545

Journal of Food Protection, Vol. 74, No. 4, 2011, Pages 545552 doi:10.4315/0362-028X.JFP-10-252
Copyright G, International Association for Food Protection

Cost of Illness and Disease Burden in The Netherlands Due to Infections with Shiga ToxinProducing Escherichia coli O157
LUQMAN TARIQ,1{ JUANITA HAAGSMA,2,3
1Centre

AND

ARIE HAVELAAR2,4*

for Prevention and Health Services Research and 2Centre for Infectious Disease Control, National Institute for Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands; 3Department of Public Health, Erasmus Medical Centre, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands; and 4Institute for Risk Assessment Sciences, Utrecht University, P.O. Box 80178, Utrecht, The Netherlands MS 10-252: Received 14 June 2010/Accepted 8 December 2010

ABSTRACT
Infections with Shiga toxinproducing Escherichia coli O157 (STEC O157) are associated with hemorrhagic colitis, hemolytic uremic syndrome (HUS), and end-stage renal disease (ESRD). In the present study, we extend previous estimates of the burden of disease associated with STEC O157 with estimates of the associated cost of illness in The Netherlands. A secondorder stochastic simulation model was used to calculate disease burden as disability-adjusted life years (DALYs) and cost of illness (including direct health care costs and indirect nonhealth care costs). Future burden and costs are presented undiscounted and discounted at annual percentages of 1.5 and 4%, respectively. Annually, approximately 2.100 persons per year experience symptoms of gastroenteritis, leading to 22 cases of HUS and 3 cases of ESRD. The disease burden at the population level was estimated at 133 DALYs (87 DALYs discounted) per year. Total annual undiscounted and discounted costs of illness due to STEC O157 infection for the Dutch society were estimated at J9.1 million and J4.5 million, respectively. Average lifetime undiscounted and discounted costs per case were both J126 for diarrheal illness, both J25,713 for HUS, and J2.76 million and J1.22 million, respectively, for ESRD. The undiscounted and discounted costs per case of diarrheal disease including sequelae were J4,132 and J2,131 , respectively. Compared with other foodborne pathogens, STEC O157 infections result in relatively low burden and low annual costs at the societal level, but the burden and costs per case are high.

Infections with Shiga toxinproducing Escherichia coli O157 (STEC O157) are an important cause of morbidity and mortality, with associated loss of life years and diminished health-related quality of life. STEC O157 induces a severe form of gastroenteritis characterized by bloody stools (hemorrhagic colitis) and may cause hemolytic uremic syndrome (HUS) in young children and sporadically in adults. HUS may result in end-stage renal disease (ESRD). Patients with ESRD are initially treated with peritoneal dialysis or hemodialysis and may later be eligible for kidney transplantation. Information on the disease consequences of STEC O157 was reviewed by Havelaar et al. (13). As in other industrialized countries, the incidence of STEC O157associated diarrheal illness in The Netherlands is low (approximately 2,000 cases per year), compared with that of illness associated with other enteric pathogens (e.g., 79,000 cases of campylobacteriosis per year) (11). Nevertheless, much effort is being made in The Netherlands and internationally to prevent diseases associated with STEC O157 because of the severity of the sequelae of this infection and the fact that it occurs more often in young
* Author for correspondence. Tel: z31 30 274 2826; Fax: z31 30 274 4434; E-mail: arie.havelaar@rivm.nl. { Present address: GlaxoSmithKline, P.O. Box 780, 3700 AT Zeist, The Netherlands.

children. Any outbreak or increase in incidence of STEC O157 infection will be significant for public health (13, 16). Various approaches have been used to quantify the societal impact of (foodborne) disease, as discussed in depth by Buzby and Roberts (3) and Mangen et al. (15). Valuation of public health impact is typically done using some form of health-adjusted life years, combining the effects of premature mortality and morbidity and accounting for healthrelated quality of life. The two most prominent healthadjusted life year metrics are quality-adjusted life years and disability-adjusted life years (DALYs). The public health impact of foodborne and zoonotic disease also can be characterized in monetary terms. The human capital approach is largely restricted to impacts on labor productivity, and willingness to pay (WTP) methods are based on revealed or stated preferences on the tradeoffs that individuals must make between health and other goods. WTP is therefore consistent with the theoretical foundation of welfare economics. The most well-known WTP measure is the value of a statistical life, a practical estimation that is not intended to place a monetary value on life itself but that refers to what an average individual in a population is willing to pay to avoid the risk of premature death. However, WTP is a function of ability to pay, and therefore results may reflect wealth and benefit. The reported value of a statistical life differs widely between and within countries;

546

TARIQ ET AL.

