Thoughts about Ambient Healthcare

Healthcare Today

Within the healthcare system, healthcare is viewed as a set of actions initiated to treat a specific illness. The treatment plan starts from initial diagnosis through a series of care steps, aiming to treat the illness or injury. In this system the process of care is the investment focus and the investment is based on tiers of prestige: provider (doctor), nurse, staff, then patients. The majority of technology investment (which includes time, money, and people) goes to the doctors' needs: personal customization of software, reducing the number of mouse clicks to perform an action, and adding increasingly niche capabilities. Less is given to the software tools used by the clinic staff for patient check-in, scrubbing doctors' schedules, and other administration tasks. Little investment is given to patient needs, though perhaps they rank above staff and nursing staff to form a distant second, third, and fourth.

My point is that within a healthcare system the provider and their execution of care is the primary focus of investment. Providers bring in money and have high prestige and power, therefore their needs have highest priority. This is economically unavoidable and entirely human: bureaucracies respond to incentives. I am not arguing this is in any way immoral, just that it is the truth. It was not until the 2009 ARRA law created Meaningful Use that most healthcare systems gave thought to patient-focused software and process investment. Meaningful Use stage I and II changed incentives and healthcare systems responded to the extent it made sense. Today most systems have a patient portal where patients can view their medical record, message their doctor, and book appointments.

However, I believe this is healthcare at the wrong end of care spectrum. The greatest gains to health and healthcare from technology will be realized by eliminating patient interactions with traditional healthcare system. Today the focus is on improving efficiency (and reducing errors) for the providers of care. Technology already existing today can diagnose and treat the vast majority of regular illnesses and problems in a person's life and greatly speed up the diagnosis and treatment of acute problems.

I would suggest that a large proportion of the ailments seen by doctors or nursing staff such as for the cold, flu, ear infection could be easily diagnosed by software. I also suggest software diagnosis will decrease the cost of healthcare, increase effectiveness and application of best practices, and create a better work environment for healthcare professionals.

The Founding Question

I pose this question: "A person realizes they have a cold four days after infection. How much earlier could technology diagnose it?"

Is the answer one day? Two days? More? What if the software could alert the human about the coming onset of symptoms and help the human deal with that intelligently. For example, to suggest calling in sick to work and suggesting palliative treatments of over-the-counter medicines or offering to arrange a video visit with a doctor. The primary thesis I will put forward in a series of blog posts is that we already have the sensors, computing power, and the software intelligence to notice things that humans health before a person could notice them.

The diagnosis capability of software and sensors can apply to many illnesses and injuries. Early diagnosis should be possible for recurring, easily diagnosed complaints such as cold, flu, pink eye, sinus infection, etc. Early diagnosis includes long-term problems that emerge slowly over time such as muscle and joint problems. The lead up to recognizing the need for knee replacement surgery is characterized by years of slow changes in a person's walking gait and also their level of mobility, the number of steps they take, speed of walking, and steadiness on their feet. The integration of software and sensors excel at something humans are not good: noticing tiny changes over a long time. To pose another question: "A person goes to the doctor when they notice walking has become regularly painful and unsteady and a knee surgery is needed. How much earlier could software notice the physical trends and alert a human?"

The starting point of this post was my frustration with the process of healthcare and the idea that most technology investment is made in the care stages of illness. I want to change this status quo and I think it is possible with the right framework of questions and practical thinking.

Introducing Ambient Healthcare

Ambient Healthcare is the current name I am using for the idea of using existing technology to achieve immediate improvements to personal healthcare and achieve a future vision of optimal health by melding the best of human and software abilities. I believe this is a new and innovative healthcare strategy. This is healthcare based on combining what software is good at with what humans are good at; joining the strengths of each side to create a much more intelligent robust and very early diagnosis and treatment process operating on a scale of days weeks months years ahead of the normal healthcare process.

The meaning of ambient in this context is the idea that we are already carrying a device with extensive sensor capabilities at all times: the mobile phone. The phone has gyroscopes, compass, microphone, data connection, and high-quality camera. It also has enormous storage and computing power. To date, in the context of healthcare this is generally used only to track a person's walking steps - a very basic count of things a person is doing throughout the day. I believe the full sensor suite can be combined to create a much more comprehensive view of how a person moves, talks, and breathes and then to notice changes. Sensing and noticing patterns is easy for software, so this is a core goal of ambient healthcare. However, the role of the software is to notice and notify the person and let a human take action and make decisions. The basic idea of Ambient Healthcare is: notice early with software, nudge the human so the human can act.