- Research article
- Open Access
- Open Peer Review
Estimating the costs of school closure for mitigating an influenza pandemic
© Sadique et al; licensee BioMed Central Ltd. 2008
- Received: 17 July 2007
- Accepted: 24 April 2008
- Published: 24 April 2008
School closure is a key component of many countries' plans to mitigate the effect of an influenza pandemic. Although a number of studies have suggested that such a policy might reduce the incidence, there are no published studies of the cost of such policies. This study attempts to fill this knowledge gap
School closure is expected to lead to significant work absenteeism of working parents who are likely to be the main care givers to their dependent children at home. The cost of absenteeism due to school closure is calculated as the paid productivity loss of parental absenteeism during the period of school closure. The cost is estimated from societal perspective using a nationally representative survey.
The results show that overall about 16% of the workforce is likely to be the main caregiver for dependent children and therefore likely to take absenteeism. This rises to 30% in the health and social care sector, as a large proportion of the workforce are women. The estimated costs of school closure are significant, at £0.2 bn – £1.2 bn per week. School closure is likely to significantly exacerbate the pressures on the health system through staff absenteeism.
The estimates of school closure associated absenteeism and the projected cost would be useful for pandemic planning for business continuity, and for cost effectiveness evaluation of different pandemic influenza mitigation strategies.
- Informal Care
- Dependent Child
- Labour Force Survey
- School Closure
- Main Caregiver
Pandemic influenza has been a national and international public health concern for many years. The continuing global spread of the H5N1 strain in birds, and associated human cases has highlighted that a pandemic can occur at any time. As a result many countries have revised and updated their pandemic plans. In such plans the use of non-pharmaceutical interventions (NPI) are proposed to help reduce the number of cases and slow the epidemic spread, particularly if vaccines or antivirals are unavailable or become ineffective because of resistance . School closure is one of the key components of many countries' non-pharmaceutical mitigation strategies [2, 3] because of the propensity of influenza epidemics to be amplified in school settings . Although there are a number of studies that have attempted to estimate the possible epidemiological impact of school closure [5–11], there are no published studies of the cost of this strategy. The major concern with closing schools is that it will result in an increase in absenteeism due to childcare needs. This could be expected to have adverse consequences for business continuity and costs to the economy. This paper aims to estimate the economic cost of school closure to the United Kingdom. In addition those sectors likely to incur the greatest costs are identified.
The impact of school closure is measured in terms of lost income from missed work of working parents as a consequence of school closure. In this study, the cost is estimated from the societal perspective which calculates the paid productivity loss of parental absenteeism during the period of school closure. The two methods that are most widely used for the valuation of productivity cost are the 'human capital method' (HCM) and the 'friction cost method' . The HCM estimates the value of potential lost production (or income) from a financial point of view . The friction cost method is based on the idea that the amount of production lost as a result of disease (or event, e.g., school closure) is confined to the period needed to replace a sick worker (often assumed to be around three months) [14, 15]. Since a pandemic is only expected to last about three months, the friction method and HCM would be expected to give similar results. We have used the HCM here because of the short time horizon .
The spring 2005 Labour Force Survey (LFS) weighted regional dataset was used  to estimate the absenteeism and its consequent cost. We first estimate the proportion of the workforce that are likely to be the main caregivers and that have dependent children. Dependent children is defined, here, as those under 16 years of age. This definition was chosen as the Labour Force Survey records the living arrangements of adults with children under 16 years. We have therefore estimated the proportion of labour force that are main care givers for the children under 16 years of age in the household (women who are either the head of the household or the spouse of the head of the household, or are cohabiting with the head of the household and who do not have other adults in the household, or men who do not have other adults in the household, but do live with children under the age of 16 years). This gives the number of working parents who are likely to be the main carer for children and may be most liable to take absenteeism to care for these children. These figures are estimated in the following way.
Total number of working mothers (single mother or cohabiting with spouse or partner) who are assumed to be the primary caregivers of children. We expect that a major part of absenteeism cost will accrue from these working mothers.
Total number of single/lone fathers who are likely to be absent from work to take care of their child during school closure. The Equal Opportunity Commission  reports that the number of families headed by lone fathers in 2005 was 180,000. Given the employment rate of 78%, we get 140,400 working lone fathers.
The above two components give us the number of parent workers who are likely to be absent from work. These estimates are made for the UK workforce as a whole, and by sector excluding those where the information is not known or classified as working outside the UK. The aggregate estimate is also adjusted by the presence of grandparents within the same household. It is assumed that grandparents are most likely to provide alternative care to their grandchild in case of emergency which will allow the respective parents to work.
