Evidence-Based Practice Paper Essay
Evidence-Based Practice Paper
The evidence-based practice guideline that I chose is titled, “Myocardial Infarction,” written by the Finnish Medical Society Duodecim. The intended users for this guideline are health care providers and physicians. The target population is individuals with suspected or known myocardial infarction. The objective of this guideline is to “collect, summarize, and update the core clinical knowledge essential in general practice” and “describe the scientific evidence underlying the given recommendations.” (Finnish Medical Society Duodecim, 2008) Clarity and Researchability of the Study’s Purpose and Question
According to the researchers, the purpose of the study is “to test how teaching format (factual versus storytelling) and restructuring the social norm of caring for others to caring for self affects how women learn to identify and respond to myocardial infarction (MI) symptoms.” (McDonald, Goncalves, Almario, Krajewski, Cervera, Kaeser, et al., 2006, p.216) I feel that the purpose of the study is significant to nursing because nurses need to educate patients about what symptoms to observe for and report to their primary care providers. Also, if this study could determine which teaching format would better assist patients in acknowledging significant symptoms of an MI needed to contact EMS, then nurses could possibly be better able to educate patients about those symptoms.
I believe that the study title of “Assisting Women to Learn Myocardial Infarction Symptoms,” is more general than the three research questions listed in the study: (a) “Are women who are taught how to recognize and respond to symptoms of an MI using a storytelling format more likely to be able to identify symptoms and plan to get help than women who are taught the same information using a factual format?”, (b)
“Does teaching women to cognitively restructure the ‘caring for others’ social norm to ‘caring for self’ make them more likely to identify symptoms of an MI and plan to call EMS than women who are not taught this form of cognitive restructuring?”, and (c) “Are women who have been taught MI symptoms and response using the storytelling format and who were taught cognitive restructuring of the ‘caring for others’ to ‘caring for self’ social norms more likely to identify symptoms of an MI and plan to call EMS than women who were provided factual information about MI symptoms and response, and who were not offered cognitive restructuring?”
(McDonald, et al., 2006, p.217-218) The two independent variables identified in the study are “teaching format (factual vs. storytelling) and social norms (caring for others first vs. caring for self)” and the dependent variable is “the posttest knowledge of MI symptoms.” (McDonald, et al., 2006, p.220)
The American Heart Association (cited in McDonald, et al., 2006, p.216) states that “heart disease remains the leading cause of death for women and kills over 248,000 women each year in the United States.” Mosca, Ferris, Fabunmi, & Robinson (cited in McDonald, et al., 2006, p.216) states that “the majority of women remain unaware that heart disease is the leading cause of death for women, despite a significant increase in awareness since 2000.” While the assumptions of the researchers are not clearly stated, I would assume that the researchers believe that women need to be further educated about the symptoms of MI in order “to avoid disabling or life-ending consequences from MI.” (McDonald, et al., 2006, p.216) Adequacy and Relevance of the Literature Review
I feel that the literature review is relevant to the problem because they discuss the differences in MI symptoms among genders and which symptoms were commonly reported by women. DeVon and Zerwic (cited in McDonald, et al., 2006, p.216-217) “reviewed studies on gender differences in MI symptoms and concluded that symptoms were similar across gender; however, in seven studies back pain, dyspnea, and nausea and vomiting occurred more frequently in women.” According to McSweeney, O’Sullivan, Cody, & Crane (cited in McDonald, et al., 2006, p.217) “women who have experienced an MI often describe additional symptoms besides chest pain, such as weakness, shortness
of breath, unusual fatigue, diaphoresis, nausea, feeling flushed or dizzy, or a heavy feeling in the arms.” The literature review also discusses the possible reasons for why women delay in contacting EMS when symptoms of MI do occur. Finnegan et al (cited in McDonald, et al., 2006, p.217) states that “women might delay responding to their own MI symptoms to meet their caregiver responsibilities.”
The literature review is logically organized, because it clearly talks about the differences among identification of MI symptoms among genders, along with the possible reasons for the delay among women in reporting their symptoms in order to receive immediate attention. The discussion about women not wanting to take care of themselves in order to continue assuming the responsibility of caring for their families, supports the research question of “cognitively restructuring the ‘caring for others’ social norm to ‘caring for self’.” (McDonald, et al., 2006, p.217)
I feel that the 23 references used were appropriate for this study. The dates of the references range from 1989 to 2005, with this study being published in the May/June 2006 issue of Public Health Nursing. Majority of the references (20/23) had to deal with heart disease, while the remaining three discussed (a) theory of planned behavior, (b) story telling as a tool, and (c) applied multivariate statistics. Both, primary and secondary sources were used as references. Agreement between Purpose, Design, and Methods
The study design described is “a pretest posttest full factorial experimental design with educational format (storytelling vs. factual) by social norms (restructuring the social norm of “caring for others” vs. no restructuring).” (McDonald, et al., 2006, p.218) According to LoBiondo-Wood & Haber (2010) “a true experimental design has three identifying properties: (a) randomization, (b) control, and (c) manipulation” (p.179). In the study, the researchers randomly assigned the participants to one of four groups, each group receiving a different type of informational MI symptom pamphlet. LoBiondo-Wood & Haber also state that “experimental designs are the most powerful for testing cause-and-effect relationships due to the control, manipulation, and randomization components” (p.185), which I think makes the design appropriate for answering the research questions in this study.
