Health and Economic Burden of Seven Foodborne Diseases in Denmark,2019

Abstract

We ranked seven foodborne pathogens in Denmark on the basis of their health and economic impact on society in 2019. We estimated burden of disease of infections with Campylobacter spp., Salmonella spp., Shiga toxin-producing Escherichia coli (STEC), Yersinia enterocolitica, Listeria monocytogenes, norovirus, and hepatitis A virus in terms of incidence, mortality, disability-adjusted life years (DALY), and economic burden in terms of direct and indirect health costs. These seven pathogens accounted for 268,372 cases, 98 deaths, and 3121 DALYs, and led to a total expenditure of 434 million Euro in 1 year in a country with 5.8 million citizens. Foodborne infections by Campylobacter, Salmonella, and norovirus caused the most DALYs, whereas Campylobacter, and norovirus and STEC had the higher costs. A combination of disease burden and cost of illness estimates is useful to inform policymaking and establish food safety priorities at the national level.

Introduction

Estimating the societal impact of foodborne diseases in a population is important for guidance of policy decisions and efficient targeting of preventive measures. However, performing such an estimation is not trivial. First, foodborne diseases are largely underdiagnosed and underreported, a phenomenon common to all countries globally and associated with the fact that only a fraction of patients seek or access health care; that doctors may not request and submit a sample of cause-specific diagnosis; or with laboratory sensitivity and results reporting (Haagsma et al, 2012; Scallan et al, 2011). Second, diseases vary largely in terms of incidence, sequelae, duration, and severity, and assessing their health impact in the population requires that these differences are accounted for.

Such differences affect the way patients experience the disease, the costs associated with treatment, and societal losses associated with reduced productivity and reduced quality of life (Jaffee et al, 2019; Mangen et al, 2015). Finally, not all diseases caused by pathogens commonly associated with foods are exclusively foodborne; exposure to many of these pathogens can also be associated with contaminated environment like water and soil, or with contact with contaminated live animals or humans (Pires et al, 2009).

The aim of this study was to estimate the burden of commonly reported foodborne pathogens in Denmark according to both the overall disease burden and associated costs in the population.

Methods

We estimated the disease and economic burden caused by seven enteric pathogens (four enteric bacteria: Salmonella, Campylobacter, Yersinia enterocolitica, Shiga toxin-producing Escherichia coli (STEC); two viruses: norovirus, hepatitis A virus (hepA); and one invasive bacterial pathogen: Listeria monocytogenes. Pathogens were selected on the basis of their public health relevance, reflected either by reported incidence, number of outbreaks, or severity of symptoms. Among these seven agents, six (Salmonella, Campylobacter, Y. enterocolitica, STEC; hepA; and L. monocytogenes) were notifiable in Denmark, that is, laboratory-confirmed cases were required to be reported to Statens Serum Institut (SSI) by the diagnosing clinical laboratory (see below).

Burden of disease measured in disability-adjusted life years

To estimate the foodborne disease burden for all pathogens, we (1) estimated the incidence and mortality of each; (2) estimated the disease burden of all health outcomes of each pathogen in terms of disability-adjusted life years (DALY); and (3) linked these estimates with estimates on the proportion of the total burden that is attributable to foods. Detailed methodology of each step is presented elsewhere (Pires et al, 2019).

Estimating incidence and mortality

In 2019 in Denmark, physicians sent specimens from suspected cases of notifiable diseases to one of the 10 existing clinical microbiology laboratories. Laboratories are required to report positive results to the National Infectious Disease Institute (SSI) within 1 week; for some agents, reporting took place by automated data capture through the Danish Microbiology Database (Voldstedlund et al, 2014). Case patients are reported in a person-identifiable format, using the Danish cpr-number, thus avoiding double registrations (Pedersen, 2011). Cases of bacterial gastrointestinal infections are recorded in the Register of Enteric Infections maintained by SSI (https://statistik.ssi.dk/). Only first-positive cases (episodes) are recorded, that is, each patient-infectious agent combination is only recorded once in any 6-month period. To adjust for underdiagnosis and underreporting of these infections, we reconstructed the surveillance pyramid (Haagsma et al, 2012; Pires et al, 2019).

We accounted for differences in the underreporting of pathogens by including a set of nonpathogen-specific and pathogen-specific parameters, informed by data collected through a population-based telephone survey conducted in 2009 (Müller et al, 2012), from National Health Registries, and literature review, defined by probability distributions (Pires et al, 2019). Estimated multipliers were applied to surveillance data from 2019 (available at www.ssi.dk/data). Reported Y. enterocolitica cases did not include cases caused by biotype 1A, which are considered nonpathogenic (EFSA, 2007). We did not differentiate between pathogens' subtypes. Data were segregated in six age categories and by sex to account for differences in the incidence and health outcomes of disease in different age and sex groups.