The absenteeism rate is also calculated in terms of work-days lost per week. This is estimated using the gender-specific employment rate in each industrial sector and patterns of employment (full time and part-time) by gender. In 2005, 39% of women with dependent children and 22% of women without dependent children in the workforce were part-timers, the corresponding proportions for men were 4% and 9% respectively .
We then translate the loss of working hours/days per worker into monetary values according to their wages. According to the neoclassical economic model, wage rates equal the value of marginal revenue generated by an additional worker under full employment, and therefore it reflects the value of lost production. We derived the average weekly wage from the sector specific wage rate (variable gross weekly wage (GRSSWKC) by variables SEX, and industry division in the main job (INDS92M)) from Quarterly LFS 2005 April  and also used average wage from a Department for Work and Pensions (DWP) survey . We also adjust for the fact that 39% of working mothers are part-timers and the rest work full-time . There is no specific number of hours that makes someone full or part-time, but we assume that a part-timer works half of the hours that a full timer works.
The cost of absenteeism was estimated from two approaches-aggregate and by industrial sector. Firstly, we calculated the cost of absenteeism using aggregate absenteeism figures from LFS and average wage from DWP survey. Alternatively in the industrial sector approach, we estimate the absenteeism for each industrial sector (based on LFS statistics). This industry specific absenteeism was then evaluated in monetary values using both sector specific wage (from LFS) and average wage from DWP. These industry costs were aggregated to give an aggregate national estimate.
The baseline estimates do not allow for any other informal care, apart from that provided by other adults in the household. A report from the Office of National Statistics  suggests that overall 54% of working mothers use informal childcare by friends, neighbours, family or childminders for all or part of their childcare. A similar estimate is also reported by the Department for Education and Skills . In our scenario analysis, the cost of absenteeism is adjusted by this figure.
It has been suggested that effective labour-time is reduced less than proportionately from absenteeism as work colleagues may be able provide some cover, and workers may be able to catch-up on some delayed tasks when they return. Although no estimates of this elasticity of production with respect to labour are available for the UK, estimates for the Netherlands suggest a value of 0.8. We have also adjusted our cost estimates by this factor in the scenario analysis. It is also probable that some individuals (particularly in certain sectors) would be able to perform some work whilst caring for their children. We were not able to find an estimate for this parameter. As a proxy we used the figure that 30% of all UK households have access to broadband , and assumed, in our scenario analysis that this would reflect the average productivity of workers at home. As a best-case (least cost) scenario we adjusted for the proportion who have access to informal care, the elasticity of production, and the proportion of parents assumed to be able to work from home simultaneously.
We assume in the base-case that, with the exception of lone-fathers, women will take absenteeism to care for children. In the scenario analysis we estimate the cost of school closure if 50% of the absenteeism results from men taking time off to care for their dependent children at home.
Estimates of the cost of school closure are presented in 2005 prices, and are given for the UK as a whole (population ~60 million) and by week. The cost of school closure is assumed to be proportional to the length of school closure.
Estimated proportion of work-days lost through school closure by sector
Absenteeism rate (in days)
Agriculture, hunting & forestry
Mining & quarrying
Electricity, gas & water supply
Wholesale, retail & motor trade
Hotels & restaurants
Transport storage & communication
Real estate, renting & business activities
Public administration & defence
Health & social work
Other community social & personal
Private households with employed persons
Extra-territorial organisations bodies
Weekly cost of school closure (in million £) under a range of different assumptions. LFS represents estimated costs based on the Labour Force Survey and DWP based on Department of Work and Pensions estimates of wages.
From aggregate figure
Informal care adjusted
Effective labour adjusted
Work from home adjusted
Effective labour, informal care, and work from home adjusted
Absenteeism of 50% working fathers
Range cost of school closure as a % of GDP.
Informal care 54%
Effective labour 80%
Work from home 30%
Effective labour, informal care, & work from home adjusted
Absenteeism of 50% working fathers
Adjusting for informal care reduces the estimated cost of school closure to between £398 million to £453 million per week, as compared to our baseline estimate of ~£1 billion per week (Table 2). Adjusting for the elasticity of production, suggests that the aggregate cost of school closure reduces to £692–£788 million per week. The cost of productivity loss when adjusted by the proportion of workforce able to work from home reduces to £605–£690 million per week. When effective labour, informal care, and working from home are adjusted simultaneously, the productivity cost of school closure falls even further (in the range of £222 million to £254 million per week). If men are as likely to take absenteeism to care for dependent children as women, then the estimated cost of school closure is increased by about 30% (Table 2) of the base case estimate, as average male wages are higher. The cost as a % of GDP for the above adjustments is also reported in Table 3.
