Twitter and CV Health

Sponsor

University of Pennsylvania (Other)

Overall Status

Completed

CT.gov ID

NCT02622256

Collaborator

(none)

611

Enrollment

Location

Arms

Actual Duration (Months)

21.1

Patients Per Site Per Month

Study Details

Study Description

Brief Summary

Cardiovascular (CV) disease is associated with significant morbidity and mortality. In the current digital age, needed is a better understanding of how information on social media sites may inform our approaches to improving CV health through novel methodologies. Investigators propose to study the conversation on Twitter about several CV diseases and their associated sequelae.

Condition or Disease	Intervention/Treatment	Phase
Cardiovascular Disease Hypertension	Behavioral: Health System: HTN Intervention	N/A

Detailed Description

Cardiovascular (CV) disease remains the leading cause of morbidity and mortality in the US and is associated with significant economic burden. To advance against this major public health problem, the American Heart Association (AHA) and others have identified the need for a targeted focus on improving the CV health of populations, emphasizing an array of new tools and competencies for implementing public health policy and population- and community-level interventions to complement the traditional, predominantly medically oriented interventions that have been promoted successfully in the past. Social media channels like Twitter offer a new opportunity to explore health related communication generated by the public and for the public.

Understanding and harnessing these new communication channels is of particular importance as the digital divide narrows and individuals across age and demographic groups increasingly share information online with known networks of friends (e.g., Facebook), and broader networks of friends and others (e.g., Twitter). Person-to-person word-of-mouth communication is one of the most enduring and persuasive ways in which people deliver and receive information. Until recently, person-to-person communication was effectively impossible to intercept, study, and alter. Central to this proposal is the recognition that these changes make some person-to-person communication observable that was previously private. It is the observability of these new communication channels that provides both innovation and promise to this area of inquiry.

This proposal evaluates social media for both its efferent and afferent pathways as a source not just to communicate to communities, but also to learn from them. There is considerable evidence that letting people know what other people do is one of the most effective ways of increasing that behavior. This social norming of behaviors is facilitated through online sharing enabling others to model behavior against broader groups whose actions would have been invisible and therefore uninfluential without these new media channels. For individuals with chronic illnesses, automated self-management support (e.g. mHealth) and online communities have been shown to improved clinical outcomes, patient satisfaction, and reduce health care costs and utilization. This proposal seeks to improve CV health and reduce the burden of CV disease by understanding how patients communicate about CV health online and improving patients ability to manage their CV disease(s).

Twitter as a global social media platform: Twitter allows users to send and receive 140-character messages referred to as tweets. Tweets may include embedded web links to information such as news articles, home pages, and pictures. Tweets originate from a single person or organization (a tweeter) and are distributed broadly to individuals with an interest in the topic of the tweet and to individuals who have voluntarily signed up to follow that tweeter. Followers can then share messages with their own followers, a process of message propagation known as re-tweeting. Tweeters can choose to share information about themselves on their profile (e.g. age, race, gender, occupation, location, likes/dislikes, picture, webpage link).

Both patients and researchers have used health-related Twitter data in novel ways. In natural disasters (e.g. Hurricane Sandy, Haiti earthquake), Twitter was used in real time to link people in need with resources. In pandemics (e.g. H1N1) Twitter was used as a surveillance tool to target flu hot-spots more rapidly than traditional data collection tools. Twitters impact in organizing individual and social attitudes was dramatically revealed in the 2011 political events in Northern Africa. In this setting Twitter, and similar social media such as Facebook, allowed the propagation and concentration of ideas sufficient to threaten and in some cases topple restrictive governments. In non-emergent settings, linguists have used Twitter to pinpoint local dialects and sociologists have used tweets to forecast the mood and emotion of specific geographic regions. Others have also used Twitter to characterize medical misconceptions (e.g. sequelae of concussions) and propagation of poor medical compliance (e.g. antibiotic use).

