Using Twitter to learn about the autism community

Beykikhoshk, Adham; Arandjelović, Ognjen; Phung, Dinh; Venkatesh, Svetha; Caelli, Terry

doi:10.1007/s13278-015-0261-5

Cited by 54 publications

(43 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…At the same time it is important to note that the value of this parameter which governs the rate of frequency drop-off as the term rank is increased, is rather different than that reported in previous work on short text analysis e.g. Twitter messages [8]. The slower drop-off in our corpus resonates with our observation that medical consultations are rather constrained semantically by their very nature.…”

Section: Resultscontrasting

confidence: 45%

Towards objective and reproducible study of patient-doctor interaction: Automatic text analysis based VR-CoDES annotation of consultation transcripts

Birkett

Arandjelović

Humphris

2017

2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

View full text Add to dashboard Cite

Abstract-While increasingly appreciated for its importance, the interaction between health care professionals (HCP) and patients is notoriously difficult to study, with both methodological and practical challenges. The former has been addressed by the so-called Verona coding definitions of emotional sequences (VRCoDES) -a system for identifying and coding patient emotions and the corresponding HCP responses -shown to be reliable and informative in a number of independent studies in different health care delivery contexts. In the present work we focus on the practical challenge of the scalability of this coding system, namely on making it easily usable more widely and on applying it on larger patient cohorts. In particular, VR-CoDES is inherently complex and training is required to ensure consistent annotation of audio recordings or textual transcripts of consultations. Following up on our previous pilot investigation, in the the present paper we describe the first automatic, computer based algorithm capable of providing coarse level coding of textual transcripts. We investigate different representations of patient utterances and classification methodologies, and label each utterance as either containing an explicit expression of emotional distress (a 'concern'), an implicit one (a 'cue'), or neither. Using a data corpus comprising 200 consultations between radiotherapists and adult female breast cancer patients we demonstrate excellent labelling performance.

show abstract

Section: Resultscontrasting

confidence: 45%

Towards objective and reproducible study of patient-doctor interaction: Automatic text analysis based VR-CoDES annotation of consultation transcripts

Birkett

Arandjelović

Humphris

2017

2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)

View full text Add to dashboard Cite

show abstract

“…As expected from the already noted unevenness in the number of publications in different fields, as illustrated by the sizes of blobs representing topics in Figure 1, most of the data is explained by relatively few topics (academic fields) with approximately 80% of the publications being in the first 20 inferred research fields. We observed a rough inverse power law distribution -observed frequently in nature across a wide range of phenomena [19,7,5,2] -in the publishing output per research field.…”

Section: Resultsmentioning

confidence: 71%

Learning nuanced cross-disciplinary citation metric normalization using the hierarchical dirichlet process on big scholarly data

Umar

Arandjelović

2017

Proceedings of the Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

Citation counts have long been used in academia as a way of measuring, inter alia, the importance of journals, quantifying the significance and the impact of a researcher's body of work, and allocating funding for individuals and departments. For example, the h-index proposed by Hirsch is one of the most popular metrics that utilizes citation analysis to determine an individual's research impact. Among many issues, one of the pitfalls of citation metrics is the unfairness which emerges when comparisons are made between researchers in different fields. The algorithm we described in the present paper learns evidence based, nuanced, and probabilistic representations of academic fields, and uses data collected by crawling Google Scholar to perform field of study based normalization of citation based impact metrics such as the h-index.

show abstract

“…Three studies used retrospective data collection to examine past tweets (Beykikhoshk et al, 2014(Beykikhoshk et al, , 2015Ramagopalan et al, 2014). Six studies collected data prospectively by setting a future date to begin collecting tweets as data (Ahlwardt et al, 2014;Greaves et al, 2014;Marton, 2012;McNeil et al, 2011;Nakhasi et al, 2016;Parsons et al, 2015).…”

Section: Mixed-methods Studiesmentioning

confidence: 99%

Using Twitter for Data Collection With Health-Care Consumers

Zhang

Albrecht

Scott

2018

International Journal of Qualitative Methods

View full text Add to dashboard Cite

Background: Twitter is one of the most popular social media platforms. The growing use of Twitter by health-care consumers creates a novel venue to understand patient experiences. To understand the potential for this platform to be utilized in patient-and family-oriented health research, this study reviewed published literature on the use of Twitter in health research. Methods: In collaboration with the research team, a research librarian designed and implemented a search strategy in eight databases. Primary and secondary screenings were conducted using predetermined criteria by one reviewer. A second reviewer verified screening decisions in 10% of the studies. Evidence tables were created to synthesize across the following study elements: research design, data collection techniques, analytic approaches, and author's insights on Twitter as a data collection method. Descriptive narrative analysis was used to synthesize data. Results: The search strategy captured 618 articles; 233 were eliminated in primary screening and 366 articles were eliminated during secondary screening. Verification by the second reviewer resulted in very good agreement (k ¼ .980). Seventeen articles were included in the final data set. Synthesis across the studies demonstrated that Twitter is currently used to search and mine research data, while active recruitment strategies on Twitter are just beginning to emerge. Conclusion: The novelty of Twitter for study recruitment and data collection with health-care consumers presents advantages and challenges that differ from traditional methods of data collection.

show abstract

Using Twitter to learn about the autism community

Cited by 54 publications

References 36 publications

Towards objective and reproducible study of patient-doctor interaction: Automatic text analysis based VR-CoDES annotation of consultation transcripts

Towards objective and reproducible study of patient-doctor interaction: Automatic text analysis based VR-CoDES annotation of consultation transcripts

Learning nuanced cross-disciplinary citation metric normalization using the hierarchical dirichlet process on big scholarly data

Using Twitter for Data Collection With Health-Care Consumers

Contact Info

Product

Resources

About