Ensuring equitable access to healthcare in the age of algorithms and AI

Yesterday, Dr. Peter Vaughan, chair of the board of directors of Canada Health Infoway, spoke at Longwoods’ Breakfast with the Chiefs.

After outlining the current state and future perspectives of digitization in healthcare, his main message was two-fold: 1. We are at risk of a “failure of imagination”, i.e. we cannot fathom all the possible futures that digital disruption might confront us with and hence fail to plan for their pitfalls adequately. 2. There is great potential for algorithms to be built in such a way as to solidify and deepen inequalities that currently exist in our system, and we need government oversight of such algorithms to prevent this from happening.

The first point is easy to understand, the second point may need little more explanation. Algorithms are used widely to determine what information is presented to us online, what choices are offered to us. We are all familiar with websites, offering us items we ‘might also like’, based on our past choices and based on what other purchasers have bought.

At a time when data from various sources can be linked to create sophisticated profiles of people, it would be easy for a healthcare organization to identify individuals that are potentially ‘high cost’ and to deny them service or to restrict access to services. Bias can creep into algorithms quickly. If people of a certain age, ethnic background or location are deemed to be ‘higher risk’ for some health issues or for unhealthy behaviours, and this is built into an algorithm that prioritizes ‘lower risk’ customers, then you are discriminated against if you share the same profile, no matter how you actually behave.

Discrimination is often systemic, unless a conscious effort is made to break the cycle of disadvantaged circumstances leading to failure to thrive leading to lower opportunity in the future. As Dr. Peter Vaughan pointed out, we in Canada value equitable access to healthcare, education and other public goods. We expect our government to put safeguards in place against discrimination based on background and circumstances. But how can this be done?

Private, for-profit enterprises have a right to segment their customers and offer different services to different tiers, based on their profitability or ‘life-time customer value’. Companies do this all the time, it is good business practice. But what about a private digital health service that accepts people with low risk profiles into their patient roster, but is unavailable to others, whose profile suggests they may need a lot of services down the line? Is this acceptable?

And if the government were to monitor and regulate algorithms related to the provision of public goods (such as healthcare) who has the right credentials to tackle this issue? People would be needed who understand data science – how algorithms are constructed and how AI feeds into them – and social sciences – to identify the assumptions underpinning the algorithms – and ethics. Since technology is moving very fast, we should have started training such people yesterday.

And how could algorithms be tested? Should this be part of some sort of an approval process? Can testing be done by individuals, relying on their expertise and judgement? Or could there be a more controlled way of assessing algorithms for their potential to disadvantage certain members of society? Or a potential for automation of this process?

I am thinking there may be an opportunity here to develop a standardized set of testing tools that algorithms could be subjected to. For example, one could create profiles that represent different groups in society and test-run them as fake applicants for this or that service.

Also, algorithms change all the time, so one would perhaps need to have a process of re-certification in place to ensure continued compliance with the rules.

And then, there would be the temptation for companies to game the system. So, if a standardized set of test cases were developed to test algorithms for social acceptability, companies may develop code to identify and ‘appease’ these test cases but continue discriminating against real applicants.

In any case, this could be an interesting and important new field for social scientists to go into. However, one must be willing to combines the ‘soft’ social sciences with ‘hard’ stats and IT skills and find the right learning venues to develop these skills.

Much food for thought. Thank you, Dr. Peter Vaughan!

Advertisements

How pharma may or may not win in a digital world

Healthcare is going digital at a rapid pace. A recent article by McKinsey & Company titled ‘How pharma can win in a digital world’ outlines emerging trends in digital health and how pharma needs to evolve to keep up with the times.

A number of predictions in this article are, I believe, misguided and reflect a common, but incorrect understanding of the potential of digital in health.

Digital health collage 2

Prediction # 1: “Patients are becoming more than just passive recipients of therapies”

Patients have certainly become more knowledgeable about their own health and about available therapies. And hopefully, health-related apps are helping people lead a healthier lifestyle and stay on top of their medical conditions and medications. However, patients have never been passive recipients of therapies. Patients have always had the choice of taking or not taking their pill, cutting it in half, skipping a dose, forgetting to take it, taking it with food when they are not supposed to etc.

Having served pharmaceutical clients for more than a decade, I have frequently observed that it is difficult for someone within the industry to understand that the medicine they are producing is not the be-all and end-all of a patient’s existence. Life is a busy thing. You work, you look after your family, you eat, you entertain yourself, and you may have a health problem that benefits from taking a medication. The act of taking a pill consumes a fraction of your time and attention. Medicines for health issues that are non-symptomatic may be forgotten because the patient does not feel sick. Medicines for chronic, life-threatening conditions may have suboptimal compliance because the patient would rather not be constantly reminded about his or her precarious situation. For acute conditions, compliance wanes as soon as the patient feels better. Side effects deter patients from taking their pill, etc. Compliance would not be such a huge unsolved problem for pharma if patients were ‘passive recipients of therapies’.

Prediction # 2: “Patients will be actively designing the therapeutic and treatment approaches for themselves with their physicians”

I have read this type of statement numerous times in articles about the future of pharma. Perhaps I am lacking understanding of what’s technologically possible nowadays, but for now let’s assume I have a pretty good handle on it. Designing a pharmaceutical product is an extremely specialized and complex process that involves scientists and labs. A chemical or biological compound with certain properties is created to address a specific health issue, and this compound cannot be easily customized. Rather, it is created and then subjected to rigorous testing, costing hundreds of thousands of dollars (or more), and if it does not hold up to scrutiny, then it’s back to the lab for more experiments and tweaking before another round of expensive testing resumes.

Physicians who spend most of their time in clinical practice do not design therapeutic and treatment approaches. They are merely the retailers of those approaches, acting as consultants to their patients and advising them which approach may be best suited for them. And patients will not be actively designing their own therapies unless they are experimenting with mixing pills and brewing up concoctions of their own invention (caution: don’t try this at home, kids!).

With substantially increased access to information patients can play a much more active role in selecting treatments, but they will not design them.

alchemist-300px

Prediction # 3: “Medicine will be personalized to address individual patients’ needs” (not in McKinsey article, but can be found in many other publications on digital health).

The move towards personalized medicine is certainly well underway. However, it does not mean that a therapy will be designed on the spot for the individual sitting in front of his or her physician. Again, the physician is the expert mechanic using existing wrenches and bolts to fix the car. The inventor who comes up with new wrenches and bolts does not deal directly with the customer whose car broke down.

Tools

Personalized medicines are medicines that target issues more precisely than was previously possible. While physicians used to set off a grenade to blast away your breast cancer, and half of your body as well, they now use a precision rifle that locks in on the malignant area and eliminates not much else. And depending on your genetic profile, there are different bullets that are most effective for your particular type of problem. So the array and precision of weaponry in the physician’s arsenal has increased vastly, and affordable genetic tests have contributed to better targeting of the weapons. But none of these things are designed on the spot, while you’re sitting in the examining room, nor will this be possible for a long, long time.

Explanation: Poor understanding of digital vs. physical contributes to common misperceptions

How do these misconceptions come about and why do smart people write these things?

The past five to ten years of our experience of living in a digital world have greatly impacted our beliefs in how easily things can get done and our feeling of agency. Want to customize your new car? Just click on the features that you want – sunroof, heated seats and the colour red – and you can get this exact model without any effort on your part. Select the perfect outfit? Choose the style, colour and size, and get it delivered to your doorstep the next day. Don’t like part of your video? Just delete and replace.

The ease of these digital experiences has gotten us into the mindset that things can be designed instantaneously, delivered rapidly and modified on the spot. We rarely think about the physical realities that enable our digital experiences. To give you the experience of ‘designing’ the perfect outfit for yourself, the maker has to come up with new styles to attract your desire, run efficient manufacturing to put the piece together with acceptable quality and at an affordable price, ensure the supply chain to enable the manufacturing, build in agility to adapt supply to demand quickly, and create a distribution system to bring the piece to you. All of these things are not done through click of a button, but through the hard work of setting up systems, negotiating agreements, fine-tuning machinery and materials and implementing physical processes.

It’s the same for pharmaceutical products. They are chemical compounds, after all.

