Evaluating the predictability of medical conditions from social media posts.

We studied whether medical conditions across 21 broad categories were predictable from social media content across approximately 20 million words written by 999 consenting patients. Facebook language significantly improved upon the prediction accuracy of demographic variables for 18 of the 21 disease categories; it was particularly effective at predicting diabetes and mental health conditions including anxiety, depression and psychoses. Social media data are a quantifiable link into the otherwise elusive daily lives of patients, providing an avenue for study and assessment of behavioral and environmental disease risk factors. Analogous to the genome, social media data linked to medical diagnoses can be banked with patients' consent, and an encoding of social media language can be used as markers of disease risk, serve as a screening tool, and elucidate disease epidemiology. In what we believe to be the first report linking electronic medical record data with social media data from consenting patients, we identified that patients' Facebook status updates can predict many health conditions, suggesting opportunities to use social media data to determine disease onset or exacerbation and to conduct social media-based health interventions.

Employment and residential characteristics in relation to automated external defibrillator locations.

BACKGROUND: Survival from out-of-hospital cardiac arrest (OHCA) is generally poor and varies by geography. Variability in automated external defibrillator (AED) locations may be a contributing factor. To inform optimal placement of AEDs, we investigated AED access in a major US city relative to demographic and employment characteristics. METHODS AND RESULTS: This was a retrospective analysis of a Philadelphia AED registry (2,559 total AEDs). The 2010 US Census and the Local Employment Dynamics database by ZIP code was used. Automated external defibrillator access was calculated as the weighted areal percentage of each ZIP code covered by a 400-m radius around each AED. Of 47 ZIP codes, only 9% (4) were high-AED-service areas. In 26% (12) of ZIP codes, less than 35% of the area was covered by AED service areas. Higher-AED-access ZIP codes were more likely to have a moderately populated residential area (P = .032), higher median household income (P = .006), and higher paying jobs (P =. 008). CONCLUSIONS: The locations of AEDs vary across specific ZIP codes; select residential and employment characteristics explain some variation. Further work on evaluating OHCA locations, AED use and availability, and OHCA outcomes could inform AED placement policies. Optimizing the placement of AEDs through this work may help to increase survival.

Linking social media and medical record data: a study of adults presenting to an academic, urban emergency department.

BACKGROUND: Social media may offer insight into the relationship between an individual's health and their everyday life, as well as attitudes towards health and the perceived quality of healthcare services. OBJECTIVE: To determine the acceptability to patients and potential utility to researchers of a database linking patients' social media content with their electronic medical record (EMR) data. METHODS: Adult Facebook/Twitter users who presented to an emergency department were queried about their willingness to share their social media data and EMR data with health researchers for the purpose of building a databank for research purposes. Shared posts were searched for select terms about health and healthcare. RESULTS: Of the 5256 patients approached, 2717 (52%) were Facebook and/or Twitter users. 1432 (53%) of those patients agreed to participate in the study. Of these participants, 1008 (71%) consented to share their social media data for the purposes of comparing it with their EMR. Social media data consisted of 1 395 720 posts/tweets to Facebook and Twitter. Participants sharing social media data were slightly younger (29.1±9.8 vs 31.9±10.4 years old; p<0.001), more likely to post at least once a day (42% vs 29%; p=0.003) and more likely to present to the emergency room via self-arrival mode and have private insurance. Of Facebook posts, 7.5% (95% CI 4.8% to 10.2%) were related to health. Individuals with a given diagnosis in their EMR were significantly more likely to use terms related to that diagnosis on Facebook than patients without that diagnosis in their EMR (p<0.0008). CONCLUSIONS: Many patients are willing to share and link their social media data with EMR data. Sharing patients have several demographic and clinical differences compared with non-sharers. A database that merges social media with EMR data has the potential to provide insights about individuals' health and health outcomes.

Using Twitter to Measure Public Discussion of Diseases: A Case Study.

