A retrospective case note analysis of the recognition and management of deteriorating patients prior to critical care admission

Abstract

This study explores the use of early warning scores (EWS) in deteriorating patients. These are widely used tools to measure vital signs and highlight abnormal physiology in acutely unwell patients. Measurements of the process in the management of the deteriorating patient includes time to first assessment of such patients. The level of clinician involved in the subsequent management is also investigated to determine whether escalation of care was appropriate. This work is a retrospective case note analysis of the recognition and management of deteriorating patients prior to critical care admission.

Research Questions

1. What violations in the optimum process are associated with sub-optimal recognition and management of deteriorating patients and delayed critical care admission in patients triggering early warning scores in acute care wards? 2. Are there independent variables which can predict the delay in the recognition and management of deteriorating patients and subsequent critical care admission?

Methods

The literature was reviewed to determine the optimum process of recognition and management of deteriorating patients in acute care wards. A data collection tool was then specifically designed and locally validated to extract objective data from the case records. A sample of 157 patients admitted to critical care from acute wards over a 6 month period were included in the study. The case records were then retrospectively reviewed and information was extracted using the data collection tool.

Results The accuracy and frequency of early warning scores were measured and findings demonstrated that 59% of Early Warning Scores (EWS) were miscalculated. The most frequent of those miscalculated were the intermediate scores (4 or 5) (error rate - 52%) followed by the higher scores (6 or more) (error rate - 32%). The least frequently miscalculated were the lower scores (0 -3) (error rate 15%).

Descriptive data from the sample such as age, ward, diagnosis, time of hospital admission, time and day of transfer / EWS triggering were included. From the total case records reviewed, 110 patients had abnormal Early Warning Scores (4 or more) and were included in the inferential data analysis.

The independent variables related to the processes objectively measurable in the recognition and management of deteriorating patients were included. After descriptive analysis the independent variables were cross-tabulated with the dependent variable using Pearson chi-square. The dependent variable was identified from the literature. This was whether time from triggering an abnormal EWS to critical care admission was delayed more than 6 hours. The subsequent predictor variables were then entered in to a binary logistic regression model for statistical analysis using SPSS version 21 software.

Binominal Logistic Regression Analysis identified three significant variables predicting delay of the recognition and management of deteriorating patients. • Frequency of EWS measurement not increased appropriately • Length of stay prior to critical care admission 12-36 hours • If no consultant review during 6 hours of abnormal EWS

Implications for Future Practice

This study highlights areas of risk in the detection of patients’ clinical deterioration in acute wards. These findings should guide quality improvement to prevent unnecessary morbidity and mortality. As a key area of patient risk included the lack of frequency and accuracy of EWS measurements, staff education is required to ensure staff are given the appropriate knowledge to understand the use of the tool. Regular review of the frequency of measurement is also required as this was statistically significant in the delay to critical care admission. The high risk time from admission of 12-36 hours needs further investigation. This study also highlights the need for senior decision makers to be involved in the care of deteriorating patients to improve outcomes.