Digital opportunities

However, the potential of digital solutions to transform the way we care for ourselves and the way healthcare is provided to us is undisputed. From life tracker apps that help you remember to take your pills on time to smart contact lenses that monitor blood glucose levels without pricking your finger to ingestible sensors that give you peace of mind that your schizophrenic brother has actually taken his medication, digital interfaces, algorithms and sensors can deliver great value to the patient.

The question is how this translates into business opportunities. Many people believe that pharmaceutical companies should transform themselves from being “a products-and-pills company to a solutions company” (see McKinsey & Company article). The idea is to not only provide medicine to the patient but also digital tools for monitoring of the patient’s condition, for communicating with the patient’s circle of care, for scheduling and reminders, for supporting rehabilitation after events and for outcomes tracking. From a patient perspective, this could certainly be a valuable offering. From a business perspective, the value proposition is less clear.

First of all, pharma companies do not typically have the expertise to develop digital solutions in house. Some form alliances with tech companies. Novartis and Google are developing smart contact lenses for people with diabetes and are scheduled to start trials this year. Otsuka and Proteus Digital Health have teamed up to embed a digital sensor into a schizophrenia medication to track compliance, and have submitted the first digitally enhanced new drug application to the FDA. J&J has set up a series of incubators and rewards startups for coming up with interesting ideas in digital health. Merck sponsors health hackathons.

What does the pharmaceutical company get out of this? Will physicians choose their medication over competitive products because it comes with a digital value add? Is the digital component just another cost factor that is necessary to stay competitive these days, or is there a revenue model somewhere? It seems that there is currently a climate of experimentation without a clear business model path ahead, not unlike many other areas of digital development.

In crowded markets with little product differentiation, it is possible that the companion app could become the deciding factor in recommending one drug over the other. However, it is hard to imagine that it would play any role if there were differences in efficacy or side effect profile between the compounds. A tricky little question is also what to do with patients who need to switch off one product and go to another. Should they be denied continued usage of the app?

To be truly solutions providers, pharma companies would need to be structured differently, around disease states, not around products. It would make more sense to form a company that is, say, a ‘cardiology broker’, offered great digital tools to manage a variety of cardiologic conditions and give patients access to the full gamut of cardiology drugs available. The sales reps for this company would not overtly or covertly ‘push’ one or two drugs, but they would advise physicians on what is new in the field and impartially discuss the merits of the different options. There are some attempts of pharma companies to become leaders in a therapeutic space and assume the role of expert provider – for example Roche or Novartis in oncology, where both companies have a large product portfolio. However, by and large, this type of business model does not apply to how pharma companies are organized and how they make money currently. It would be more applicable to private payors, and we see some organizations in the U.S. moving in this direction.

Low-hanging but sour fruit

The obvious area where digital tools can be used very effectively to drive engagement is patient-related. Arguably, a more engaged patient will likely be more compliant and stay on therapy longer, resulting in immediate benefit to the bottom line.

However, while many companies try to be patient-centric, any direct engagement with a patient carries the risk of an adverse event report with it. While adverse event reporting systems have been set up to keep patients from harm, unfortunately, reporting requirements are ridiculously broad. Nobody is keen on generating massive amounts of adverse event reports for their drugs. So digital engagement of patients has to be done with all sorts of caveats to reduce the risk of learning about an adverse event. Some companies stay away from direct engagement with patients altogether for that reason; others have taken the plunge and struggle to come up with creative ways around the problem.

Another challenge in engaging with patients through digital tools and platforms is finding appropriate engagement formats for particular audiences. A platform that has been designed to help kids with pain through gamified challenges and ‘levels’ may not be the right approach to engage a 70-year old cancer patient. Very little testing and research has been done to date to find out what tools best support patients with certain conditions. The key here is to be open to a multi-platform approach. While a game may be great at motivating one audience, a combination of text reminders and phone support may be best suited to keep another audience adherent to their treatment. Unfortunately, many of the vendors that design patient engagement tools on behalf of pharma are either all digital or not digital at all. What would be needed is a new type of vendor who can pull together various types of tools and customize them for a particular target patient population.

