Accuracy or bias in nurses' ratings of patient compliance: a comparison of treatment modality.

OBJECTIVE: The study examined the accuracy of nurses' assessments of patient compliance and identified factors influencing these assessments, including possible biases. DESIGN: Nurses' assessments of compliance, lab serum levels of dietary compliance, and interdialytic weight gain (IWG) were collected and compared to each other. End-stage renal disease (ESRD) patients on peritoneal dialysis (PD) and hemodialysis (HD) were compared on these measures and their compliance. SETTING: The study was conducted at the tertiary care university hospital at Stony Brook. PATIENTS: Data were collected for 62 ESRD patients (38 male, 24 female; mean age 54 years). The sample consisted of 26 HD and 36 PD patients. INTERVENTIONS: Nurses rated patients' compliance with fluid restrictions (HD patients only) and overall dietary compliance, as well as individual indicators of compliance including protein, potassium, and phosphorus compliance on a 7-point rating scale. Interdialytic weight gain, dietary (serum BUN and K levels), and medication compliance were recorded from charts for a 3-month period for each patient. MAIN OUTCOME MEASURES: The main outcome measures were the correlations between nurses' ratings of compliance and medical compliance data and the regression coefficients, which indicate the relative importance of each of the factors that nurses use to make their compliance ratings. RESULTS: Nurses' ratings for patients in both treatment modalities were highly correlated with the medical data for measures of fluid (r = 0.66, p < 0.001), potassium (r = 0.36, p < 0.01), and phosphorus (r = 0.36, p < 0.01). A regression analysis indicates that potassium (beta = 0.48, p < 0.001), phosphorus (beta = 0.19, p < 0.05), and protein (beta = 0.31, p < 0.01) all significantly contributed to nurses' assessments of patients' overall compliance. However, nurses' ratings of education levels for patients in both treatment modalities were not associated with phosphorus (r = 0.07, p < 0.61), protein (r = 0.18, p < 0.23), or potassium (r = 0.03, p < 0.85) measures. Finally, regressions revealed that nurses used personal knowledge of the patients when rating noncompliant patients (beta = 0.49, p < 0.05) but not when rating compliant patients (beta = 0.05, p < 0.75). CONCLUSIONS: Nurses rely heavily on medical records to rate patients' compliance and to make accurate assessments. Nurses also use several individual indicators (lab values and IWG) to rate overall dietary compliance, suggesting a thorough assessment. While assessments are not biased by personal factors such as nurses' perceptions of patients' education levels, nurses do rely on personal knowledge when rating noncompliant patients.