workplace absenteeism and duration of employment are growing concerns among health care organizations, in terms of high costs, lost productivity, and job turnover. the aim of this study was to identify sociodemographic and work-related factors that predict workplace absenteeism, duration of workplace absenteeism, and duration of employment in cancer care workers. this study included 244 cancer care workers employed by cancer care ontario at the northeastern ontario regional cancer centre (neorcc), sudbury, ontario, between january 1 st, 1998 and december 31st, 2003. sickness absence data were obtained from employee attendance records and human resource databases. logistic regression analyses were used to estimate the risk of high sickness absence (three or more absence events per year), high sickness absence duration (five or more days of absence per year), and high mean duration of absence (1.5 or more days absent per event). a cox proportional hazards regression model, using duration of employment (in years) as the outcome measure (n=182), was used to estimate the risk of job separation at the neorcc. the highest rates of sickness absence were observed during the winter months (january to march) and lowest during the summer months (july to september) (p=0.001). the daily mean rate of sickness absence was lowest on fridays and highest on tuesdays, however the difference in sickness absence rates between the days of the week was nonsignificant. low job level (odds ratio (or)=2.54, 95% confidence interval (cl): 1.22- 5.27) was associated with a significantly increased risk of sickness absence of three or more absence events per year. male gender (or=0.28, 95% cl: 0.13-0.60) and workers employed for a short-term period (less than five years) (or=0.40, 95% cl: 0.22-0.72) displayed a significantly lower risk of high sickness absence. low job level (or=3.38, 95% cl: 1.57-7.28) was associated with a significantly increased risk of sickness absence of five days or more per year. male gender (or=0.32, 95% cl: 0.15-0.70) and workers employed for a short-term period (less than five years) (or=0.39, 95% cl: 0.21-0.71) had a significantly lower risk of sickness absence duration of five days or more per year. low job level workers (or=2.20, 95% cl: 1.01-4.81) had a significantly increased risk of high mean duration, that is, sickness absence of 1.5 days or more per absence event, compared to high job level workers. clinical workers (or=9.55, 95% cl: 2.03-44.95) and non-clinical workers (or=6.01, 95% cl: 1.23-29.38) displayed a significantly increased risk of high mean duration sickness absence compared to administration workers. workers employed for a short-term period (less than five years) had a significantly lower risk of high mean duration (or=0.45, 95% cl: 0.25-0.82). at the end of the study period, 62 (25%) employees had their job separated. duration of employment from the start date of employment at the neorcc ranged from 1.06 years to 30.13 years. the mean duration of employment in cancer care workers who remained employed at the neorcc was 8.98 (sd 5.28) years. permanent employment (or= 0.21, 95% cl: 0.12-0.38) was associated with a significantly decreased risk of job separation.