By no means all of the workforce would be affected by a policy of school closure. We estimate that about 16% of the workforce may take absenteeism because they are probably the main care-giver of dependent children. Nevertheless this level of absenteeism could have a significant impact on the economy. Our estimates suggest that the cost of school closure would be £0.2 bn to £1.2 bn per week. This amounts to around 0.2–1% of GDP for school closure lasting the duration of a pandemic wave (around 12 weeks). These estimates exclude the possible inflationary effects that could result from prolonged school closure, from the increasing costs to firms . Although the motivation for estimating these costs is pandemic planning, they are not specific to an epidemic of influenza, and would arise from other sources of school closure (such as a national teacher's strike).
Our estimates are dependent on a number of assumptions and are subject to significant uncertainty, as shown by the scenario analysis. It is clear from this, that the proportion of individuals who may be able to access informal care is a critical determinant of the cost of a school-closure policy. However, it should also be remembered that the use of informal care over prolonged periods may be difficult to arrange, and could reduce the desired benefits of school closure, if children are kept in relatively large informal groups. A further assumption is that workers who are at home caring for their dependent children are unable to work. It is likely that many workers in certain sectors of the economy, would be able to do some work from home, at least on a part-time basis. We were not able to obtain specific estimates of what proportion of the workforce might be able to work from home, and what level of productivity would likely result from such arrangement. The most appropriate proxy measure we could find was the proportion of homes that have access to broadband (with the implicit assumption that those that do have access are able to work at full capacity). Our estimates of the impact of this are correspondingly speculative.
The health-care sector employs a high proportion of women, compared to other sectors of the economy, and home-working is likely to be virtually impossible for many staff. As a result, this sector is likely to be most badly affected by a policy of school closure. Our estimates suggest that about a third of the workforce in this sector have dependent children under 16 years in the home, and might therefore have to take time off to care for their children. This is absenteeism in the absence of illness. At the peak more than 10% of the workforce would be expected to be absent through pandemic influenza  and about 5% from other causes . That is, school closure, plus illness absenteeism, could reduce the workforce of the health sector by 45% at the peak of the epidemic – up to a third through school closure (assuming no adjustments due to informal care arrangements etc), and around 15% of the remainder from illness. Such levels of absenteeism, coupled with the dramatic increase in demand for health services that would be expected during a pandemic, would put the remainder of the health-related work force under severe strain. Note also, that if our estimates are broadly applicable, then the health service will be severely stretched throughout the period of school closure, and that even at the peak of the epidemic the majority of absenteeism in this sector might result from school closure rather than staff illness.
There are a number of limitations of this study. These include significant uncertainty in a number of important parameters (as mentioned above). In addition, however, we have not included any possible long term cost of school closure, such as extra teaching efforts required to bring the children up to level, or the cost associated with later entry into the labour force, or the costs resulting from a poorer educated workforce. Such costs would be difficult to estimate, and by ignoring them out analysis will be conservative in this respect.
Epidemiological  and modelling studies [5–7, 9, 10] have suggested that school closure can help in mitigating the effects of a pandemic, – particularly by reducing illness rates in children. This is the first study that we are aware of that has attempted to estimate the cost of such a policy. The estimates of school closure associated absenteeism and the projected cost would be useful for pandemic planning for business continuity, and for cost effectiveness evaluation of different pandemic influenza mitigation strategies. The potential benefits of school closure need to be weighed against the possible costs, to determine the best course of action. That is, although the costs of school closure, as estimated here, might appear to be large, the benefits of the policy, in terms of cases and deaths prevented and consequent savings to the health sector and society, might be acceptable. Only a full economic analysis can shed light on whether such a policy should be adopted.
This work was part of "SARSControl: Effective and acceptable strategies for the control of SARS and new emerging infections in China and Europe", a European Commission project funded within the Sixth Framework Programme, Thematic Priority Scientific Support to Policies, Contract number: SP22-CT-2004-003824.
