Access Type

Open Access Dissertation

Date of Award

January 2016

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Education Evaluation and Research

First Advisor

Shlomo Sawilowsky

Abstract

The purpose of this study was to characterize patients hospitalized with acute decompensated heart failure with and without low systolic blood pressure using exploratory factor analysis (EFA). Direct and surrogate measurements were measured. The aim was to use EFA for data reduction to elicit a parsimonious set of factors summarizing the relationships between variables by measuring intercorrelations of the clinical variables collected as part of standard care, and abstracted from electronic medical records.

A better understanding of the characteristics and outcomes of the target group could potentially lead to individualized treatment modalities tailored to effectively and economically improve care. Patients hospitalized are at a high risk for adverse outcomes after discharge.

Prospectively collected new data is expensive, labor-, and time- intensive while the use of existing data allows a quicker, more efficient and less expensive source. A large urban, academic teaching hospital was the study site. Wayne State University Human Investigation Committee and Henry Ford Internal Review expedited review approval was obtained.

Eligible cases were patients hospitalized with a primary diagnosis of acute decompensated heart failure for the 2014 year. Variables collected were identified based on review of the literature, Framingham criteria, clinical relevance, and were routinely availability.

As is the case in empirical studies, determining sample size in EFA, a large sample size technique, is based on the minimum necessary to obtain reliable results from the analysis. Guidelines or a rule of thumb by expert opinions such as Gorsuch (1983) and Kline (1994) include absolute numbers of at least 100 cases.

Dimension reduction of factors via SPSS (ver 23) was conducted on all cases regardless of presenting systolic blood pressure (Group 1), cases with normal to high systolic blood pressure (Group 2) and cases with low systolic blood pressure (Group 3) separately, for a total of groups. All cases were screened for entry criteria and the first 300 chronologically dated cases were identified.

EFA was conducted on the data abstracted from 300 electronic medical records. The major findings of the study were that two factors, Anemia and Kidney Function were seen across the three groups. Several individual factors that affect kidney function were found. Data reduction using EFA is a highly pragmatic function. Computer software programs such as SPSS® allow for quick and easy computations and a large number of variables can be directly imported from databases such as Excel®. However, EFA is a complex procedure with fewer absolute guidelines or rules for selecting options compared to other statistical approaches. The steps taken were detailed, justified by the literature reviewed and alternate choices were discussed. The seven stages in factor analysis design as outlined by Hair et al. (2006) were employed in this analysis.

The factors identified in each group using EFA can be tested in a future confirmatory factor analysis study. Once these factors are the confirmed, an Acute Decompensated Heart Failure Risk Model can be developed for Emergency Room Resident Training within the context of evidence-based medicine.

The pedagogical approach in medical education where instruction is provided by the experienced physician to the novice, namely the medical resident, is in conflict of adult learning theory leading to a contributing factor to the success or failure of teaching evidence-based medicine. Risk models are powerful tools for assessing biomedical significance but the importance of how to teach and use a risk model cannot be underestimated. Building on what emergency room residents may know, or determining whether there is a knowledge deficit is extremely important. A step-by-step process layering information on what is already known (present level of understanding) by the leaner o the required knowledge level is needed. The results of the EFA conducted indicates that patients with and without low systolic blood pressure share common factors. These factors, anemia and kidney function also directly affect blood pressure. If emergency room residents do not know that these factors are shared, then the first step would be to educate them about this finding. If emergency room residents do know from prior knowledge, then the teacher would be adding to their knowledge base when teaching the residents the use of the risk model as is described by Knowles, Holton, and Swanson (2005) as the first underlying assumption. The shift to student centered learning is based on adult learning theory (Spencer, 1999) and transformational learning should be employed.

Share

COinS