Identification of a reliable subset of process indicators for clinical audit in stroke care: an example from Australia

Abstract

Abstract:

In 2007, the National Stroke Foundation (Australia) conducted the first national audit of acute inpatient services for stroke with >30 indicators. Routine collection of many variables can be a burden for clinicians, and methods to identify practical subsets are needed.

Purpose:

To identify a subset of indicators to facilitate meaningful longitudinal comparisons, and to reliably represent the full suite of indicators.

Methods:

3 steps taken:

1. Value-based judgments to establish a subset were made by the National Advisory Committee (of the National Stroke Foundation) given: level of evidence, clinical relevance, consumer importance, and ability for international comparisons

2. Statistical analyses were used to identify subsets that could predict patient outcome and total process score

3. Comparisons of steps 1 and 2: logistic regression and estimation of agreement using intraclass correlation coefficient and Lin’s concordance coefficient.

Results:

Value-based judgments resulted in 14 indicators being selected; parametric methods identified 12. Six indicators were consistently selected: stroke unit care; aspirin, physiotherapy assessment, and speech pathology assessment within 48 hours; a care plan; and antihypertensive medication at discharge. The scoring method based on the value-based indicator subset demonstrated excellent agreement with total process scores of hospitals.

Conclusion:

Selection of an indicator subset requires consideration of several factors. Indicators selected by experts were robust.

Keywords::

• stroke
• processes of care
• performance
• acute care
• Australia