RESUMO
BACKGROUND: Current atherosclerotic cardiovascular disease (ASCVD) predictive models have limitations; thus, efforts are underway to improve the discriminatory power of ASCVD models. OBJECTIVE: We sought to evaluate the discriminatory power of social media posts to predict the 10-year risk for ASCVD as compared to that of pooled cohort risk equations (PCEs). METHODS: We consented patients receiving care in an urban academic emergency department to share access to their Facebook posts and electronic medical records (EMRs). We retrieved Facebook status updates up to 5 years prior to study enrollment for all consenting patients. We identified patients (N=181) without a prior history of coronary heart disease, an ASCVD score in their EMR, and more than 200 words in their Facebook posts. Using Facebook posts from these patients, we applied a machine-learning model to predict 10-year ASCVD risk scores. Using a machine-learning model and a psycholinguistic dictionary, Linguistic Inquiry and Word Count, we evaluated if language from posts alone could predict differences in risk scores and the association of certain words with risk categories, respectively. RESULTS: The machine-learning model predicted the 10-year ASCVD risk scores for the categories <5%, 5%-7.4%, 7.5%-9.9%, and ≥10% with area under the curve (AUC) values of 0.78, 0.57, 0.72, and 0.61, respectively. The machine-learning model distinguished between low risk (<10%) and high risk (>10%) with an AUC of 0.69. Additionally, the machine-learning model predicted the ASCVD risk score with Pearson r=0.26. Using Linguistic Inquiry and Word Count, patients with higher ASCVD scores were more likely to use words associated with sadness (r=0.32). CONCLUSIONS: Language used on social media can provide insights about an individual's ASCVD risk and inform approaches to risk modification.
RESUMO
Importance: There are areas of skilled nursing facility (SNF) experience of importance to the public that are not currently included in public reporting initiatives on SNF quality. Whether patients, hospitals, and payers can leverage the information available from unsolicited online reviews to reduce avoidable rehospitalizations from SNFs is unknown. Objectives: To assess the association between rehospitalization rates and online ratings of SNFs; to compare the association of rehospitalization with ratings from a review website vs Medicare Nursing Home Compare (NHC) ratings; and to identify specific topics consistently reported in reviews of SNFs with the highest vs lowest rehospitalization rates using natural language processing. Design, Setting, and Participants: A retrospective cross-sectional study of 1536 SNFs with online reviews on Yelp (a website that allows consumers to rate and review businesses and services, scored on a 1- to 5-star rating scale, with 1 star indicating the lowest rating and 5 stars indicating the highest rating) posted between January 1, 2014, and December 31, 2018. The combined data set included 1536 SNFs with 8548 online reviews, NHC ratings, and readmission rates. Main Outcomes and Measures: A mean rating from the review website was calculated through the end of each year. Risk-standardized rehospitalization rates were obtained from NHC. Linear regression was used to measure the association between the rehospitalization rate of a SNF and the online ratings. Natural language processing was used to identify topics associated with reviews of SNFs in the top and bottom quintiles of rehospitalization rates. Results: The 1536 SNFs in the sample had a median of 6 reviews (interquartile range, 3-13 reviews), with a mean (SD) review website rating of 2.7 (1.1). The SNFs with the highest rating on both the review website and NHC had 2.0% lower rehospitalization rates compared with the SNFs with the lowest rating on both websites (21.3%; 95% CI, 20.7%-21.8%; vs 23.3%; 95% CI, 22.7%-24.0%; P = .04). Compared with the NHC ratings alone, review website ratings were associated with an additional 0.4% of the variation in rehospitalization rates across SNFs (adjusted R2 = 0.009 vs adjusted R2 = 0.013; P = .003). Thematic analysis of qualitative comments on the review website for SNFs with high vs low rehospitalization rates identified several areas of importance to the reviewers, such as the quality of physical infrastructure and equipment, staff attitudes and communication with caregivers. Conclusions and Relevance: Skilled nursing facilities with the best rating on both a review website and NHC had slightly lower rehospitalization rates than SNFs with the best rating on NHC alone. However, there was marked variation in the volume of reviews, and many SNF characteristics were underrepresented. Further refinement of the review process is warranted.