Big Data

Rustam Gilfanov: The ethical dilemma with big data in healthcare and how to solve it

big data in healthcare

The coronavirus pandemic puts greater emphasis on the problem of digitalization for many, if not all, aspects of present-day life, with the healthcare industry being in the vanguard for obvious reasons. The amount of medical data that need prompt collection, processing, transfer, and interpretation saw a tenfold or even a hundredfold increase. Not only has this caused the workload of healthcare personnel to soar and revealed several similar issues, but it also demonstrated the vast opportunities MedTech can offer.

Nowadays, big data activities belong not only to business operations but also to medical routines. The only difference is that the management of large data arrays affects, apart from revenues, the life and well-being of clinic patients. Big data operations enable optimization of almost all working stages of a medical facility, from verifying medical data and integrating diagnostics results into a unified system to making diagnostics and treatment planning more precise. In theory, such upscaling is not going to put more workload on the medical staff. Well, the intentions seem to be good – but how are they put to practice?

Underequipped medical facilities, imperfect software solutions, lacking competencies (particularly, digital literacy) among both patients and medical employees are the major obstacles to the effective implementation of big data technologies in healthcare. But clearly, this is not the news. What is more interesting, even the most promising MedTech projects will not deliver the intended results unless they meet certain ethical requirements.

Here is the minimal, albeit incomplete, set of conditions for success. Firstly, the attitude to IT infrastructure must be professional, with products covering basic functional needs, leaving no gaps, and “speaking the same language” without any contradictions or overlaps. Secondly, proper implementation and support must ensure protection from potential black swans — unpredictable game-changing events. Thirdly, the ongoing personnel training and dialogue with patients need to be maintained. Finally, probably the most essential requirement is to work out an ethical code that is put to paper and encompasses all business processes bar none. Successful MedTech startups all over the world, from Silicon Valley to the Dead Sea, have proved this approach to be the most effective.

Painful big data

To better understand the ethical aspect of big data activities in healthcare, let us take a look at the infamous online leak of information about 300,000 Moscow residents who suffered from the coronavirus.

So, the database of patients diagnosed with COVID-19 went public. The leaked archive contained PII, personal identifiable information (full names, phone numbers, DOBs, addresses, and passport numbers) of all persons, including children and foreign citizens, medical reports, screenshots, and photographs of 1C datasheets. Any user could get all the info on a patient in a few seconds simply by searching through the archive.

According to the media, this accident and similar minor leaks were followed by public discrimination or downright harassment of people, whose personal information had become available.

Thus, it turns out healthcare digitalization and optimized medical data management require a new, morality-based perspective. MedTech involves not only technical support and legal protection but also strict ethical and personal limits.

Healthy and unhealthy conservatism

The reasonable attitude to personal data is to treat them like body parts, i.e., taking good care and demonstrating them only to reliable and responsible people. As most people all over the world have not adopted this mindset yet, significant risks are posed for both patients and the professional community. In this case, healthy conservatism can be helpful, as it urges not to reveal personal information needlessly.

Now let us talk about unhealthy conservatism. Most probably, any story like the one mentioned above stems from the lack of data literacy, general awareness, and basic knowledge on cybercrimes and digital hygiene. Besides, there is no secret that healthcare remains one of the most conservative structures not only in Russia – this is a global tendency. However, while the Western counties began to adopt system digitalization processes as far as in the 1990s, Russia showed some real progress in this area only a few years ago.

The growing implementation of online technologies, coupled with the new coronavirus pandemic, caused the healthcare system to quickly adapt to emerging demands. While foreign countries have prepared the groundwork for such rapid changes, a lot has to be done from scratch in Russia. However, this gives us a certain advantage: we have the opportunity to follow the trends and develop a sophisticated model while avoiding mistakes made by other nations in the past.

According to the 2020 survey made by Medical Information Solutions company, half of the respondents (3,264 doctors from 85 Russian federal subjects) still use paper medical charts. This tendency was confirmed by the 2021 survey by Vademecum medical journal, with 66% of the respondents keeping soft copies of digital charts and only 86% having a computer with Internet access at their workplace.

Meanwhile, it is the systemic aggregation of medical data via digital charts that can address at least three essential tasks: improving the quality of clinical services, reducing risks of medical errors, and paving the way for target-driven data management that would benefit the healthcare system.

So, we are getting back to the beginning of my article and the prerequisites for success that I listed. The IT infrastructure of medical facilities will surely correspond to present-day needs, with standardization rates rising higher and vendor products becoming more accessible and of better quality. However, this requires not only professional integrators that work for serious money but also the crucial resource of time. Some local specifics will remain but it is nothing to worry about: this happens everywhere, be it healthcare or, say, power industry.

Public reaction to big data application in medical facilities is somewhat ambiguous. On the one hand, people are willing to use telemedicine technologies, as well as book appointments and exchange documents online. On their hand, when it comes to providing personal information, their attitude changes for the opposite.

