Dr. Melanie A Sutton

In 2024, Google rolled out AI Overviews, placing an AI chatbot in front of every user during Google searches. Given the public's awareness of these AI tools, it is important for public health professionals to incorporate these tools into their data visualization workflows to create effective visualizations for public-facing communications.
Compare methods for effective priming of AI tools to generate improvement suggestions for created maps. Demonstrate tips and tricks to attain the full potential of AI tools in the map-making process for epidemiological data. The Department of Public Health formed a multidisciplinary faculty/student team to explore the capabilities of a suite of AI tools in designing and evaluating epidemiological maps derived from data coded at country, state, city, and census tract-level levels. Representative data sets included variables for crime, disease occurrence, and medical testing facility locations.
Strategies for effectively using Chat GPT 3.5 and Google Gemini were developed based on several map types created by faculty and students using Google Geo Charts or Excel 3D Maps. Tips for priming AI tools in the process of designing and interpreting public-facing epidemiological maps to ensure equity and accessibility were created, alongside guidance on recognizing AI hallucinations.
Epidemiological data is commonly visualized with publicly distributed maps, and utilizing AI tools in the map-making process will ensure thorough, well-designed displays. Incorporating AI tools into data visualization workflows provides students, faculty, and public health professionals with additional foundational knowledge to ensure generated maps contribute to building trust in public health.

Objective: The use of complimentary and integrative medicines has emerged as significant area of research both in the United States and internationally. However, much remains to be learned about the true magnitude of these products/this practice and the rationale for their use, especially among persons with a combination of health problems. The purpose of our study is twofold: 1) estimate the prevalence of CAM use in menopausal women with a concurrent cancer diagnosis.

Methods: Data was for 19,241 adult American women from the 2012 National Health Interview Survey (NHIS) and CAM supplement. The outcome variable was ever use of CAM and CAM use in the women who reported either menopausal problems in 12 months prior, cancer diagnosis or both. Data were weighted and analyzed using Stata software version 14 for Windows. In addition to descriptive and bivariate statistics, multivariate logistic regression was performed to estimate the odds of CAM use by selected covariates and disease dyads.

Results: The average age of the women in the study was 48.9 years (SD = 18.58). Approximately 35.7% of the women experienced menopausal problems in the preceding 12 months, 9% of them had a cancer diagnosis. Overall CAM use was reported in 30.7% of respondents. At the multivariate level, age,
race, marital status, education, income, region of residence, high cholesterol, having insurance, smoking, and alcohol consumption were significant (p <0.05). However, the concurrent presence of menopausal problems and cancer diagnosis did not appear to influence CAM use, when adjusting for other covariates (p>0.05).

Conclusion: Health care professionals should be aware of CAM use in this segment of the population due to potential for treatment compliance challenges such as potential drug/CAM interaction