J. Food Prot., Vol. 74, No. 4

FIGURE 1. Disease model for infection with Shiga toxinproducing Escherichia coli O157. Reprinted from Havelaar et al. (13) with permission.

one landmark study found a range of $0.7 million to $20.8 million (30). An alternative approach to monetary valuation is the cost of illness (COI) method, which traces the economic flow associated with an adverse health outcome through the quantification of measurable monetary costs. Mangen et al. (15) stated, it seems a reasonable conclusion that prioritization efforts should consider using a combination of approaches, namely, the use of COI with either WTP or health-adjusted life years, both of which capture the intangible costs of adverse health states missing from COI. In this study, we evaluate the societal impact of STEC O157 infection in The Netherlands using a combination of DALYs and COI and compare this approach to that used commonly in the United States with the value of a statistical life. Foodborne illnesses generate a considerable disease burden and major economic losses. According to Buzby and Roberts (3), foodborne illnesses generate costs that are borne by households whose members become ill, the food industry, and the regulatory and public health sectors. For example, the costs of human Campylobacter infections in The Netherlands were estimated at J21 million (90% credible interval [CI] of J11 million to J36 million) per year (16). Little is known about the economic costs imposed on the Dutch or other societies by disease associated with STEC O157. Frenzen et al. (8) estimated the COI due to STEC O157 in the United States at $405 million (in 2003 dollars). Buzby and Roberts (3) updated this estimate to $459 million (in 2007 dollars). Scharff et al. (23, 24) suggested that STEC O157 infections cost about $990 million (in 2009 dollars) to U.S. residents. In these studies, valuation of death was the largest component of total costs, which was estimated using the value of a statistical life for premature deaths. In contrast, The Netherlands uses the friction cost method to account for productivity loss due to disease because of temporary absence from work, disability, and/or premature mortality (14).

In the present study, we extended the previous estimates of the burden of disease associated with STEC O157 infection with estimates of the COI in The Netherlands, using a previously developed detailed epidemiological model (13). We also present the impact of discounting on these estimates. MATERIALS AND METHODS
Disease model. The disease model for STEC O157 infections is provided in Figure 1. Infections may be asymptomatic or may lead to diarrheal illness, including hemorrhagic colitis. HUS occurs mainly in young children and may lead to death during the acute phase, to ESRD, or to other sequelae. In this study, we attributed and calculated costs and disease burden estimates for each disease state. In contrast to the model presented by Havelaar et al. (13), we assumed that patients with ESRD can maximally receive two kidney transplantations and that patients older than 70 years are not eligible for the second kidney transplantation. Simulation model. To estimate the updated disease burden and COI associated with STEC O157 infections, a second-order stochastic simulation model was built in Excel (Microsoft, Redmond, WA) and implemented with the add-in software @Risk (version 5.0, Palisade Corp., New York). As explained previously in detail (13), the model simulates the annual incidence of STEC O157associated gastroenteritis and death and the incidences of HUS and ESRD. The model was run with 250 simulations and 1,500 iterations per simulation to account for uncertainty and variability, respectively. Each iteration within one simulation calculated a possible set of life histories of HUS and ESRD patients. All transitions from one health state to the next in the outcome tree are represented by conditional probabilities, resulting in different realizations of the system in each iteration. Together, the 1,500 iterations represent the variability of the system and consequently of the disease burden and COI, for a particular set of uncertain input parameters. Different simulations use different samples from the uncertainty distributions of the input parameters and hence quantified the uncertainty in the model results. As found in a previous study (13), in this model uncertainty dominates

J. Food Prot., Vol. 74, No. 4

COSTS AND BURDEN OF STEC O157 IN THE NETHERLANDS

547

TABLE 1. Summary of input data

Estimatesa Assumptions 1996 2009 Reference

Total Dutch population Discount rate Exchange rate from J to NLGb Exchange rate from $US to J Consumer price indices from 1996 to 2009 Friction costs Hemolytic uremic syndrome Full care center hemodialysis (HD) First year Follow-up Peritoneal dialysis (PD) First year Follow-up Kidney transplantation First year Follow-up Ratio HD:PD
a b

