1. Tija et al. (2007) developed and tested guidelines for testing and monitoring high-risk drugs. First, they contacted health care providers regarding a list of the high-risk drugs, reaching a consensus after 2 rounds of an Internet-based Delphi process (a structured method of methodical communication using surveys completed by experts). Additionally, participants were asked to rate drugs on efficacy, toxicity, and drug interactions using a 5-point Likert scale. Once the high-risk drug list was completed, recommendations for laboratory testing were created by research pharmacists and consensus of local physicians. Examination of records at a large group practice showed that high-risk drugs were less likely to be prescribed than low-risk drugs, but when high-risk drugs were used, patients often did not receive appropriate monitoring.
The resulting guidelines were tested through a pilot study.
Fisher et al. (2008) tested a new method, conjoint analysis, to evaluate patterns of decision making by nurses who made decisions by proxy for individuals with intellectual disability (ID). Conjoint analysis is a method that determines how individuals make decisions — what factors they consider and how much importance or weight the factor has in the final decision. It uses simulations and multiple experiments. In this case, 23 emergency department (ED) nurses participated in the scenario of an individual with ID and a dental abscess coming to the emergency room. Factors including person’s age, mental competence, family input, and functional status were varied in creating the simulations. Individuals with family approval, good future health status, and ambulatory status were most likely to be approved for care, whereas those with family disapproval, deteriorating health, and bedfast status were less likely to be approved.
Use your promo and get a custom paper on
"Statistical Analysis in Nursing".
2. The use of rating scales (such as Likert scales) and other types of rank-ordering are common in nursing literature. Data gathered in this way is ordinal, so that the statistics commonly used are Χ2, Mann-Whitney U, or Analysis of Variance (ANOVA).
3. The nursing literature typically uses statistical techniques that are acceptable for data that does not meet standard parameters, i.e., non-parametric tests like the F-test and Mann-Whitney U. This is due to the fact that some important variables in nursing care are difficult or impossible to define in an interval fashion (e.g., demographic characteristics such as ethnicity or socioeconomic level, pain rating scales, self-report questionnaires). Less commonly, if the data are interval level, t-tests and linear regression are used.
3a. As indicated above, in nursing research it is common for one or more variables to be nominal or ordinal. This prohibits the use of most parametric statistics for those variables.