BACKGROUND: Twitter is increasingly used to estimate disease prevalence, but such measurements can be biased, due to both biased sampling and inherent ambiguity of natural language. OBJECTIVE: We characterized the extent of these biases and how they vary with disease. METHODS: We correlated self-reported prevalence rates for 22 diseases from Experian's Simmons National Consumer Study (n=12,305) with the number of times these diseases were mentioned on Twitter during the same period (2012). We also identified and corrected for two types of bias present in Twitter data: (1) demographic variance between US Twitter users and the general US population; and (2) natural language ambiguity, which creates the possibility that mention of a disease name may not actually refer to the disease (eg, "heart attack" on Twitter often does not refer to myocardial infarction). We measured the correlation between disease prevalence and Twitter disease mentions both with and without bias correction. This allowed us to quantify each disease's overrepresentation or underrepresentation on Twitter, relative to its prevalence. RESULTS: Our sample included 80,680,449 tweets. Adjusting disease prevalence to correct for Twitter demographics more than doubles the correlation between Twitter disease mentions and disease prevalence in the general population (from .113 to .258, P <.001). In addition, diseases varied widely in how often mentions of their names on Twitter actually referred to the diseases, from 14.89% (3827/25,704) of instances (for stroke) to 99.92% (5044/5048) of instances (for arthritis). Applying ambiguity correction to our Twitter corpus achieves a correlation between disease mentions and prevalence of .208 ( P <.001). Simultaneously applying correction for both demographics and ambiguity more than triples the baseline correlation to .366 ( P <.001). Compared with prevalence rates, cancer appeared most overrepresented in Twitter, whereas high cholesterol appeared most underrepresented. CONCLUSIONS: Twitter is a potentially useful tool to measure public interest in and concerns about different diseases, but when comparing diseases, improvements can be made by adjusting for population demographics and word ambiguity.

Use of social media across US hospitals: descriptive analysis of adoption and utilization.

BACKGROUND: Use of social media has become widespread across the United States. Although businesses have invested in social media to engage consumers and promote products, less is known about the extent to which hospitals are using social media to interact with patients and promote health. OBJECTIVE: The aim was to investigate the relationship between hospital social media extent of adoption and utilization relative to hospital characteristics. METHODS: We conducted a cross-sectional review of hospital-related activity on 4 social media platforms: Facebook, Twitter, Yelp, and Foursquare. All US hospitals were included that reported complete data for the Centers for Medicare and Medicaid Services Hospital Consumer Assessment of Healthcare Providers and Systems survey and the American Hospital Association Annual Survey. We reviewed hospital social media webpages to determine the extent of adoption relative to hospital characteristics, including geographic region, urban designation, bed size, ownership type, and teaching status. Social media utilization was estimated from user activity specific to each social media platform, including number of Facebook likes, Twitter followers, Foursquare check-ins, and Yelp reviews. RESULTS: Adoption of social media varied across hospitals with 94.41% (3351/3371) having a Facebook page and 50.82% (1713/3371) having a Twitter account. A majority of hospitals had a Yelp page (99.14%, 3342/3371) and almost all hospitals had check-ins on Foursquare (99.41%, 3351/3371). Large, urban, private nonprofit, and teaching hospitals were more likely to have higher utilization of these accounts. CONCLUSIONS: Although most hospitals adopted at least one social media platform, utilization of social media varied according to several hospital characteristics. This preliminary investigation of social media adoption and utilization among US hospitals provides the framework for future studies investigating the effect of social media on patient outcomes, including links between social media use and the quality of hospital care and services.

Hidden in plain sight: a crowdsourced public art contest to make automated external defibrillators more visible.

OBJECTIVES: We sought to explore the feasibility of using a crowdsourcing study to promote awareness about automated external defibrillators (AEDs) and their locations. METHODS: The Defibrillator Design Challenge was an online initiative that asked the public to create educational designs that would enhance AED visibility, which took place over 8 weeks, from February 6, 2014, to April 6, 2014. Participants were encouraged to vote for AED designs and share designs on social media for points. Using a mixed-methods study design, we measured participant demographics and motivations, design characteristics, dissemination, and Web site engagement. RESULTS: Over 8 weeks, there were 13 992 unique Web site visitors; 119 submitted designs and 2140 voted. The designs were shared 48 254 times on Facebook and Twitter. Most designers-voters reported that they participated to contribute to an important cause (44%) rather than to win money (0.8%). Design themes included: empowerment, location awareness, objects (e.g., wings, lightning, batteries, lifebuoys), and others. CONCLUSIONS: The Defibrillator Design Challenge engaged a broad audience to generate AED designs and foster awareness. This project provides a framework for using design and contest architecture to promote health messages.

EHR in emergency rooms: exploring the effect of key information components on main complaints.