16,486,000 4% for costs, 1.5% for burden 2.20371 0.68 1.34298 J48,207 J25,000 NLG 152,666 NLG 145,757 NLG 129,456 NLG 114,547 NLG 90,000 NLG 18,000 2.75 J93,000 J88,000

25 7

25 14 3 6

6 J79,000 J70,000 6 J55,000 J11,000 26

All cost estimates are mean costs per case. NLG, Dutch guilders. Indirect health care costs (costs of unrelated medical care in life years gained during the extended life expectancy (27)) were not included in this study, as recommended by the Dutch Pharmacoeconomic guidelines (4). Contrary to these guidelines, direct non health care costs (including patients travel costs and patients copayments for costs such as informal care) were not included because they were relatively low in previous studies (16). Derivation of the model input parameters. The epidemiological model and DALY calculations were based on parameters described by Havelaar et al. (13). To quantify the COI caused by STEC O157 infection, data were gathered from several sources. We used consumer price indices from Statistics Netherlands (25) taking the year 1995 as an index and an exchange rate of J0.68 to $1 to adjust foreign currency to a 2009 price level. According to the method of Oostenbrink et al. (19), productivity losses were considered only for patients 15 to 64 years of age, using the friction cost method. We defined a friction period of 154 days as proposed by Koopmanschap et al. (14). The assumed productivity losses per day per person were based on the salary of an average person in that age class (in 2009). We calculated productivity losses per day and per friction period in the year 2009 for both an average person and an average working person, depending on the age group of the patients. In the absence of Dutch data, we used the data published by Buzby and Roberts (3), who suggest that a patient who had HUS but not ESRD and survived had an average cost of $38,695 (2007 price). For the costs of hemodialysis and kidney transplantation, Dutch estimates from de Wit et al. (6) for 1996 were used. Hemodialysis costs about 152,000 Dutch guilders (J93,000 in 2009 currency) in the first year of treatment and about 145,000 Dutch guilders (J88,000 in 2009 currency) in later years. Peritoneal dialysis costs about 130,000 Dutch guilders (J79,000 in 2009 currency) in the first year of treatment and about 114,000 Dutch guilders (J70,000 in 2009 currency) in later years. Kidney transplantation costs about 90,000 Dutch guilders (J55,000 in 2009 currency) in the first year and about 18,000 Dutch guilders

variability; in this article we present the results of only the uncertainty analysis. Disease burden. Disease burden is expressed in DALYs. By using the DALY methodology, morbidity (expressed in years lived with disability) and mortality (expressed in years of life lost) are summed into one metric (18). The number of reported cases of STEC O157 infection in The Netherlands from 1999 through 2008 was on average 40 per year (0.25 cases per 100,000 person years), without a clear trend (9, 29). Hence, the epidemiological estimates made by Havelaar et al. (13) were still considered valid. Serotype O157 represents only 20% of all STEC isolates from clinical material in The Netherlands. However, until now non-O157 STEC strains have not been associated with HUS in The Netherlands (11). Hence, we estimated that the additional burden and costs of non-O157 STEC strains would be low, and we did not include these strains is our analysis. COI. Following the Dutch Pharmacoeconomic guidelines from the Dutch Health Insurance Board (4), the following cost categories were included in the COI estimates: (i) direct healthcare cost, including costs such as medical consultations, hospitalization, drugs, rehabilitation, and other medical services, and (ii) indirect nonhealth care costs, including costs associated with productivity loss due to disease because of temporary absence from work, disability, and/or premature mortality. Productivity losses were estimated according to the friction cost method. Mainly, two methods are used to value productivity: the human-capital method and the friction cost method. The human-capital method takes the patients perspective and counts any hour not worked as an hour lost. The friction cost method takes the employers perspective and counts as lost only those hours not worked until another employee takes over the patients work (28).

548

TARIQ ET AL.

