DNP-830A Data Analysis

Assessment Description

Power analysis is used to determine if a sample size is adequate to evaluate project outcomes prior to initiating any project. This is also called a priori. Justify the project sample size for your DPI Project by using the sample size calculator to conduct a power analysis. Discuss how your sample size may affect the validity of your project. Determine the sampling method used in your project. Compare the convenience sample to other sampling methods. Provide evidence supporting your response.

Bonnie Flores (In 200 Words)

As doctoral learners about to start on our direct practice improvement (DPI) project it is vital to determine the appropriate sample size for the DPI. If the sample size is too small, then this could lead to decreased statistical power, reduced reliability, increased probability of a Type II Error, and research that is unreliable in predicting future patient outcomes (Grand Canyon University (GCU), n.d.). Additionally, the data could be skewed and inaccurate; therefore, leading to research that does not adequately reflect the general population (GCU, n.d.). Conversely, a sample size that is too large may lead to statistical results that inflate patient outcomes (GCU, n.d.). Hence, when embarking on our DPIs, we must conduct a G Power to determine the correct sample size that would yield statistical significance with a p < 0.05. Initially, I determined that a sample size of 30 patients would suffice; however, when completing my 10 Strategic Points, I conducted a G Power which resulted in a minimum sample size of 54 patients for a 0.95 confidence interval and a margin of error of 0.05; the actual G Power was 0.9502120. While we are attempting to improve patient outcomes with clinical significance, we are also testing if the intervention resulted in statistical significance. P values that do not fall < 0.05 are not considered to be statistically significant (Sharma, 2021). The time we are spending on our project is extraordinary; therefore, we want the work that we put in to result in improved patient outcomes. If we do not have an adequate sample, whether too small or too large, then the DPI may not have valid results, meaning that the intervention measures what is intended (Sylvia & Terhaar, 2018).

 

Sampling refers to taking a sample of the population for the intervention since testing the entire population is not feasible. A sample size that is too small will not yield results that are generalizable to the population, which is why it is necessary to have an adequate number. There are two main types of sampling, probability, and convenience (Stratton, 2021). Types of probability sampling include random sampling (e.g., simple, systematic, stratified, and cluster); randomization reduces the risk of bias (Stratton, 2021). Convenience sampling (a non-probability sampling) is not as objective as probability sampling but is a common type of sampling that occurs in a population, prehospital, and disaster research due to its reduced cost and time (Stratton, 2021). For my DPI it would be most feasible to use a convenience sample as I will need to find 54 patients at the practice site who meet the inclusion criteria. I thought 30 patients was a good number, hence finding 54 patients will be a challenge in and of itself without also adding randomization, clustering, and stratification on top of that. If the project lasted longer than eight weeks, then it may be feasible to use a different sampling method.

 

References

 

Grand Canyon University, Center for Innovation in Research and Teaching. (n.d.). Sample size. Retrieved on April 20, 2023, from https://cirt.gcu.edu/research/develop/realworld/samplesize

 

Sharma, H. (2021). Statistical significance or clinical significance? A researcher’s dilemma for appropriate interpretation of research results. Saudi Journal of Anaesthesia, 15(4), 431–434. https://doi.org/10.4103/sja.sja_158_21

 

Stratton, S. J. (2021). Population research: convenience sampling strategies. Prehospital and Disaster Medicine, 36(4), 373-374.

 

Sylvia, M. L. & Terhaar, M. F. (Eds). (2018). Clinical analytics and data management for the DNP (2nd ed.). Springer Publishing Company.

 

 

Shabnampreet Kaur (In 200 Words)

Given a specified degree of power, alpha, and effect size, power analysis is a statistical approach to finding the minimal sample size needed to detect a statistically significant impact. Before the study starts, a priori power analysis is carried out to ensure the sample size will be sufficient to assess project outcomes. We can use a sample size calculator that considers variables like the intended degree of significance, effect size, and power to determine the appropriate sample size for the DPI project. Before completing the ten strategic points, I chose a sample size of 50 for the DPI project would be sufficient. While working on the ten strategic points, I conducted a power analysis. The confidence interval was 95%, and the margin of error was 0.05%. The ideal sample size calculated was 45 on the power analysis (Qualtrics,2020). Choosing participants who fit the study’s eligibility requirements is crucial when researchers have established the ideal sample size. The sample size may significantly impact the project’s validity. Even if a statistically significant effect exists, researchers can only see it if the sample size is more meaningful (Center for Innovation in Research and Teaching,2022). Also, the sample size may significantly impact a project’s validity. The degree to which a study accurately assesses what it is supposed to measure is called validity. External validity problems may result from a sample size that needs to be more significant since it may not fairly represent the population from which it is derived (Sharma,2021). The term “external validity” describes how well research findings can be applied to different people and environments. For instance, a study with a limited sample size might not accurately represent the whole range of population variability, leading to incorrect estimations of the effect size or prevalence of particular phenomena. This may restrict the findings’ generalizability and the capacity to make significant conclusions. A limited sample size might affect internal validity in addition to external validity. Internal validity is the extent to which the results of a study can be attributed to the intervention or exposure being studied rather than other unrelated factors. The accuracy and precision of the results may be impacted by confounding, bias, or random error due to a small sample size. On the other hand, a sample size that is too large might not offer any further benefits in terms of validity and may be a waste of time and money. To ensure that the project has sufficient power to detect meaningful effects while reducing the risk, it is crucial to determine an appropriate sample size based on a priori power analysis.

