Digitizing Healthcare, Because Our Lives Matter

This article originally appeared on IBM Big Data & Analytics Hub.

The United States spends around 17-18% of its GDP on healthcare every year. When you put this into dollar numbers, it is a mind-boggling $2.9 trillion.

Unfortunately, that spending will grow at a faster rate now due to baby boomers becoming an aging population, and they are the largest demographic in the U.S. (Baby boomers are about 76 million, which accounts for 25% of the population of the U.S.). The healthcare related spending is expected to grow at a faster pace than the under 5% annual rate it grew over the last decade.

Unless the U.S. gets this spiraling healthcare spending under control, in a few short years we will be spending almost 25% of our entire GDP in healthcare trying to fix people’s failing health, instead of spending it somewhere else where it is desperately needed. Obviously, we can’t stop the aging population, but we can make the healthcare system more efficient. Overall, chronic diseases account for about 86% of the health care spending in USA. Severe chronic conditions such as heart disease, arthritis, asthma and diabetes alone cost 33% of the total spending.

In the healthcare system, “mass medicine” without individual diagnosis is not a solution unless there is an epidemic that needs to be localized and cured. The physician is the authoritative person to give us the personalized medicine. But the unfortunate fact is that wait times to see doctors are getting longer, and the typical doctor/patient visits are getting shorter – typically lasting for 30 minutes or less. Even during that short visit, the diagnosis is based on how well the patient tracks the symptoms and, more importantly, how effectively the symptoms are communicated to the physician. Most times, the physicians try to associate the symptoms based on the local happenings to figure out the root cause. While this works most of the time, with the increasing load, there are times the opportunity for right diagnosis might be missed in the first visit. This could lead to costly hospitalization later. Especially when you take into account the monitoring of the symptoms and the effective communication of that to the physician is based on the patient’s knowledge and self-monitoring the results can lead to subjective decision making. Most times, symptoms won’t surface when at doctor’s office, much the same way car problems never surface when at a repair shop :).

Solving this problem was not easy, until now. This is because we were faced with mostly closed down patient record systems, no remote monitoring devices that can effectively and accurately monitor the patient vitals, near impossible retrieval and return of information back into patient records, heavy government and insurance regulations on what can be monitored, collected, communicated and stored, etc.

In the past, digitization of patient records was just for record keeping purposes. But with the advances in IoT, predictive analytics, cognitive computing and the mighty APIs to connect them all together, things have changed dramatically.

While the goal is not to replace the doctor in the process, the “quantified self” concept will get the right information for future analysis, and it can also predict and prescribe a treatment in appropriate time to save lives in a much more cost-effective manner.

Predicting Patient Readmission via Remote Patient Monitoring: Readmission for patients with heart failures is very high in USA.

In a simple study conducted with 1095 patients, just two automated phone calls were made in the space of 30 days to check on the status of their health. Of those reported having a negative response trend, 37% were readmitted; whereas, only about 16% from positive trend and 14% from neutral trend were readmitted. The only problem with this test was that the self-monitoring and assessment were left to the patient which makes the results subjective.

But instead, a predictive analysis can identify patients at a higher risk for readmission even before discharge. This allows for the doctors, nurses and other caregivers to train the patients about managing their health and reporting the status back on a regular basis. An educated patient is the best patient to control his/her own destiny.

With the advances in wearable/mobile health technologies, health vitals such as heart rate, blood pressure, glucose levels, respiration rate, blood oxygen saturation levels, and even ECG patterns can all be monitored almost to the clinical grade and the information can be fed back into the patient records in real time. With that the physician-directed patient self-monitoring becomes a reality.

This takes the subjective nature out of the equation, and fairly accurately predicts the path towards readmission very early. This detection will allow the patients to be put on customized, proactive programs where they will be constantly monitored. They will be guided to participate in healthy lifestyles by doing behavioral modifications and constant in-person coaching will be provided as necessary. This eliminates the re-admission occurrences.

Telehealth for the veterans: We all know the fiasco the VA (Veterans Affairs) went through recently. But they seem to be doing the telehealth portion of it more effectively now. Over ½ million veterans get healthcare via telehealth home monitors every year. Most of these veterans live in remote areas and may not have access to healthcare as needed. According to Dr. Darkins, VA’s chief consultant for telehealth services, they can take digital pictures of retina, skin lesions and send them securely for analysis and recommended next steps. This service keeps about 40,000 people living at their homes instead of nursing homes or other assisted care facilities. A quarter of a million patients are screened for eye diseases without a need to over crowd eye clinics. According to him, the home telehealth program has reduced the hospital admission by over 30%.

Remote monitoring of patient’s health and safety: M2M technologies, a customer of IBM API Economy solutions, decided to take their security and comfort monitoring for the elderly, called Kizuna One, to the next level. Kizuna One, which collects data from wearable health devices to a repository, has decided to expose the information as APIs to their business partners to create new business opportunities. Essentially, keep the inflexible backend systems, but convert them into flexible and customized APIs to their business partners as their Chief Architect suggests. This allows them to scale up their business and IT systems without a very costly engagement. Those resold APIs allowed their business partners to create newer, modern and engaging digital apps and business models using the same baseline service that is offered by M2M.

To make all this happen, IBM is working with Mayo Clinic, MD Anderson cancer center, MSK cancer center, Apple, CVS, Epic and numerous other companies to make the above a reality in collecting the data from IoTs, getting that data securely into the existing patient medical record systems, analyzing the data using predictive analytics, integrating the data movement easily with existing restricted systems using APIs, and recommending smarter decisions using our Watson cognitive systems to move the healthcare systems from a siloed legacy systems into a mainstream data economy.

All of those changes mean going forward we can diagnose right the first time, predict right all the time, treat efficiently, monitor/follow-up effectively, and reduce costs along the way. It is the care we all deserve!

Reach out to me at @AndyThurai to find out more about the above usecases or how APIs are the foundational piece for accelerating digital healthcare.