J. Food Prot., Vol. 74, No. 4

TABLE 2. Detailed burden of disease and cost of illness per case associated with STEC O157 infection in The Netherlands a
Burden of disease (mean values/yr)c Incidence rate (no./million people/yr) YLD 0% 1.5% No. of fatal cases 0% YLL 1.5% DALYs 0% 1.5% DALYs/case 0% 1.5%

Illness typeb

Incidence

GE HUS clinical phase ESRD First dialysis period First transplantation Life with healthy graft Dialysis after first graft rejection Second transplantation Life with healthy graft Dialysis after second graft rejection Total HUS and ESRD Total GE, HUS, and ESRD
a b

2.114 22 3 2 2 2 1 1 1

128.23 1.33 0.18 0.12 0.12 0.12 0.06 0.06 0.06

7 3 4 0 3 3 0 2 7 21 28

7 3 3 2 2 1 0 1 4 15 22

1 2 1 0

8 58 10 8

4 38 5 5

14 61 14 9

11 41 8 5

0.1 6.1

0.0 4.0

0 0

9 5

6 4

12 5

7 4

0 3 4

6 97 105

5 62 67

13 114 133

8 73 87

Values are given at nondiscounted rate (0%) and discounted rates (1.5 or 4%). GE, gastroenteritis; HUS, hemolytic uremic syndrome; ESRD, end-stage renal disease. c YLD, years lived with disability; YLL, years of life lost; DALYs, disability adjusted life years. d DHC, direct health care costs; INHC, indirect nonhealth care costs.

(J11,000 in 2009 currency) in later years. The fraction of patients receiving hemodialysis or peritoneal dialysis was derived from Renine (26), the national registry of data concerning dialysis and kidney transplantation patients in The Netherlands. We calculated proportions of patients receiving hemodialysis and peritoneal dialysis per age category. Results are presented both undiscounted and discounted, using the formula K ~ k/r[1 2 (1 2 r)T], where K is the discounted cost, k is the undiscounted cost, r is the discount rate, and T is the time interval (years) (7). DALYs were discounted at 1.5%, and COI was discounted at 4%, as recommended in the Dutch guidelines (4). Table 1 shows a summary of model input parameters.

RESULTS In Table 2, mean values per year of the burden of STEC O157associated illness and the COI estimates in The Netherlands are shown. The average burden per case and average COI associated with STEC O157 in The Netherlands are shown in Table 3. Disease burden. The estimated undiscounted (discounted) mean disease burden of STEC O157associated illness in the Dutch population was 133 (87) DALYs per year (90% CI, 97 to 182 DALYs) compared with a previous estimate of 116 DALYs (13). Death in the clinical phase of HUS was the largest single contributing factor, with 58 (38) DALYs. Overall, fatal outcomes accounted for 105 (67) DALYs, and morbidity accounted for 28 (22) DALYs. COI. The estimated undiscounted (discounted) COI per year associated with STEC O157 infection in The Netherlands were J9.1 million (J4.5 million) (90% CI, J5.1

million to J13.5 million). Direct health care costs accounted for J8.8 million (J4.3 million) (90% CI, J4.8 million to J13.3 million). Indirect nonhealth care costs were not discounted and accounted for J0.3 million (90% CI, J0.05 million to J0.60 million). The cost category with the highest numbers in the estimated COI were direct health care costs, specifically costs within the first dialysis period, at an estimated J1.9 million (J835,000), and costs of dialysis after the second graft rejection, at an estimated J4.0 million (J1.8 million). Reasons for these high costs include the number of years spent in these disease states and the high costs of dialysis. Indirect nonhealth care costs were minor (3%). Total costs in the HUS clinical phase were estimated at J560,000, whereas total costs of ESRD accounted for J8.8 million. Figure 2 shows the distribution of the estimated COI associated with STEC O157 infection in The Netherlands without and with discounting, respectively. Uncertainty analysis. The undiscounted (discounted) 90% CI of the burden of disease estimate per year was 97 to 182 (64 to 125) DALYs (Table 4). Years of life lost ranged from 77 to 145 (50 to 96) per year. Life years lived with disability ranged from 17 to 41 (11 to 29) per year. Total cost estimates per year ranged from J5.1 million to 13.5 million (J2.8 million to 7.3 million). Direct health care costs per year ranged from J4.8 million to J13.3 million (J2.6 million to J6.5 million). Indirect nonhealth care costs were not discounted because these costs occur within 1 year and ranged from J54,293 to J606,766. Detailed statistics for the output data are available upon request from the corresponding author.

