In an 850-bed Dutch hospital a study was earned out to evaluate the effects of the implementation of a Dutch form of primary nursing on five job characteristics (viz autonomy, responsibility, feedback/clarity, complexity/difficulty and Job demands) and on the communication and power processes in the organization of the nursing unit The study consisted of three measuring moments one preintervention (t1, March 1992) and two post-intervention moments (t2, November 1992 & t3, May 1993) Primary nursing was introduced in the two experimental units (the experimental group) after t1 No changes were made to the control group After t2, this quasi-experimental design changed, because primary nursing was also implemented in the three original control units (the control group) after this moment The research variables were measured by means of questionnaires No significant effects were found for the five job characteristics Only the improvement in the communication with the team head and the decrease in the influence of physicians on patient care in the experimental group and the improved communication with physicians in the control group were in line with expectations Changes in the amount of influence of different groups on patient care were inconsistent with what had been expected according to the nurses the influence of the nursing management increased, whereas the influence of the nurses decreased and the influence of the patients did not change In other words, the shift of power was not in the expected direction