Thursday, December 12, 2019

What is a Platform?

This month I'll deliver several keynote addresses.  In my presentations, I'll use terms such as platform enterprises, platform thinking, and platform strategy.  But what is a platform?

Is it just a collection of standards?   If so,  is a USB flash drive a platform, since I can transfer a file from my Chromebook to someone else's Macbook using it, in a low effort, low cost fashion?

Not exactly—in the USB example, there is no agreement about what file types are preferred, what data those files may contain and what security controls will be used to protect the integrity and privacy of the data.

In my view, a platform is a combination of technology (data standards, APIs, security controls), policy (who can do what for what purpose with what privacy controls), and process (what workflow is supported by what people and what automation).   In short, it is a way to use knowledge and technology to facilitate connections, and create value in the process.

For example, Unity Farm Sanctuary is entirely controlled with the Google Home platform. My locks, lights, thermostats, cameras, and mobile devices are linked via a set of APIs and security controls (OAuth). I can delegate rights to use selected devices for selected functions to authorized collaborators but the general public cannot gain access to my heat, light and power controls.     I use a combination of approaches to support device and workflow integration - Google Assistant routines, proprietary apps, and secure websites.     The end result is that I can monitor and manage the well being of 250 animals from my phone.

How does this apply to healthcare?

We know that the CMS Interoperability Proposed Rule and the ONC Information Blocking Proposed Rule  are likely to be finalized into active regulations. This means that in 2020 or 2021, hospitals and clinician offices will be required to exchange data via APIs to patient controlled apps. Increasingly healthcare must be a competent data business as well as an empathetic care delivery business. Supporting new regulations with point solutions will create a chaotic collection of heterogeneous user experiences and security vulnerabilities. Why not create a single, managed and well supported front door which enables quality, safety and efficiency solutions to be deployed more quickly?  That's Platform thinking.

The transition from a collection of products to a platform strategy is a journey.  How will technology services be delivered - via Google Cloud platform, Amazon Web Services, Azure, or other provider? How will access be granted, managed, and monitored? What workflows for what use cases will be supported and when? How will it be paid for? How will the effort be communicated so that all stakeholders understand privacy protections, ethical use of data, and possible participants in a platform ecosystem?

There are many questions to be answered while on the road to becoming a platform enterprise.  Over the past ten years, I've written nearly 2000 posts about Life as a CIO.   I feel the next 2000 posts about the platform transformation of healthcare, my dispatches from the digital health frontier,  will be even more important.

Wednesday, December 4, 2019

What's Next?

After nearly 25 years in Boston, I'm beginning a new journey at Mayo Clinic in the role of president, Mayo Clinic Platform.  Many colleagues have asked me about the transition.   

First, I have profound thanks for my mentors and collaborators in Boston.   I could easily fill an entire blog post with the names of hundreds of people who worked with me since 1996 on cloud services, mobile applications, machine learning, connected devices, and data standards.

Those innovations  made a positive impact on many people.    At Mayo, I believe I can scale the lessons learned in Boston to stakeholders around the world.  How?

As an adviser to many startups, incubators, and accelerators around the world, I've experienced the barriers and enablers to innovation.    Challenges include lack of standardized technology (APIs with sufficient data granularity and workflow integration),   policies (templates for security, privacy, risk analysis, ethical use of data, and communication),   and people (sufficient staffing to run pilots and focus on collaborators).  Launching a pilot can take 6 months just to work through approval processes.     Sometimes academic medical centers can take as long as 18 months to formalize a proof of concept project.   What if a Platform of technology, policies and people were able to radically shorten the time to evaluate emerging companies and created an "innovation factory"  for  collaboration?    That's how I think about the Mayo Clinic Platform opportunity.

Although I've visited Mayo many times, I've just scratched the surface in my understanding of the culture, capabilities, and colleagues.  A good way for anyone to understand what makes Mayo uniquely Mayo is to watch the Ken Burns documentary

How will I spend my first 100 days at Mayo?   Although I do not begin the role until January 1, 2020, I'm taking personal time off to volunteer at Mayo during December, meeting dozens of stakeholders at every level of the organization.   I need to listen to their hopes and needs for Mayo Clinic Platform projects.    There is a remarkable interim team leading the Platform today and they will guide me through the work to date, the critical decisions ahead, and the refinement of the strategic plan.    Together, we'll advance the strategy, structure, and staffing for the Platform.   Outstanding support teams in legal, development, compliance, IT, and public affairs will help.    I'll speak about our early decisions at JP Morgan and HIMSS.     Mayo CEO, Gianrico Farrugia,  will keynote the HIMSS conference .