Patients will behave more consciously and responsibly when clear “printed” ethical codes appear and medics themselves improve their data literacy. Socially favored behavior will eventually be established for the digital culture; disregard for personal data will seem as embarrassingly downgrading as the habit of throwing garbage out of your window.

Medical, but no longer private

The ethical issue of implementing Big Data into the medical assistance system focuses primarily on medical privacy protection.

According to the Federal Law “On Basics of Health Protection of the Citizens in the Russian Federation” the principle of medical privacy applies to the fact of requesting medical treatment, diagnosis, health status, and other information obtained during medical examinations and therapy. This also encompasses non-medical data, e.g., patient’s personal information, address, etc. Consequently, all information collected in a medical facility is regulated by this federal law. This includes not only textual records kept in the patients’ register but also CT scans, X-ray images, biopsy and lab test results, etc. So many different formats make it hard to systematize and analyze medical data, as well as to ensure their proper protection.

To bring together these incongruous pieces of information, the IT infrastructure applies several standards for medical data management and communication (HL7, ASTM, DICOM, GDT). These standards do not depend on medical equipment manufacturers and facilitate data transfer between hospitals, regions, and countries. It must be noted that establishing a reliable infrastructure to meet all requirements and ensuring its protection is as important as protecting the information it contains. These two things are connected – that is what we learned, as we were working on Longenesis, the tool for digitalizing biomedical research results. No matter how simple and logical a data management technology is, it will be completely useless unless it is properly protected against hacking and has a well-structured access system.

The policy of giving access to information on individual patients is regulated by the law. It stipulates that such data can be disclosed or transferred at the patient’s approval or, for instance, after the patient’s death to protect their relatives’ right to life and well-being. When it comes to mass transfers of personal medical records, the situation becomes more complicated. In March 2021, the Ministry of Digital Development announced that Russian citizens’ medical data would become available to private AI companies so that they could develop a new diagnostics system. Although the ministry officials emphasized the data would be depersonalized, this news led to a huge public and media controversy.

Apart from the legal challenges the project will surely face, several other aspects deserve attention as well. Firstly, how will the information be depersonalized, and is there a possibility to restore the original version of the code? Secondly, does not this offer seem too generous? Analyzing patients’ economic behavior, preferences, and other statistical evidence will be useful for businesses; nevertheless, it is impossible to find any patterns for pathology diagnostics without information on individual patients, their medical history, and body specifics.

If we compare Russian practices with international cases, we can see that Russia, as always, is trying to work out its own, “Eurasian” solution. On the one hand, we are interested in our Asian neighbor’s experience – it is well-known that analytical data can be collected way easier in China than, for example, in Canada. On the other hand, we belong to the Western information market both structurally and technologically, adopting its mindset and approaches. From the legislative standpoint, Russia has much more in common with Germany, France, or Australia than with China. This applies, in particular, to federal regulation of personal data, requirements for business entities, and biometric protection standards. However, the case of Israel is an interesting one, as this country has stringent requirements for data protection, while effectively using these data for scientific and medical purposes.

“The West will help us?”

I have changed Ostap Bender’s famous line from Ilf and Petrov’s famous picaresque novel, turning it into a question that I will answer myself. No, it will not. There is no point in waiting for some higher power to solve our problems with healthcare digitalization and its ethical aspects. We have to roll up our sleeves and learn from our colleagues’ best practices.

The USA is the global leader of the MedTech industry. The active introduction of information technologies and big data in healthcare is driven mainly by economic interests: all the processes are focused on optimization and cost reduction. Since the 1990s, patient registers have been maintained in the US for specific diseases. Now, they provide unique information on over 15 million patients. One of the first smart databases, also created in the US, comprises tumor and brain imaging scans, anamnesis data, etc. Data management is well-organized even on the level of local hospitals. Safer laser therapy treatment, better death rate forecasting for surgical interventions, more effective drug therapy – these are just a few clinical achievements of the American healthcare system. Economic benefits are also self-evident: a rational approach helps to avoid additional expenses on equipment or laboratory research.

Such a boost is obviously not possible without government support and investments. However, it should be noted that the success and effectiveness of Big Data in the healthcare industry rely on well-designed IT architecture, safe and obstacle-free data exchange between medical facilities, pharmaceutical companies, the government, and patients themselves, as well as on strict adherence to human right for privacy, moral and ethical standards – the latter factor being the pivotal one.

As I began to study foreign practices of using big data in healthcare, I was surprised by how openly the data are collected, how systemically they are processed, how promptly transferred and interpreted. It took me some time to find an explanation. Of course, well-established access policies for data arrays, sophisticated IT infrastructure, and multiple levels of data protection are of great value; still, the major role in this staggering success belongs to data literacy of all participants: doctors, nursing personnel, patients, administrative workers, and IT experts and developers.