- Mascona A: Oseltamivir resistance-disabling our influenza defences. N Eng J Med. 2005, 353 (25): 2633-2636. 10.1056/NEJMp058291.View ArticleGoogle Scholar
- Institute of Medicine: Modeling Community Containment for Pandemic Influenza-A Letter Report. 2006, The National Academy PressGoogle Scholar
- Glass RJ, Glass LM, Beyeler WE, Min HJ: Targeted social distancing design for pandemic influenza. Emerging Infectious Disease. 2006, 11: 1671-1681.View ArticleGoogle Scholar
- Aledort J, Lurie N, Wasserman J, Bozzette SA: Non-pharmaceutical public health interventions for pandemic influenza: an evaluation of the evidence base. BMC Public Health. 2007, 7: 208-10.1186/1471-2458-7-208.View ArticlePubMedPubMed CentralGoogle Scholar
- Bootsma MC, Ferguson NM: The effect of public health measures on the 1918 influenza pandemic in U.S. cities. PNAS. 2007, 18: 7588-93. 10.1073/pnas.0611071104.View ArticleGoogle Scholar
- Hatchett RJ, Mecher CE, Lipsitch M: Public health interventions and epidemic intensity during the 1918 influenza pandemic. PNAS. 2007, 18: 7582-7. 10.1073/pnas.0610941104.View ArticleGoogle Scholar
- Vynnycky E, Edmunds WJ: Analyses of the 1957 (Asian) influenza pandemic in the United Kingdom and the impact of school closures. Epidemiol Infect. 2007, 1-14.Google Scholar
- Heymann A, Chodick G, Reichman B, Kokia E, Laufer J: Influence of school closure on the incidence of viral respiratory diseases among children and on health care utilization. Pediatr Infect Dis J. 2004, 7: 675-7.View ArticleGoogle Scholar
- Ferguson N: Strategies for mitigating an influenza pandemic. Nature. 2006, 442: 448-452. 10.1038/nature04795.View ArticlePubMedGoogle Scholar
- German TC, Kadau K, Longini IM, Macken CA: Mitigation strategies for pandemic influenza in the United States. PNAS. 2006, 103: 5935-40. 10.1073/pnas.0601266103.View ArticleGoogle Scholar
- Ferguson N, Cummings D, Cauchemez S, Fraser C, Riley S, Meeyai A, Lamsirithaworn S, Burke D: Strategies for containing an emerging influenza pandemic in Southeast Asia. Nature. 2005, 437: 209-214. 10.1038/nature04017.View ArticlePubMedGoogle Scholar
- Berger ML, Murray JF, Xu J, Pauly M: Alternative valuation of work loss and productivity, Health and Productivity Management Center. 2001, American College of Occupational and Environmental MedicineGoogle Scholar
- Borghouts JA, koes BW, Vondeling H, Bouter LM: Cost-of-illness of neck pain in the Netherlands in 1996. Pain. 1998, 80: 629-636. 10.1016/S0304-3959(98)00268-1.View ArticleGoogle Scholar
- Koopmanschap MA, Rutten FH, van Ineveld BM, van Roijen L: The friction cost method for measuring indirect costs of disease. Journal of Health Economics. 1995, 14: 171-189. 10.1016/0167-6296(94)00044-5.View ArticlePubMedGoogle Scholar
- Koopmanschap MA, Rutten FH: A practical guide for calculation of indirect cost of disease. Pharmacoeconomics. 1996, 10: 460-466.View ArticlePubMedGoogle Scholar
- Sadiraj K, Groot I, Heyma A: How to measure production loss in economic evaluations?. 2006, Discussion paper no 49, Research for the Coronel Institute, AmsterdamGoogle Scholar
- Office of National Statistics: 2005 Spring Labour Force Survey. [http://www.statistics.gov.uk/STATBASE/Product.asp?vlnk=14248]
- Equal Opportunities Commission: Facts about men and women in Britain. 2006, [http://www.eoc.org.uk/pdf/facts_about_GB_2006.pdf]Google Scholar
- Office of National Statistics: Focus on gender. 2006, [http://www.statistics.gov.uk/statbase/product.asp?vlnk=10923]Google Scholar
- Office of National Statistics: 2005 Quarterly Labour Force Survey April. 2008, Colchester, Essex: UK Data ArchiveGoogle Scholar
- Women & Equality Unit: National Statistics Publication on Individual Incomes of men and women. 2006, [http://www.womenandequalityunit.gov.uk/indiv_incomes/index.htm]Google Scholar
- Office of National Statistics: Social Trends 34. 2004, [http://www.statistics.gov.uk/downloads/theme_social/Social_Trends34/Social_Trends34.pdf]Google Scholar
- Department for Education and Skills. Sure Start Research and Evaluation Division and National Centre for Social Research: Childcare and Early Years Provision: Parents' Use, Views and Experiences, 2003–2005. 2007, Colchester, Essex: UK Data ArchiveGoogle Scholar
- Office of Communication: More broadband, more digital, more mobile – a common picture of accelerating take-up of new products and services. [http://www.ofcom.org.uk/media/news/2005/07/nr_20050713]
- Keogh-Brown , Wren-Lewis S, Edmunds WJ, Smith RD: Calculating the macroeconomic effects on the UK of an influenza pandemic (submitted).Google Scholar
- Information Centre for health and social care: Sickness absence rates of NHS staff in 2004. [http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsStatistics/DH_4112061]
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2458/8/135/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.