Personally, I will live in an apartment in Rochester, Minnesota from Monday-Thursday, then return to Unity Farm Sanctuary for weekend animal care and farm maintenance.     The flights are easy (2.5-3 hours, 3 times per day on Jet Blue), and even with weekly commuting I may actually travel less in 2020 than in 2019 (400,000 miles in 40 countries).

During times of great challenge and change, I've blogged on a daily basis, sharing my successes and failures transparently with government, academia and industry colleagues.  Recently I've focused on writing articles and books.   The pace of the Mayo Clinic Platform effort necessitates frequent blog posts.  I'm renaming my blog from "Life as a CIO"  to "Dispatches from the Digital Health Frontier".  My hope is those dispatches will help others with the path forward, following the best trails and avoiding pitfalls. 

Am I excited by the work ahead?  Most definitely.    Am I daunted by the responsibility and accountability of shaping the future of Mayo's digital businesses?   Of course.   I call this, excited anxiety.   During those stages of life when there is a perfect storm for innovation , I find that a little adrenaline really maximizes focus, especially when you're not sure where the path ahead will lead.   As my colleague Prof. Yitshak Kreiss, M.D., Director General, Sheba Medical Center  told me "innovation is when you have an urgency to change but don't know exactly how to get there.   If you know where you're going, it's just implementation not innovation."

With today's Mayo announcement, I believe the next perfect storm is beginning and I'm ready.   And I'll make sure my readers have a front row seat.

Thursday, February 7, 2019

Remote Patient Monitoring and Self-Responsibility

At HIMSS next week, I'll be doing 5 presentations about the future of healthcare IT, focusing on patient directed data exchange, internet of things, and telemedicine.     Remote patient monitoring,  which combines all three, will be increasingly important.

Remote patient monitoring can take numerous forms,  and the evidence supporting these tools is mixed. Here’s another excerpt from our new book—The Transformative Power of Mobile Medicine—co-authored by Paul Cerrato that dives into the issues.  For those interested in reading the entire book, the publisher is offering a deep discount until March 31, 2019; coupon code: HIMSS2019.

Many thought leaders are convinced that remote patient monitoring improves patient care, but surveys suggest that health-care professionals are still not convinced. An analysis from the New England Journal of Medicine Catalyst Insights Council asked respondents to rate various patient engagement initiatives. “Remotely monitoring using wireless devices/wearable” was listed as the least effective way to engage patients while having physicians, nurses, or other clinicians spend more time with patients was listed as the most effective tactic. [1] There is also uncertainty about the benefits of remote patient monitoring in the scientific literature. Of course, remote patient monitoring can take so many different forms that it’s impossible to make a blanket statement about its effectiveness. But a randomized controlled trial (RCT) that included more than 1400 patients (median age 73 years) who had been hospitalized for heart failure generated less than encouraging results. Michael Ong, MD, from the University of California, Los Angeles, and his colleagues divided the group into an intervention arm, which received health coaching phone calls along with the collection of vital signs that included blood pressure, heart rates, symptoms, and weight with the help of electronic devices, and a control arm that received the usual care [2]. Ong et al. couldn’t find any significant difference in hospital readmission rates 180 days after discharge for any cause: 50.8% were readmitted despite having all the extra attention and access to all the high-tech monitoring devices versus 49.2% in the usual care arm. Similarly, the investigators detected no difference in 30-day readmission or 180-day mortality. The experimental group did, however, report better quality of life at 180 days.

 On the other hand, Essentia Health, a Minnesota-based system that includes 16 hospitals and 68 clinics, has been using home telemonitoring with a body weight scale to keep track of CHF patients. Patients weigh themselves every morning and answer a few basic questions about their symptoms. Their responses are transmitted via telephone line to the computers that triage the incoming data and alert clinicians to those in need of additional attention. Essentia has been able to reduce 30-day readmission rates to less than 2% with the program. The average readmission rate for CHF patients
is 25%. [3-4]

Detecting statistically and clinical significant benefits for remote patient monitoring is complicated. Unfortunately, Americans are used to being passive recipients of health care. When they see their physician, they expect to receive a pill or have a procedure performed. The only demand on their time and attention is taking the pill or undergoing the operation. Asking patients to take on a more active role in their care, including weighing themselves daily, taking blood pressure readings, and so on, requires a stronger sense of self-responsibility and better cognitive skills. It also requires a deep, long-term commitment from the health-care care organization launching the program. Providers cannot expect to “patch” a remote patient monitoring system into the mix without a great deal of planning and commitment from physician leaders and clinicians in the trenches. Tracy Walsh, a senior consultant with the Advisory Board, sums up the issue succinctly: providers need to “track program metrics that closely map to the organization’s broader strategic objectives.” [5]. Walsh provides a detailed graphic to help providers choose wisely. It addresses three basic questions regarding remote patient monitoring:

·     Is it technically feasible?

