note the minimum contents that economic outbreaks of respiratory studies should include to help inform future and upcoming research, especially in light of the COVID-19 pandemic. These include clearly noting of the study year, the population at risk and, the population infected, the type of economic perspective (ie, healthcare, societal, etc), the study timeframe and discounting, as well as detailed reporting of the direct and indirect costs of the respiratory outcome and the interventions applied.

Strengths and limitations

A significant strength of this review is the comprehensive approach that was followed and the assessment of data quality—which indicated that the majority of the studies included were of high quality. Second, the synthesis of the results was performed using the DRM approach, which allowed for a direct comparison of the cost-effectiveness of each intervention to the null intervention.

However, there are a few limitations: first, costs and resources varied between different countries, different regional settings and over time, making the cost component comparison of cost-effectiveness measures complex to interpret. Moreover, we only focused on EU and OECD analogous high-income countries excluding Japan and South Korea, and hence our cost-effectiveness analyses are not applicable and generalisable to other countries and particularly middle-income and low-income countries. Additionally, discrepancies in context and populations likely affect the implementation and efficacy of interventions, undermining even the effectiveness elements comparability in the cost-effectiveness measures, especially in complex multi-component public health interventions. In addition, our study did not include studies published before 2003 or after 2019. Also, it should be noticed that publication bias may exist due to the English language restriction applied. Another limitation to be noted is that this review excluded seasonal influenza outbreaks since these occur on a yearly basis. Furthermore, this study was performed before the impact of COVID-19 and hence reflects the published knowledge before the current pandemic. Thus, the results cannot be directly extrapolated to the COVID-19 pandemic.

Conclusion

The value of this systematic review of econometric studies is to provide a synthesis of the evidence of the cost of respiratory infectious disease outbreaks and the cost-effectiveness of specific interventions that can be applied in response. Furthermore, our assessment identifies a minimum number of econometric measures which should be recorded during the reporting of respiratory infectious disease outbreaks that would aid future decision making. Our cost analysis results give evidence to public health policymakers, primarily in the EU or the USA, as to the cost-effectiveness of a range of pharmaceutical and non-pharmaceutical intervention strategies which may be applied to mitigate or respond to infectious respiratory disease outbreaks.