The reporting system for norovirus infections in Denmark only captures outbreak-related cases, which represent a small fraction of total cases and are insufficient to describe the epidemiology and public health impact of the disease (Korcinska et al, 2020). To estimate the total incidence of norovirus infections in Denmark, we collected the total national diarrhea incidence and mortality envelopes as published by the Global Burden of Disease Study (available at http://ghdx.healthdata.org/gbd-results-tool). Then we estimated the proportion of diarrhea cases/deaths attributable to norovirus based on data extracted from a systematic review (Ahmed et al, 2014), and estimated the number of norovirus-specific cases and deaths by multiplying the incidence and mortality envelopes with the attributable proportions. To account for asymptomatic carriage, we defined a norovirus attributable factor based on prevalence studies with healthy controls (Ahmed et al, 2014). Finally, to account for norovirus illnesses not associated with diarrhea, but which were vomiting only, we defined a vomiting inflation factor based on prevalence data from the same study.

For hepA infections, we assumed a fixed underreporting multiplier, extracted from Havelaar et al (2012). We estimated incidence of hepA infections by multiplying the age- and sex-categorized reported cases by this multiplier.

We assumed that all invasive listeriosis patients in the population were diagnosed and notified to the public health surveillance system. Age and sex-specific incidence and mortality of listeriosis cases in 2019 were collected from the National Listeria Surveillance database. Under Danish surveillance, pregnancy-associated infections are notified as a single case (the woman), and fetal loss or still-born babies are thus not recorded.

Disability-adjusted life years

We accounted for pathogen-specific sequelae as presented in Table 1. The possible health outcomes of infection with bacteria and associated probabilities were identified through literature review (Pires et al, 2019). For listeriosis and hepA infection, we followed the model defined by the European Center for Disease Control and Prevention (ECDC, 2015). For more details on how we estimated the incidence of specific sequalae for each pathogen, see Supplementary Appendix Table SA1. We calculated DALYs as the sum of years lived with disability (YLD), and the years of life lost due to premature death (YLL) caused by a disease (Devleesschauwer et al, 2014). For each disease, we combined the estimated incidence of each health outcome with disability weights (DW) collected from Salomon et al (2015) with duration of disease and life expectancy statistics as published by Statistics Denmark (available at http://www.statbank.dk/HISB8) to estimate YLD.

Table 1.

Input Parameters for the Calculation of Disability-Adjusted Life Years of Five Foodborne Pathogens (Salmonella, Campylobacter, Yersinia enterocolitica, Shiga Toxin-Producing Escherichia coli, and Norovirus)

Health outcome	Duration (days)	DW				References
Diarrhea Salmonella, Campylobacter, and Yersinia	PERT (1.0, 3.0, 7.0)	Mild: 0.074 (0.049–0.104)	Moderate: 0.188 (0.125–0.264)	Severe: 0.247 (0.164–0.348)	Overall: PERT (0.0817, 0.05, 0.123)	Salomon et al (2015)
Diarrhea STEC	PERT (5.0, 7.0, 10.0)	Mild: 0.061 (0.036–0.093)	Moderate: 0.202 (0.133–0.299)	Severe: 0.281 (0.184–0.399)	Overall: PERT (0.0817, 0.05, 0.123)	Majowicz et al (2014); Salomon et al (2015)
Diarrhea gastroenteritis	PERT (1.0, 3.0, 7.0)	0.072				Kirk et al (2015)
Reactive arthritis	222	Not visiting GP: 0.127	Visiting GP: 0.21	Hospitalized: 0.37		Salomon et al (2015)
IBS	1825	0.042				Haagsma et al (2010)
GBS (Campylobacter)	Life-long	Mild: 0.090	Severe: 0.280	Residual symptoms: 0.160		Havelaar et al (2000)
HUS (STEC)	PERT (14.0, 28.0, 42.0)	0.21				Kirk et al (2015)
End-stage renal disease (STEC)	Life-long	PERT (0.397, 0.573, 0.749)				Havelaar et al (2004)

DW, disability weights; GBS, Guillain–Barré syndrome; GP, general practitioner; HUS, hemolytic uremic syndrome; IBS, irritable bowel syndrome; PERT, PERT distribution; STEC, Shiga toxin-producing E. coli.

To estimate YLL, we combined the estimated mortality with Standard Expected Years of Life Lost (WHO, 2018). Table 1 presents all input parameters for DALY calculations. To estimate the associated uncertainty, we applied a stochastic model using the DALY Calculator interface developed in R (http://daly.cbra.be/).

Foodborne burden of disease

To estimate the burden of disease that was due to consumption of contaminated foods, we applied the attributable foodborne proportions to the total disease burden, previously estimated by Hald et al (2016). Because estimates were not produced at a national level, we applied estimates for the subregion that includes Denmark (WHO subregion EUR-A), which corresponds to European countries with very low child and adult mortality (https://www.who.int/choice/demography/mortality_strata/en/). For simplification purposes, we multiplied point estimates and did not quantify associated uncertainties in this model step.

Cost of illness

The basic model

The societal economic burden of the seven pathogens was estimated using an extended cost of illness (in short, COI) method, building on the sex- and age-specific incidence and mortality estimates as described above, incidence, and mortality estimates for different health outcomes. Traditional COI analyses include direct cost in the health sector and indirect costs of lost productivity (see e.g., Kuchler and Golan, 1999). However, there is an increasing focus on the hidden costs for the individual and for society associated with the unpaid time lost to illness (Sendi and Brouwer, 2004; Verbooy et al, 2018). Time that is spent on going to the doctor or being ill at home or at the hospital may not only replace working hours, but also time spent on leisure activities, unpaid work for people, or other valuable activities outside the paid workforce. Consequently, we included estimates of the value of unpaid time lost to morbidity and premature deaths in excess of the working hours lost. The economic cost of premature death was estimated as the value of statistical life years lost, based on findings from the literature.

For all pathogens, we distinguished between five categories of severity of illness [inspired by Hoffmann et al (2015)]:

Mild cases that are not in contact with a general practitioner (GP)

A GP is visited

Hospitalization (mild symptoms with recovery)

Hospitalization (severe symptoms with recovery)

Hospitalization (nonrecovery with death)

We assumed that all cases visiting a GP or hospitalized that were diagnosed with one of the infections were registered, whereas mild cases that were not registered were captured by corrections for underreporting (see above). The only exception was hepA, for which we assumed that around 80% of the nonregistered mild cases visited a GP but were not diagnosed with hepA, as assumed in the Netherlands by Havelaar et al (2012). We assumed that a patient always visits the GP before going to hospital. The duration of hospitalization and subsequent recovery period was assumed to vary across pathogens, whereas premature death was assumed to be preceded by 7 days of hospitalization for all pathogens.

The cost of illness for any pathogen j was estimated as:

where n_j is the total number of cases of pathogen j and p_ji is the conditional probability of pathogen j being in severeness category i. The cost item c_ji is the medical expense per case of pathogen j for a person in severeness category i, h is the hospitalization cost per day, and k_ji is the average number of days at hospital for pathogen j in severeness category i; w is the average daily wage rate and d_ji is the average number of days off work (including days at hospital) for pathogen j in severity category i. Unpaid time lost to illness l_ij was measured as number of hours lost for pathogen j in severity category i, multiplied by the value of an unpaid hour u. m_ji is an indicator variable assuming the value 1 if pathogen j in severeness category i is fatal and 0 otherwise. V is the value of remaining statistical life years, where the expected life expectancy depends on the age of the individual patient.

Direct costs of health care

Direct costs of health care included expenditures to medicine, GP visits, laboratory diagnostics, and hospital expenses. These costs capture the direct financial burden for the health care system. As there rarely exist precise registrations of medicine expenditures, the number of consultations with a GP or hospital expenses at patient levels for individual diseases, we estimated these costs using available data from the literature (Table 2).

Table 2.

Overview of Assumptions of Illness Characteristics Used to Estimate Economic Burden of Disease

		Pathogen
		Salmonella	Campylobacter	STEC	Yersinia	Norovirus	Hep A	Listeria
Underreporting		7.7	10.9	19.6	10.9	30.1	1.2	1.0
Probability of health outcomes
Mild cases—no GP	%	87.0	90.9	94.9	90.8	96.7	14.6	0.0
GP-only cases	%	7.8	7.0	3.7	7.4	3.0	79.0	0.0
Mild hospitalization cases	%	5.0	2.0	1.1	1.5	0.3	4.0	14.5
Severe hospitalization cases	%	0.0	0.1	0.2	0.3	0.0	2.0	57.2
Fatal cases	%	0.1	0.1	0.1	0.1	0.0	0.4	28.3
GP visits per reported case	no. visits	1.0	1.0	1.0	1.0	1.0	2.1	1.0
Hospitalization days per hospitalized case	days	2.1	2.3	8.1	2.9	2.3	4.8	15.0
Duration of illness (days)
Per mild case	days	2	2	2	2	2	14	3
Per GP case	days	5	5	5	5	4	14	5
Per hospitalized case	days	9.1	9.6	12.8	11.7	10.1	35.0	33.0
Unpaid time lost
Per mild case	hours	22	22	22	22	22	154	33
Per GP case	hours	55	55	55	55	44	154	55
Per hospitalized case	hours	101	105	141	128	111	385	363

Cost of GP consultations: €18.50 per visit (Lægeforeningen, 2014), prescription of drugs €26.85 per prescription (own estimate), cost of sample analysis €10.55 per sample (Sundhedsstyrelsen, 2006), cost of specialist treatment: €65.77 per treatment (own estimate), cost of hospitalization: €537 per day (Sundhedsstyrelsen, 2006).

Sources: Christensen and Jensen (2017), Pires et al (2019).

GP, general practitioner; STEC, Shiga toxin-producing Escherichia coli; Hep A, Hepatitis A virus.

For each pathogen and for each of the five severity categories, we estimated average costs of medication, number and costs of GP visits, duration, and costs of hospitalization, expressed in 2019 price level. Descriptions of disease outcomes and distribution across categories of illness were based on international (Gadiel, 2010; Havelaar et al, 2012; Hoffmann et al, 2015; Korsgaard et al, 2009; Mangen et al, 2013a; RL, 2012; Scallan et al, 2011; Sundström, 2010; Tam and O'Brien, 2016; Williams et al, 2019) and Danish literature (Lægeforeningen, 2014; Pires, 2014; Pires et al, 2019; Sundhedsstyrelsen, 2006).

The costs of lost time

The costs of lost productivity due to illness causing temporary or permanent reductions in working ability or premature death are often estimated in present value terms using a discount rate of 4%. To estimate the costs of productivity loss, we used public statistics (Statistics Denmark, n.d.) on wage rates and labor market participation rates for men and women in different age intervals. For each category of illness, the associated duration of absence from work was estimated based on the literature (Gadiel, 2010; Hoffmann et al, 2015; Mangen et al, 2013b; RL, 2012). To calculate the productivity loss associated with death or permanent reduction in work ability, we assumed a retirement age of 65 years—and that life years lost after the age of 65 did not contribute to the estimated productivity loss. For illnesses of children, absence from work of one parent was included. Table 2 shows an overview of assumptions used to estimate economic burden of disease. For further details, see Christensen and Jensen (2017).

Time costs related to loss of unpaid activities during illness were calculated using estimates of leisure time. We assumed that an average person in Denmark has 11 h of leisure time per 24 h, consisting of nearly 3 h of housework and 8 h available for active or passive leisure activities (DØR, 2016). Leisure time was valued in monetary terms using the average wage rate after tax [as suggested in Johannesson (1996)] estimated as 19.19 Euro per hour. As this valuation is subject to uncertainty, we performed a sensitivity analysis, where this rate was reduced by 40%.

The value of a statistical life year

Determining the association between the value of a statistical life and the value of statistical life years involves ethical considerations, as well as empirical evidence on, for example, their dependence on age, type of illness, and income (Ananthapavan et al, 2021). We followed the line of thinking of DØR (2016) and Ananthapavan et al (2021), and assumed that the value of a statistical life year (VOLY) is constant, whereby the value of a statistical life is assumed to be a decreasing function of age—the older a person is, the lower is the value of the remaining life.

We extracted the VOLY from a Danish study carried out by DØR (2016). Using a stated preference for reducing the risk of dying prematurely, they estimated the value of a statistical life of ∼4 million Euro, and the costs of premature death to be ∼200,000 Euro per year lost. We estimated the economic costs of premature death using national statistics of expected age at death and disease statistics concerning the age distribution of premature death cases. We conducted a sensitivity analysis where this cost rate was reduced to one third.

Results

Burden of disease

We estimated that the pathogens causing the higher relative total burden of disease in Denmark in 2019, measured in DALY, were Campylobacter, Salmonella, and norovirus (Table 3). In contrast, norovirus caused the higher number of cases, and Campylobacter the higher number of deaths. The contribution of YLD and YLL to total DALY estimates also varied (Fig. 1). While YLD contributed to over 60% of total DALY of Campylobacter and Y. enterocolitica, YLD contributed to a smaller proportion of the burden of STEC (12%), L. monocytogenes (6%), and hepA (23%).

FIG. 1.

Total burden of disease measured in DALY due to YLD and YLL, and foodborne DALY caused by seven foodborne pathogens in Denmark, 2019. YLD, years lived with disability; YLL, years of life lost due to premature death.

Table 3.

Estimated Annual Burden of Disease for Seven Foodborne Pathogens as Described by Six Different Indicators, Denmark 2019

Pathogens	Registered cases	Estimated cases	Estimated^or registered^* deaths	DALY	Total costs (mill. €)	Direct health costs (mill. €)	Foodborne share (%)
Campylobacter	5389	58,983	41^*	1691	124	1.8	76
Salmonella	1122	8657	10^*	548	30	0.6	76
Norovirus	—	185,161	27^*	334	185	0.8	26
Listeria monocytogenes	60	60	17^**	185	43	0.5	100
Yersinia enterocolitica	221^a	3249	2^*	99	5	0.4	100
STEC	619	12,136	10^**	255	46	7.8	60
Hepatitis A	34	126	1^*	9	1	0.0	42

Registered cases exclude Y. enterocolitica biotype 1A.

DALY, disability-adjusted life years; STEC, Shiga toxin-producing Escherichia coli.

Cost of illness

The costs of illness are shown as total costs and direct costs in Table 3. The highest total health costs were estimated for norovirus (185 million €), Campylobacter (124 million €), followed by STEC (46 million €) and L. monocytogenes (43 million €); the costs of Salmonella were estimated to be 30 million €. The pathogen leading to highest direct health costs was STEC (7.8 million €, which represented 16% of total costs). This high contribution is explained by the relatively larger number of complicated hospitalizations associated with STEC infections (Supplementary Appendix Table SA2). The contribution of direct health costs to the total economic burden of the remaining six foodborne pathogens was substantially lower (<3%).

Table 4 shows the average total health costs per case by sex and age for the four most important pathogens: Campylobacter, Salmonella, norovirus, and STEC. While there was no clear sex pattern in costs, there was a clear age-related pattern for all pathogens. The average cost per registered case for 65+ years was substantially higher than in younger age groups, mainly due to a relatively large mortality at older ages. There was also an increasing share of more severe cases with higher risk of hospitalization and longer absenteeism from work with age. The age pattern for STEC was slightly different, mainly due to an apparently higher share of complicated or fatal cases in the 5–14-year age category, which can imply loss of many years for those affected.

Table 4.

