In this issue, Simonsen and colleagues1 present an elegant analysis of the incidence of laboratory-confirmed gastrointestinal infections in Denmark over a period of 11 years. They focus on bacterial pathogens for which foodborne transmission is known to play an important role in human infection. The Danish setting enables them to link data from civil registration, labor market research, and national surveillance of laboratory-confirmed infections for the entire population—something possible in only a few other countries—and compare the incidence of the various pathogens among demographic, socioeconomic, and cultural background groups. They report 5 main findings: (1) incidence tends to increase with increasing socioeconomic status (as measured by average adult income and years of schooling); (2) incidence is higher among married compared with single individuals; (3) incidence is higher among adults with no children; (4) incidence is lower among foreign-born compared with Danish-born individuals; and (5) when comparing Danish-born individuals born to foreign versus indigenous parents, incidence among those born to foreign parents is lower for Yersinia, but higher for Salmonella and Shigella.
The interpretation of studies based on laboratory reporting data is subject to a number of caveats. Firstly, reported cases of gastroenteritis comprise a minority of all cases ocurring in the community. They represent only the proportion of cases who sought healthcare, provided a specimen for microbiological investigation, had a pathogen identified, and were reported to national surveillance. In England, national surveillance statistics capture approximately 1 in every 3 cases of sporadic Salmonella infection and 1 in 8 cases of Campylobacter infection.2 For norovirus, this ratio is 1 in 1562, a figure known to be an underestimate because of the considerable improvement in the sensitivity of diagnostic tests for this pathogen in recent years.3 Secondly, reported cases comprise a highly selected subset of cases. Studies in England and the Netherlands have shown that disease severity,4,5 recent foreign travel and lower socioeconomic status4 are strongly associated with health care use among community cases of infectious gastrointestinal illness. As Simonsen and colleagues point out, it is likely that some of the observed differences of disease incidence by social class are at least partly due to differences in health care usage and, hence, likelihood of a case being reported.
With this in mind, the authors’ findings still support the existence of real associations between socioeconomic status and incidence of bacterial gastroenteritis. To the extent that diseases caused by different pathogens are clinically indistinguishable, any reporting bias should be systematic across all pathogens, but the authors find that associations with average adult income are positive for some pathogens, but negative for others, suggesting real differences. The authors conclude that these patterns are likely to reflect differences in dietary habits and travel activity, with people in higher socioeconomic groups having higher risk diets and being more likely to travel abroad. The role of diet and travel appears to be supported by the fact that incidence is different (albeit lower) among foreign-born individuals, and that among Danish-born residents with foreign parents, incidence is higher for some infections but lower for others.
Despite these differences, it is difficult to know how such findings should be interpreted, for a number of reasons. Firstly, in none of the comparisons do the most extreme groups in which the strongest associations are seen comprise more than one-tenth of the person-time of follow-up compared with the baseline group; in only a few notable exceptions (such as Shigella cases among Danish-born children under 5 years with foreign parents) do these subgroups give rise to the majority of cases. It is a common paradox in epidemiology that the strongest associations are often seen when conditioning the data into ever more extreme and uncommon subsets, while the most effective interventions bring about a small benefit to the majority.
Secondly, we currently have a very poor understanding of the development and prevalence of immunity to these pathogens. The relatively high burden of some of these infections suggests that the development of immunity must be important. This is supported by the fact that in highly-endemic countries, clinical disease due to many of these pathogens is largely restricted to young children. Indeed, the lower prevalence of exposure in higher-income countries may result in a greater risk of illness among those exposed, because they have fewer opportunities to develop protective immunity. There is evidence for this from studies of persistently-exposed groups such as poultry abattoir workers6 and farm-resident children,7 in whom IgG antibody titer against Campylobacter rises over time, while risk of disease is inversely proportional to duration of exposure. Knowledge of the distribution of immunity and susceptibility in the population, as well as their determinants, is crucial to further understanding of the epidemiology and transmission of these pathogens.
