The less online pandemic-related information you seek out the less anxiety you may have, according to a new baseline study from the University of California Institute for Prediction Technology. The research was undertaken by UCIPT’s Dominic Ugarte and written by Parvati Singh. It was published in the Journal of Medical Internet Research in September. 406 adults with pre-existing mental health conditions were recruited through Facebook, Google, and Reddit and became participants in the study. The participant’s GAD (generalized anxiety disorder) scores were analyzed, along with the amount of time they spent online. The results of this study were that spending four or more hours online searching about COVID-19 is associated with increased anxiety. The research also concludes that telepsychiatry, telemedicine, and social media-based interventions may be of benefit to people that use social media and the internet frequently. Further studies on this topic are being undertaken by UCIPT now.
In other news, the Institute is developing an app to detect COVID-19 misinformation. A paper on this topic, co-authored by Dominic Ugarte, was accepted in an Empirical Methods in a Natural Language Processing (EMNLP) Workshop. The research found that existing misinformation detection datasets and models are not effective for evaluating systems designed to detect misinformation due to novel language and the rapid change of information on this topic. A significant takeaway from our work on this topic is that we have created our own dataset that we are using to teach our models. Further, the study found that misinformation detection can be divided into two sub-tasks: (i) retrieval of misconceptions relevant to posts being checked for veracity, and (ii) stance detection to identify whether the posts Agree, Disagree, or express No Stance towards the retrieved misconceptions. Work toward developing an app continues.
In partnership with Wholistic, Dominic Ugarte is conducting our Cannabinoid & Anxiety Relief Education Study (C.A.R.E.S.) survey. This first-of-its-kind survey is expected to yield unique insights about the effects of the current pandemic on anxiety levels and sleep quality, and whether cannabis and/or hemp-derived CBD use during this time has had any impact on users’ health and well-being. The research findings will be broadly shared with the general public to aid in cannabis and CBD education.
Researcher Romina Romero from our group is working on viewpoint papers relating to a variety of topics, including Blockchain and HIV; Blockchain and Opioids; Ethics and Technology; Ethics, Technology, and HIV; and Adolescents and Opioids. Updates will be provided on the publication of these papers in the future.
The UCIPT HOPE HIV study is underway and is currently in week 4. This study is being conducted to study whether peer leaders on an online social network can help to increase health behaviors (HIV testing, PrEP use, etc.) among men who have sex with men. It is also even more relevant now, as it will also give us insight into how the pandemic has affected health behaviors and access to care and resources in this population. On a related study, we had also previously sent a survey specific to these issues to participants in previous waves and are currently in the process of analyzing that data.
UCIPT has received funding from the NIH for phase II of our HOPE MOUD (Medications for Opioid Use Disorder) study. This research has become more relevant recently and will facilitate insights into how COVID-19 has affected access to MOUD. The IRB has now given its approval to begin the study, and we hope to start recruitment by end of February/early March.
Lidia Flores from our group is using Twitter data to identify methods that may help predict public health outbreaks and assist health departments in their community interventions. This research involves exploring the impact of events, such as tweets from the President or government announcements regarding COVID-19, and whether the public’s opinion and behavior changes as a result. Flores is interested in determining whether a public announcement on Twitter relating to the death rate for COVID-19 makes individuals more inclined to support the usage of facemasks.
We are also working on a qualitative analysis of tweets implying sexual risk behaviors. The goal of this exploration is to determine whether big data can be used in detecting regions with higher HIV risk. Risky behaviors involving the use of substances such as alcohol or marijuana are also being investigated.
Also relating to the topic of HIV, is our work on how events such as a celebrity announcement can impact perspectives on pre-exposure prophylaxis. Lidia Flores is exploring why people choose to use or not use PrEP and the brands people are using — for example Descovy and Truvada — in various regions. Whether the availability of a drug within a particular region can make the public more likely to try PrEP is also being researched.
Work is also underway relating to the opioid epidemic. Specifically, how opioid consumption methods vary across populations of different sizes, and which regions have higher mentions of smoking, pill use, etc. The goal of this exploration is to determine whether big data can be used in detecting regions with higher opioid overdose risk. This can be very helpful for emergency and public health departments by gathering real-time data on regions that may have a higher influx of overdoses.