Low-hanging sweet, sweet fruit

One area where pharma could employ digital innovation easily and with sustained impact is in the way companies communicate with physicians. While almost everyone has switched to iPads for detailing over the past few years, pharma companies (in Canada, my home turf) still have limited understanding of how digital can be used to improve access and deliver value to physicians. Knowledge about different forms of digital engagement is lacking in marketing departments where people think Twitter and Instagram are for self-absorbed teenagers with too much time on their hands. Also, there is a feeling that digital is not important to the physicians who are core to the business. However, as one year after another go by and younger physicians become key opinion leaders and high prescribers, companies may find that they have missed the boat in establishing a digital rapport with these individuals.

Only recently have some companies started to think about conducting media audits and finding out from their core target how they use digital tools and what might be of value to them. Putting some effort and resources into understanding the myriad of different ways digital can be used, and physician preferences in this regard is relatively simple and will almost certainly have a payoff within a five-year timeframe. There will likely be some resistance from the sales folks who tend to see alternatives to face-to-face engagements as a threat to their position. However, I believe that the 21st century sales rep needs to be an expert in offline and online relationship building. Pharmaceutical companies need to figure out how to integrate different forms of digital and non-digital engagement optimally, and create internal structures and tools to maximize value for the customer.

Cartoon

McKinsey & Company article source:

http://www.mckinsey.com/insights/pharmaceuticals_and_medical_products/how_pharma_can_win_in_a_digital_world

Image sources:

‘Digital health collage’: Made the image myself

‘The alchemist’: https://openclipart.org/detail/222415/alchemist

‘Tools’: DeWalt DEWALT DWMT72163 118PC MECHANICS TOOL SET on http://toolguyd.com/dewalt-ratchets-sockets-mechanics-tool-sets/

‘Cartoon’: I’ve seen this cartoon on the web many times, but don’t know who made it originally. I’ve copied it from https://effectivesoftwaredesign.files.wordpress.com/2015/12/wheel.png?w=640

 

Wearable Tech and Health – not quite there yet?

Wearable tech is revolutionizing healthcare delivery – at least that’s what the pundits have been predicting for a number of years. The array of devices that are under development or already commercially available is stunning.

Looking only at devices that are used by patients on a day-to-day basis, there are three different areas of usage for wearable tech:

  1. Continuous monitoring of chronic illnesses

Many patients with chronic illnesses need to monitor certain biophysical parameters that indicate how well they are doing, if their medications are working and when an exacerbation of their condition would warrant visiting a healthcare professional.

Wearable devices that can sense and accurately measure heart rhythm, breathing rate, or blood glucose levels enable continuous tracking of critical markers and can help alert patients and their healthcare providers early to any arising problems.

  1. Improving the lives of people with disabilities

This is an area in which assistive devices have had a long history (think: hearing aids, wheel chairs, etc.). Digital sensor technology is now making devices more accurate, more personalized and more helpful.

Some examples of new technology that improves daily living include:

  • eSight Eyewear: A device consisting of a high-end camera, video processing software and processing unit and highest quality video OLED screens which project a real-time image that allow legally blind people to see.
  • Sensimat Systems: A series of pressure sensors that are placed under a wheel chair cushion. The sensors use a proprietary algorithm to monitor the seating pattern of the wheel chair user, and send a notification via smart phone when it is time to change position to minimize the risk of pressure sores.
  • TAPS Wearable: Velcro touch pads that can be worn on top of clothing or on the wheel chair. Each pad is a trigger for a smart phone app to play a pre-programmed phrase. This helps people who have difficult speaking (for example due to ALS or cerebral palsy) to communicate more easily.
  1. Recovery and rehabilitation devices

Also an area in which assistive devices have had their place for a long time, digital enhancements now tailor these types of wearables more to the patient’s needs. A number of companies are working on solutions to increase patients’ mobility – typically using some form of exoskeleton, together with sensors and algorithms to help with movement and recovery.

 

How do these new technologies fit into our healthcare system and how accessible will they be to patients who can benefit from their use?

Our healthcare system is already set up to evaluate new assistive devices, and potentially pay for them. Device makers would have to prove that their inventions are useful and enable patients to live more independently and / or return to work earlier and save or reduce disability payments or insurance costs.

Those who develop the wearables have to figure out which ones of the many institutions that share healthcare costs in our country they should approach to be considered for funding.

