Original Article
The Dutch diagnostic model for laboratory animal allergen sensitization was generalizable in Canadian apprentices

https://doi.org/10.1016/j.jclinepi.2008.06.020Get rights and content

Abstract

Objective

To assess the transportability of an existing diagnostic questionnaire model for the sensitization to laboratory animal (LA) allergens.

Study Design and Setting

The model was externally validated in 414 Canadian animal health apprentices. Several approaches were used: (1) no adjustment; (2) recalibration of the intercept of the model; (3) re-estimation of the intercept and the regression coefficients of predictors; and (4) model revision, by excluding the existing predictor(s) and/or including new predictor(s). The bootstrapping procedure was done following the third and fourth methods. The calibration was assessed graphically and with the Hosmer–Lemeshow (HL) test. Discriminative properties were determined by the area under the receiver operating characteristic curve (ROC area).

Results

When applied without adjustment, the model's discriminative ability was adequate (ROC area was 0.74 vs. the original ROC area of 0.76); the calibration was poor (HL test P < 0.001). The other methods yielded models with good calibration (P > 0.10) and reasonable discrimination (ROC area ranged between 0.73 and 0.75). The refitted and revised model showed a good internal validity (correction factor from the bootstrapping procedure was more than 0.90).

Conclusion

Once updated, the diagnostic model is valid and can be applied with reasonable performance in an animal health apprentice setting.

Introduction

At baseline, in a cohort study of Canadian apprentices beginning animal health and veterinary medicine career programs, the skin reactivity to work-specific laboratory animal (LA) proteins was 13.8%. The study suggested that sensitization could ensue even if a very brief specific occupational exposure had occurred [1].

To demonstrate work-related sensitization, standardized allergen preparations are required, but these are costly and may not always be available in an occupational health practice. Therefore, Meijer et al. developed a diagnostic questionnaire model for specific sensitization to LA allergens among Dutch laboratory workers exposed to rats, mice, and other rodents [2]. They demonstrated that a diagnostic model based on questionnaire items could be accurately used to predict the presence of workers at high or low risk of being sensitized without having to perform the more advanced reference test. Furthermore, they demonstrated that additional information from skin-prick test (SPT) responses to animal and non-animal common allergens improved the diagnostic performance. In a more recent study, it was shown that it is possible to develop a generic model for sensitization to occupational high molecular weight (HMW) allergens with some modifications for specific work environments [3]. The use of these diagnostic models can increase the efficiency of health surveillance, by allowing an occupational physician to predict for an individual worker the probability of being sensitized to LA allergens. On the basis of this prediction, a decision can be made to conduct additional specific tests to diagnose occupational allergies among workers with a high probability [4].

The utility of predictive models depends on how well they perform when applied to a population, which may be different from, but related to the individuals used to develop a model. In general, prediction models show a lower performance in populations other than where the model was derived [5]. For this reason, external validation of the model is necessary to address the accuracy of a model in different, but related, workers [5], [6]. Canadian apprentices at 1.4 months after entry showed a different distribution of personal and exposure characteristics compared with Dutch workers. However, both Dutch workers and Canadian apprentices are exposed to the same occupational allergens, and thus, are at risk of developing the same occupational allergies [1], [2].

Therefore, the objective of this study was to predict at an early phase the likelihood of sensitization to LA allergens in Canadian animal health technology apprentices, who had been exposed for about 1.4 months. We, thus, used the known predictors from the existing questionnaire model and externally validated this model in these trainees [5], [6], [7]. The second objective was to derive a diagnostic questionnaire model for the sensitization to LA allergens from the Canadian apprentices, and evaluate if whether or not the inclusion of predictors that were available in the Canadian setting could improve the performance of the existing model.

Section snippets

Populations

The existing diagnostic questionnaire model for the sensitization to LA allergens was derived from the first period of a cohort study investigating exposure–response relationships among 472 Dutch LA workers [2], [8]. Questionnaire items, exposure determinants, IgE serology, SPT, and lung function tests were collected from all workers. The self-administered questionnaire was based on a Dutch version of an internationally accepted respiratory questionnaire [9]. The questionnaire included

Results

The prevalence of sensitization to LA allergens among the Dutch LA workers and Canadian animal health apprentices were 27.3% and 13.8%, respectively (Table 1). The Dutch workers were older, exposed longer to LA allergens, and most of them were males. Contrarily, most of the Canadian apprentices were female, and they had just started their vocational training in animal health technology. Twelve point that 5% of the Dutch worked 20 hr/wk or more with rats, whereas all Canadian apprentices were

Discussion

We aimed at assessing the transportability of an earlier obtained diagnostic questionnaire model for sensitization to LA allergens derived from LA workers population to the animal health apprentices' setting. Our analyses suggest that with local adjustment, the diagnostic model is externally valid, and has a reasonable performance in predicting the presence or absence of sensitization to LA allergens in the Canadian animal health apprentices [5], [7].

Acknowledgments

This study was supported by the Canadian Institutes of Health Research (Grant MOP-53118), the Medical Research Council of Canada (Grant MT-12256), and the Institut de recherche Robert-Sauvé en santé et en sécurité du travail (Grant 099-164). One author (E.S.) is a research fellow with a grant from the Center for Asthma in the Workplace-Canadian Institutes of Health Research.

Conflicts of interest: none declared.

References (19)

  • S.E. Bleeker et al.

    External validation is necessary in prediction research: a clinical example

    J Clin Epidemiol

    (2003)
  • A. Hollander et al.

    Cat and dog allergy and total IgE as risk factors of laboratory animal allergy

    J Allergy Clin Immunol

    (1996)
  • P. Peduzzi et al.

    A simulation study of the number of events per variable in logistic regression analysis

    J Clin Epidemiol

    (1996)
  • D. Gautrin et al.

    Specific IgE-dependent sensitization, atopy, and bronchial hyperresponsiveness in apprentices starting exposure to protein-derived agents

    Am J Respir Crit Care Med

    (1997)
  • E. Meijer et al.

    Detection of workers sensitised to high molecular weight allergens: a diagnostic study in laboratory animal workers

    Occup Environ Med

    (2002)
  • E. Suarthana et al.

    Diagnostic model for sensitization in workers exposed to occupational high molecular weight allergens

    Am J Ind Med

    (2005)
  • E. Meijer et al.

    A strategy for health surveillance in laboratory animal workers exposed to high molecular weight allergens

    Occup Environ Med

    (2004)
  • E.W. Steyerberg et al.

    Validation and updating of predictive logistic regression models: a study on sample size and shrinkage

    Stat Med

    (2004)
  • A.C. Justice et al.

    Assessing the generalizability of prognostic information

    Ann Intern Med

    (1999)
There are more references available in the full text version of this article.

Cited by (12)

View all citing articles on Scopus
View full text