Telephone audit for monitoring stroke unit facilities: a post hoc analysis from PROSIT study.

Background: Although several valid approaches exist to measure the number and the quality of acute stroke units, only few studies tested their reliability. This study is aimed at establishing whether the telephone administration of the PROject of Stroke unIt ITaly (PROSIT) audit questionnaire is reliable compared with direct face-to-face interview.

Methods: Forty-three medical leaders in charge of in-hospital stroke services were interviewed twice using the same PROSIT questionnaire with 2 different modalities. First, the interviewers approached the medical leaders by telephone. Thereafter, they went to the hospital site and performed a direct face-to-face interview. Six independent couples of trained researchers conducted the audit interviews. The degree of intermodality agreement was measured with kappa statistic.

Results: We found a perfect agreement for stroke units identification between the 2 different audit modalities (K = 1.00; standard error [SE], 1.525). The agreement was also very good for stroke dedicated beds (K = 1.00; SE, 1.525) and dedicated personnel (K = 1.00; SE, 1.525), which are the 2 components of stroke unit definition. The agreement was lower for declared in use process of care and availability of diagnostic investigations.

Conclusions: The telephone audit can be used for monitoring stroke unit structures. It is more rapid, less expensive, and can repeatedly be used at appropriate intervals. However, a reliable description of the process of care and diagnostic investigations indicators should be obtained by either local site audit visit or prospective stroke register based on individual patient data.