Funding is more difficult for wearable devices used in monitoring chronic illness. In most cases, there is no precedent for continuous patient monitoring.  Not only the patient’s engagement in the process is required, but a whole new infrastructure approach to healthcare is needed on the provider side. Currently, neither private practices nor hospitals are set up to receive, monitor and act upon myriad patient data coming in through wearable devices.

Many barriers impede adoption of new technologies for patient monitoring:

  • Concern about the reliability of incoming data – how accurate is the wrist-mounted heart monitor, are there differences between different devices and who is at fault if the device either gives a false positive and triggers an unnecessary medical intervention, or a false negative that puts the patient’s health at risk?
  • Integration with existing technology – how will data come into the clinic, will it be compatible with currently used IT solutions, how can staff easily access the data and how will confidentiality and privacy be safeguarded?
  • Integration into existing work flows – who will review the data, at what intervals, and which actions should follow particular cues? Will healthcare professionals need special training on how to read the data? Is extra staff required? How can incoming data be standardized to avoid confusion?
  • And last, but not least, who pays for the extra time that clinic staff spends on continuous patient monitoring?

Many of us are still in the phase of excitement over the wealth of possibilities that wearable tech affords us for delivering better healthcare. The successful players will be the ones who figure out how the possible can be turned into the doable, and profitable, within the constraints of our infrastructure and funding environment.

What should patients know about their disease?

A recent survey by the Dana-Farber Cancer Institute showed that many women who had surgery for their breast cancer did not know basic facts about their own tumour, such as their HER2 status, their ER status, their tumour stage and grade. @DanaFarber

The lead author of the study, Dr. Rachel Freedman, expressed surprise and concern at these results.

http://www.dana-farber.org/Newsroom/News-Releases/many-women-lack-basic-understanding-of-their-breast-cancer-study-finds.aspx

Why should patients know or care about these things? I have been immersed in the field of healthcare and pharmaceutical market research for more than ten years, but when it comes to my own health, I find I am not really interested in learning about the intricate details of my ailments.

Professionally, I often talk with physicians who make treatment decisions on life-threatening diseases and marketing executives from pharmaceutical companies who promote therapies to treat these diseases. Both of these groups focus a great deal on the minute differences between this drug and that, between different genetic mutations and cellular targets and markers that guide therapy choices. This knowledge is essential to their livelihood and their job. Should it be essential for patients as well?

Let’s draw an analogy to other areas of life.

When my car breaks down I bring it to the mechanic. What I am most interested in is: A) How serious is it? B) Is it fixable? C) How much will it cost to fix it and how much time will it take? I believe the mechanic to have all the knowledge about the different parts in my car and the tools that are needed to fix them, should they break down. I trust my mechanic to give me an honest assessment of the damage and the different options for repairing it (yes, I have an honest one). I don’t need to know anything else. Why should it be different for medicine?

DSC_0651

Or here’s another one:

When I decide whether to buy an Apple smartwatch or a Samsung Gear, I base my decision on things like what apps they offer, how the screen is laid out and which smart gestures they support. How the people at Apple or Samsung get the watch to have these features is irrelevant for me. All I care about is what I can and cannot do with the watch. And whether it looks cool. And how much it costs.

Samsung Gear S

And now back to medicine:

Is it fixable or will I die (soon)? How long do I need to take this therapy for? Will it make me nauseous? Does my insurance cover it? Will it prolong my life? How much of a hassle is it to take it? These are the types of questions that are relevant for a patient.

That’s hard to understand for someone who is focused on the science of medicine. But healthcare is more than science. Extending someone’s life or halting disease progression are only surrogate endpoints. The real endpoints are enabling patients to lead the life that they want, how they want it. These endpoints are, of course, immeasurable.

The Right to Choose

Recently, the CBC reported on a ten-year old girl, Makayla Sault, who was diagnosed with acute lymphoblastic leukemia, a cancer of the blood. The girl experienced severe side effects from the chemotherapy she was receiving and decided, together with her parents, to discontinue therapy.

Now, the Children’s Aid Society is getting involved with the intent of convincing the family to complete the course of therapy. The parents and the community they live in also fear that Makayla may be forcibly removed from home and given therapy against her will.