Incidence per 100,000 Inhabitants and Average Total Costs per Case by Four Pathogens Commonly Transmitted Through Foods in Denmark, 2019, by Sex and Age

	Campylobacter		Salmonella		Norovirus		STEC
	Reported incidence	Average cost (€/reported case)	Reported incidence	Average cost (€/reported case)	Reported incidence	Average cost (€/reported case)	Reported incidence	Average cost (€/reported case)
Females, total	85	17,052	18	21,937	828	9214	11	53,990
0–4 years	85	8758	44	9169	5489	2230	35	15,843
5–14 years	34	3808	13	4229	851	2230	10	225,957
15–44 years	103	4782	15	5281	406	1674	9	15,144
45–64 years	92	16,985	18	19,441	393	10,072	8	15,224
65+	69	54888	20	59,481	963	26,999	12	93,206
Males, total	101	15,967	20	22,201	828	8307	10	54,636
0–4 years	103	7242	42	9169	5489	2230	48	15,843
5–14 years	52	3808	22	4229	851	2230	11	225,957
15–44 years	110	4815	14	5283	405	1674	8	15,143
45–64 years	115	16,786	21	19,464	390	10,055	7	15,225
65+	90	50,531	25	62,045	950	27,097	9	101,550

Ranking of diseases

A comparison of the total DALY attributed to each pathogen against their total costs associated showed that Campylobacter caused the highest overall burden, and hepA the lowest (Fig. 2). Norovirus, STEC, and L. monocytogenes cause a relatively lower disease burden, but higher economic burden. The sensitivity analysis reducing the value of life years lost to premature death by two thirds led to lower costs for all pathogens, but did not alter their ranking in terms of economic costs. Likewise, the sensitivity analysis with 40% lower economic value of leisure hours did not change the cost ranking of the pathogens.

FIG. 2.

Plot of the burden of disease at the population level (total DALYs) and the economic burden (total costs) for seven infectious pathogens commonly transmitted through foods in 2019, Denmark.

Discussion

We estimated the overall societal burden in Denmark of seven pathogens commonly transmitted through foods. We accounted for the disease burden caused by these pathogens using six different indicators, including the composite measures, DALY and economic burden, as translated into direct and indirect health costs. Results showed that Campylobacter leads the ranking of foodborne pathogens in almost all indicators explored. In 2019, Campylobacter caused nearly 59,000 cases, 41 premature deaths, and over 1690 DALYs, and led to combined costs of 124 million Euro. The only exception was direct health costs, which were higher for STEC due to the on-average higher severity and need for medical treatment associated with severe STEC infections. HepA ranked lowest on all indicators.

The ranking of other pathogens appear to depend on the chosen indicator. The ranking of the burden of disease was also sensitive to whether we included all cases or only cases that were estimated to be foodborne. Salmonella and norovirus ranked second and third in burden of disease, whereas Campylobacter, STEC, and L. monocytogenes took that place when focusing on costs. Invasive listeriosis is a severe disease that carries a high mortality and substantial health costs. In general, costs associated with complicated hospitalization cases of Campylobacter, Salmonella, and Y. enterocolitica were around 10-fold higher than for mild cases. STEC, which ranked high in the costs' burden and had a high DALY severity and mortality, was the pathogen with the highest contribution of the direct health costs to the total economic burden. The high costs were associated with the fraction of cases estimated as severe, since mild STEC cases—typically mild gastroenteritis—have a low impact on health costs. In our analyses, the costs of complicated hospitalization cases of STEC were more than 100 times larger than for mild cases.

The total cost of STEC infections as estimated in this study should, however, be interpreted with caution, as it is likely that the proportion of severe cases was overestimated. In 2019, the number of cases reported through surveillance in Denmark was almost twice the number of that reported in 2017. We believe this reflects at least partly the introduction of more sensitive diagnostic methods in Danish clinical laboratories, in particular fecal polymerase chain reaction methods without prior culture steps. Such methods may lead to the detection of relatively more infections by strains not typically associated with hemolytic uremic syndrome or bloody diarrhea. If so, this would have led to us overestimating the total cost of STEC infections, as well as the total STEC burden.

The DALY and COI approaches have in common that they integrate morbidity and mortality to obtain a single valued indicator of the societal burden of disease. Also, both approaches can demonstrate that diseases with very distinct incidence can cause similar disease burden. One of the reasons for this is that one type of pathogen may cause few cases with severe symptoms, whereas another pathogen may cause many cases of illness with mild symptoms. Likewise, a short period with severe symptoms can result in the same burden of illness as a longer period with mild symptoms.

The DALY and COI approaches use the same kind of building blocks in terms of incidence; mortality; well-described health outcomes with different severity, duration of each possible health outcome, and assigned probability of each outcome. Both approaches are transparent and enable comparisons with other (nonfoodborne) diseases. They are widely used within their field, and thus the obtained estimates of disease burdens can be compared with burden of disease values in other countries and parts of the world. Differences in rankings based on the two methods are not contradictory, but reflect for instance the impact of the duration of illness in the overall health costs. For example, if an illness has low disability (low DW) but long duration, leading to long absence from work, it may be more expensive than a short-duration illness, even if the latter has a higher DW.