Thirdly, even if the patterns in incidence reported by Simonsen and colleagues1 are real, it is unclear what kind of action is warranted. Our frameworks for studying foodborne illness focus primarily on food production at one end and food consumption at the other, with the rather naive assumption that there is a linear relationship between the amount of microbiological contamination at production and the risk of illness. Very little research is dedicated to studying the role of behavioral and cultural and societal factors in determining risk of foodborne illness. Relative-risk statistics are strangely seductive in that they invite us to segregate the population into groups and characterize them on the basis of post hoc hypotheses. However, many of these hypotheses remain untested. Thus, to a reader working in the field, the higher incidence among Danish-born residents with foreign parents might be explained by differences in diet, or by the fact that such individuals are more likely to travel to foreign countries and thus more likely to be infected, while not having had the chance to develop immunity. These or other explanations may indeed turn out to be true, but there is little (if any) empirical evidence to support them. We do not know, for example, how immunity develops with age, or what frequency or dose of exposure is required to acquire long-term immunity. There is also considerable evidence that changes in dietary habits toward those of the adoptive country are an important part of the acculturation process among immigrants,8–10 and that processes related to acculturation are both generational and dependent on time since migration.9,11 Similarly, the authors suggest that a higher risk diet, possibly involving more frequent consumption of fresh poultry, may explain the greater incidence seen in higher socioeconomic groups, but they do not comment on an earlier study from Denmark (also published in this journal) that indicated that lower parental socioeconomic status is a risk factor for diarrheal illness among children under five.12 An alternative explanation involving a higher-risk diet among lower socioeconomic groups might be proposed in that instance, again with little evidence to support it. We should thus be wary of untested hypotheses and the consequent risk that our theories of foodborne illness are shaped by relative risks, anecdote, and misclassification.
I do not intend to detract in any way from the work of Simonsen and colleagues,1 which is a thorough and rigorous analysis rarely possible in other settings and a good starting point for further research. I merely remind readers that an epidemiology of foodborne illness constitutes more than the sum of the epidemiologies of its constituent pathogens—more than incidence estimates and microbiological typing of pathogens. An epidemiology of foodborne illness should include an epidemiology of food: of food production, distribution, and marketing; of food choice and availability; of consumer and societal attitudes to food; and of food preparation practices; as well as how all these contribute to influence risk of illness. It should include an understanding of the modifying effect of immunity: of how it develops, how long it lasts and how it is distributed in the population. Only then can we begin to understand the complexities of foodborne illness and make effective progress toward its control.
ABOUT THE AUTHOR
CLARENCE TAM is Lecturer in Epidemiology at the London School of Hygiene & Tropical Medicine. His main research activities include the epidemiology of infectious intestinal disease and long-term consequences of diarrheal disease, as well as the epidemiology of dengue fever.
1. Ethelberg S, Simonsen J, Frisch M. Socioeconomic risk factors for bacterial gastrointestinal infections in Denmark. Epidemiology
2. Wheeler JG, Sethi D, Cowden JM, et al. Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ
3. Amar CF, East CL, Gray J, et al. Detection by PCR of eight groups of enteric pathogens in 4,627 faecal samples: re-examination of the English case-control Infectious Intestinal Disease Study (1993–1996). Eur J Clin Microbiol Infect Dis
4. Tam CC, Rodrigues LC, O'Brien SJ. The study of infectious intestinal disease in England: what risk factors for presentation to general practice tell us about potential for selection bias in case-control studies of reported cases of diarrhoea. Int J Epidemiol
5. de Wit MA, Kortbeek LM, Koopmans MP, et al. A comparison of gastroenteritis in a general practice-based study and a community-based study. Epidemiol Infect
6. Cawthraw SA, Lind L, Kaijser B, et al. Antibodies, directed towards Campylobacter jejuni antigens, in sera from poultry abattoir workers. Clin Exp Immunol
7. Belongia EA, Chyou PH, Greenlee RT, et al. Diarrhea incidence and farm-related risk factors for Escherichia coli O157:H7 and Campylobacter jejuni antibodies among rural children. J Infect Dis
8. Mellin-Olsen T, Wandel M. Changes in food habits among Pakistani immigrant women in Oslo, Norway. Ethn Health
9. Wandel M, Raberg M, Kumar B, et al. Changes in food habits after migration among South Asians settled in Oslo: The effect of demographic, socio-economic and integration factors. Appetite.
2007. Epub ahead of print.
10. Hammar N, Hakala P, Jorgensen L, et al. Migration and differences in dietary habits—a cross sectional study of Finnish twins in Sweden. Eur J Clin Nutr.
2007. Epub ahead of print.
11. Van Hook J, Glick JE. Immigration and living arrangements: moving beyond economic need versus acculturation. Demography
12. Ethelberg S, Olesen B, Neimann J, et al. Risk factors for diarrhea among children in an industrialized country. Epidemiology