Validity and reproducibility of a screening examination for neurological abnormality in persons exposed to methylmercury

The validity and reliability of methods of screening for neurologic abnormality were assessed as part of an investigation of an outbreak of methylmercury exposure in two northern Canadian communities. Four hundred and forty-five Cree Indians were examined by one of five neurologists in a complete neurologic examination and by a trained paramedical observer in a short screening examination which included a selection of tests from the complete examination. The screening examinations were recorded on videotape and those for 176 men were reviewed by the five neurologists and the paramedical observer 1 year after the field studies. The prevalence of abnormality assessed in the field screening examination was greater than that assessed during the complete neurologic examination, for neurologic features included in both examinations. However, agreement between examinations in identifying individuals with abnormality was poor with the sensitivity of the screening examination falling under 50% for half of the neurologic features examined. In contrast, specificity was over 80% for 14 of 18 features. Review of the videotapes revealed marked interobserver variation in the assessment of the prevalence of neurologic abnormality and poor agreement with the neurologic examinations in the identification of abnormality in individuals, with κ < 0.2 for most neurologic features. The levels of agreement between the neurologic examinations and the screening examinations conducted in the field and by videotape review suggest that neither screening examination provides equivalent information in the identification of the presence of abnormality to that obtained in the neurologic examination.