PIP: Between January and August 1992 in Greece, researchers conducted telephone interviews with parents of children with leukemia (136 cases, most in Attica [Athens and its environs] and the others on the island of Crete) and with parents of children not afflicted with leukemia (187 controls) to determine whether childhood leukemia may be a result of a subclinical infection at an earlier age. They controlled for place of residence when they conducted the multiple logistic regression analyses. They used attendance at a day care facility as a proxy for earlier infection because children come in close contact with each other at day cares, thereby exposing them to many infectious agents. Children who attended a day care had a lower relative risk of developing leukemia than those who did not attend day care (.67), but attendance did not have a significant protective effect. It did appear to have a significant protective effect for children who attended day care during infancy (for at least 3 months during the first 2 years of life), however, (relative risk = .28; p = .03). Significance remained even when the researchers considered different operational definitions of early attendance. Exposure to magnetic fields appeared to have a protective effect also against the development of childhood leukemia, but this effect did not reach significance (p = .07). The relative risk for 100 m from an electricity substation was 0.35. There was a slight, but insignificant increase in the relative risk for children living close to power lines, however (1.19 for 5 m; p = .63). Significant risk factors included young age at diagnosis and mothers with less than high school education.