This study characterizes the information components associated with improved medical decision-making in the emergency room (ER). We looked at doctors' decisions to use or not to use information available to them on an electronic health record (EHR) and a Health Information Exchange (HIE) network, and tested for associations between their decision and parameters related to healthcare outcomes and processes. Using information components from the EHR and HIE was significantly related to improved quality of healthcare processes. Specifically, it was associated with both a reduction in potentially avoidable admissions as well as a reduction in rapid readmissions. Overall, the three information components; namely, previous encounters, imaging, and lab results emerged as having the strongest relationship with physicians' decisions to admit or discharge. Certain information components, however, presented an association between the diagnosis and the admission decisions (blood pressure was the most strongly associated parameter in cases of chest pain complaints and a previous surgical record for abdominal pain). These findings show that the ability to access patients' medical history and their long term health conditions (via the EHR), including information about medications, diagnoses, recent procedures and laboratory tests is critical to forming an appropriate plan of care and eventually making more accurate admission decisions.

Using a mobile app and mobile workforce to validate data about emergency public health resources.

BACKGROUND: Social media and mobile applications that allow people to work anywhere are changing the way people can contribute and collaborate. OBJECTIVE: We sought to determine the feasibility of using mobile workforce technology to validate the locations of automated external defibrillators (AEDs), an emergency public health resource. METHODS: We piloted the use of a mobile workforce application, to verify the location of 40 AEDs in Philadelphia county. AEDs were pre-identified in public locations for baseline data. The task of locating AEDs was posted online for a mobile workforce from October 2011 to January 2012. Participants were required to submit a mobile phone photo of AEDs and descriptions of the location. RESULTS: Thirty-five of the 40 AEDs were identified within the study period. Most, 91% (32/35) of the submitted AED photo information was confirmed project baseline data. Participants also provided additional data such as business hours and other nearby AEDs. CONCLUSIONS: It is feasible to engage a mobile workforce to complete health research-related tasks. Participants were able to validate information about emergency public health resources.

Crowdsourcing--harnessing the masses to advance health and medicine, a systematic review.

OBJECTIVE: Crowdsourcing research allows investigators to engage thousands of people to provide either data or data analysis. However, prior work has not documented the use of crowdsourcing in health and medical research. We sought to systematically review the literature to describe the scope of crowdsourcing in health research and to create a taxonomy to characterize past uses of this methodology for health and medical research. DATA SOURCES: PubMed, Embase, and CINAHL through March 2013. STUDY ELIGIBILITY CRITERIA: Primary peer-reviewed literature that used crowdsourcing for health research. STUDY APPRAISAL AND SYNTHESIS METHODS: Two authors independently screened studies and abstracted data, including demographics of the crowd engaged and approaches to crowdsourcing. RESULTS: Twenty-one health-related studies utilizing crowdsourcing met eligibility criteria. Four distinct types of crowdsourcing tasks were identified: problem solving, data processing, surveillance/monitoring, and surveying. These studies collectively engaged a crowd of >136,395 people, yet few studies reported demographics of the crowd. Only one (5 %) reported age, sex, and race statistics, and seven (33 %) reported at least one of these descriptors. Most reports included data on crowdsourcing logistics such as the length of crowdsourcing (n = 18, 86 %) and time to complete crowdsourcing task (n = 15, 71 %). All articles (n = 21, 100 %) reported employing some method for validating or improving the quality of data reported from the crowd. LIMITATIONS: Gray literature not searched and only a sample of online survey articles included. CONCLUSIONS AND IMPLICATIONS OF KEY FINDINGS: Utilizing crowdsourcing can improve the quality, cost, and speed of a research project while engaging large segments of the public and creating novel science. Standardized guidelines are needed on crowdsourcing metrics that should be collected and reported to provide clarity and comparability in methods.

TV Audience Measurement with Big Data.

TV audience measurement involves estimating the number of viewers tuned into a TV show at any given time as well as their demographics. First introduced shortly after commercial television broadcasting began in the late 1940s, audience measurement allowed the business of television to flourish by offering networks a way to quantify the monetary value of TV audiences for advertisers, who pay for the estimated number of eyeballs watching during commercials. The first measurement techniques suffered from multiple limitations because reliable, large-scale data were costly to acquire. Yet despite these limitations, measurement standards remained largely unchanged for decades until devices such as cable boxes, video-on-demand boxes, and cell phones, as well as web apps, Internet browser clicks, web queries, and social media activity, resulted in an explosion of digitally available data. TV viewers now leave digital traces that can be used to track almost every aspect of their daily lives, allowing the potential for large-scale aggregation across data sources for individual users and groups and enabling the tracking of more people on more dimensions for more shows. Data are now more comprehensive, available in real time, and cheaper to acquire, enabling accurate and fine-grained TV audience measurement. In this article, I discuss the evolution of audience measurement and what the recent data explosion means for the TV industry and academic research.

