Monday, October 27, 2014

The #EMR, #Ebola and #Bigdata - what Can We Learn

After all the hype and knee jerk politics and media I was delighted to read this piece Ebola US Patient Zero: lessons on misdiagnosis and effective use of electronic health records by Upadhyay,  Sittig and Singh (PDF file here)

A thoughtful piece that drilled in to the detail of events surrounding the arrival and subsequent consultations, admission and treatment of Thomas Eric Duncan

who sadly died on October 8 succumbing to the ravages of the Ebola virus

As the authors state
The mishandling of US Patient Zero is receiving widespread media attention highlighting failures in disaster management, infectious disease control, national security, and emergency department (ED) care.....also brought decision-making vulnerabilities in the era of the Electronic Health Record (EHR) into the public eye
Much of the commentary generated "fear, uncertainty, and doubt about the competence of our health care delivery system" and while there were problems I agree with the authors that this is a “teachable moment”  and a chance to identify the missed opportunities and key issues that we can learn from

The authors used the publicly available documents and testimony in their quest and it is important to note that they did not have access to the full record, the EMR used or indeed all the pieces of the puzzle and made up for this in some areas with educated guesses.

It is interesting to note that in the first visit to the ED the patient's temperature spiked to 103 degrees accompanied by pain described by the patient as 8/10 in severity.
This from a Malaria Study but typical of the Spike in Temperature found with this disease

He was diagnosed on initial discharge included sinusitis but "but that CT scans of “head and abdomen” ordered during the ED visit showed no evidence of sinusitis" and perhaps with more attention and importantly time made available to the clinical staff would offer them the opportunity to focus on the history and examination and less on high tech investigation. In many cases clinicians are forced into their use not by clinical practice but rather to meet the production pressures - as the authors put it
A host of system-related factors detract from optimal conditions for critical thinking in the ED, leading clinicians to lose situational awareness. These include production pressures, distractions, and inefficient processes
The upshot was a discharge and subsequent return days later at and even then:
even after the second ED visit which led to hospitalization, strict Ebola isolation precautions were not followed for 2 days, until the diagnosis was confirmed by the CDC
Offering a window into the events that is made so much easier with the benefit of 20/20 hindsight

The authors offer some learning opportunities that are worth highlighting

Top of the list in would be working with software developers to improve EHR usability
As this case illustrates, EHR-based clinical workflows often fail to optimize information sharing amongst various team members, leading to lapses in recognizing specific clinical findings that could aid in rapid and accurate diagnosis
As an interesting addition none of the systems (or incentives) have any form of feedback loop built in to allow clinicians to learn from their actions.

As for the process of information capture - we have lost site of the information that is relevant in the fog of billing and regulatory driven template driven charting.
Condition-specific charting templates, drop-down selection lists, and checkboxes developed in response to billing or quality reporting requirements potentially distort history-taking, examination, and their accurate and comprehensive recording.. Clinicians also tend to ignore template-generated notes in their review process; often the signal-to-noise ratio in these notes is low. EHRs can lead to less verbal exchange, which is all the more needed and more effective when dealing with complex tasks and communicating critical information 
Right on except to say this does not "potentially distort history-taking" - it does distort history-taking and
not "EHRs can lead to less verbal exchange" - EHRs do lead to less verbal exchanges

The data entry requirements place an enormous burden on our clinical professionals

who are tasked and measured not on clinical practice and the delivery of great care but on specific content of documentation that is mandated to capture clinical information in specific ways determined by the reimbursement, coding and regulatory system.
Other factors, such as heavy data entry requirements and frequent copy-and-paste from previous notes, detract from critical thinking during the diagnostic decision-making process... For EHRs to be most effective, they need to be able to automatically sort through patient data, identify the pertinent findings, and present them in an easy to understand manner. Computer algorithms could combine patient-specific information with the latest evidence-based clinical knowledge to help clinicians reach the correct diagnosis
This is the next frontier of Healthcare technology and in particular clinical documentation - we know we can sort through patient data, identify the pertinent findings - focused in these examples on quality of care and evidence based guidelines and we know computer algorithms can use patient specific information combined with evidence based knowledge to help

Technology can help but there are some fundamental flaws in the design and management of healthcare that are fed by the current incentives. Many initiatives attempting to improve patient safety and value-based purchasing but don't focus on accuracy and timeliness of diagnosis and in particular Outpatient reimbursement policies do not reward diagnostic decision-making, teamwork, or quality time spent with the patient in making a diagnosis.

What you incent is what you get and this needs to be changed as well.

Wednesday, October 22, 2014

Tracking #Ebola Effectively hindered thanks to #ICD10 (double) delay

This graphic
Offers a timely reminder that the US Government delayed a second time the implementation of ICD10 coding system that is used in the rest of the world

There is no code for Ebola in ICD9 - just a non-specific 078.89: Other specified diseases due to viruses which covers:

Disease Synonyms
Acute infectious lymphocytosis
Cervical myalgia, epidemic
Disease due to Alpharetrovirus
Disease due to Alphavirus
Disease due to Arenavirus
Disease due to Betaherpesvirinae
Disease due to Birnavirus
Disease due to Coronaviridae
Disease due to Filoviridae
Disease due to Lentivirus
Disease due to Lone star virus
Disease due to Nairovirus
Disease due to Orthobunyavirus
Disease due to Parvoviridae
Disease due to Pestivirus
Disease due to Polyomaviridae
Disease due to Respirovirus
Disease due to Rotavirus
Disease due to Spumavirus
Disease due to Togaviridae
Duvenhage virus disease
Ebola virus disease
Epidemic cervical myalgia
Infectious lymphocytosis
Lassa fever
Le Dantec virus disease
Marburg virus disease
Mokola virus disease
Non-arthropod-borne viral disease associated with AIDS
Pichinde virus disease
Tacaribe virus disease
Vesicular stomatitis Alagoas virus disease
Viral encephalomyelocarditis
Applies To
Epidemic cervical myalgia
Marburg disease

