Predictive analytics moves beyond standard forecasting or estimating and is a form of data science (What is predictive analytics? The PAW resource guide, 2021). Prediction is at the center of big data, and the whole point of data is to learn from it to predict. Predictions drive and render organizational and operational decisions. Rather than solely providing insights, a predictive model generates a predictive score for each individual, which directly drives or informs decisions for that individual, e.g., whether to apply a specific medical treatment. Decision-makers can allocate budgets based on per-person predictions, assisting health leaders in the challenge of resource allocation (Giga, 2017). A predictive model supports early identification, allowing preventative interventions to begin earlier while possibly decreasing the need for invasive investigative procedures later in the continuum of care.
The mission of Predictive Analytics World (PAW) is to foster breakthroughs in the value-driven operationalization of established deep learning methods (What is predictive analytics? The PAW resource guide, 2021). Their mission aligns for this evidence-based practice (EBP) project, and to process the pilot implementation data before deciding whether the change is appropriate for adoption into practice and if the process should be hard-wired and integrated system-wide. Ultimately, producing better patient outcomes by helping to target and treat high-risk patients is the goal (Giga, 2017). Predictive analytics technology learns from the data to predict or infer an unknown, resulting in improved outcomes, lower costs, and higher patient satisfaction. The data will determine if the unknown, whether conducting routine sleep screening increases the discovery and treatment of obstructive sleep apnea (OSA). From this starting point, the data collection will build on evaluating the potential rewards against expenditures while providing high-value patient outcomes (Giga, 2017).
In my first discussion post for this week, I identified that quantitative research designs would be better suited to my evidence-based project because of the need for subjective data as evidence to support my proposed intervention. As my project is centered around trends in patient vital signs, particularly blood pressure, in response to either intravenous fluid resuscitation or vasopressor administration, much of the data will be gathered from patient electronic medical records (Bertelsen et al., 2020). Knowing this, the independent variables that my project will examine are patient blood pressure and medication administration (either fluids or vasopressors); the patients’ need for flap revision surgery being the dependent variable. A chi-square test to determine whether each intervention correlates with increased rates of flap failure and need for revision surgery can be conducted. This would help determine whether negative outcomes from each intervention have any relation with each other, which might indicate complications other than hypotension necessitating flap revision arising in the studied population (Burkhard et al., 2021). In order to determine the overall effectiveness of the PICOT intervention, I would use a multiple linear regression test to examine the relationship between both groups of patients in the study, as this would allow the consideration of other factors that affect flap outcomes in correspondence with my studied intervention (Bertelsen et al., 2020).
Though, as statistical analysis is not my strong suit, I will be doing additional research to consider different methods of analyzing my data as well.
References
Bertelsen, C., Hur, K., Nurimba, M., Choi, J., Acevedo, J. R., Jackanich, A., Sinha, U. K., Kochhar, A., Kokot, N., & Swanson, M. (2020). Enhanced Recovery After Surgery-Based Perioperative Protocol for Head and Neck Free Flap Reconstruction. OTO open, 4(2), 2473974X20931037. https://doi-org.lopes.idm.oclc.org/10.1177/2473974X20931037
Order this paper