J. Food Prot., Vol. 74, No. 4

COSTS AND BURDEN OF STEC O157 IN THE NETHERLANDS

549

TABLE 2. Extended
Cost of illness (mean euros/yr)d DHC 0% 4% 0% INHC 4% 0% Total costs 4% 0% Costs/case 4%

26,276 560,455 1,830,254 112,223 458,683 1,481,267 73,901 279,058 4,005,670 8,801,512 8,827,788

26,276 560,455 834,132 112,223 155,876 613,168 73,901 100,474 1,808,738 4,258,968 4,285,244

241,136 6,166 53,262

241,136 6,166 53,262

267,412 566,622 1,883,516 112,223 458,683 1,481,267 73,901 279,058 4,005,670 8,860,941 9,128,353

267,412 566,622 887,394 112,223 155,876 613,168 73,901 100,474 1,808,738 4,318,396 4,585,808

4,318 421,845 3,198,238

2,169 211,922 1,446,661

59,429 300,565

59,429 300,565

DISCUSSION This study was conducted to estimate annual COI in The Netherlands due to disease associated with STEC O157 infection and provided updated estimates of the disease burden associated with STEC O157 infection in the Dutch population, e.g., HUS and ESRD. COI was J9.1 million per year undiscounted and J4.5 million per year discounted. Mean disease burden was 133 DALYs per year undiscounted and 87 DALYs per year discounted. STEC O157 infection has a relatively low burden of disease and COI on a population basis. For example, Mangen et al. (16) estimated the costs of human Campylobacter infections and sequelae in The Netherlands at J21 million with a burden of 1.200 DALYs per year. However, on a per case basis, STEC O157 infection has a .12-fold higher health impact (55 DALYs per 1,000 cases of diarrheal illness) than does Campylobacter infection (4.6 DALYs per 1,000 cases) and a 16-fold higher COI (J4,300 per case of diarrheal illness versus J262 per case of Campylobacter infection). Notably, most of the COI associated with STEC O157 infection occurs within the health care sector, whereas this is not the case for other pathogen infections where indirect nonhealth care costs

dominate. The indirect nonhealth care costs are much lower for STEC O157 infections (J0.2 million per year) than for Campylobacter infection (J14 million per year). However, cost categories such as cost associated with productivity losses because of (temporary) absence from work by parents of sick children are not included in this estimate. In previous studies performed in the United States, much higher COI estimates of STEC O157 infection were found. Frenzen et al. (8) estimated the COI due to STEC O157 infection in the United States at $405 million (in 2003 dollars), including $370 million for premature deaths. Buzby and Roberts (3) updated this estimate to $459 million (in 2007 dollars), including $265 million for premature deaths. Scharff et al. (23, 24) suggested that STEC O157 infections cost about $990 million (in 2009 dollars) to U.S. residents, including $627 million due to HUS. Taking into account differences in population sizes in 2009 (The Netherlands, 16 million; United States, 301 million) and the fact that the estimated incidence rate in the United States is twice that in The Netherlands (0.28 compared with 0.14 cases per 1,000 people, respectively) (see data provided by Mead et al. (17)), the U.S. costs estimates are up to twice those for The Netherlands. In the

TABLE 3. Average burden of disease and cost of illness per case associated with STEC O157 infection in The Netherlands a
Illness typeb Total nondiscounted (discounted) DALYsc Total nondiscounted (discounted) costs (euros) Nondiscounted (discounted) DALYs/case Nondiscounted (discounted) costs/case (euros)

No. of cases

Cost categories

GE HUS ESRD Total

a b c

2,114 22 3 2,114

GE HUS ESRD GE z HUS z ESRD

14 60 58 132

(14) (40) (37) (91)

267,416 565,688 8,283,931 9,117,035

(267,416) (565,688) (3,671,994) (4,505,098)

0.006 2.7 19.33 0.06

(0.006) (1.8) (12.33) (0.04)

126 25,713 2,761,310 4,312

(126) (25,713) (1,223,998) (2,131)

Values are given at nondiscounted rate (0%) and discounted rates of 1.5% for DALYs and 4% for costs per case. GE, gastroenteritis; HUS, hemolytic uremic syndrome; ESRD, end-stage renal disease. DALYs, disability adjusted life years.

550

TARIQ ET AL.

J. Food Prot., Vol. 74, No. 4

FIGURE 2. Undiscounted and discounted cost of illness associated with STEC O157 infection in The Netherlands.