There are various options for the sampling technique employed in research, including convenience, stratified, and random (Singh,2018). Choosing participants for convenience sampling entails selecting those who are readily available and willing to participate. It is frequently utilized in research investigations with little funding or time. Convenience sampling may not entirely represent the population, and there might be bias in the selection process. Other sampling techniques, like random sampling, offer a more representative sample of people but can be more expensive and time-consuming. For my DPI project, I will use convenience sampling, a non-probability sampling technique in which participants are selected based on the inclusion and exclusion criteria and their willingness to participate. Also, it will be a cost-effective way of selecting participants.

A priori power analysis should be used to estimate the sample size for the DPI project, considering variables including the desired degree of significance, effect size, and power. Considerations like resources, time, and representativeness should all be considered when choosing the sampling technique for the project.

References

Center for Innovation in Research and Teaching. (2022, December). Sample Size. CIRT. https://cirt.gcu.edu/research/develop/realworld/samplesize

Sharma, H. (2021). Statistical significance or clinical significance? A researcher’s dilemma for appropriate interpretation of research results. Saudi Journal of Anaesthesia, 15(4), 431–434. https://doi.org/10.4103/sja.sja_158_21

Singh, S. (2018, December 25). Sampling Techniques – Towards Data Science. Medium. https://towardsdatascience.com/sampling-techniques-a4e34111d808

Qualtrics. (2020). Calculating sample size: A quick guide (calculator included). https://www.qualtrics.com/blog/calculating-sample-size/

 

 

DQ 2

Assessment Description

While the goal of the DPI Project is to achieve clinical significance, you will also need to discuss if the data analysis also has statistical significance. Statistical significance (p = .05) helps determine whether a result is by chance. On the chance that your DPI Project does not result in statistical significance, explain how you can justify clinical significance.

Bonnie Flores (In 200 Words)

During the analysis of the direct practice improvement (DPI) project, it will be incumbent upon us to determine if our intervention achieved statistical significance (p = 0.05) with a confidence interval of 95% and a margin of error of 5% (Qualtrics, 2023). This data will show whether the intervention was significant in proving what we sought out with the implementation of our intervention. Therefore, we will need to ensure that the sample size is adequate to show a significant relationship; a larger sample size will reduce the chance of error (Sharma, 2021). Statistical significance is not the only measure of a good project. The DPI will examine if there is statistical significance, but equally, if not more important, is whether the project results in clinical significance, which equates to the improvement of physical, mental, or emotional improvement in a patient’s health and well-being (Sharma, 2021). The two do not necessarily correlate with one another; it is possible to have clinical significance without statistical significance and vice versa (Davis et al., 2021). In fact, there is a school of thought that posits that too much emphasis is placed on statistical significance and that it should not be the threshold for value, but rather one piece of evidence along with other factors (McShane et al., 2019).

Determining whether the intervention was clinically significant will depend on how that significance is interpreted (Davis et al., 2021). The doctoral learner must ascertain what they deem to be enough of a magnitude of change (Davis et al., 2021). For example, for my DPI, I will see whether the Intensive Care Unit (ICU) Liberation Bundle impacts ventilator days. Does a clinically significant change mean that a patient is off the ventilator for half of one day less, one day less, or more? The guidelines for the DPI suggest that we must show clinical significance for at least one patient. It would be easy to argue that reducing ventilator time for one patient by one day is clinically significant for that patient. The longer patients stay on mechanical ventilation, the more deconditioned they get and the higher the risk for delirium, prolonged length of stay, and increased mortality (Pun et al., 2019). Hence, reducing this risk in even one person will mean a potential increase in the quality of life for that patient and their family. As a nurse, it is always my goal to improve the quality of life for my patients and families; therefore, any measurement of improvement no matter how small, that leads to an improved life will be deemed as clinically significant.

References

Davis, S. L., Johnson, A. H., Lynch, T., Gray, L., Pryor, E. R., Azuero, A., … & Rice, M. (2021). Inclusion of effect size measures and clinical relevance in research papers. Nursing Research, 70(3), 222.

McShane, B. B., Gal, D., Gelman, A., Robert, C., & Tackett, J. L. (2019). Abandon statistical significance. The American Statistician, 73(sup1), 235-245.

Pun, B. T., Balas, M. C., Barnes-Daly, M. A., Thompson, J. L., Aldrich, J. M., Barr, J., Byrum, D., Carson, S. S., Devlin, J. W., Engel, H. J., Esbrook, C. L., Hargett, K. D., Harmon, L., Hielsberg, C., Jackson, J. C., Kelly, T. L., Kumar, V., Millner, L., Morse, A., Perme, C. S., … Ely, E. W. (2019). Caring for critically ill patients with the ABCDEF bundle: Results of the ICU Liberation collaborative in over 15,000 adults. Critical Care Medicine, 47(1), 3–14. https://doi.org/10.1097/CCM.0000000000003482

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6298815/

Qualtrics. (2023, March 21). Calculating sample size: A quick guide (calculator included). Retrieved on April 22, 2023, from https://www.qualtrics.com/blog/calculating-sample-size/

Sharma, H. (2021). Statistical significance or clinical significance? A researcher’s dilemma for appropriate interpretation of research results. Saudi Journal of Anaesthesia, 15(4), 431–434. https://doi.org/10.4103/sja.sja_158_21