Jonas Tana, Ph.D., senior lecturer, Department of Healthcare, Arcada UAS, firstname.lastname@example.org
Pauleen Mannevaara, M.Sc., senior lecturer, Department of Healthcare, Arcada UAS
Ira Jeglinsky-Kankainen, Ph.D., principal lecturer, Department of Health and Wellbeing, Arcada UAS
Camilla Wikström-Grotell, Ph.D., principal lecturer, Department of Health and Wellbeing, Arcada UAS
The rise of search engines, discussion forums and social media has changed the way people access and engage with health information (Tana, 2019). The enormous possibilities to engage with online health information have since the birth of the Internet grown to an unprecedented extent, and the internet has become a platform to express individual moods and feelings of daily life. People share their thoughts and concerns in relation to a myriad of things in social media, discussion forums and search engines. For health-related matters this behaviour is rather extensive, exemplified by the fact that four thousand Google searches per second are health related (Seifter et al., 2010). Within the health and illness continuum, the health issues and threats that individuals face have been suggested to trigger health information behaviour. This behaviour can be defined as how people seek, obtain, evaluate, categorise and use health-related information in relation to their health and health threats (Ek, 2013). This behaviour is essential in regards to coping with illness and maintaining proper health and health behaviours (Ek, 2013; Lambert and Loiselle, 2007). More and more people also take greater responsibility in maintaining their health, and people even try to diagnose or treat themselves before seeking medical advice from professionals (Erikainen et al., 2019; Tana, 2019). Today, most of this health information behaviour happens online, and internet use as a source for information related to health or illness in Finland is, and has for quite some time already been ubiquitous (Tana, 2019). It seems that in some cases people prefer to interact and ask questions online, since members of online communities might be more knowledgeable about the condition or have dealt with or survived similar experiences (Halder, Poddar and Kan, 2017; Naslund et al., 2016).
This behaviour leaves vast amounts of digital traces behind, in search engines, social media and on websites. Billions of digital footprints from nearly all parts of the world provide a powerful opportunity to expand the evidence base across health. Compared to surveys, which apart from time lags suffer from well-known limitations and biases, web data is available almost immediately (Ayers, Althouse and Dredze, 2014). Web data can be an important source for proposing new health related hypotheses and insights. This is especially useful when studying phenomena where little or no traditional data exist, or when studying stigmatizing or sensitive health topics (Tana, 2019; Tana, Eirola and Eriksson-Backa, 2019a; Tana, Eirola and Eriksson-Backa, 2019b). This type of behavioural data will only continue to accumulate, and the use of data-rich, novel sources such as search engine and social media data is becoming a fast and cost-effective way to identify population needs and predict or even prevent healthcare emergencies.
While the impact on health of these massive amounts of data has yet to be fully appreciated, there is a clear need to utilise novel methods to maximise the gains that this new data has to offer (Ayers, Althouse and Dredze, 2014). Advances in artificial intelligence (AI), machine learning (ML) and natural language processing (NLP) present new possibilities to utilize these data for monitoring public health and health related issues, as well as to improve health promotion. AI applications have already entered health care and have been predicted to have a profound impact on health, and how healthcare is delivered (Dredze and Paul, 2014).
Utilising these novel methods can help in providing both early warning signals as well as prediction of different health trends, threats and issues. This is crucial in planning early, or so-called just in time, interventions, which can be highly beneficial for positive health outcomes, intervention impacts or disease containment (Ayers, Althouse and Dredze, 2014; Nuti et al, 2014; Tana, 2019; Yang et al, 2010). It could also help ensure that individuals, especially those who do not seek treatment or cannot be reached by traditional means, are provided evidence-based timely information, aid and interventions.
Opportunities in mental health issues
One such condition, where individuals do not always seek medical attention or help, is mental health disorders. The complexity, stigma, and barriers to care that mental illness present, increases the possibility for individuals to seek information about their problems online (Tana, 2019; Tana, Eirola and Eriksson-Backa, 2019b). Yet, depression and mental health issues are regarded by the World Health Organization to be among the most burdensome diseases in the world and deemed a public health priority (Marcus et al., 2012). Thus, knowledge about online health information behaviour in relation to mental health issues has the potential to present not only new and meaningful insights, but also potentially provide new solutions in prevention and treatment of these disorders.
Digital traces from online health information behaviour in relation to mental health issues can be harnessed to provide insights into mental health issues. Within the Arcada project AI driven Nordic Health and Welfare (https://www.arcada.fi/en/article/research/2021-01-19/research-programme-ai-driven-nordic-health-and-welfare) these digital traces are being utilized for research with the aim to build an Artificial Intelligence (AI) powered Mental Health Index. This index applies supervised machine learning algorithms that utilize web data from search engines and social media to analyse and predict trends in the manifestation of mental health concerns, like depression. The index enables the discovery of trends, patterns and signals to gain novel insights from large data sets.
