The problem with using computer programmes to assign ethnicity: Immigration decreases sensitivity

Objectives: It is recognized that ethnic group is important in describing differences in infection and disease, but is often not routinely available to surveillance systems. Computerized programmes, such as NamPehchan, can assign ethnicity according to name; however, sensitivity and positive predictive value (PPV) can vary. The aim of this Study was to assess whether the sensitivity and PPV of NamPehchan had changed, after an observation that surnames previously associated with South Asians were increasingly reported as Black. Study design: Cross-sectional. Methods: NamPehchan was used to classify women as South Asian using name, and compared with the gold standard (midwife-reported ethnicity). Sensitivity and PPV were calculated overall and by year. Frequency of infection by ethnic group was estimated. Results: A total of 627 women positive for hepatitis B surface antigen were identified. The majority were from minority ethnic groups, particularly Asian. The overall sensitivity of NamPehchan was 74.5% and PPV was 68.5%. Almost 50% of Black African women were classified as South Asian by NamPehchan. Conclusions: Immigration from African countries has reduced the sensitivity of NamPehchan in this group. Care is needed when using NamPehchan for groups which include Africans from Muslim areas, as misclassification is likely to occur. (C) 2009 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.