U.S. studies, valuation of death was the largest component of total cost, which was estimated using the value of a statistical life for premature deaths. In contrast, we used the friction cost method to account for productivity loss due to temporary absence from work, disability, and/or premature mortality, all the result of disease. This approach was proposed by Koopmanschap et al. (14) and has been widely accepted and applied in many European countries. HowevTABLE 4. Summary statistics of output distributions a
Nondiscounted Variableb Mean of means SD of means 5%

er, we also ran our model using an updated estimated value of a statistical life of J3.6 million (3). This resulted in a total COI of J23 million undiscounted and J19 million discounted. Direct health care costs account for J8 million undiscounted and J4 million discounted. Indirect health care costs rose to J15 million (see Table 5). These estimates are similar to the U.S. estimates provided by Frenzen et al. (8) and Buzby and Roberts (3). Clearly, the valuation of premature deaths is a very influential factor in COI estimates of foodborne and other diseases. In this study, statistical uncertainty was quantified with the second-order stochastic simulation model run with 250 simulations and 1,500 iterations. However, uncertainty regarding other input parameters remains. Havelaar et al. (13) quantified uncertainty of disease burden estimates with scenario analyses. Regarding COI estimates, uncertainty exists around costs of HUS and dialysis periods. Costs of HUS were derived from a U.S. study (3), with uncertainty about the transferability of these costs to The Netherlands. For costs during dialysis periods, Dutch estimates based on valid data were used from 1996 (6). Because costs of HUS are relatively low and have little influence on the outcomes of this study and costs of dialysis have substantial influence because they are relatively high, we used the dialysis estimates. However, more recent estimates are desirable. The strength of this study lies in the use of the DALY and the COI methods to quantify intangible and monetary costs associated with STEC O157 infection. As argued by Mangen et al. (15), this combined approach provides the most complete basis currently available for comparison of and prioritizing foodborne hazards. Results of this study can be used for any cost-benefit analysis of STEC O157 infection. Using a threshold on the costs to avert 1 DALY, the costs of implementing interventions to reduce exposure to STEC O157 can be evaluated in a cost-utility framework as presented previously for human Campylobacter infections (12, 31). The present study was restricted to the physical consequences of STEC O157 infection. However, STEC O157associated gastroenteritis and its sequelae, HUS and ESRD, also may have indirect (psychological) impact. Some children hospitalized at the pediatric intensive care unit experienced negative behavioral changes, elevated

Discounted 95% Mean of means SD of means 5% 95%

DALYs YLD YLL Cost of illness (euros) DHC (euros) INHC (euros)
a b

133 28 105 9,128,353 8,827,788 300,565

28 11 23 2,800,415 2,738,266 346,175

97 17 77 5,075,498 4,859,552 54,329

182 41 145 13,562,175 13,292,749 606,766

87 22 67 4,585,808 4,285,244 300,565

20 10 16 1,259,352 1,163,748 346,175

64 11 50 2,801,733 2,643,933 54,329

125 29 96 7,363,168 6,499,605 606,766

Values are given at nondiscounted rate (0%) and discounted rates of 1.5% for DALYs and 4% for cost of illness. DALYs, disability adjusted life years; YLD, years lived with disability; YLL, years of life lost; DHC, direct health care costs; INHC, indirect nonhealth care costs.

J. Food Prot., Vol. 74, No. 4

COSTS AND BURDEN OF STEC O157 IN THE NETHERLANDS

551

TABLE 5. Cost-of-illness estimates based on friction cost method and value of statistical life a
Friction cost method (euros) Variableb Undiscounted Discounted Value of statistical life (euros) Undiscounted Discounted

