Kristensen, P., Hilt, B., Svendsen, K. et al. Eur J Epidemiol (2008) 23: 11. doi:10.1007/s10654-007-9203-5
We investigated a suspected cluster of lymphohaematopoietic cancer in departments of biology and chemistry at a Norwegian university. A historical cohort of students and employees (n = 7,189) were followed for cancer for 96,936 person-years in the Cancer Registry of Norway. Overall, 12 cases of lymphohaematopoietic cancer were identified, which was close to the expected number (standardised incidence ratio 1.1, 95% confidence interval (CI) 0.6–1.9). The SIR for leukaemia was 1.8 (95% CI 0.6–4.2) whereas incidences for non-Hodgkin lymphoma and Hodgkin’s disease incidence were close to the expected. Distinct rate differences were observed in subgroups, and the strongest adjusted associations were found for Ph.D. students (rate ratio (RR) = 4.8, 95% CI 1.1–20.3) and students who participated in basic organic chemistry laboratory course (RR = 4.4, 95% CI 1.2–18.8). The latter category was identified a priori because it was the only student activity with known use of benzene according to the university administration. This is an observational study with some limitations, and hence, caution is warranted when drawing inferences on the causal relationship between university activities and cancers under study. The study provides an example that cluster investigations can answer questions of public importance despite a more limited value regarding causal inference.