I started my blog over 15 years ago. Yes – it’s been less active in recent years, and as I reflect on why I’ve been less active – only part of the reason is that I was working for a publicly traded company from 2008-2011 .. and a federal agency from 2011 – 2014. Both of these organizations have reasons to control the messages of their employees. I needed to be cautious about what I blogged. So I didn’t blog publicly very much.
But since November, I’ve had no excuse. And yet nothing much flowed from these fingertips.
It should have. Back in the day – THCB and Docnotes – and a handful of other sites offered bookmarks and observations on health care delivery, the convergence of health care and IT, and random observations. These days – there is a tidal wave of these things on the Internet. I sometimes question whether MY contributions are of any value now that there is so much out there. I remember when Dave Winer toyed with killing his blog. He didn’t. Nor should I. This post celebrates the not-killing of my new blog, and the beginning of the NEXT 15 years of my public observations.
Here goes ..
I’ve been learning so much. I owe it to you to share what I’m seeing out there, beyond the ivory towers of government, beyond the walnut paneled board rooms of hospitals and health IT companies, and beyond even the ping-pong tables of Silicon Valley start-ups and the crisp offices of their venture capitalists.
I started blogging so that I could make the thought processes of a physician transparent to “normal humans.” 15 years ago, concepts like “shared decisions” were neither mentioned nor understood. Real-world applications of genomics were a twinkle in the eyes of Watson and Crick (and perhaps some others). Physician blogs are now plentiful. We know how they feel about thir long days, the meaningful use EHR incentive program, Maintenance of Certification, SGR fixes (or attempts at least) and electronic health records.
So here’s some of what I’ve observed recently – to make the thought process of this physician more transparent:
- There is growing interest in DPC models (see this recent Time article about QLiance and others – experimenting with new payment models) – and “full risk” medicare advantage care delivery models, such as what Chris Chen and his family/team have done in Miami. “Patient Centered” is more than just a catch-phrase to these folks. They live it every day. I witnessed the T’ai chi class at ChenMed – which turns out to be much more effective at managing patients’ osteoarthritis than lortab.
- Other organizations like Iora Health, Landmark Health, and Caremore are finding ways to really achieve the triple aim.
- None of this was possible 20 years ago, and the difference today is that technology is better, smaller, faster, more ubiquitous, and easier to implement, iterate and optimize.
- Yet the IT we have in health care remains “dumb.” The systems remain focused on the capture, aggregation, and (ideally) transmission of data. Yes – there are sometimes decision support interventions that remind us to do the things we forgot – or NOT to do the things that might be harmful – but these capabilities are not adaptive or dynamic in the way that Google Maps or Waze adapt to changing conditions ahead, or even a change in the destination. How do we know what this patient’s optimal goal really is? (I vividly remember Eric Dishman’s anecdote about his cancer treatment: his goal was to maximize time on the ski slopes – not necessarily the length of his life. How could the oncology team manage toward THIS endpoint? It wasn’t in the guideline or their training. But it was the most important endpoint to the patient. Current processes, systems can’t do this – in part because they can’t capture and understand “maximize skiing.”)
- Consumer applications leverage machine learning and even deep learning so that they can anticipate our needs. Amazon’s patent filing for “anticipatory shipping” is just one example of the many ways in which we increasingly leverage intelligent IT in our day-to-day lives.
- In health care, we have seen enhanced interest in analytics, predictive modeling, and machine learning. Indeed – my first “real job” in health IT was at Medremote, where we applied machine learning to transcribed progress notes in order to create structured data from narrative documents. But as an industry, we’ve failed to make scalable applications of IT that really help improve health – that leverage tools capable of predicting how this individual will find the path toward better health. The EHRs remain repositories of information. Analytics platforms generate reports of care gaps, but are rarely actionable. Predictive models abound in the financial sector, but are rarely leveraged in the average clinician’s daily routine.
These vectors will converge:
- Care providers who are focused on the health of an individual and a population will manage their decisions, their information, their processes, and their communications differently from how care providers who are paid to maximize volume.
- We will see continued “shifting left” of health care. What used to happen in the hospital will now happen in the clinic. What used to happen in the clinic will now happen in the home. Practices will increasingly share (or own) risk. Employers will drive providers to improve quality, just as Intel has done.
- Innovative care models will drive new requirements for IT that focus on care quality, collaboration, communication, shared decisions, and patient empowerment (not just “engagement”).
- Machine learning tools will provide real-time insight into the best choices for patients, and these decision support tools will be the core of next-generation health IT systems (rather than “bolted on” as we have today). Our technology infrastructure in health care will (finally) provide some of the adaptive, anticipatory assistance that we have come to expect from the rest of our interactions with the technical tools we use every day.
- Because of the enhanced intelligence of the IT infrastructure, safety of health IT systems will become increasingly important, and the government agencies such as the FDA and ONC (and their international counterparts) will need to find the right risk based regulatory framework to allow for innovation while protecting public safety.
What does this mean?
I’ve expressed above what I believe to be a set of self-evident observations. Yes – I’ve over-simplified a bit, and one can find exceptions to many of the statements I make about both existing systems and the absence of 21st-century solutions in health care. Both would be valid critiques.
But I stand by my statements as good representations of the majority of current state health IT, and the obvious progression toward the future. These systems will be more intelligent. Care delivery will evolve – finally – both due to the federal government’s deep interest in reaching the triple aim – and from the private sector finally understanding that we simply deserve better, as do our children, our parents, and our neighbors.
So it’s obvious (to me) what I need to work on next. I need to weave together my passion for reforming care delivery with my interest in usability/user experience, and my interest in decision support and shared decisions between care providers and individuals/families.
I’ll blog about my personal work to skate to where this puck will be over on docnotes – for the next 15 years.
This is an exciting time. 2015 will mark the real transition as we start to turn the corner from “fee for volume” – and the motivators that maximize volume of care (with the harm that comes along for the ride) – to “fee for value” and the use of intelligent systems that will get us to where we need to be.