Cost estimates used DHC INHC Total cost of illness

a b

48,000 8,827,788 300,564 9,128,353

48,000 4,285,243 300,564 4,585,808

3,600,000 8,000,000 15,000,000 23,000,000

3,600,000 4,000,000 15,000,000 19,000,000

Discounted rate of 4% for cost of illness. DHC, direct health care costs; INHC, indirect nonhealth care costs.
8. Frenzen, P. D., A. Drake, and F. J. Angulo. 2005. Economic cost of illness due to Escherichia coli O157 infections in the United States. J. Food Prot. 68:26232630. 9. Friesema, I., B. Schimmers, O. Stenvers, A. Heuvelink, E. de Boer, K. van der Zwaluw, C. de Jager, D. Notermans, I. van Ouwekerk, R. de Jonge, and W. van Pelt. 2008. Intensieve surveillance van Shigatoxin-producerende Escherichia coli (STEC) in Nederland. Infect. Bull. 21:2834. 10. Gudmundsdottir, H. S., A. Elklit, and D. B. Gudmundsdottir. 2006. PTSD and psychological distress in Icelandic parents of chronically ill children: does social support have an effect on parental distress? Scand. J. Psychol. 47:303312. 11. Haagsma, J. A., B. van der Zanden, L. Tariq, W. van Pelt, Y. T. van Duynhoven, and A. H. Havelaar. 2009. Disease burden and costs of selected foodborne pathogens in the Netherlands, 2006. RIVM report 330331001/2009. RIVM, Bilthoven, The Netherlands. 12. Havelaar, A. H., M. J. Nauta, M. J. J. Mangen, A. G. de Koeijer, M.-J. Bogaardt, E. G. Evers, W. F. Jacobs-Reitsma, W. van Pelt, J. A. Wagenaar, G. A. de Wit, and H. van der Zee. 2005. Costs and benefits of controlling Campylobacter in the Netherlandsintegrating risk analysis, epidemiology and economics. RIVM report 250911008. RIVM, Bilthoven, The Netherlands. 13. Havelaar, A. H., Y. T. van Duynhoven, M. J. Nauta, M. Bouwknegt, A. E. Heuvelink, G. A. de Wit, M. G. Nieuwenhuizen, and N. C. van de Kar. 2004. Disease burden in The Netherlands due to infections with Shiga toxinproducing Escherichia coli O157. Epidemiol. Infect. 132:467484. 14. Koopmanschap, M. A., F. F. Rutten, B. M. van Ineveld, and L. van Roijen. 1995. The friction cost method for measuring indirect non health care costs of disease. J. Health Econ. 14:171179. 15. Mangen, M. J., M. B. Batz, A. Ka sbohrer, T. Hald, J. G. Morris, Jr., M. Taylor, and A. H. Havelaar. 2010. Integrated approaches for the public health prioritization of foodborne and zoonotic pathogens. Risk Anal. 30:782797. 16. Mangen, M. J. M., A. H. Havelaar, R. A. J. A. M. Bernsen, R. van Koningsveld, and G. A. de Wit. 2005. The costs of human Campylobacter infections and sequelae in the Netherlands: a DALY and cost-of-illness approach. Acta Agric. Scand. 2:3551. 17. Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607625. 18. Murray, C. J., and A. K. Acharya. 1997. Understanding DALYs (disability adjusted life years). J. Health Econ. 16:703730. 19. Oostenbrink, J. B., C. A. M. Bouwmans, M. A. Koopmanschap, and F. F. H. Rutten. 2004. Standaard kostprijzen en rekenwaarden. Bijlage bij Handboek voor Kostenonderzoek, methoden en standaard kostprijzen voor economische evaluaties in de gezondheidszorg. College voor Zorgverzekeringen, Diemen, The Netherlands. 20. Pollock, K. G., E. Duncan, and J. M. Cowden. 2009. Emotional and behavioral changes in parents of children affected by hemolyticuremic syndrome associated with verocytotoxin-producing Escherichia coli: a qualitative analysis. Psychosomatics 50:263269. 21. Rennick, J. E., and J. Rashotte. 2009. Psychological outcomes in children following pediatric intensive care unit hospitalization: a systematic review of the research. J. Child Health Care 13:128149.

anxiety, and increased symptoms of posttraumatic stress disorder (20, 21, 32). Research conducted with parents of these children revealed that as many as one in three parents met the Diagnostic and Statistical Manual of Mental Disorders IV criteria for acute stress disorder, and after discharge, one in five parents experienced symptoms indicative of posttraumatic stress disorder (1, 2, 10, 21). These findings suggest that the results reported in the present study may underestimate the actual burden of disease and COI of STEC O157 infection because patients with psychological syndromes often continue to have symptoms for more than 1 year and the impact on health-related quality of life and health care is substantial (5, 22, 33, 34). In conclusion, the annual undiscounted (discounted) COI due to infections with STEC O157 in The Netherlands is J9.1 million (J4.5 million), with a mean disease burden of 132 DALYs (91 DALYs). The findings of this COI study suggest that additional efforts to control this pathogen will reduce health care costs and may be cost-effective from a societal point of view. ACKNOWLEDGMENT
We thank Dr. G. A. de Wit for providing input estimates for the model and for critically reading the manuscript.