As the DALY metric is independent of whether a person is employed or not, the loss measured in DALY's will be relatively higher for retired people than the economic loss estimated using COI. Haagsma et al (2008) introduced a “relevance criterion” to investigate the importance of mild cases when estimating DALY's. The authors proposed that if 50% of respondents did not want to trade-off any time with perfect health to avoid a mild case of illness, then these mild cases should not count in the burden of disease measure. They found that employing this criterion reduced the DALY for Campylobacter by 25% but only by 5% for STEC. An alternative approach for the COI estimations (which to our knowledge has not been carried out so far) could be to investigate to what extent respondents would go to work and how effective they would be in doing their job given mild symptoms of disease.

Because even mild cases may stay home from work for a few days, the COI approach is able to reflect the societal costs of diseases that lead to low burden of disease through inclusion of lost productivity. This happens for norovirus, which has a high incidence and thus leads to the loss of a high number of productive days across the population and a higher ranking using COI than DALY. Also, a relatively high share of norovirus cases is <50 years of age and are thereby likely to be on the labor market, resulting in higher costs compared with illnesses that are more prevalent among retired people.

There are large differences in the reported incidence of the various foodborne diseases, because some are more common, some are notifiable, and some cause more severe health outcomes. As an example, most invasive listeriosis cases are severe, and thus listeriosis has a low degree of underreporting and relatively high costs per case. For norovirus, we have the complete opposite. The underreporting factors for four pathogens were estimated using data from a population survey conducted in 2009 (Müller et al, 2012). While that survey was representative of the population and provided crucial data to estimate care-seeking behavior of diarrhea patients, possible changes in these behaviors in the Danish population in recent years would not have been captured. Furthermore, the estimated proportion of clinical laboratories that tests for STEC may now be outdated.

In recent years, more and more laboratories include STEC in their testing protocol. For norovirus, we estimated total incidence, based on data extracted in a global systematic review by Ahmed et al (2014). While we focused on data from the EU, assuming that those are representative of Denmark, data were also collected in earlier time periods, which mainly add uncertainty to our estimates. Hepatitis A cases are notifiable by physicians in Denmark. Because the infection's clinical picture is often mild, particularly in children, mild infections are likely to be substantially underdiagnosed; this is not the same for severe cases. However, hepA infections are not endemic in Denmark (nor Europe), and most cases are acquired abroad, except for the relatively rare outbreak cases—a fact, which is considered by clinicians for the diagnosis.

We have identified the possible health outcomes associated with each of the seven pathogens through a review of the literature, as reported in Pires et al (2019). A recent review revisited the literature and was able to collect more recent evidence of the health outcomes of various foodborne infections, including by six of the pathogens that we included in our study (all except hepA) (Pogreba-Brown et al, 2020). While the output of this review largely supports the choices of health outcomes in our model, it also shows that there may be other health outcomes of foodborne infections that could be considered in burden and cost calculations in future iterations.

We addressed the uncertainty of the burden of disease estimates by using probability distributions to describe input parameters. In contrast, cost estimates were formulated as best available economic estimates for the individual components of the disease burden. To assess the uncertainty of results, we have carried out sensitivity analyses for important parameter values. These indicated that the VOLY assumption is particularly important for the cost estimates.

Our results show that there may be differences in the ranking of the burden of foodborne pathogens, depending on the choice of metric used. However, these metrics complement each other, and a combination of disease burden and cost of illness estimates is useful to inform policy making and establish food safety priorities at the national level.

Footnotes

Disclosure Statement

No competing financial interests exist.

Authors' Contributions

S.M.P., T.C., and J.D.J. designed the study, performed the analysis, and wrote the first draft of the article. All authors interpreted the results and contributed to the article.

Funding Information

No funding was received for this study.

Supplementary Material

Supplementary Appendix Table SA1

Supplementary Appendix Table SA2

References

Ahmed

, Hall

, Robinson

, et al. Global prevalence of norovirus in cases of gastroenteritis: A systematic review and meta-analysis. Lancet Infect Dis, 2014; 14(8):725–730; doi: 10.1016/S1473-3099(14)70767-4.

Ananthapavan

, Moodie

, Milat

, et al. Systematic review to update ‘value of a statistical life’ estimates for Australia. Int J Environ Res Public Heal 2021, Vol 18, Page 6168, 2021; 18(11):6168; doi: 10.3390/IJERPH18116168.

Anonymous. Statistiks Denmark. n.d. Available from: www.statbank.dk (last accessed May 1, 2022 ).

Christensen

, Jensen

. Estimation of the Societal Costs of Illness Due to Salmonella, Campylobacter, Listeria and VTEC [Beregning Af de Samfundsøkonomiske Omkostninger Ved Sygdomstilfælde Forårsaget Af Salmonella, Campylobacter, Listeria Og VTEC.] Institut for Fødevare- Og R. Copenhagen: Copenhagen, Denmark; 2017.

Devleesschauwer

, Havelaar

, Maertens de Noordhout

, et al. Calculating disability-adjusted life years to quantify burden of disease. Ó Swiss Sch Public Heal, 2014; 59(3):565–569; doi: 10.1007/s00038-014-0552-z.