ICD-10 has one specific code for Ebola: A98.4 - Ebola Virus Disease
Clinical Information
A highly fatal, acute hemorrhagic fever, clinically very similar to marburg virus disease, caused by ebolavirus, first occurring in the sudan and adjacent northwestern (what was then) zaire.

Accurate tracking and reporting stop at the border of the United States

This is one of many examples of codes "missing" in ICD9 for conditions and care we are already delivering and dealing with

Wednesday, October 15, 2014

Connected Health and Accelerating the Adoption of #mHealth

I attended the Connected Healthcare Conference in San Diego yesterday
Accelerate mHealth Adoption: Deliver Results through Data Driven Business Models for End-User Engagement

Never has there been so much to play for in the mobile health landscape, a revolution is just round the corner with key players from the health care and consumer markets coming together to develop the mHealth industry. This Connected Health Summit will create a bridge bringing together hospitals, clinicians, providers, payers, software and hardware innovators, consumer groups and the wireless industry.

You can find the agenda here and the organizers will be publishing the presentations - there were many interesting insights

Andrew Litt, MD (@DrAndyLitt) (Principal at Cornice Health Ventures, LLC) opened the conference with a great overview of the industry and a slew of challenges and opportunities.

He sees our industry in Phase 1 - the Capture and Digitization of records
and we have yet to really move and explore Phase 2:
Move and Exchnage Data AND Analyze and Manage Data that is linked to Information Driven decision Making
And Phase 3:
Managing Patient Health
In our need to move from data to analysis and information he cited a statistic from a white paper: Analytics: The Nervous System of IT-Enabled Healthcare that sadly puts 80% of data in the EMR unstructured.
This is a fixable problem today with Clinical Language Understanding and we are seeing some results and a change in the industry to stop looking to doctors to be data entry clerks
He also cited Hospitals:
Technology offers tremendous scope to not only fix these problems but get ahead of the problem (as is done in other industries like the Airline industry that has rebooked your flights before you even land and miss your connection). As he suggested could we use data to understand who is likely to develop a heart attack in the next 2 hours and try and change this outcome

But integrating mHealth into our workflow requires an mHealth Ecosystem:

mHealth needs an ecosystem that improves workflow and integrates data to reduce clinicians workload. This is why doctors and clinicians are resisting mHealth - they don’t like the change to the workflow that has little if any positive effect (for the doctor - they may have a positive effect for the individuals health) of reducing clinicians workload

Interesting comment on wearables and the perspective of doctors on these devices:
What bothers the doctor - mostly the people who are buying and using wearable fitness/activity trackers are the people that are young healthy fit and want to prove to (themselves/others) that they are young fit and healthy?
His graphic on Security and privacy was on the money:

Essential to balance Privacy of Health with interoperability but trust is the imperative
The stats he presented were troubling (at best)

  • 96% - Percentage of all healthcare providers that had at least one data breach in the past two years
  • 18 Million - Number of patients whose protected health information was breached between 2009 and 2011
  • 60% - Proportion of healthcare providers that have had 2 or more breaches in the past 2 years
  • 65% - Proportion of breaches reported involving mobile devices
  • $50 - Black market value of a health record

The healthcare industry is under attack and is the most attacked industry today:

You might find these figures of the value of Healthcare data as it is valued on the black-market

Another interesting data point:

HIMSS records a total of 11,000 Healthcare Technology companies - less than 100 are large size and the balance of 10,900 are small business that are essentially capturing and scattering your data across many systems and data repositories...
Multiple other presentations and panelists that were all insightful. As always Jack Young (@youngjhmb) from Qualcomm Life Venture fund had some great insights - impossible to capture all of them but here are some:

Healthcare is moving out of the hospital into the home for many reasons but cost is a big driver:

and he suggested there was at least $1.5 Trillion in economic value as the industry shifts (shifting vs replacement?)

Many were surprised by his stat that users check their smart phone at least 150 times per day (just looking around my world this seems low) - in fact a quick check online suggests this is no longer valid and it is probably 221 times per day. Given this device is the one thing we will not leave home without and it now contains a range of sensors including:

  • Accelerometer
  • Gyroscope
  • Magnetometers
  • GPS
  • Cameras
  • Infrared
  • Touchscreen
  • Finger print
  • Force
  • NFC
  • WiFi/Bluetooth/Cellular

We have the potential for more passive compliance with our patients (and as many stated in their presentations likely more accurate as self reported data is notoriously inaccurate)
He predicted a a 10x growth in wearables from 2014 - 2018 with 26% of this growth attributable to smart watches (I know hard to believe at this point but I think if you looked back 4 years ago the iPad had nothing like the level of penetration it does today)
iPad Growth Rate

I liked his assessment of the werable market place by researching the eBay Discount against the price of the new device:
and even worse for Smart Watches

I also presented “mHealth Reimbursement - Who Will Pay:
You can see it here at Slideshare or below: