Using Data for Better Cancer Treatment
By Åsa Cajander, Uppsala University, Department of Information Technology, Division of Visual Information and Interaction, Christiane Grünloh, TH Köln, Germany and KTH Royal Institute of Technology, Stockholm, Jonas Moll, Uppsala University, Department of Information Technology, Division of Visual Information and Interaction
In high-income countries, many types of cancer are nowadays curable, or treatable in a way that they can be considered as a chronic rather than a fatal disease. Nevertheless, the number of patients being diagnosed with cancer every year is increasing and survival and effective treatment is dependent on early detection as well as on continuous monitoring. This highlights the importance of collecting and analyzing large amounts of data in an effective manner.
At the same time, there are already large amounts of health-related data that are readily accessible to healthcare professionals today, e.g. electronic health records, biobanks, and knowledge banks. Cancer patients, however, do not necessarily have access to this kind of information despite their growing interest. Furthermore, patients often compile information themselves, e.g. regarding nausea, pain, medication etc. These “medical logbooks” are currently not used on a regular basis, e.g. discussed during a visit or used to continuously monitor progress.
The workshop will address these issues and is aimed at creating a common understanding of how the future might be, using a vision seminar process approach.
The main focus areas for the workshop will be:
Joint analysis of critical incidents related to the diagnosis and treatment of cancer. The critical incidents will be related to all stakeholders involved in the process. What are the enablers, barriers, and learning opportunities?
To create visions of how diagnosis and treatment of cancer can be informed by more effective and integrated use of existing data by different stakeholders.
Cancer care has come a long way and treatments are constantly improving. Some cancer conditions are today considered as chronic and not necessarily fatal. The number of people diagnosed with cancer globally is, however, constantly increasing, particularly in low- and middle countries. For economic and logistical reasons, many advances in cancer diagnostics and treatment are not yet accessible in these countries, resulting in a high number of fatalities. Equal access to healthcare is still a long way off. Demographic differences, especially when it comes to diagnosis and outcome, can be seen between children and adults. Cancer in children is often not discovered early enough. Moreover, cancer treatment is facing the same issues as healthcare at large, including more elderly patients, gender differences, more patients with multiple chronic conditions and a major lack of personnel and considerable differences globally in access to care. eHealth has emerged as one possible solution to some of these problems, and increased access to and use of healthcare data is often mentioned as one way forward. Such things have the potential to improve quality, make treatment more efficient, and expand capacity.
There are three main data user groups in cancer care. Healthcare professionals, who need to see data related to their patients, such as blood tests, radiology results, and notes from previous meetings with doctors. Patients, who want to see not only their own data but also aggregated data from other patients (available, for example, on platforms such as www.patientslikeme.com). And researchers, who both generate and use extensive amounts of data coming from biobanks and quality registries, or clinical notes. There are other kinds of users, such as statisticians, managers, drug industry representatives, and politicians, but these seldom interact with the data on a daily basis and their needs are not addressed in the workshop.
New technology, new opportunities
New technologies are not only rapidly changing society, but also opening up for new possibilities in healthcare. Various data sources have been available for some time now, and it is likely that both the number of sources and the amount of data will increase even more. Digitalization in healthcare is on top of the agenda all over the world, though implementation capacity varies greatly. Eventually all patient data will be digital and enable analysis that can advance not only research and treatment, but also cancer prevention.
High-performance computing opens new door for analysis
New technologies also bring new opportunities to engage in innovative research, for example making use of large quantities of data to identify patterns in specific types of cancer. Previously, this was not possible as the data were either not available, or high-performance computing was not sufficiently advanced. Advancements in distributed computing aim to increase processing power, for example by making use of idle computers and even smartphones. The technology knows no limits!
The ever-increasing data points have also given rise to artificial intelligence solutions, like cognitive computing in diagnostics . This technology accesses vast amounts of data sources and finds patterns that can be vital for cancer diagnostics. This is one example of how artificial intelligence has moved into the medical domain, powered by existing data.
Technology and medical developments also go hand in hand. For example, due to the decrease in cost, genetic sequencing has become more popular. This is particularly interesting for oncology in terms of precision medicine, such as identifying personalized treatment options for the patient. Technology can help medical professionals find appropriate evidence-based treatment options, explore clinical trial opportunities, and recruit participants for research studies much more easily.
Today, genetic testing is even available as a consumer product. Furthermore, the widespread use of advanced devices, such as smartphones and wearables, allow users to collect a vast amount of data about their own health and wellbeing. This is not only relevant to current patients, but to everyone who might become one. In some incidents, data from fitness trackers have proven useful in preventing or detecting heart attacks, for instance. These data could also be used to measure the continuous health effects of ongoing cancer treatments.
