Another area where AI can be hugely helpful is in presenting data and helping patients make sense of it. After all, they are confronted with the same challenge as their doctors: quantities of specialized information that can be life saving, or just the opposite.  The patient’s informed consent is required in every aspect of care – from their contributing biological samples preparing for complex surgeries, to accepting a given course of treatment. How do we empower patients to make truly informed decisions, while allowing developers access to the streams of data? 


Respect for a patient’s individual autonomy is an established principle in modern medicine. In the past half century, the concept of autonomy has promoted patients from passive recipients of care to partners in planning their own treatments.

The notion of patient empowerment is reflected by developments in regulation and guidance. Think of phrases such as “patient led care,” “patient engagement” and “shared treatment decision-making”. 

The new strategy of patient involvement in health services is emerging in conjunction with initiatives for deepening inter-professional collaboration, all with the further aim of providing team-based care. Policy-makers envision some non-physician providers’ expanding scope of practice going hand-in-hand with this increased focus on patient-centered care. Can AI help deepen the trust between patients and the professionals in their care circle?

Research shows that high-performing teams are not built piecemeal. They achieve superior levels of cooperation because their members trust one another. Trust and a strong sense of group identity build confidence in their effectiveness as a team. In other words, such teams possess high levels of group EI (emotional intelligence).

The right amount of information, shared at the right time, can promote significant improvement in patients’ ability to choose the right treatments.

In a randomized controlled trial at two hospitals in Boston researchers studied the impact of a video-based decisions support tool for patients in the hospital. Decisions about cardiopulmonary resuscitation (CPR) and intubation are core parts of advance care planning, particularly for seriously ill, hospitalized patients. However, these discussions are often avoided. Seriously ill patients who viewed a video about CPR and intubation were less likely to want these treatments. Better informed about their options, they gave orders to forgo CPR/ intubation, and discussed their preferences with providers. This study brings into question what represents Informed Consent in the emerging world of AI and Big data? (El-Jawahri, et al., 2015)

Anonymizing data is a recognized pre-condition for its collection. It allows access to health data without compromising the patient’s right to privacy or security. Sayo, founder of Self-Care Catalysts, makes a compelling case for codifying trust between the givers and the gatherers of data: patients who give their health data should have greater access to it themselves. They should be able to track it, and see how their data is being used. There should also be an available, clearly delineated process of opting out of data sharing. (Nuffield Council on Bioethics, 2015)
Researchers in the United Kingdom have tested the concept of Dynamic Consent. (Spencer, K. et al., 2016 ). Here, information technology is used to determine just what patients are consenting to share. A succession of digital interface screens is presented to the patient. Information and choices about data gathering, and its potential uses, are delineated. Patients are thus enabled to tailor consent according to their own preferences. 
However, this concept is bound to come into conflict with the demand from AI for access to large pools of patient data. This is useful in training AI’s predictive function, which is not always related to a specific patient’s care. These secondary uses of identifiable patients’ data, (eg. to develop commercial products,) represent unchartered waters for public payers, patient advocates and commercial vendors. Google’s company, DeepMind, launched several projects with large health systems that bring out important questions about the commercial exploitation of individuals’ data and publicly funded health systems. These systems are fertile ground for “schooling” AI programs owned by powerful global corporations. (, 2016)
There is great potential to be realized by combining data with analytics, and technology with expertise. Many ills can be more efficiently cured, once value is understood and paid for. However, a patient centric, value-defining framework – one capable of informing data-sharing in the new healthcare sphere – does not yet exist. 

The Office of the National Coordinator for Health Information Technology, United States Government Health and Human Services (ONC) has identified patient-generated health data (PGHD) as an important issue for advancing patient engagement. It has initiated a series of activities to gain more information about PGHD’s value, and the various approaches to its implementation. The PGHD policy framework project is integral to Stage 3 of the Meaningful Use Rule, and the U.S. administration’s Precision Medicine Initiative. (, 2016)


AuthorVeronika Litinski

Where can AI improve health services? The short answer is, wherever Big Data lives. Policy makers and healthcare administrators are grappling with the recent emergence of Big Data in healthcare. These can include large linked data (from electronic patient records,) streams of real-time geo-located health data (collected by personal wearable devices, etc.) and open data (from shared datasets.) Together these form Big Data, a realm rich in new research opportunities and avenues for commercial exploitation. (Kostkova, 2015)

AI In Health Systems

It is nearly impossible for doctors to stay abreast of all the new and changing rules governing their fields, on top of the constant innovations taking place therein.

In the paper, Analysis of Questions Asked by Family Doctors Regarding Patient Care, (Ely, J. W.  et al. 1999) observed 103 physicians over one workday. Those physicians asked 1,101 clinical questions during the day. The majority of those questions (64%) were never answered. Among questions that did get answered, the physicians spent less than two minutes looking for their answers.


Obviously, providing quick answers to clinical questions will always improve the quality of healthcare. No wonder the Chief Health Officer at IBM Corporation, Rhee Kyi, a physician earlier in his career, recognizes the role IBM Watson will play in healthcare delivery. Watson, and other commercial solutions, promise to provide insights, reveal patterns and relationships across data sets. The allure of Watson lies in its being designed to work with unstructured data, such as genetic data and the free text portions of electronic health records.


The expectation is that research on large, shared medical datasets will provide radically new pathways for improving health systems as well as individual care. Facilitating personalized or “stratified medicine,” such open data can shed light on causes of disease, and the effects of treatment.  


Some of the most powerful applications can be found in Public Health, where data sets from communities of practice, social networks, and wearable devices can be mined for a wide spectrum of public health monitoring, and launch of persuasive technologies for public health interventions.  However, these fascinating opportunities develop against a backdrop of decades of under-investment in public health systems, which lack the resources to tackle the full range of health threats, from potential chemical or biological attacks, to serious chronic disease epidemics, or emerging infectious diseases like Zika. (Trust For America’s Health, 2016)


The allure of analytics here is obvious. It can help health systems crunch data to improve care quality and reduce costs, especially for organizations that aim to profit under shared savings or financial risk contracts. How can we balance cost, value and liability in a regulated healthcare industry?



AuthorVeronika Litinski