The ability to handle data, draw conclusions from them, and use them for professional activities, coupled with clear legal regulation and ethical rules, predetermines productive usage of Big Data for the public good.

American analysts Bhumil Shah and Rachel Capan cooperated with QLIK to share their experience of managing medical data during the coronavirus pandemic. Their main objective was to provide equal access to data for all system users, with no exceptions, and make big data as transparent as possible so that users could make reasonable decisions basing on them. I think we can agree that despite their importance, legal or technical aspects of data protection are less relevant. The leading role should be given to the ethical, moral understanding of what is acceptable and what is not.

European nations are keeping up with their transatlantic partners: they have been actively using big data in healthcare, although opting for a less democratic approach. The SNDS system that originated in France and was eventually implemented all over Europe collects information on patients from various sources (e.g., databases on disability or lethal cases). The right to access this data array is provided to government structures and other agencies related to the healthcare system. Data safety is ensured by a French Internet regulator that operates according to the laws on personal information protection.

This systemic and structured approach to big data application in healthcare is not typical for this country – however, the sky is not the limit.

What’s the treatment?

In theory, it is possible to use administrative funds and give computers with Internet access to all doctors and even patients – but the real problem is the mindset. Unfortunately, this mindset is shared by many in professional and civil communities, requiring a search for a consensus.

The European Union offers an example of reasonable legislation based on practice and public discussion, as it is actively implementing the General Data Protection Regulation (GDPR) provisions. Since its adoption in 2016 and coming into effect in 2018, it has raised many questions and caused many disputes due to the involvement of multiple interested parties. Citizens and companies are engaged in the dialog on improving this essential document – and we will see these improvements coming to life.

The precursor to the GDPR was Direction 95/46/EC of October 24, 1995, titled “Directive on the protection of individuals with regard to the processing of personal data and on the free movement of such data”, made by the European Parliament and the European Council. Upon its adoption, the rights of data subjects were significantly expanded while the operators’ responsibilities and penalties for their non-fulfillment were largely increased.

Summing up, it seems logical that ethical questions, including ones related to the IT industry, need to be addressed according to the three aspects: legal, technological, and social – with the latter being the most troubling one. Nothing can be harder than changing people’s mindsets.

The medical community plays an important role in dealing with this challenge. Why do doctors tend to keep traditional paper charts instead of digital ones? Medical education is conservative by its nature and cannot keep up with the best practices. Of course, modern educational programs involve various digital solutions, yet a skeptical attitude to them is passed viva voce from academic elders to aspiring Hippocrateses. This situation can be changed only by changing the attitude to the educational process, introducing digital and data awareness disciplines. Equipping every doctor with a computer is not enough: medics need to be taught how to collect the data and draw conclusions from them. This is the only way to make medical workers of all levels realize the value of the information they handle, as well as the moral and legal responsibility imposed on them.

Maybe there is a more down-to-earth explanation to this conflict between medics and information technologies: the idea that professionals can be substituted by machines does not sound particularly heart-warming. This has to be kept in mind to avoid overwhelming digitalization; besides, not all processes and medical operations can be “algorithmized”. The joint research conducted by Harvard Medical School, Beth Israel Deaconess Medical Center in Boston, and Massachusetts Institute of Technology found out that in the case of brand cancer diagnostics, the error rate is 7.5% for the neural network and 3.5% for doctors – but if they work together, the rate drops down to 0.5%. This is why promoting a hybrid medical system, where people and IT are equally important, sounds more promising.

This knowledge can help to improve diagnostics quality, particularly in telemedicine. This is especially relevant for remote, scarcely populated regions, where there it is impossible to physically localize required medical specialists.

The analysis of recent trends demonstrates that the industrial application of AI and Big Data is becoming more important, as it makes diagnoses more precise and saves time. This tendency will definitely continue.

The patients’ fear of keeping digital records about their diagnosis and other personal things should also be treated by raising awareness and understanding. When it comes to digital medical charts, it is better to explain than to enforce. There is an opinion that the outcome will be similar to credit card introduction, with initial overwhelming mistrust eventually turning into universal usage. However, credit card users risk nothing except their money, while unlawful disclosure of medical records or diagnoses puts people’s reputations at stake. We all can agree the damage is definitely not equal.

Sophisticated information technologies and their infrastructure also form an essential element of the big picture titled “Effective big data in healthcare”. Even the most data-literate doctors cannot work with vast arrays of medical information without the right tools: advanced equipment and software capable of reliable data collection and protection.

Healthcare is a very specific aspect of life that is human-oriented at its core. Setting goals for digitalization and following fashionable big data trends is not enough. At every step, we must ask ourselves: “on the personal level, how is it going to impact people?”. It is the matters of life and health that we are discussing here, so ethical and moral issues must be at least as important as technological approaches to healthcare modernization.

About the Author

Rustam Gilfanov is an IT entrepreneur and a venture partner of the LongeVC fund.

To Top

Pin It on Pinterest

Share This