Studying person-to-person communication: An estimated 400 million tweets are posted daily by more than 200 million active users. Twitter is representative of big data that are increasingly being explored to better understand online information from large, broad populations of patients. Twitter offers promise as a research tool due to its immense scale, its immediacy (for example, emergency departments in Boston learned about the tragic marathon bombings faster through Twitter than through news or established emergency service communication channels), and the systematic searchability of its content. Also of interest is that the site is not focused on health and so it draws people by their interest in communicating more generally, and yet includes public discourse on a broad array of health topics, from the perspective of patients, providers, policymakers, organizations and others. Although the Twitter user base is not a nationally-representative sample it has a surprisingly deep representation across age, geography, and social distributions. African-Americans, Latinos, and those in urban populations are in fact overrepresented on Twitter relative to the general population.

This proposal reflects early work, but work that is fundamental to developing a base for understanding the scientific uses and limitations of Twitter and related social media. This proposal aims to analyze CV health behaviors being discussed online and evaluate new approaches for improving access to CV health information and implementing behavior modification.

Study Design

Study Type:

Interventional

Actual Enrollment :

611 participants

Allocation:

Randomized

Intervention Model:

Parallel Assignment

Masking:

None (Open Label)

Primary Purpose:

Other

Official Title:

Twitter & Cardiovascular Health

Actual Study Start Date :

Jun 1, 2016

Actual Primary Completion Date :

Oct 1, 2018

Actual Study Completion Date :

Nov 1, 2018

Arms and Interventions

Arm	Intervention/Treatment
Experimental: Health System: HTN Intervention Investigators will identify patients with known hypertension (ICD-9 code 401.9), from the Penn Data Store (PDS). Participants will be asked to complete 3 short surveys.	Behavioral: Health System: HTN Intervention Interested participants can enroll online via a Twitter link. Patients will consent to having their electronic health records accessed to validate clinical data. Participants will complete short surveys. The project Twitter account would follow tweeters with high impact CV messages and tweet daily high impact and accurate CV health messages (identified in aim 3). Participants will follow the study team & may receive daily private heart health messages via Twitter. This would allow participants to see CV health messages posted in the words and context of patients that may be similar to them, participate in online CV health discussions, and access CV health networks that they may not otherwise know about. Participants will also tweet heart health messages weekly.
No Intervention: Health System: HTN Control, survey only Investigators will identify patients with known hypertension (ICD-9 code 401.9), from the Penn Data Store (PDS). Participants will be asked to complete 3 short surveys. Participants in this arm will be exposed to daily messages about heart health and asked to tweet about health.

Arm

Intervention/Treatment

Experimental: Health System: HTN Intervention

Investigators will identify patients with known hypertension (ICD-9 code 401.9), from the Penn Data Store (PDS). Participants will be asked to complete 3 short surveys.

Behavioral: Health System: HTN Intervention

Interested participants can enroll online via a Twitter link. Patients will consent to having their electronic health records accessed to validate clinical data. Participants will complete short surveys. The project Twitter account would follow tweeters with high impact CV messages and tweet daily high impact and accurate CV health messages (identified in aim 3). Participants will follow the study team & may receive daily private heart health messages via Twitter. This would allow participants to see CV health messages posted in the words and context of patients that may be similar to them, participate in online CV health discussions, and access CV health networks that they may not otherwise know about. Participants will also tweet heart health messages weekly.

No Intervention: Health System: HTN Control, survey only

Investigators will identify patients with known hypertension (ICD-9 code 401.9), from the Penn Data Store (PDS). Participants will be asked to complete 3 short surveys. Participants in this arm will be exposed to daily messages about heart health and asked to tweet about health.

Outcome Measures

Primary Outcome Measures

Blood Pressure [6 months]
5mm difference pre- study compared with post- study

Secondary Outcome Measures

Patient Activation [6 months]
Baseline and follow-up surveys will include questions regarding patient activation from the patient activation measure assessment (PAM-13). We will observe for a 2-3 point change in the index PAM score (average 58)