REFERENCES
1. Ballufi, A., N. Kassam-Adams, A. Kazak, M. Tucker, T. Dominguez, and M. Helfaer. 2004. Traumatic stress in parents of children admitted to the pediatric intensive care unit. Pediatr. Crit. Care Med. 5:547553. 2. Bronner, M. B., A.-M. Kayser, H. Knoester, A. P. Bos, B. F. Last, and M. A. Grootenhuis. 2009. A pilot study on peritraumatic dissociation and coping styles as risk factors for posttraumatic stress, anxiety and depression in parents after their childs unexpected admission to a pediatric intensive care unit. Child Adolesc. Psychiatry Ment. Health 3:33. doi:10.1186/1753-2000-3-33. 3. Buzby, J. C., and T. Roberts. 2009. The economics of enteric infections: human foodborne disease costs. Gastroenterology 136: 18511862. 4. College voor Zorgverzekeringen. 2005. Richtlijnen voor farmacoeconomisch onderzoek; evaluatie en actualisatie. College voor Zorgverzekeringen, Diemen, The Netherlands. 5. Davidson, J. R., D. J. Stein, A. Y. Shalev, and R. Yehuda. 2004. Posttraumatic stress disorder: acquisition, recognition, course, and treatment. J. Neuropsychiatry Clin. Neurosci. 16:135147. 6. de Wit, G. A., P. G. Ramsteijn, and F. T. de Charro. 1998. Economic evaluation of end-stage renal disease treatment. Health Pol. 44:215 232. 7. Drummond, M. F., M. J. Schulper, G. W. Torrance, B. J. OBrien, and G. L. Stoddart. 2005. Methods for the economic evaluation of health care programmes, 3rd ed. Oxford University Press. Oxford.

552

TARIQ ET AL.

J. Food Prot., Vol. 74, No. 4

22. Saarni, S. I., J. Suvisaari, H. Sintonen, S. Pirkola, S. Koskinen, A. Aromaa, and J. Lonnqvist. 2007. Impact of psychiatric disorders on health-related quality of life: general population survey. Br. J. Psychiatry 190:326332. 23. Scharff, R. 2009. Health related costs from foodborne illness in the United States. Produce Safety Project, Georgetown University, Washington, DC. 24. Scharff, R. L., J. McDowell, and L. Medeiros. 2009. Economic cost of foodborne illness in Ohio. J. Food Prot. 72:128136. 25. Statline. 2010. Statistics Netherlands. Centraal Bureau voor de Statistiek, Den Haag, The Netherlands. 26. Stichting Renine Nederland. 2008. Verdeling over de behandelingsvormen, op 1 Januari 2008, per leeftijdsgroep en totaal. Available at: https://www. renine.nl/static?id~prev_distr&render~png. Accessed 18 June 2009. 27. van Baal, P. H. M., T. L. Feenstra, R. T. Hoogenveen, G. A. de Wit, and W. B. F. Brouwer. 2007. Unrelated medical care in life years gained and the cost utility of primary prevention: in search of a perfect cost-utility ratio. Health Econ. 16:421433. 28. van den Hout, W. B. 2010. The value of productivity: human-capital versus friction-cost method. Ann. Rheum. Dis. 69:i89i91.

29. van Duynhoven, Y. T., C. M. de Jager, A. E. Heuvelink, W. K. van der Zwaluw, H. M. Maas, and W. van Pelt. 2002. Enhanced laboratory-based surveillance of Shiga-toxinproducing Escherichia coli O157 in The Netherlands. Eur. J. Clin. Microbiol. Infect. Dis. 21: 513522. 30. Viscusi, W. K., and J. E. Aldy. 2003. The value of a statistical life: a critical review of market estimates throughout the world. J. Risk Uncertainty 27:576. 31. Vosough Ahmadi, B., K. Frankena, J. Turner, A. G. J. Velthuis, H. Hogeveen, and R. B. M. Huirne. Unpublished data. 32. Ward-Begnoche, W. 2007. Posttraumatic stress symptoms in the pediatric intensive care unit. J. Spec. Pediatr. Nurs. 12:84 92. 33. Wells, K. B., M. A. Burnam, W. Rogers, R. Hays, and P. Camp. 1992. The course of depression in adult outpatients. Results from the Medical Outcomes Study. Arch. Gen. Psychiatry 49:788 794. 34. Zatzick, D. 2003. Posttraumatic stress, functional impairment, and service utilization after injury: a public health approach. Semin. Clin. Neuropsychiatry 8:149157.