Healthcare and Insurance – Should we try to incentivise the ‘right’ behaviours through technology? Part 2/2
This is part 2 of the interview. Please find part 1 of the interview here.
Uli Kleber, TDI News Correspondent for Germany and Switzerland, got together with Professor Henriette Neumeyer and Lukas Naab recently to discuss developments in healthcare and insurance. Henriette is a medical doctor and professor in charge of the Master of Healthcare Management program at NORDAKADEMIE in Hamburg, Germany while Lukas is the co-founder of MINDS medical – a startup from Frankfurt, Germany that uses Artificial Intelligence and Machine Learning to streamline administration and processes in the insurance and healthcare system, e.g. hospitals.
Uli: Focus on patient care and treatment results are great points and I would actually like to go back to another point that you have mentioned: preventive care. To me this topic seems to start getting a bit more traction from the insurer side. However, while everybody agrees that preventing is better than curing, it still does not seem to have the importance that it should have. What is your view on this?
Henriette: Yes, if I look at my medical education, it was already geared very much towards our healthcare system. So, everybody learns that it would be better to prevent instead of curing, but the incentive structure is rather set up to the contrary. Take smoking cessation as an example: this has a tremendous effect on long-term patient health. But, if I am a doctor that would like to support my patient in this endeavor, I will not get paid for it. I could do it on a separate, privately-invoiced basis or as my hobby, but the healthcare system will not reward me for my efforts.
So, why not reward certain results, for example for achieving a behavioral change in my patients? I think this could have tremendous positive effects for the
whole system. But unfortunately, at the current point in time, the budget for prevention in our healthcare system is very limited.
Lukas: Yes, it all goes back to a patient view that has become more compartmentalised and less holistic. For example, I would like to add the transparency of patient history to this topic. In today’s fast-moving world, fewer and fewer people have a single doctor who oversees the whole patient history. Now, if you add the little time available that doctors have to establish their diagnosis, then you have an ever-growing number of people that move from specialist to specialist, sometimes getting treatments and drugs that have a countereffect on each other. This lack of transparency is a huge cost driver and, most importantly, detrimental to patient health because a certain drug might be used to treat your liver but has a negative effect on your heart.
We are not able to change the speed of change in our society and the time pressure on health care professionals, but what we can do, is to make better use of data to bridge that gap and reestablish a holistic patient view. To me this is also a form of prevention.
Uli: I think that the lack of data transparency is incredibly important. I have heard stories about doctors that had to use taxis to drive patient files from one hospital to another to get access to important patient history. Do you think that an electronic patient file would be the solution to all of this?
Henriette: I think that we need to have more technology that supports our lives and health in a very practical way. As an example, there is a digital test case for heart patients used in Mecklenburg-Vorpommern, Germany. These patients have a caretaker who constantly monitors the person’s situation through digital devices. If something is off, they will receive an alert and can directly see the patient’ situation. The important thing is that they will not only have the real time data but also have access to the full patient history. Of course, this takes quite a bit of work in the beginning to establish the full picture. Very often the challenge lies in the details, e.g. a patient might not take his drugs regularly because of certain side effects or some misunderstanding. So, there may also be some discussions around patient compliance, side effects and the like. But once patient and caretaker have taken those initial hurdles together, they are off to a great start with enormous long-term benefits. And in all these cases, the supporting use of data will be key. Particularly, if the patient is later entering the hospital for whichever reason and all this collected data is immediately available to find the optimal treatment plan.
In my view, there should be a lot more of these initiatives and we could also be quicker in creating the required digital infrastructure in Germany, but at least I can see that the intention is there to move in this direction. Also on the political level.
Lukas: To me one of the big challenges for this kind of initiatives, is the unstructured nature of data and compartmentalised access. Take other types of insurance, like life or disability insurance. For these, the initial risk assessment is very important to understand the type of risk they are taking on. But often the process is quite cumbersome and sometimes takes up to five weeks because data is not available or unstructured, that means distributed across hundreds of pages of patient records from different providers. Through the use of technology, like image recognition and machine learning this process can be accelerated and made much more accessible to clients and insurers alike.
Uli: Now, we get to an interesting point in the discussion: Henriette says that we need to collect much more data, while Lukas rightly indicates that the more data is collected by different providers, the harder it gets to turn it back into a coherent picture of the patient situation. It sounds like there is still a lot of ground work to be done in our data infrastructure.
Henriette: Well, a first step in the right direction would be data transparency. Let’s take Lukas’ example of insurance medical underwriting. Very often, the data exchange is taking place in the background and people do not even get to see the type of data that is exchanged about them. As a client, I would like to see this information first, so that I can take an informed decision on whether this may be a good moment to apply for an insurance or not. We do not have that type of transparency for the consumer at the moment and that does not really help to create trust in the process from a client perspective.
Lukas: I totally agree with your view on this point. The process is not transparent and it takes forever. We need instead to empower people with their own data, also from a data protection point of view. Right now, every doctor, every provider keeps the patient data in their own silos. We need to break up these silos and empower the patient. If people know what data has been collected about them and then can decide how they would like to use this data, we get to a point where they will be more than willing to share the data in their own best interest.
Uli: What types of use cases come to your mind with this type of patient data?
Lukas: Medical underwriting for insurers is definitely a hot topic as mentioned before. Instead of going through 200 pages of medical records, a machine learning-based system will quickly report back what types of risks are NOT present in the files at hand. Another added benefit is that these systems are learning systems that get better the more they are used. This means that over time you will be able to create fully digital online underwriting systems that will only need very little human intervention. We can see this already in less complex insurance products today, take car insurance as an example. Of course, you will still want and need specialist underwriters for highly complex cases but even they can be supported by technology like an underwriting dashboard that already delivers 80% of the data. Then the underwriter can concentrate on analysing the data instead of wasting their time on gathering it.
