Faces of digital health

View Original

Why is it Difficult to Make a Business Case in Healthcare? (Karim Kershavjee)

  • Software development in healthcare is difficult because of the lack of returns on investment.

  • Healthcare technology has to work at the level of policy, technology and interoperability.

  • Apps shouldn’t try to trick people into changing their behavior.

Karim Karshavjee is the Program Director of the Masters of Health Informatics program at the University of Toronto. He is also the CEO of InfoCli, which is a Canadian primary care health IT Consultancy. In this discussion, he shared his experience with mHealth apps, interoperability and software design for healthcare.

mHealth apps are poorly used

Among his areas of research is the usability of mHealth apps. In the chapter on Designing Disease-Specific mHealth Apps for Clinical Value in the book Smart and Pervasive Healthcare, which he co-authored, authors write that while investments in mHealth is continues to grow, usage of mHealth apps continues to be low.  “Based on the data in 2017, the use of apps was especially low amongst patients with chronic disease who are most likely to benefit from their use [4]. While 38% of respondents with no health condition had downloaded 1–5 mHealth apps in a survey in 2017, only 6.6% with hypertension had done the same. “ The chapter offers an overview of the current usability and use of mHealth apps and what to take into account if you are developing one.

Approximately three years ago the count of mHealth apps stopped at around 350.000. It seems as if the counting stopped since. Additionally, we don’t hear as much about mHealth apps anymore, but about digital therapeutics or simply healthcare apps. The development of healthcare apps is still challenging.

As Karim Keshavjee et al. write in the article, mHealth apps need to be built on a well-researched behavior theory. There are as many as 89 different behavior theory models that have been developed over the last several decades. 

Behavior psychology is tricky

Understanding behavior is essential for creating useful solutions. But too many solutions try to be manipulative. “Many things I saw are trying to get you to do something that you are not naturally inclined to do or would want to do. And it's saying: you are wrong, you don't know how to behave property here. Let me show you how to behave. But, I don't want to really show you how to behave. Let me just trick you into behaving this way. And then all of a sudden you will exhibit the right behaviors and you will become healthier,” Karim Keshavjee observes.

How can we improve healthcare data management?

Karim Kershavjee is also a Family Physician with over 25 years of experience designing, developing, and implementing Electronic Health Records/Electronic Medical Records and helping clinicians use them effectively. He believes one of the key things we need to figure out in healthcare IT is how to tag and track data and how it moves around. “If you think about it, any bank, like a large bank, has many different branches, but all those branches belong to that bank. In healthcare we have many outlets; primary care clinics, we have hospitals, we have other outlets, but they are not under one umbrella. They all exist independently. And if we want to be able to share data between them we need to be able to move it safely. And that's going to require more sophisticated technology than we have right now.”

For technology to succeed, it has to succeed at three levels: technology, policy, interoperability, Kershavjee says. The policy level requires the solution to solve a valuable problem and have good economics. It also has to be something that policymakers understand and are able to make policy about. Then technology has to be interoperable with other systems. “The product in healthcare is knowledge. And if you don't have data to support that knowledge-making that system is going to be incomplete. And when it is incomplete, it will make incorrect recommendations. And if it makes incorrect recommendations Incorrect recommendations are very easily detectable by humans, by clinicians. And when they detect an incorrect recommendation, they will not trust that system and they will stop using it. So it has to work at the level of interoperability and then it also has to work at the level of the technology itself, usability, ability to connect healthcare providers and patients. So there's a lot of complexity in making an app work in our healthcare.”

This is only an excerpt. Tune in to the full discussion in Spotify or iTunes.


Questions addressed:

  • We’re going to talk a lot about healthcare IT today, so the first thing I’m going to ask you is your thoughts about healthcare in the metaverse? Many futuristic ideas are floating in the air, yet we more often than not live in a very stark reality: battling paper and fax machines, or manual transcriptions from one IT system to another. Interoperability is progressing very slowly. 

  • You have over 25 years of experience designing, developing, and implementing Electronic Health Records/Electronic Medical Records. Tell us more about that journey - EHRs from the past two decades are very infamous. :) How has the field advanced in your eyes? What’s changed, what has stayed the same in terms of the problems with usability and burnout? 

  • There is a common belief that regardless of where you go, western medicine will be the same. But the reality is that cultures and institutions have different guidelines. One doctor will prescribe you a painkiller, the other one won’t. Doctors work differently. What have you learned about how to create products applicable to a broad range of users? Is a high level of configurability inevitable? 

  • What do you see as the most difficult part of software development for healthcare?

  • You’re the CEO of InfoCli, which is a Canadian primary care health IT Consultancy, so you have a very deep understanding of primary care, you even architected Canada’s Primary Care Chronic Disease Surveillance System (www.cpcssn.ca). How would you describe the current state of primary care in Canada? In 2019, 14.5% of Canadians aged 12 and older (roughly 4.6 million people) reported that they did not have a regular health care provider they see or talk to when they need care or advice for their health. 

  • You co-authored a book titled Smart and Pervasive Healthcare, to which you contributed with a chapter on Designing Disease-Specific mHealth Apps for Clinical Value. The chapter offers an overview of the current usability and use of mHealth apps and what to take into account if you are developing one. To quote one of the findings -  while investments in mHealth is continues to grow, usage of mhealth apps continues to be low.  Based on the data in 2017, the use of apps was especially low amongst patients with chronic disease who are most likely to benefit from their use [4]. While 38% of respondents with no health condition had downloaded 1–5 mhealth apps in a survey in 2017, only 6.6% with hypertension had done the same. What are in your view some of the most important finding from this research?  

  • How can we bridge the gap between apps, patients and healthcare providers? For example, a patient can find and app, loves it, but stops using it because his healthcare provider doesn’t see the value in the app or can’t access data of the app? In an ideal scenario, the doctor will know and recommend an app. But it’s hard for physicians to know about all the novelties on the market, let alone know them well enough to make recommendations. 

  • How do you see patient data collection with sensors? In my observation wearables were hyped six years ago, then the excitement subsided for a short while with ideas about implantable sensors or IoT sensors in our environment, then with the Apple watch and sensors inside products primarily aimed for another purpose. 

  • In the book chapter, you mention that “mhealth apps need also to be built on a well-researched behavior theory. There are as many as 89 different behavior theory models that have been developed over the last several decade….” How can one go about behavior theory when there’s so many of them? 

  • What has work on technology for diabetes prevention taught you about behavior health and behavior theory? 

  • What would your advice be to healthcare providers regarding healthcare IT adoption and implementation?