Data was collected by having participants respond in writing first to a demographic form, and then to a Heart Attack Survey, both before and after, reading an informational pamphlet about MI symptoms. “The Heart Attack Survey consisted of two questions. The first question was open ended, ‘list all of the signs and symptoms of a heart attack that you are aware of.’ The second question used a 0-10 scale, with the corresponding anchors definitely would not call and definitely would call, to measure the behavioral intention of calling 911 within 30 min if heart attack symptoms occur.” (McDonald, et al., 2006, p.219) I feel that their data collection procedure is logical and practical because paper-and-pencil instruments “are most useful for collecting data on variables that cannot be directly observed or measured by physiological instruments.” (LoBiondo-Wood & Haber, 2010, p.274)
The researchers did not clearly discuss their instruments in terms of content validity and reliability. LoBiondo-Wood & Haber (2010) define validity as “the extent to which an instrument measures the attributes of a concept accurately” and reliability as “the ability of an instrument to measure the attributes of a concept or construct consistency” (p.286). I think that both of their instruments were valid since the demographic form only measured demographic info, and the Heart Attack Survey only measured the number of MI symptoms the participants knew along their intent to call 911.
However, I do not think the Heart Attack Survey was very reliable, since the researchers did mention that many of the women did not answer the same MI symptoms on the posttest that they had mentioned on the pretest. If the instrument was reliable, I would believe that the posttest would have the same MI symptoms as listed on the pretest, along with any new symptoms the women learned from reading the informational pamphlets. Suitability of the Sampling Procedure and the Sample
I feel that the researchers used a convenience sample consisting of 113 adult women. The women were recruited by graduate nursing student data collectors in shopping malls, restaurants, and other public areas. “Inclusion criteria included of (1) female, and (2) age 25 years or older, and (3) able to speak, read, and understand English or Spanish. Exclusion criteria included (1) previous MI, (2) physician or nurse, or (3) current or past EMS worker.” (McDonald, et al., 2006, p.218) Descriptive characteristics of the sample include: (a) a mean age of around 42 years, (b) majority of participants having a high school education, with about an additional 36% having higher education, (c) most women were white, with next largest racial group consisting of black Americans, (d) nearly more than half being of non-Hispanic ethnicity, (e) few women having personal history of heart disease, while around half reported family history of heart disease, and (f) half reporting having caregiver responsibilities.
I think the sample size was adequate for the study. The researchers state that “a small effect size for the intervention effect was anticipated based on the REACT findings” and “the sample size needed for a four-group multivariate analysis of variance (MANOVA) with a power of 0.80, a significance level of .05, two dependent variables, and a small anticipated effect size was n=115.” (McDonald, et al., 2006, p.218) There were an additional seven participants, but were not included in the final sample because they provided incomplete data, by not completing both the pretest and the posttest. The researchers state that “there were no significant differences between women completing the study and women not completing the study for age, ethnicity, race, marital status, education, having health insurance, a personal history of heart disease, a family history of heart disease, or responsibility for caring for others.” (McDonald, et al., 2006, p.220)
The researchers state that “the study was approved for human subjects’ protection by the university internal review board.” (McDonald, et al., 2006, p.219) The researchers also state in their study that “each participant was provided verbal informed consent and a copy of the study information sheet.” (McDonald, et al., 2006, p.219) Correctness of Analytic Procedures
The statistical procedures named in the study are: (a) analysis of variance (ANOVA), (b) χ2 (chi-square), (c) Pearson’s r correlation, (d) t test, and
(e) analysis of covariance (ANCOVA). According to LoBiondo-Wood & Haber (2010) “analysis of covariance (ANCOVA) is a statistic that measures differences among group means and uses a statistical technique to equate the groups under study in relation to an important variable” (p.574). In the study, McDonald, et al., (2006) tested the three research questions through ANCOVA: The two independent variables—teaching format (factual vs. storytelling) and social norms (caring for others first vs. caring for self)—were entered as the grouping variable. The pretest knowledge of MI symptoms served as the covariate, and the posttest knowledge of MI symptoms was entered as the dependent variable. (p.220)
Since the study wanted to know the type of MI symptoms that the women could identify, the level of measurement for this study would be nominal. According to LoBiondo-Wood & Haber (2010) “when data are at the nominal level and the researcher wants to determine whether groups are different, the researcher uses the chi-square (χ2)” (p.326). The MI symptoms would be considered categorical variables because they could have “more than two true values”; also, since only one point was given to each symptom identified, “there was no order” to the variables. (LoBiondo-Wood & Haber, 2010, p.312)