DØR, Årsrapport 2016, Det Miljøøkonomiske Råd, (2016 Annual Report, The Danish Economic Councils). Available from: https://dors.dk/vismandsrapporter/oekonomi-miljoe-2016 (last accessed July 19, 2022 ).

ECDC. ECDC BCoDE Toolkit [Software Application]. Version 1.4 Stockholm: European Centre for Disease Prevention and Control. 2015. Available from: https://ecdc.europa.eu/en/publications-data/toolkit-application-calculate-dalys (last accessed December 5, 2018 ).

EFSA. Monitoring and identification of human enteropathogenic yersinia spp.—Scientific opinion of the panel on biological hazards. EFSA J, 2007; 5(12):595; doi: 10.2903/j.efsa.2007.595.

Gadiel

. The Economic Cost of Foodborne Disease in New Zealand, Report Prepared for New Zealand Food Safety Authority. 2010. Available from: https://www.mpi.govt.nz/dmsdocument/25814-The-economic-cost-of-foodborne-disease-in-New-Zealand (last accessed July 19, 2022 ).

10.

Haagsma

, Geenen

, Ethelberg

, et al. Community incidence of pathogen-specific gastroenteritis: Reconstructing the surveillance pyramid for seven pathogens in seven European Union member states. Epidemiol Infect, 2012; 141(8):1–15; doi: 10.1017/S0950268812002166.

11.

Haagsma

, Havelaar

, Janssen

, et al. Disability adjusted life years and minimal disease: Application of a preference-based relevance criterion to rank enteric pathogens. Popul Health Metr, 2008; 6:7; doi: 10.1186/1478-7954-6-7.

12.

Haagsma

, Siersema

, De Wit

, et al. Disease burden of post-infectious irritable bowel syndrome in The Netherlands. Epidemiol Infect, 2010; 138(11):1650–1656; doi: 10.1017/S0950268810000531.

13.

Hald

, Aspinall

, Devleesschauwer

, et al. World Health Organization estimates of the relative contributions of food to the burden of disease due to selected foodborne hazards: A structured expert elicitation. PLoS One, 2016; 11(1):1–35; doi: 10.1371/journal.pone.0145839.

14.

Havelaar

, Haagsma

, Mangen

, et al. Disease burden of foodborne pathogens in the Netherlands, 2009. Int J Food Microbiol Elsevier, 2012; 156:231–238.

15.

Havelaar

, de Wit

, van Koningsveld

, et al. Health burden in the Netherlands due to infection with thermophilic campylobacter spp. Epidemiol Infect, 2000; 125(3):505–522; doi: doi:10.1017/S0950268800004933.

16.

Havelaar

, Van Duynhoven

, Nauta

, et al. Disease burden in The Netherlands due to infections with Shiga toxin-producing Escherichia coli O157. Epidemiol Infect, 2004; 132(3):467–484; doi: 10.1017/S0950268804001979.

17.

Hoffmann

, Maculloch

, Batz

. Economic Burden of Major Foodborne Illnesses Acquired in the United States. 2015. Available from: https://www.ers.usda.gov/webdocs/publications/43984/52807_eib140.pdf (last accessed July 19, 2022 ).

18.

Jaffee

, Henson

, Unnevehr

, et al. The safe food imperative: Accelerating progress in low- and middle-income countries. Agric Food Ser, 2019; doi: 10.1596/978-1-4648-1345-0.

19.

Johannesson

Theory and Methods of Economic Evaluation of Health Care. Vol. 4. Berlin, Germany: Springer Science & Business Media, 1996.

20.

Kirk

, Pires

, Black

, et al. World Health Organization estimates of the global and regional disease burden of 22 foodborne bacterial, protozoal, and viral diseases, 2010: A data synthesis. PLoS Med, 2015; 12(12); doi: 10.1371/journal.pmed.1001921.

21.

Korcinska

, Bjerre

, Rasmussen

, et al. Detection of norovirus infections in Denmark, 2011–2018. Epidemiol Infect, 2020; 148:7; doi: https://doi.org/10.1017/S0950268820000461.

22.

Korsgaard

, Madsen

, Feld

, et al. The effects, costs and benefits of salmonella control in the Danish table-egg sector. Epidemiol Infect, 2009; 137(6):828–836; doi: 10.1017/S0950268808000903.

23.

Kuchler

, Golan

Assigning Values to Life: Comparing Methods for Valuing Health Risks. Washington, DC: US Department of Agriculture; 1999.

24.

Lægeforeningen. Overenskomst Om Almen Praksis | Læger.Dk. 2014. Available from: https://www.laeger.dk/Overenskomst-om-almen-praksis (last accessed October 29, 2021 ).

25.

Majowicz

, Scallan

, Jones-Bitton

, et al. Global incidence of human Shiga toxin-producing Escherichia coli infections and deaths: A systematic review and knowledge synthesis. Foodborne Pathog Dis, 2014; 11(6):447–455; doi: 10.1089/fpd.2013.1704.

26.

Mangen

M-JJ

, Bouwknegt

, Friesema

IHM

, et al. Cost-of-illness and disease burden of food-related pathogens in the Netherlands, 2011. Int J Food Microbiol, 2015; 196:84–93; doi: 10.1016/j.ijfoodmicro.2014.11.022.