Another suitable method to study digital traces in mental health related behaviours and issues is Natural Language Processing (NLP). Utilizing NLP and sentiment analysis as methods can broaden our understanding of so called pre-clinical behaviours that affect health, an area where a large knowledge gap is present. Analysing various open data sources, like social media or discussion forum data as well as survey data could help in identifying symptoms of anxiety or depression. Analysing the language or the emotional tone in online messages to identify, measure and detect depression or depressive symptoms as expressed in text could assist in constructing early detection/warning systems on various online platforms.
Modern healthcare is reliant on the individual to manage and to some extent selfcare of health and illness related issues and problems, a behaviour that is rather extensive. Yet these “pre-clinical” behaviours, that happen without the involvement of health professionals, have not previously been studied to a larger extent. Reasons for this have partly been the methodological limitations in studying and monitoring this kind of behaviour (Tana, 2019). As these barriers reduce, thanks to the novel approaches that have evolved in recent years, these need to be harnessed to study health related behaviours and phenomena, with the aim to fill the knowledge gaps that barriers in research methods previously posed. As more and more people are inclined to seek and share information about health-related issues online, changes in health status are often reflected by changes in different patterns on the internet. Utilizing web data for both quantitative and qualitative health behaviour research could serve as a means to identify and discover interesting insights that can function as a base for further hypothesis testing. It can also broaden our understanding of health-related behaviours in relation to different issues and threats. However, it needs to be emphasised that utilising novel approaches and methods for research in different health related areas should be seen as a complement to traditional health research, but should in no way be seen as a replacement. By utilizing novel approaches, we can hopefully provide more optimal and effective interventions and reach people in need of care more quickly.
Ayers, J. W., Althouse, B. M. and Dredze, M. (2014). Could behavioral medicine lead the web data revolution?. JAMA, 311(14), 1399-1400.
Dredze, M. and Paul, M. J. (2014). Natural language processing for health and social media. IEEE Intelligent Systems, 29(2), 64-67.
Ek, S. (2013). Gender differences in health information behaviour: a Finnish population-based survey. Health Promotion International, 30(3), 736-745.
Erikainen, S., Pickersgill, M., Cunningham-Burley, S. and Chan, S. (2019). Patienthood and participation in the digital era. Digital Health, 5, 1–10.
Halder, K., Poddar, L. and Kan, M. Y. (2017). Modeling Temporal Progression of Emotional Status in Mental Health Forum: A Recurrent Neural Net Approach. In Proceedings of the 8th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis (pp. 127-135).
Lambert, S. D. and Loiselle, C. G. (2007). Health information—seeking behavior. Qualitative Health Research, 17(8), 1006-1019.
Marcus, M., Yasamy, M. T., van Ommeren, M., Chisholm, D. and Saxena, S. (2012). Depression: A global public health concern. WHO Department of Mental Health and Substance Abuse, 1, 6-8.
Naslund, J. A., Aschbrenner, K. A., Marsch, L. A. and Bartels, S. J. (2016). The future of mental health care: peer-to-peer support and social media. Epidemiology and Psychiatric Sciences, 25(2), 113-122.
Nuti, S. V., Wayda, B., Ranasinghe, I., Wang, S., Dreyer, R. P., Chen, S. I. and Murugiah, K. (2014). The use of google trends in health care research: a systematic review. PloS one, 9(10), e109583.
Seifter, A., Schwarzwalder, A., Geis, K., & Aucott, J. (2010). The utility of “Google Trends” for epidemiological research: Lyme disease as an example. Geospatial health, 135-137.
Tana, J. 2019. Infodemiology: Studying rhythmicity in online health information behaviour. Åbo Akademi University. Doctoral dissertation.
Tana, J., Eirola, E. and Eriksson-Backa, K. (2019a). The aspect of time in online health information behaviour: a longitudinal extensive analysis of the Suomi24 discussion forum. Informaatiotutkimus 38(2019): 2.
Tana, J., Eirola, E. and Eriksson-Backa, K. (2019b). Rhythmicity of health information behaviour Utilizing the infodemiology approach to study temporal patterns and variations. Aslib Journal Of Information Management, 71(6), 773-788.
Vehko, T., Ruotsalainen, S. and Hyppönen, H. (2019). E-health and e-welfare of Finland: Check Point 2018. National Institute for Health and Welfare (THL). Helsinki, Finland.
Yang, A. C., Huang, N. E., Peng, C. K. and Tsai, S. J. (2010). Do seasons have an influence on the incidence of depression? The use of an internet search engine query data as a proxy of human affect. PloS one, 5(10), e13728