The researchers listed several p values in their study: (a) “the factual information with the social norms restructured group had more White participants (26.6%) and less non-White participants (0.9%) than the other conditions, χ2(3) = 7.94, p < .05”, (b) “women responding to the English instruments reported significantly more symptoms (M = 6.4; SD = 2.73) than women (n=29) responding to the Spanish instruments (M = 4.7; SD = 3.26), t (111) = 2.75, p < .007”, and (c) “for the pretest the number of MI symptoms and the 911 response score were unrelated, r (113) = 0.16, p < .09, and slightly related, r (113) = 0.20, p < .04, on the posttest.” (McDonald, et al., 2006, p.220) According to LoBiondo-Wood & Haber (2010) “the minimum level of significance acceptable for nursing research is 0.05” (p.322). Clarity of Findings
The findings described under the results section state that “the women identified significantly more MI symptoms after reading the MI pamphlet.” (McDonald, et al., 2006, p.220) “Table 1contains frequencies for the entire sample of the most frequently identified MI symptoms on the pretest and posttest.” (McDonald, et al., 2006, p.220) Table 1 highlights the finding that a majority of the MI symptoms were identified more frequently on the posttest when compared to the pretest. “Table 2 contains group frequencies for identified MI symptoms, with pretest frequencies for each of the four groups preceding the posttest frequencies,” which highlights the finding that “no significant group differences emerged for storytelling and social norms.” (McDonald, et al., 2006, p.220)
While the results did show that “women generally learned three more MI symptoms to add to their previous knowledge of the commonly identified MI symptoms, chest pain, shortness of breath, and arm pain,” the results also discuss that “the storytelling versus factual format for teaching women about MI symptoms did not affect how women learned MI symptoms.” (McDonald, et al., 2006, p.221) I feel that these results show that more research is needed in order to find a way to adequately teach women the symptoms of MI and which symptoms would require them to quickly contact 911. While I do not think the conclusions are generalized beyond the sample, I agree with the researchers when they state that it is important for everyone to know that “MI symptoms are ambiguous, and that even when people are unsure about their MI symptoms they should always call 911.” (McDonald, et al., 2006, p.222)
The researchers state in their study that there were “several potential limitations” that could have impacted their results: (a) participants were recruited from the community and might not have carefully read the informational pamphlets, “potentially reducing the amount of information learned”, (b) the words ‘heart attack’ were not included in “the content of the factual group pamphlets,” which might have led the women in that group to not associate the “symptom information with MI symptoms,” (c) “the pretest and posttest were identical and separated only by the time taken to read the intervention pamphlet” which as a result “might not reflect symptoms that the women later remember and identify as potential MI symptoms,” (d) “many women did not include all of the MI symptoms that they wrote on the pretest, decreasing the score that they received for
identifying MI symptoms,” (e) “the MI symptom score did not reflect if the same symptom was included on the pretest and posttest, or if commonly occurring but less frequently recognized MI symptoms were learned,” and (f) “the 911 scale proved to be an inadequate measure of response to MI symptoms.” (McDonald, et al., 2006, p.222)
According to LoBiondo-Wood & Haber (2010) “a research study using a true experimental design is commonly called a randomized control trial (RCT)” (p.179). LoBiondo-Wood & Haber also state that “an individual RCT generates Level II evidence because of the minimal bias introduced by this design through use of randomization, control, and manipulation” (p.179). Since this study used a “pretest posttest full factorial experimental design” and “the participating women were randomly assigned, using a web-based random number generator, to one of the four experimental conditions,” I would classify this study as an RCT generating Level II evidence. (McDonald, et al., 2006, p.218)
According to the evidence-based practice guideline that I chose, “instead of chest pain, acute dyspnoea may be the primary symptom” and “the diagnosis should be made without delay since early therapy improves the prognosis decisively.” (Finnish Medical Society Duodecim, 2008) I feel that the results of this study support the guideline because it is important for women to be able to recognize the symptoms of MI early on and “be aware of their risk for an MI so that they can secure immediate emergency medical care when symptoms of an MI occur.” (McDonald, et al., 2006, p.216)
Finnish Medical Society Duodecim. (2008). Myocardial infarction. Retrieved from the National Guideline Clearinghouse website. http://www.guideline.gov/summary/summary.aspx?ss=15&doc_id=12794&nbr=006596&string=myocardial+AND+infarction LoBiondo-Wood, G., & Haber, J. (2010). Nursing research: Methods and critical appraisal for evidence- based practice, 7th Ed. St. Louis: Mosby.
McDonald, D. D., Goncalves, P. H., Almario, V. E., Krajewski, A. L., Cervera,