27.

Mangen

M-JJ

, Plass

, Havelaar

, et al. The pathogen- and incidence-based DALY approach: An appropriated methodology for estimating the burden of infectious diseases. Kirk M. ed. PLoS One 2013a;8(11):e79740; doi: 10.1371/journal.pone.0079740.

28.

Mangen

MJJ

, Plass

, Havelaar

, et al. The pathogen- and incidence-based DALY approach: An appropriated methodology for estimating the burden of infectious diseases. PLoS One, 2013b;8(11); doi: 10.1371/journal.pone.0079740.

29.

Müller

, Korsgaard

, Ethelberg

. Burden of acute gastrointestinal illness in Denmark 2009: A population-based telephone survey. Epidemiol Infect, 2012; 140(02):290–298; doi: 10.1017/S0950268811000471.

30.

Pedersen

CB.

The Danish civil registration system. Scand J Public Heal, 2011; 39(7 Suppl.):22–25; doi: 10.1177/1403494810387965.

31.

Pires

SM.

Burden of disease of foodborne pathogens in Denmark, Technical Report, 1. edition. Søborg, Denmark: National Food Institute, Technical University of Denmark; 2014.

32.

Pires

, Evers

, Van Pelt

, et al. Attributing the human disease burden of foodborne infections to specific sources. Foodborne Pathog Dis, 2009; 6(4); doi: 10.1089/fpd.2008.0208.

33.

Pires

, Jakobsen

, Ellis-Iversen

, et al. Burden of disease estimates of seven pathogens commonly transmitted through foods in Denmark, 2017. Foodborne Pathog Dis 2019:fpd.2019.2705; doi: 10.1089/fpd.2019.2705.

34.

Pogreba-Brown

, Austhof

, Armstrong

, et al. Chronic gastrointestinal and joint-related sequelae associated with common foodborne illnesses: A scoping review. Foodborne Pathog Dis, 2020; 17(2):67–86; doi: 10.1089/FPD.2019.2692.

35.

Economic burden from health losses due to foodborne illness in the United States. J Food Prot, 2012; 75(1):123–131; doi: 10.4315/0362-028X.JFP-11-058.

36.

Salomon

, Haagsma

, Davis

, et al. Disability weights for the global burden of disease 2013 study. Lancet Glob Heal, 2015; 3(11):e712–e723; doi: 10.1016/S2214-109X(15)00069-8.

37.

Scallan

, Hoekstra

, Angulo

, et al. Foodborne illness acquired in the United States-major pathogens. Emerg Infect Dis, 2011; 17(1):7–15; doi: 10.3201/eid1701.P11101.

38.

Sendi

, Brouwer

. Leisure time in economic evaluation: Theoretical and practical considerations. Expert Rev Pharmacoecon Outcomes Res, 2004; 4(1):1–3; doi: 10.1586/14737167.4.1.1.

39.

Sundhedsstyrelsen. Evaluering Af Patientforløb for Patienter, Som Gennemgår Primær Total Hofte- Eller Knæalloplastikoperation 2006. 2006. Available from: https://www.sst.dk/da/udgivelser/2006/evaluering-af-patientforloeb-for-patienter-som-gennemgaar-primaer-total-hofte—eller-knaealloplastik (last accessed October 29, 2021 ).

40.

Sundström

. Samhällskostnader För Salmonellos, Campylobacterios Och EHEC. 2010. Available from: https://docplayer.se/16586052-Samhallskostnader-for-salmonellos-campylobacterios-och-ehec.html (last accessed July 19, 2022 ).

41.

Tam

, O'Brien

. Economic cost of campylobacter, norovirus and rotavirus disease in the United Kingdom. PLoS One, 2016; 11(2); doi: 10.1371/journal.pone.0138526.

42.

Verbooy

, Hoefman

, van Exel

, et al. Time is money: Investigating the value of leisure time and unpaid work. Value Heal, 2018; 21(12):1428–1436; doi: 10.1016/J.JVAL.2018.04.1828.

43.

Voldstedlund

, Haarh

, Mølbak

, et al. The Danish microbiology database (MiBa) 2010 to 2013. Eurosurveillance, 2014; 19(1):20667; doi: 10.2807/1560-7917.ES2014.19.1.20667.

44.

WHO. Department of information, evidence and research. WHO methods and data sources for global burden of disease estimates 2000–2016. Global Health Estimates Technical Paper WHO/HIS/IER/GHE/2018.4. Geneva: WHO; 2018.

45.

Williams

, O'Brien

. Assessing the contribution made by the food chain to the burden of UK-acquired norovirus infection. Report, 2019. Available from: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwj3r675ipT5AhXrpYsKHTBeDiIQFnoECAMQAQ&url=https%3A%2F%2Fwww.food.gov.uk%2Fsites%2Fdefault%2Ffiles%2Fmedia%2Fdocument%2Fassessing-the-contribution-made-by-the-food-chain-to-the-burden-of-uk-acquired-norovirus-infection.pdf&usg=AOvVaw1ZRZT-tkvbHX6XqVsgy9SU (last accessed July 25, 2022 ).

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.01 MB