Development of clinical quality indicators for health care-associated infection in Taiwan

Background: Inter-hospital benchmarking of health care-associated infection (HAI) has being conducted with the objective of improving the quality of health care and patient safety. A large body of literature has addressed issue relating to select the quality indicators of HAI. There are limited previous studies have evaluated the appropriateness of HAI indicators.

Objective: To evaluate the existing HAI indicators and to determinate which indicators could be applicable and reliable for inter-hospital comparison.

Methods: Candidates of HAI indicator were drawn from various existing quality indicator projects, action plans, evidence-based intervention measurements, and related publications. A structured anonymous questionnaire was designed to investigate how the reviewer rates each candidate indicator against four principal dimensions: importance to patient safety, scientific soundness, potential feasibility, and comparability on benchmarking. Opinions from the multidisciplinary expert panel were collected via two rounds of the modified Delphi method.

Results: One hundred and twenty nine indicators were obtained from literature review. Among these, 25 non-duplicated indicators with operational definition and could be analyzed from existing databases in Taiwan were chosen for further evaluation. Of the 25 candidate indicators, 18 (72%) indicators were rated as importance to patient safety by the 12 experts, 6 (24%) indicators were rated as comparable on benchmarking. Only one indicator, staffing level of infection control practitioners, was found to meet all criterions that defined by Rand appropriateness method for four principal dimensions. As only one structured indicator was included, seven additional indicators, including central line-associated bloodstream infection rate, ventilator-associated pneumonia rate, urinary catheter-associated urinary tract infection rate, knee arthroplasty superficial incision site infection rate, knee arthroplasty deep incision site infection rate, rate of occurrence of specific hospital-acquired antimicrobial resistance strain infections, and staffing level of infection control physicians, failed to fulfill the “agreement” criterion were considered as the second tier of our indicator list.

Conclusions: It is a need to set the HAI quality indicators to assure patient safety. However, under current situation, there are no enough reliable and applicable indicators which can be applied to hospital benchmarking. This study is an initial step towards developing HAI indicators that will able to serve for inter-hospital comparison. Multi-source monitoring system and new indicators explored are needed for national benchmarking in Taiwan’s hospital.