Variability in Language used on Social Media prior to Hospital Visits

Abstract

forecasting healthcare utilization has the potential to anticipate care needs, either accelerating needed care or redirecting patients toward care most appropriate to their needs. While prior research has utilized clinical information to forecast readmissions, analyzing digital footprints from social media can inform our understanding of individuals' behaviors, thoughts, and motivations preceding a healthcare visit. We evaluate how language patterns on social media change prior to emergency department (eD) visits and inpatient hospital admissions in this case-crossover study of adult patients visiting a large urban academic hospital system who consented to share access to their history of facebook statuses and electronic medical records. An ensemble machine learning model forecasted eD visits and inpatient admissions with out-of-sample cross-validated AUCs of 0.64 and 0.70 respectively. Prior to an ED visit, there was a significant increase in depressed language (Cohen's d = 0.238), and a decrease in informal language (d = 0.345). Facebook posts prior to an inpatient admission showed significant increase in expressions of somatic pain (d = 0.267) and decrease in extraverted/social language (d = 0.357). These results are a first step in developing methods to utilize user-generated content to characterize patient care-seeking context which could ultimately enable better allocation of resources and potentially early interventions to reduce unplanned visits.

Studying social media language changes associated with pregnancy status, trimester, and parity from medical records

“…We took a case-crossover approach 12 to evaluate if machine learning models would distinguish time periods within individuals relevant to pregnancy—primarily the first trimester where developing interventions to improve early prenatal care can substantially improve pregnancy outcomes. We used the LDA topics as features to train a Random Forest classifier to distinguish linguistic cues associated with first-trimester pregnancy compared with a random time period and 3 months before conception (see Figure 1).…”

“…There is evidence that social media language is indicative of individuals’ psychological stress, 7,8 loneliness, 9 depression, 10 suicide risk, 11 hospital utilization, 12 and medical conditions in general. 13 Prior works on studying pregnancy using social media focused on predicting the risk of postpartum depression 14 and identifying women who have announced their pregnancy on Twitter to detect large numbers of cohorts.…”

Social Media Reveals Psychosocial Effects of the COVID-19 Pandemic

“…For the above research questions, we measure psychosocial effects in terms of symptomatic mental health expressions of anxiety, depression, stress, and suicidal ideation, and expressions seeking emotional and informational support. Our work is founded on a large body of work on studying mental health and psychosocial dynamics with social media data [24][25][26][27][28][29][30]. Several studies have leveraged Twitter (which is also the data of our current study) to study health attributes and public health [30], including symptoms related to diseases [31], disease contagion [32], obesity and physical health [33], mood and depressive disorders [28,34], mental health self-disclosures [27], post-traumatic stress disorder [35], addictive behaviors and substance use [36,37], etc.…”

“…Towards our first research aim of understanding the psychosocial impacts of the COVID-19 outbreak, we conduct two types of analysis on our Twitter dataset, which we describe below. Our work builds upon the vast, rapidly growing literature studying mental health concerns and psychosocial expressions within social media data [19,21,[24][25][26][27][28]34,48,[50][51][52].…”

“…Social media provides a real-time platform where people often candidly self-disclose their concerns, opinions, and struggles during this pandemic [25]. Our research is, therefore, founded on prior work to understand people's psychosocial distress in terms of their symptomatic mental health expressions of anxiety, depression, stress, and suicidal ideation, and their expressions seeking emotional and informational support [20,[26][27][28][29].…”