·     Is it clinically relevant?

·     Is it cost-effective?


1.     Volpp KG, Mohta NS. Patient engagement survey: improved engagement leads to better outcomes, but better tools are needed. NEJM Catal May 12, 2016;. Available from: https://catalyst.

2.     Ong MK, Romano PS, Edgington S, et al. Effectiveness of remote patient monitoring after discharge of hospitalized patients with heart failure: the Better Effectiveness After Transition_Heart Failure (BEAT-HF) randomized clinical trial. JAMA Intern Med. 2016;176:310-18.

3.     Siwicki B. Essentia Health slashes readmissions with population health initiative, telehealth. Healthcare IT News. March 15, 2016.

4.     Agency for Healthcare Research and Quality. Heart failure disease management improves outcomes and reduces costs. ,

5.     Walsh T. Studies are conflicted about remote patient monitoring—here's what we think. Advisory Board March 31, 2016;. Available from:

Tuesday, January 29, 2019

What's Next for Healthcare?

Today I had the honor of keynoting at a major announcement from Apple, Aetna and CVS/Caremark.   I'd summarize the message as "Digital Health has arrived and is now mainstream, fully embraced by the major stakeholders in the healthcare ecosystem."

Here's the upshot.

The history of healthcare and payment has been risk pool based and focused on paying for sickness.

The future belongs to wellness, personalization, and a team-based approach to keeping you healthy in your home. 

The announcement of Aetna Attain, a collaboration of Aetna, CVS/Caremark, and Apple, is based on a few key ideas:

1.  Personalization is essential
Asking everyone to walk 10,000 steps a day is unrealistic.   Some should walk 5000 and some should walk 20,000.    Performance goals should be set based on past performance, current health state, and patients like you.

2.  Privacy is foundational
Managing an ecosystem of apps, interfaces, and cloud services requires that privacy is designed in from the ground up.   Everything around Attain is based on opt-in, patient controlled privacy preferences, de-identification, continuous authentication (how are you holding your phone now), and aggregation of information.

3.  Machine learning based on patients like you is empowering
Randomized controlled trials are great but impossible to personalize.  Ideally, the past data from millions of patients like you will inform your wellness strategies for the future.

4.  Incentives/alerts are important to motivate wellness
We're all busy people.   There are many tasks competing for our attention.   If we're reminded at just the right time, maybe we'll be more likely to follow through on positive behaviors.   If we're given an incentive payment (or even compelling, avoiding a penalty) for achieving our goals, we're more likely to focus.

5.   A combination of historical medical experience plus telemetry is powerful
Raw data in the absence of context is rarely helpful.   A few years ago I had a debate with a major industry executive who argued all healthare data is "atomic" i.e. all we need to know is systolic blood pressure then we can act.  I countered that we need to know history, current symptoms, medications, how the blood pressure was taken, and patient care preferences before action can be taken.   The Attain approach is to combine all these factors before suggesting wellness behaviors.

As I said in my keynote today, I'm very optimistic about 2019.   Attain is a great example of  emerging technology convergence - internet of things, cloud, machine learning, telecare, and security coming together.    The industry will learn a great deal from the Attain project about patient generated healthcare data, patient data stewardship and alignment of incentives for wellness.

I'm glad to be part of the exploration.

Tuesday, January 22, 2019

Embracing ChromeOS (and the cloud mindset required)

2019 will bring many changes.   After more than 20 years as a Chief Information Officer, I will pivot to lead innovation as part of the senior leadership team for the newly merged Beth Israel Lahey Health on March 1, 2019.
Here's the Boston Business Journal article about it.

My innovation role will focus on 5 areas:

A front door/liaison to government/industry/academia for digital health collaboration at Beth Israel Lahey Health

Exploring new technologies, especially those arising from outside healthcare, to assess their role in provider/patient/payer workflow

Mentoring startups and internal faculty seeking to create new products/services especially in the areas of machine learning, mobile, telecare, internet of things and blockchain

Lecturing/writing to broadly disseminate lessons learned about innovation

Hosting of international visiting groups from around the world which want to learn about our innovation efforts.

