Data Analysis Techniques and Statistical Tests’ Worth

What Are the Data Saying?

Medical research can be developed in a variety of ways, depending on a researcher, his or her background knowledge, and available resources. As soon as a topic is chosen, and research questions are developed, the establishment of research design and size is required.

Statistics are used to make and analyze interventions and their role for a population. There are many quantities to be calculated during statistical research, including p-values, deviations, errors, and intervals, and the choice of a statistical test is a significant step to be taken. In this paper, data analysis techniques, the types of studies, and the worth of statistical tests will be discussed through the evaluation of parametric and nonparametric qualities, as well as reliability and validity measures.

Study Types

To develop a direct practice involvement (DPI) project on the use of capnography during resuscitation of patients in the coronary care unit (CCU), different studies have to be chosen and analyzed. Qualitative articles like the one written by Iyer, Koziel, and Langhan (2015) help explore the perceptions on the topic and understand all the important techniques, reasons, and opinions. Quantitative studies aim at investigating phenomena and practices via the analysis of statistical and computational techniques.

Turle, Sherren, Nicholson, Callaghan, and Shepherd (2015) are the authors who show one of the possible examples of capnography applications during cardiac arrests in patients within the frames of qualitative research. Mixed methods studies are used when the collection and analysis of qualitative and quantitative data are required to achieve the best results. The goals of this type of study depend on the topic and the authors’ intentions. For example, Langhan, Kurtz, Schaeffer, Asnes, and Riera (2014) investigated the reasons for capnography limited use in acute care and discussed the major barriers to its implementation. Qualitative, quantitative, and mixed methods data say a lot about capnography and its use during resuscitation.

Statistical Tests

A decision of what statistical test should be used in research depends on the chosen design, the intentions to distribute the data, and the types of variables offered for the study. Chi-square tests are used to identify the association between variables and paired or independent t-tests help analyze the difference between variables. If the information does not meet parametric measures, nonparametric tests like Wilcoxon tests are implemented to calculate the differences between variables. Langhan et al. (2014) used Fisher exact test and student t-test to analyze the demographics of participants and the differences between capnography use and nonuse. The sample size of the study was small, and it was reasonable to compare variables in terms of their values.

However, it is not the only test that could be effectively used in the studies about capnography. For example, Arafa, Abouzkry, and Mohamady (2018) proved the worth of a paired t-test along with a Chi-square test. Such an approach properly works when a researcher has enough evidence to prove the correctness of a null hypothesis. When the sample size is large, Chi-square tests are used to check a nominal variable and analyze it through the already established theoretical expectations.

Many ways can be offered to check variables and focus on pre- and post-intervention conditions. In the study by Lin, Huang, Hsiao, Chiang, and Jeng (2017), independent and paired t-tests were used to compare pre- and post-interventional treatment. There is almost the same statistical significance in the study by Langhan et al. (2014). Finally, the Mann-Whitney U test was a nonparametric alternative in Lin et al. (2017) study. This statistical test evaluates two independent variables with equal means on a final dependent outcome. In general, Chi-square tests and t-test are the most frequent options for quantitative researchers.

Applicability of a Statistical Test

To introduce a strong and credible quantitative study, it is necessary to use a combination of several statistical tests. The experience of Langhan et al. (2014) and the analysis of other quantitative projects prove the worth of Chi-square tests used along with paired t-tests. The applicability of the Chi-square test may be explained by the presence of one nominal variable (capnography use during resuscitation) with several values. Values are the dependent variables that include the recovery dynamics that are observed after cardiac arrests, patient satisfaction about the quality of care offered by nurses, and the possibility to monitor the level of end-tidal carbon dioxide (ETCO2).

The paired sample t-tests are effective to determine the difference between observations before and after capnography is used. There are two types of hypotheses in the DPI project. One of them is that cartographers may influence the efficacy of resuscitation in patients with cardiac arrests, and another is that the same tool does not lead to evident changes. The necessity to test the idea through the application of several statistical procedures cannot be ignored in this study.

Parametric vs. Nonparametric Tests

In the majority of statistical analyses, researchers prefer to use parametric tests to obtain clear results. However, special attention should also be paid to an effective alternative, known as nonparametric tests. They are also used as distribution-free tests if a study contains weak evidence to further specific distribution. In other words, if the data does not meet enough assumptions required for paramedic tests like t-tests, analysis of variance (ANOVA), or Pearson coefficient of correlation, nonparametric ones like Wilcoxon rank-sum or signed-rank tests, Kruskal-Wallis test, or Spearman’s rank correlation, should be used.

As a rule, parametric tests have bigger statistical power compared to nonparametric tests. However, when the sample size is not big, and assumptions (outliers) cannot be removed, nonparametric tests are accessible. In other words, when parametric tests are characterized by statistical data distribution, and nonparametric tests have no data distribution and include the analysis of medians’ differences, which prevents the rejection of a null hypothesis. In this case, a t-test is a parametric test, and a Chi-square test is its supportive alternative to the study about the use of capnography.

Reliability and Validity

Three types of studies have their reliability and validity factors. These principles are the fundamentals of medical or scientific research. Reliability is the idea of repeating the same investigation under the same or similar conditions and the possibility to achieve new results and contribute to the discussion of the chosen topic. The researchers of quantitative articles under consideration use the same statistical tests to prove their hypotheses. In qualitative and mixed methods articles, the authors rely on human judgments and opinions about the issue, obtaining information from face-to-face interviews and observations.

