Adopting Ai in Methodological Practices: Transforming Predictive Analytics in Health

Abstract

Background: Conventional predictive analytics in health research has the potential to benefit from the integration of Artificial Intelligence (AI) thereby enhancing accuracy, efficiency, and decision-making processes. Yet, a deeper analysis is needed of their implications for predictive analytics (and the challenges involved in adopting them).

Objective: We embarked on this study to better understand how AI adoption can affect predictive analytics in health research, as well as the interrelations between perceived benefits, challenges, and broader outcomes.

Methods: We surveyed 300 healthcare professionals who use predictive analytics as part of their work, using a cross-sectional approach. Medium-sized data through a Likert-scale survey instrument of perceptions around AI and predictive analytics benefits, and challenges, among other measures. Data normality was tested using the Shapiro-Wilk test and reliability of scale with Cronbach's Alpha. These data were analyzed using non-parametric tests, which included Spearman's rank correlations because they violated normality.

Results: The internal consistency was poor, with a Cronbach's Alpha of 0.144 In all variables, non-normality was confirmed by the Shapiro-Wilk test results (p < 0.001), therefore we used non-parametric methods for comparison The Spearman correlation analysis indicated that the AI BETTER score had a significant positive across-tabs association with raters I on technological expertise. Accuracy and Privacy Concern (ρ = 0.14, p = 0.02), while most other relationships were weak and non-significant.

Conclusion: Findings show AI benefits research but integration challenges persist regarding privacy, finances, and technical nuance. The survey's internal contradiction demands improved tools to capture AI perceptions. Additional inquiry can promote respectful progress overcoming barriers, and realizing prediction's potential through partnership, not disruption.