Facebook
Twitter
LinkedIn
Email
 |
Checking for direct PDF access through Ovid | |
Excerpt
This response concerns doctor Harrison and Wells' comments regarding "Predicting Pressure Ulcer Risk: Multisite Study of the Predictive Validity of the Braden Scale," by Bergstrom, Braden, Kemp, Champagne, & Ruby, Nursing Research, Vol. 47, No. 5.
The letter writers note: a) we looked at various populations within one study rather than comparing findings with other populations; b) using the Braden Scale to assist in allocation of resources (including care) requires careful consideration; and c) research needs to get beyond psychometric testing to applications and translating data for local decision making. We will comment briefly on each of these observations.
Comparing findings with other populations. The purpose of the paper was to evaluate the predictive validity of the Braden Scale in a variety of settings to determine the critical cutoff point for classifying risk and to determine if the cutoff point differed between settings. We studied six settings in three states and summarized the data in detail, providing lengthy tables (thanks to the generosity of Nursing Research). Following such a lengthy report, we did not then summarize all other work that had been done to date (the letter suggests a few of those studies), as we have in other articles. It is an interesting exercise and most readers find it informative, but space prevailed. We are familiar with the studies cited and have shared our research protocol with many clinicians and investigators; therefore, the similarities between this study and others is more than happenstance.
Allocation of resources. When using a risk-assessment tool, the cutoff point for assessing risk is a major decision faced by each agency. There must be balance between the sensitivity and specificity when deciding the cutoff point. If the line is conservatively drawn to prevent a larger proportion of pressure ulcers, more preventive care is required and health care dollars rise; if the line is drawn to reduce the cost of preventive strategies, more ulcers and the concomitant costs of treatment may result. The high NPV reported demonstrates that care is not being withheld from those who need it; while the low PPV demonstrates that there may be both over prediction and effective care in place. When effective care is in place, patients at risk are less likely to get ulcers. In all of our studies, nurses continue to give usual preventive care while we determine the predictive validity of the tool. This no doubt influences the PPV, yielding lower PPV. Kudos may belong to the risk reduction effort. Each agency has to determine the risk tolerance and cost/benefit ratio.
Beyond psychometric testing to translating data for local decision making. We are the first to acknowledge that the Braden Scale is not a perfect tool; however, we are pleased that it has proven to be useful in many clinical agencies. We have no doubt that variables other than those identified in the Braden Scale may influence pressure ulcer risk, but those variables in the Braden Scale reflect patient risk factors common to most populations. Adding other variables, in our studies and those of others, has not proven to be effective or parsimonious. Statistically, the likelihood of improving risk prediction in a cost effective manner is less likely to be fruitful than studying methods for reducing risk once it is identified. Furthermore, testing interventions based on level of risk is more likely to assist local experts in planning care for their institution.
A key ingredient in the success of any protocol for reducing pressure ulcer risk is the philosophy of continuous improvement.