Barriers to overcome
Although there seems to be no limit to the opportunities, there are many barriers to overcome regarding the successful use of technologies in healthcare. Sadly, between 50 and 70 per cent of information and communication technologies (ICT) development projects fail to reach their goals. In addition, healthcare professionals and patients often struggle with the systems implemented in their own organization. Although automation, big data, smart technologies, and distributed computing offer great possibilities for healthcare, they need to be incorporated in an ecology of computer systems and people that is already quite complex and becoming more so all the time. The barriers to the successful implementation of eHealth initiatives can be categorized in many different ways. Furthermore, they are interrelated, overlapping, and situated on different levels in society and organizations.
The barriers to eHealth are often organizational
Many would argue that the challenge for successful eHealth implementation is not related to technical but rather organizational and infrastructural aspects. Organizational barriers include managing organizational change and development so that technology is efficiently incorporated into work. Changes may be required in structures, business processes, and culture. This can be very challenging, especially in developing countries that are still establishing a basic working infrastructure. Infrastructural barriers can include access to electricity and Internet as well as to hi tech IT solutions.
ICT must be integrated not isolated
A barrier related to ICT is usability. Far too many systems are not effectively integrated into everyday work practices, and score badly when it comes to effectiveness, efficiency, and satisfaction. The use of data needs to become part of a work environment that supports the professional development of healthcare personnel while providing useful support in their work.
Technical and ethical aspects of interoperability
Another problem is related to system interoperability, i.e. programs are not compatible with one another and thus there is no information exchange. This makes it more difficult to link up databases to find new patterns and gather data from different systems to discover correlations. Problems with interoperability result in data redundancy as the same information is documented in several systems, inefficient work, and lack of information from other areas or countries. The establishment of interoperability is challenging, as the system architecture has to allow for efficient exchange of information between systems with different levels of confidentiality, some of which is collected by patients (e.g., through self-monitoring or wearables), while at the same time preventing unauthorized access. In other words: Storing sensitive information in the same place can be seen as an invitation for hackers. Therefore, standards and architectures have to be developed that ensure efficient and trustworthy exchange of information.
Under-representation of certain cancer types in existing data banks
Moreover, eHealth innovations often run into legal and ethical challenges. Laws and regulations need to be adjusted to accommodate new initiatives, while also addressing integrity and privacy issues. Health data used in research are usually anonymized, however, anonymization is extremely difficult when it comes to big data or data used in genetics research. Inequality between developed and developing countries is still prevalent as regards access to data and advanced treatment methods. Large amounts of data are gathered through, for example, patient communities and in biobanks and genetic banks in most European countries and in the United States. Since the infrastructure and knowledge necessary to gather and make use of these quantities of data are not yet in place in developing, low-income countries, cancer types which are most frequent in these countries are under-represented in the available repositories. This is problematic for both healthcare professionals and researchers. This inequality, as well as eliminating the digital divide, needs to be addressed when discussing how to utilize existing data in diagnostics and treatment. Another ethical dilemma relates to artificial intelligence solutions. Although they provide great opportunities, it is still open to question how much the conclusions drawn by a machine can be trusted.
A watershed in the doctor-patient relationship thanks to new technology
The introduction of new technologies also has an effect on relationships within healthcare. Patients today are more involved (see, e.g. Chen, Y., Elenee Argentinis, J.D., and Weber, G), want to access their medical data, and also contribute, for example, by sharing data from self-monitoring. Although this provides many opportunities to increase the quality of care, it also has consequences. These have to be accounted for by technical and organizational infrastructures, laws, and medical education, as roles and responsibilities in the relationship between patients and healthcare professionals are changing.
Using Critical Incidents and visions of the future
This workshop will focus on practical issues, challenges, and opportunities related to using existing data for the diagnosis and treatment of cancer. In the first phase of the workshop, real-life Critical Incidents will be presented from the perspectives of doctors, nurses, patients and researchers, respectively, and used to inspire discussions on how existing data are being used today, and what the problems and opportunities are.
The second phase of the workshop will begin with a keynote on visions of the future, and what visions are good for in relation to societal change. Different visions from the stakeholders’ perspectives will then be developed by the participants. These will include scenarios and descriptions related to how various types of existing data can be used for the diagnosis and treatment of cancer.
Chen, Y., Elenee Argentinis, J.D., and Weber, G. (2016). IBM Watson: How Cognitive Computing Can Be Applied to Big Data Challenges in Life Sciences Research. Clinical Therapeutics (38: 4), pp. 688-701.
Grünloh, C., Hallewell Haslwanter, J.D., Kane, B., Lee, E., Lind, T., Moll, J., Rexhepi, H., and Scandurra, I. (2017). Using Critical Incidents in Workshops to Inform eHealth Design. In: proceedings of the 16th IFIP TC 13 International Conference on Human-Computer Interaction, INTERACT 2017 (Mumbai, India, September 2017). pp. 364-373.
Riggare, S., and Unruh, K.T. (2015). Patients organise and train doctors to provide better care. BMJ (351). H6318.
 Chen, Y., Elenee Argentinis, J.D., and Weber, G