Where are lifesaving automated external defibrillators located and how hard is it to find them in a large urban city?

OBJECTIVES: Automated external defibrillators (AEDs) are lifesaving, but little is known about where they are located or how to find them. We sought to locate AEDs in high employment areas of Philadelphia and characterize the process of door-to-door surveying to identify these devices. METHODS: Block groups representing approximately the top 3rd of total primary jobs in Philadelphia were identified using the US Census Local Employment Dynamics database. All buildings within these block groups were surveyed during regular working hours over six weeks during July-August 2011. Buildings were characterized as publically accessible or inaccessible. For accessible buildings, address, location type, and AED presence were collected. Total devices, location description and prior use were gathered in locations with AEDs. Process information (total people contacted, survey duration) was collected for all buildings. RESULTS: Of 1420 buildings in 17 block groups, 949 (67%) were accessible, but most 834 (88%) did not have an AED. 283 AEDs were reported in 115 buildings (12%). 81 (29%) were validated through visualization and 68 (24%) through photo because employees often refused access. In buildings with AEDs, several employees (median 2; range 1-8) were contacted to ascertain information, which required several minutes (mean 4; range 1-55). CONCLUSIONS: Door-to-door surveying is a feasible, but time-consuming method for identifying AEDs in high employment areas. Few buildings reported having AEDs and few permitted visualization, which raises concerns about AED access. To improve cardiac arrest outcomes, efforts are needed to improve the availability of AEDs, awareness of their location and access to them.

Online discussion of drug side effects and discontinuation among breast cancer survivors.

PURPOSE: While patients often use the internet as a medium to search for and exchange health-related information, little is known about the extent to which patients use social media to discuss side effects related to medications. We aim to understand the frequency and content of side effects and associated adherence behaviors discussed by breast cancer patients related to using aromatase inhibitors (AIs), with particular emphasis on AI-related arthralgia. METHODS: We performed a mixed methods study to examine content related to AI associated side effects posted by individuals on 12 message boards between 2002 and 2010. We quantitatively defined the frequency and association between side effects and AIs and identified common themes using content analysis. One thousand randomly selected messages related to arthralgia were coded by two independent raters. RESULTS: Among 25 256 posts related to AIs, 4589 (18.2%) mentioned at least one side effect. Top-cited side effects on message boards related to AIs were joint/musculoskeletal pain (N = 5093), hot flashes (1498), osteoporosis (719), and weight gain (429). Among the authors posting messages who self-reported AI use, 12.8% mentioned discontinuing AIs, while another 28.1% mentioned switching AIs. Although patients often cited severe joint pain as the reason for discontinuing AIs, many also offered support and advice for coping with AI-associated arthralgia. CONCLUSION: Online discussion of AI-related side effects was common and often related to drug switching and discontinuation. Physicians should be aware of these discussions and guide patients to effectively manage side effects of drugs and promote optimal adherence.

Decoding twitter: Surveillance and trends for cardiac arrest and resuscitation communication.

AIM OF THE STUDY: Twitter has over 500 million subscribers but little is known about how it is used to communicate health information. We sought to characterize how Twitter users seek and share information related to cardiac arrest, a time-sensitive cardiovascular condition where initial treatment often relies on public knowledge and response. METHODS: Tweets published April-May 2011 with keywords cardiac arrest, CPR, AED, resuscitation, heart arrest, sudden death and defib were identified. Tweets were characterized by content, dissemination, and temporal trends. Tweet authors were further characterized by: self-identified background, tweet volume, and followers. RESULTS: Of 62,163 tweets (15,324, 25%) included resuscitation/cardiac arrest-specific information. These tweets referenced specific cardiac arrest events (1130, 7%), CPR performance or AED use (6896, 44%), resuscitation-related education, research, or news media (7449, 48%), or specific questions about cardiac arrest/resuscitation (270, 2%). Regarding dissemination (1980, 13%) of messages were retweeted. Resuscitation specific tweets primarily occurred on weekdays. Most users (10,282, 93%) contributed three or fewer tweets during the study time frame. Users with more than 15 resuscitation-specific tweets in the study time frame had a mean 1787 followers and most self-identified as having a healthcare affiliation. CONCLUSION: Despite a large volume of tweets, Twitter can be filtered to identify public knowledge and information seeking and sharing about cardiac arrest. To better engage via social media, healthcare providers can distil tweets by user, content, temporal trends, and message dissemination. Further understanding of information shared by the public in this forum could suggest new approaches for improving resuscitation related education.

