Increased mortality is often reported among workers in short term employmen
t. This may indicate either a health-related selection process or the prese
nce of different lifestyle or social conditions among short term workers. T
he authors studied these two aspects of short term employment among 16,404
Danish workers in the reinforced plastics industry who were hired between 1
978 and 1985 and were followed to the end of 1988. Preemployment hospitaliz
ation histories for 1977-1984 were ascertained and were related to length o
f employment between 1978 and 1988. Workers who had been hospitalized prior
to employment showed a 20% higher risk of early termination of employment
than those never hospitalized (rate ratio (RR) = 1.20, 95% confidence inter
val (Cl) 1.16-1.29), and the risk increased with number of hospitalizations
. For workers with two or more preemployment hospitalizations related to al
cohol abuse or violence, the rate ratios for short term employment were 2.3
0 (95% Cl 1.74-3.06) and 1.86 (95% Cl 1.35-2.56), respectively. An unhealth
y lifestyle may also be a determinant of short term employment. While it is
possible in principle to adjust for lifestyle factors if proper data are c
ollected, the health-related selection of workers requires careful consider
ation when choosing a reference group for comparative studies of cumulative
occupational exposure.