A new rating form for use by nurses in assessing residents' humanistic behavior

To determine the reliability and validity of an evaluation form for assessing the humanistic behavior of internal medicine (IM) housestaff. The form is for use by nurses.

Evaluations were gathered three times during the 1987-88 academic year. Generalizability coefficients (interpreted like traditional reliability coefficients) were generated to establish the form's reliability, while data from attending physicians and from housestaff evaluation committee members were used to help establish its validity.

Three hospitals in central Ohio: a large university tertiary care center, a large private hospital, and an urban community hospital.

The nurse raters were volunteers solicited by their head nurses. The criteria governing their participation were two years of postgraduate experience in nursing and regular contact with residents, which was self-determined. All IM residents who had worked on a medicine inpatient service at least once during the months under study were included. A total of 493 nurses and 116 residents participated.

Sixty-four percent of the generalizability coefficients were 0.90 or higher, and 82% were above 0.75, indicating stable, reliable ratings. The nurses' ratings were positively and significantly correlated with attending faculty's and evaluation committee members' ratings (r = 0.38, p less than 0.01; r = 0.49, p less than 0.001).

The evaluation form and the nurses provided consistent, reliable information about medical residents' humanistic behavior; data from five to six nurses should provide statistically reliable ratings using this form. Also, nurses' data yielded information somewhat different from those provided by physicians, suggesting that the form is a useful instrument for assessing this dimension of residents' performance.