LESSONS LEARNED ABOUT PUBLIC HEALTH FROM ONLINE CROWD SURVEILLANCE.

The Internet has forever changed the way people access information and make decisions about their healthcare needs. Patients now share information about their health at unprecedented rates on social networking sites such as Twitter and Facebook and on medical discussion boards. In addition to explicitly shared information about health conditions through posts, patients reveal data on their inner fears and desires about health when searching for health-related keywords on search engines. Data are also generated by the use of mobile phone applications that track users' health behaviors (e.g., eating and exercise habits) as well as give medical advice. The data generated through these applications are mined and repackaged by surveillance systems developed by academics, companies, and governments alike to provide insight to patients and healthcare providers for medical decisions. Until recently, most Internet research in public health has been surveillance focused or monitoring health behaviors. Only recently have researchers used and interacted with the crowd to ask questions and collect health-related data. In the future, we expect to move from this surveillance focus to the "ideal" of Internet-based patient-level interventions where healthcare providers help patients change their health behaviors. In this article, we highlight the results of our prior research on crowd surveillance and make suggestions for the future.

Identifying potential adverse effects using the web: a new approach to medical hypothesis generation.

Medical message boards are online resources where users with a particular condition exchange information, some of which they might not otherwise share with medical providers. Many of these boards contain a large number of posts and contain patient opinions and experiences that would be potentially useful to clinicians and researchers. We present an approach that is able to collect a corpus of medical message board posts, de-identify the corpus, and extract information on potential adverse drug effects discussed by users. Using a corpus of posts to breast cancer message boards, we identified drug event pairs using co-occurrence statistics. We then compared the identified drug event pairs with adverse effects listed on the package labels of tamoxifen, anastrozole, exemestane, and letrozole. Of the pairs identified by our system, 75-80% were documented on the drug labels. Some of the undocumented pairs may represent previously unidentified adverse drug effects.

A system for de-identifying medical message board text.

There are millions of public posts to medical message boards by users seeking support and information on a wide range of medical conditions. It has been shown that these posts can be used to gain a greater understanding of patients' experiences and concerns. As investigators continue to explore large corpora of medical discussion board data for research purposes, protecting the privacy of the members of these online communities becomes an important challenge that needs to be met. Extant entity recognition methods used for more structured text are not sufficient because message posts present additional challenges: the posts contain many typographical errors, larger variety of possible names, terms and abbreviations specific to Internet posts or a particular message board, and mentions of the authors' personal lives. The main contribution of this paper is a system to de-identify the authors of message board posts automatically, taking into account the aforementioned challenges. We demonstrate our system on two different message board corpora, one on breast cancer and another on arthritis. We show that our approach significantly outperforms other publicly available named entity recognition and de-identification systems, which have been tuned for more structured text like operative reports, pathology reports, discharge summaries, or newswire.

Natural supplements for H1N1 influenza: retrospective observational infodemiology study of information and search activity on the Internet.

BACKGROUND: As the incidence of H1N1 increases, the lay public may turn to the Internet for information about natural supplements for prevention and treatment. OBJECTIVE: Our objective was to identify and characterize websites that provide information about herbal and natural supplements with information about H1N1 and to examine trends in the public's behavior in searching for information about supplement use in preventing or treating H1N1. METHODS: This was a retrospective observational infodemiology study of indexed websites and Internet search activity over the period January 1, 2009, through November 15, 2009. The setting is the Internet as indexed by Google with aggregated Internet user data. The main outcome measures were the frequency of "hits" or webpages containing terms relating to natural supplements co-occurring with H1N1/swine flu, terms relating to natural supplements co-occurring with H1N1/swine flu proportional to all terms relating to natural supplements, webpage rank, webpage entropy, and temporal trend in search activity. RESULTS: A large number of websites support information about supplements and H1N1. The supplement with the highest proportion of H1N1/swine flu information was a homeopathic remedy known as Oscillococcinum that has no known side effects; supplements with the next highest proportions have known side effects and interactions. Webpages with both supplement and H1N1/swine flu information were less likely to be medically curated or authoritative. Search activity for supplements was temporally related to H1N1/swine flu-related news reports and events. CONCLUSIONS: The prevalence of nonauthoritative webpages with information about supplements in the context of H1N1/swine flu and the increasing number of searches for these pages suggest that the public is interested in alternatives to traditional prevention and treatment of H1N1. The quality of this information is often questionable and clinicians should be cognizant that patients may be at risk of adverse events associated with the use of supplements for H1N1.