One hundred and thirty-four older adults completed the 17-item questionnaire twice within a 2-week interval in random order; 98 complete questionnaires were available for full analysis. Interviewer administration was completed over the telephone. Intraclass correlation and kappa coefficients were calculated to determine agreement.
There is very good agreement between overall self- and telephone-administered SCREEN II scores (intraclass correlation coefficient = 0.74). When using a cut-point (score <50) to identify those at risk of malnutrition, agreement is good (κ = 0.53). Increasing age appears to consistently affect agreement. Individual questions which had poorer agreement included: food/fluid intake questions, supplement use, and intention to change weight.
The self- and interviewer-administered SCREEN II can be used interchangeably in practice or research, if total scores or the risk cutoff are used. Comparison of individual questions completed in different modes is not recommended.