Researchers in China have created a prediction model for acute respiratory distress syndrome (ARDS) patients who require mechanical ventilation, according to a recent study published March 30 on the Public Library of Science website.

Lead author Zhongheng Zhang with the Jinhua Municipal Central Hospital in Zheijian, China wrote that the model contains eight covariates readily available in routine clinical practices and can be applied in all manners of critical care setting.

ARDS is a severe form of lung injury most often observed in intensive care units, with an incidence rate of approximately 40 to 80 per 100,000 patients. The mortality rate of ARDS is between 50 and 60 percent.

In an effort to reduce mortality rates associated with ARDS, risk factors must be identified that lead to poor clinical outcomes.

“In the present study, we aimed to develop a prediction model for ARDS patients requiring mechanical ventilation,” the authors wrote. “The principal in developing the model is a balance between parsimony and model fitting. Furthermore, the variables included in the model should be readily available in routine clinical practices.”

The researchers conducted the study through secondary analysis of data collected from 33 hospitals between August 2007 and July 2008. In all, 282 ARDS patients were included in the model development.

From the dataset variables like age, gender, body mass index and co-morbidities were extracted.

“Univariate logistic regression model was performed to screen factors associated with mortality in ARDS patients requiring mechanical ventilation,” Zhonghen and team wrote. “The dependent variable was a binary outcome with ‘1’ indicated death and ‘0’ indicated survival.”

The authors wrote that for a model to be useful in a clinical setting, it needed to be easy to use.

“In the study we incorporated variables that were readily available in routine clinical practice,” they wrote.

Some of the independent risk factors for death that the study found included old age, vasopressor use and platelet count.

Overfitting was a major concern for the authors as the model was established with a single cohor and no external validation.

“We employed bootstraps procedure to shrink coefficient and chose model with the principal of parsimony. However, the result showed that the bootstrap procedure did not change the coefficient, indicating that the estimated coefficient is less likely to be biased,” they wrote.