In many ways, this next step combines the best of my youthful experiences researching and writing in the early 1980's Silicon Valley with my 30 years of experience as an IT leader and professor.    I look forward to it.

In the early 1980's while writing for Infoworld, I had the opportunity to personally evaluate emerging products - the IBM PC, Compaq's portable, Wordstar, Microsoft compilers, and various dial up modem services.  Nearly 40 years after evaluating the first PCs, I'm now evaluating Chromebooks and ChromeOS as the next frontier in personal computing.   Here's the experience thus far.

Think of ChromeOS as not just an operating system replacement for Windows or MacOS/ioS but an entirely new approach to computing.   It's essentially a cloud viewer, consuming data and services available on the internet combined with limited offline replication of data just in case the internet is not available.    What does that mean?  Instead of using a local file system to store my documents, media, and data, I'm using Google Drive.   If my device is lost, stolen or damaged, there is nothing to hack on the device.    My Chrome applications are all web services with nothing running on the local Chromebook - Gmail, Gsuite, and Outlook Web Access (or Office 365).   

Why is this a useful concept?    A few months ago, I was in Tel Aviv and I damaged my computer.  There was no way to replace/repair it easily.   If I had been using a Chromebook, I could buy any $200-300 Chromebook and immediately have access to all my data and services.

The Pixelbook supports Android apps in addition to ChromeOS, so I can run local software with local data if I chose.   I've added a few such apps such as Nest (cameras, smart home controls), Gmail offline (local replication of email), and Weatherlink (to control my weather station), but I really don't need them.   The combination of my phone for apps and a Chromebook for cloud hosted services works extremely well.

Are there downsides?   Absolutely.

Cloud/web native apps may not be as sophisticated or usable as locally installed apps.   Moving from local storage to cloud storage requires some planning and adaption.    I installed sync software on my previous computer and synced all my files into Google drive so cloud migration was one step.    I set up offline files so that every document I edit in the cloud is replicated into offline storage on my Chromebook for easy access when I'm on an airplane without wifi.     I previously managed photos and media on devices but now I manage them in the cloud.    All of this is change and requires getting used to.

For my use case - productivity applications, email, media management - a Chromebook works perfectly well.    I imagine there may be tasks/high intensity computing  use cases for which the cloud application and file system approach may not be optimal.   But for me, it works.

Truly, about the only thing I would like to see improved is that the Beth Israel Deaconess version of Outlook Web Access (2013) is not as full featured as Gmail or Office 365.    Once we upgrade or migrate, then the final piece of my cloud-based computing environment will fall into place.

I know that I may be edgy by suggesting that thin client, cloud-centric computing is the future, but from a security, cost, and maintainability perspective, it certainly seems like the right direction to me.

Thursday, January 17, 2019

Exploring the Connected Medical Home

Over the past few weeks, I've been exploring the combination of internet of things, artificial intelligence, and ambient listening with a focus on how these technologies might improve care management, patient/family navigation of the health system, and wellness.

Google, Apple, and Amazon all have ecosystems that include the functionality I'm writing about.  Purely because I'm spending January investigating the Android/ChromeOS environment, my first exploration has been with Google products.    I'll explore Apple next.

Here's my test bed:

Unity Farm Sanctuary heating and cooling is controlled by Nest Thermostats.

The animal areas are streamed from Nest Cameras.

In the living room, I've installed a Google Home Hub, a Chromecast Generation 3, and Chromecast audio (now discontinued).

I'm currently carrying a Google Pixel 3 phone running Android Pie.   My personal computing platform this month is a Google Pixelbook.

What does this infrastructure enable me to do?

Here's a sample dialog (Google product responses are in italics)

Ok Google, what is the temperature in the living room?

The farm living room is at 66 degrees.

Ok Google, increase the living room temperature to 68 degrees

The farm living room temperature has been increased to 68 degrees.

Ok Google, show the cows on TV

Showing the Sanctuary Cows on the Farm Living Room TV

Ok Google, play music on the speakers

Playing selections from Google music on the farm kitchen speakers

Ok Google, pause.

Music paused.

Ok Google, call Mom

Calling Mom

This internet of things, artificial intelligence, and ambient listening example illustrates the many possibilities for any internet connected home.    As 5G cell phone technology is deployed in 2020, gigabit internet will exist over the air throughout the country - no wires/fiber needed.   The potential is only limited by our imagination.