Reliability usually determines the validity of the study and enhances the results. Validity identifies the way in how a concept is discussed, and how the results help answer research questions. It includes the choice of a sample for the study, the description of the population, and the appropriateness of the ideas offered in terms of intervention. In the chosen articles, communication with participants shows nurses’ and patients’ attitudes towards capnography. Sampling convenience and the tools for data analysis are the common characteristics of the studies.

Summary

All three articles with different types of study and three additional articles that include the results of statistical investigations in medical research can be applied as good examples in practice. The idea of statistical methods was described by Turle et al., Arafa et al., and Lin et al., Iyer et al. showed how to explore the topic. Finally, Langhan et al. explained how to combine both qualitative and quantitative data in research. Chi-square and paired t-tests are the best alternatives for discussion of the data about capnography use during patient resuscitation.

References

Arafa, S., Abouzkry, A., & Mohamady, A. (2018). Accuracy of bedside upper airway unltrasonography vs. standard auscultation for assuring the location of endotracheal tube after tracheal intubation: Comparative study controlled by quantitative waveform capnography. Journal of Anesthesia & Clinical Research, 9(2). Web.

Iyer, N. S., Koziel, J. R., & Langhan, M. L. (2015). A qualitative evaluation of capnography use in paediatric sedation: Perceptions, practice and barriers. Journal of Clinical Nursing, 24(15-16), 2231–2238. Web.

Langhan, M. L., Kurtz, J. C., Schaeffer, P., Asnes, A. G., & Riera, A. (2014). Experiences with capnography in acute care settings: A mixed-methods analysis of clinical staff. Journal of Critical Care, 29(6), 1035–1040. Web.

Lin, H. J., Huang, C. T., Hsiao, H. F., Chiang, M. C., & Jeng, M. J. (2017). End-tidal carbon dioxide measurement in preterm infants with low birth weight. PloS One, 12(10), e0186408. Web.

Lin, T. Y., Fang, Y. F., Huang, S. H., Wang, T. Y., Kuo, C. H., Wu, H. T.,… Lo, Y. L. (2017). Capnography monitoring the hypoventilation during the induction of bronchoscopic sedation: A randomized controlled trial. Scientific Reports, 7(1), 8685-8687.

Turle, S., Sherren, P. B., Nicholson, S., Callaghan, T., & Shepherd, S. J. (2015). Availability and use of capnography for in-hospital cardiac arrests in the United Kingdom. Resuscitation, 94, 80–84. Web.

Appendix

Table 1. Comparison of Articles.

Type of Article Title of Article Type of Analysis and Rationale (can be more than one) Applicability of Test Reliability and Validity
Quantitative “Availability and use of capnography for in-hospital cardiac arrests in the United Kingdom”. Microsoft Excel spreadsheet software was used to analyze data of the study. The goal of this quantitative research was to quantify the availability of ETCO2monitoring among patients who have cardiac arrests. Monitoring evaluation proves that waveform capnography was frequently used in the majority of emergency cases, compared to other methods, such as disposable colorimetric devices or quantitative devices. A description of the current application of capnography during cardiopulmonary resuscitation was developed. An application was properly validated with a high level of sensitivity. A telephone questionnaire helped to clarify the outcomes of capnography use in EDs and ICUs. Convenience of sampling, clearly established goals and the comparison of techniques within the same cardiac context.
Qualitative “A qualitative evaluation of capnography use in paediatric sedation: Perceptions, practice and barriers”. An inductive approach was used to analyze the information. The first step was a review of interview transcripts to gain an overall understanding of the themes, identify emerging topics for discussion, and clarify the connection between different concepts. A grounded theory approach was based on a coding structure with line-by-line transcripts. A comparative method was applied to refine codes and encourage a systematic review. Manual analysis was performed to promote outcome evaluation using HyperRESEARCH version 3.5.2 for defining thematic categories. Inductively generated theories and themes were analyzed until theoretical saturation. As soon as capnography technology was introduced, it could be applied to a variety of patients in different healthcare settings. Interviews helped to test people’s opinions and attitudes towards the chosen technique. Criterion sampling (past participation in randomized controlled trials) and the presence of a team of researchers who reviewed discrepancies and developed consensus-based decisions.
Mixed Methods “Experiences with capnography in acute care settings: A mixed-methods analysis of clinical staff”. First, a grounded theory approach was applied through interactive data analysis and a comparative method where ideas were inductively generated, and theories were built. Then, a coding team worked on interview transcripts line by line. An iterative process occurred to edit themes. HyperRESEARCH qualitative analysis software was used to organize gathered data. Finally, an audit trial was performed to ensure that all systematic documentation was added to the study. This provoked analytical decision-making among the researchers. Cohen’s khelped to control reporting of capnography use. The Fisher exact test was applied to analyze demographic data of participants. Student t test helped to analyze the differences between capnography use and nonuse. Inductive analysis enhanced reliability, and interviews were used to obtain credible information from participants. A purposeful sample contributed to a better discussion of the topic. Physicians and nurses from different departments were invited to participate in semi-structured interviews and share their opinions about future implementation of capnography in their practice.