Do parents have the right to choose whether or not a potentially life-saving treatment is given to their child? Does the child have a say in this decision? Does the state have the responsibility to act in what is perceived as the best interest of the child, against the parents’ and the child’s wishes?

Let’s consider all the factors.

Deadliness of the disease. The form of leukemia that Makayla has can progress very rapidly and lead to death within a few months, if not treated. If it was a slower progressing disease, would it seem reasonable to let the parents decide on the course of treatment?

Chance of cure. If treated, there is a high chance of long-term remission or cure for a patient with acute lymphoblastic leukemia. Some quote the likelihood of treatment success as 75% or even higher. How does this information influence your view on the case? Does the parents’ decision seem responsible? Should the state step in?

What if treatment success was estimated at 50%? …at 30%? …at 10%? How would this influence your view on whose right it is to decide what to do? With a low chance of remission or cure, would it seem reasonable to allow parents to, basically, let their child die without having to go through the agony of chemotherapy?

Cultural context. The news coverage focused in on the point that Makayla and her parents belong to the New Credit First Nation, based in Ontario. Her family decided to try traditional remedies instead chemotherapy, and their local community has shown great support for their decision. The interference of the Children’s Aid Society is seen by some as another attempt of a government agency to take away native children, as had been done during the era of forced residential schooling.

How does the cultural factor influence your view on the case? Does being First Nation give Makayla’s parents more of a right to decide her destiny than being of Irish decent, being Jewish or being Iranian? What if the family belonged to a religious group that was viewed as being ‘extremist’? Would you feel the same about the case or different?

To what extent is the state responsible for the well-being of our children, and to ensure their well-being in the face of parental opposition? Laws and mechanisms to protect children against abusive parents certainly seem appropriate. How about protecting children against well-meaning but ill-informed parents? How about protecting children against well-meaning, well-informed parents who adhere to a different belief system? Difficult decisions.

On Chemotherapy

My friend had breast cancer twenty years ago. After surgery and radiation, she has been in remission and free of health complaints. Now that she is almost 80 years old, another lump was found in the same breast. She has access to excellent medical care, and she was given the following advice:

While the lump was small, she should have a mastectomy, just in case. As she was not eligible for another round of radiation therapy, it was recommended that she go through chemotherapy, because physiologically, she was deemed to be perhaps ten years younger than her actual age.

This is typical advice to elderly adults who are in good physical shape. It is based on the assumption that adding years to your life trumps every other consideration, provided you are in good health otherwise.

At age 80, how long does the average person have to live? Two years, five years, ten years tops? With all advances in medicine, we have not managed to extend life much beyond that. We are getting more and more people to reach their eighties, but we are not living significantly longer than that.

My friend was struggling with the advice she was given, and what was presented to her as the best medical solution. She did not want to lose her breast, and she did not want to put up with the side effects of chemotherapy. So she tried hard to find second and third opinions to support the view that it was not absolutely necessary to conduct a mastectomy and chemotherapy, and that she was not carelessly jeopardizing her chances of survival by refusing these options.

As we spoke about the different considerations that come into play in these decisions, she kept saying “I will get chemotherapy if I have to, but I’d rather not”. To me, this indicated an obligation that she felt to justify her choices to her friends and family, and perhaps to her medical advisors. Nobody wants to be seen as reckless with regard to one’s own health.

However, isn’t that bizarre and plain wrong? A person’s life is a person’s life, and having a lump in your breast is just one of many things you will consider and care about. So a woman who has carried two breasts through her entire life wants her life to end with both of them in place. So a woman who may have five, six or seven more years to live, probably deteriorating progressively as she ages, does not want to ruin her sense of well-being right now with an aggressive therapy regimen. So what?

Why is she made to feel guilty or irresponsible? Why does she need to justify her course of action?

It is understandable that friends and family of a person in this situation may advocate the more aggressive therapy, as they may fear losing this person. Fair enough.

However, it is my view that medical professionals should state very clearly that the choice is entirely up to the patient, that every choice has its trade-offs and that an informed choice, whatever it is, represents a responsible and acceptable way of managing one’s health. To have this sort of back-up from the medical community would make dealing with a difficult situation easier for the patient.