Another use case is diagnostics coding as it is taking place in hospitals today. This is also a very cumbersome process as data often has to be entered numerous times. An idea would be to leverage the doctor’s letter which is written at the time of patient exit from the hospital to prepopulate the hospital bill with the corresponding DRGs.
Uli: Thank you for mentioning the DRGs, Lukas. We often hear that there is a lot of criticism around the economisation of the healthcare system and the DRGs are at the center of the discussion. Some consider them a driver of unnecessary procedures that are only performed because they generate revenue for the hospitals or serve to justify in-patient stays. Do you see any alternatives to the current system or maybe how we could use technology to improve it?
Henriette: I think this is a classical principal-agent issue. Still, I would argue that whoever is working in the medical field did enter it with the goal of helping people and is interested in providing good care to their patients. But, of course, it is also good to provide some checks and balances. To me technology can play a role in this by reducing the administrative burden, for example by automating routine controls so that audits only concentrate on the cases where you really find important discrepancies.
Lukas: Actually, this is another great use case for an electronic patient file as it will allow you to access the patient history and procedures performed over time.
Also, looking at the current system, I feel that it is time to end the cold war between hospitals and the statutory health insurance providers. An enormous amount of resources goes into arguing about which procedures were necessary and what would be the right reimbursement amounts. Instead, I could image a system where you have a technology provider that serves as a smart system-based clearing house for hospital bills. By using artificial intelligence, you could establish fraud checks and economic thresholds that detect large deviations and then focus on these instead of arguing over nearly every single bill. I would love to run a proof of concept on such an approach – in my view the opportunity costs that could be saved for the whole system would be tremendous.
Henriette: Yes, and these findings could then also be used to improve the quality of care. For example, if you find that certain procedures are repetitively performed, but do not provide the expected outcomes in the patient histories over time. You could then play these learnings back to the healthcare providers and everybody would benefit.
Uli: So, it is again about gathering data and using technology to provide best practice approaches that help to improve the efficiency and quality of care in the system. Now, if we take all these findings and go back to our question from the beginning and close the loop: what should a hospital and care structure in Germany look like in the future?
Henriette: I would like to point out two main actions: Firstly, I think that we have learned, particularly through the current Corona crisis, that there is value in having a „strategic reserve“ in the system which allows to compensate for unforeseen situations. We know that digitalization can contribute positively to efficiency, but even in Spain, which is one of the foremost European countries in terms of digital healthcare approaches, we see that there is a certain limit to what can be achieved.
My second approach would be to recreate trust between the healthcare providers and the statutory health system that pays the bills. If trust helps us to get to a point where we can have a more constructive discussion, we could even come up with novel approaches. For example, by looking at which procedures are useful to be performed and which are not, we might even come up with an approach that may reward a provider for NOT performing a certain procedure. I think this would be a revolutionary outcome.
Uli: Yes, and while we talk about innovation, let us also look to the future. Corona has led to an important acceleration in the field of virtual collaboration and telework. What kind of use cases do you see for the application of telemedicine? How could this change the future of treatment?
Lukas: I think that initial patient triage and the guidance of patient flows through the system would be a great use case for telemedicine. This could make a lot of sense for non-emergency cases. I would like to share a personal example. When my wife was pregnant about a year ago, we had a case where she was not feeling well and we were debating on whether to go to the hospital. Then I remembered a German start up called ‘kinderhelden’ which has certified nurses on call. We called them up and they were able to give us a quick assessment which provided a lot of reassurance. The result was that we decided to stay at home and go see our regular gynecologist the next day.
I think that people often just need to get a quick assessment and maybe some reassurance that they are doing the right thing. If we now combine this quick response with some further digital means, like access to your fitbit or apple health data, then you might already be able to cover quite a couple of cases reducing people’s worries and also alleviate the pressure on hospital emergency rooms.
Henriette: The digital caretaker principle is definitely a model for the future of care and we need to see many more examples like the heart patient pilot case which I mentioned before. I would even go as far as making it a basic patient right to have access to such a service. Not only for emergency cases but as a part of the standard outpatient care system. There is a tremendous need for long-term caretaking in many different areas. Take strokes as an example. You might consider this to be a pure emergency business. But in fact, we see that even here, the long-term outcomes are all about aftercare and having someone who checks in regularly with the patient over time. Particularly in the first year after returning from rehabilitation, when patients are back at home or in a nursing home environment
This is also going to be a new career field in the making, which is the digitally-empowered caretaker. So, I hope to see a lot more initiatives in this field going forward and we definitely need the support from the political establishment to push regulation and support the rollout of the required digital infrastructure.
Uli: I think that this is really a great conclusion for our conversation in which we have covered a lot of ground. Starting with a look at efficiency, we quickly saw that the patient and quality of care need to be at the center of the discussion. Also, we have looked at some good use cases in which digitalisation can help to provide better patient care, particularly in the fields of prevention and in making medicine more patient-centric. Finally, we will need to move to a model that integrates medical support and advice more tightly into peoples’ everyday lives, like through the use of telemedicine or the digitally-enabled caretaker model that we have discussed. For this we need the required political support for regulation and the much-needed investment in the digital infrastructure.
Thank you, Henriette and Lukas, for your time and the great insights!