What exactly is the potential?   Google Home enables the definition of routines - a kind of macro that links commands together, including "if this, then that" kinds of controls.

For example

Ok Google, Good Morning

Good Morning John

It's 22 degrees outside and today will be dry with a high of 32 degrees.

Your commute to work today will take 47 minutes because traffic is heavy

Increasing the temperature of the kitchen and decreasing the temperature of the bedroom

Playing morning music

Ok Google, Good Night

Turning off the lights

Activating the security system

Reducing the temperature of the kitchen and increasing the temperature of the bedroom

Sleep well

All of this has worked so well, that it makes me believe the future of computing is not limited to phones and apps, but increasingly a voice driven integrated ecosystem that requires very little technical expertise to use.

The artificial intelligence components can be startling.

Ok Google, play NCIS on TV

NCIS is available on CBS All Access and The CW

CBS All Access

Playing NCIS from CBS All Access on the Farm Living Room TV.  I will play from CBS All Access next time you ask for NCIS.

This required identifying that NCIS is a television program, offered by certain vendors, each with a different way to play it.   Notice also that I was vague about which TV to use, but Google Home figured it out.

Fast forward to healthcare. As we think about the integration of wearables and other in home wellness devices, this technology can integrate devices, routines, and voice commands to measure activities of daily living, suggest healthy behaviors, evaluate compliance with care plans, and communicate with care teams.

Today, only those with technological literacy and dexterity can perform these functions with a collection of apps, but it takes diligence, planning, and a steep learning curve.

Assembling all the connected home functions described above was done by me, in a few minutes, by unboxing devices and doing minimal setup - a one time only event.

As we move from fee for service to value-based purchasing, reimbursement reform will align incentives for wellness in the home rather than the treatment of sickness.  I predict that healthcare delivery organizations will restructure themselves for success  by shifting work from building more hospital beds to empowering patients outside of the hospital.  This will require training clinicians in  telemedicine  (let's call the new medical specialty "virtualists"), home support people (both visiting nurses and connected home technicians), and care managers who ensure all services are coordinated to maximize quality while reducing total medical expense.

To me, this tech works so well, and is so affordable (compared to treating sickness), that the future of the connected medical home looks very bright.

Tuesday, January 15, 2019

Dispatch from India

I spent last week in Bihar, an area of Northern India near Nepal.  The best way to describe the journey is in pictures.

Our small team visited villages along the Ganges to the east of Patna, tracing the path of patients from seeking care to diagnosis to treatment to compliance to wellness.   We met with patients, providers, field officers (think of them as care managers), chemists (pharmacists), and labs.    Here's what we experienced:

The villages had hand pumped water supplies, electricity and 4G cellular connections.  Cows and goats were a part of many households.

A unique telemedicine program from World Health Partners (WHP) provided access to experts, connecting each village to trained clinicians in urban areas.    We participated in such a consultation.

We visited patients in their homes to hear their stories.  All of us were touched by Pooja, a 25 year old new mother who spent 70,000 rupees (about $1000) on unnecessary medical care due to a misdiagnosis.   She had to sell her land and her cow to pay for healthcare.   We've started a go fund me to help rebuild her life.

We reviewed medical records and imaging studies, which in India are maintained by patients and families.  In this photo, I'm reviewing the records of a TB patient who is feeling better after treatment, but appears to have a negative initial chest X-ray.

We visited a local lab which offered a menu of diagnostic tests ranging in price from $.70 to $14.00. Diagnostics included GeneExpert TB testing and 3D doppler ultrasound.   All lab data is manually recorded on paper and carried by the patient.

Local chemists make available a range of medications at very low cost.

Medical record keeping is done via a brief note which is the property of the patient.  Prescriptions are often abbreviated in a way that can be hard to decipher but a local chemist can understand.   Registry data is entered for tuberculosis and is one of the few electronic workflows, completed on low cost android phones by provider support staff.

I came away with a better understanding of the cultural, political, and clinical workflow in the state of Bihar.   Next steps will be designing the digital health services which are most likely to serve the stakeholders, now that we have experience with their requirements and constraints.  We'll do everything possible to leverage the remarkable national cloud hosted services available in India including identity management via Aadhaar, payments via UPI  and the rest of the "India Stack".

As I wrote last week, the next 30 years of my life will be dedicated to purposeful causes that I hope
will make a difference.    India and China, which comprise more than 1/3 of the humans on this planet,  seem like the right focus for 2019.