Tuesday, November 25, 2008

Automating the Pull of Information from EHR's

The Social Security Administration announced they wanted to set up a project to test the concept of pulling information automatically form EHR's to help them deal with the 2.5 million disability claims it receives each year

Currently the process cost ~ $500 million a year to retrieve paper copies of records and then process them. The initial request asked vendors, health providers and payers to suggest a process and referred to a trial currently in process using the Continuity of Care Record (CCR). The framing by SSA as:
a fully automated Personal Health Records prototype system
is a little misleading and submissions can come from all areas. If these records are produced by dictation and transcription (and given the high proportion of information that currently is it is a fair bet that most is) then the opportunity here is for the value to the information by creating these documents in CDA format to be made available to SSA for those records to allow them to process the information automatically.

Information on the request can be found here

This project could be a great demonstration vehicle to show the value of moving all documents into a CDA format making the information instantly more useful and available for processing.

Tuesday, November 11, 2008

Magical Thinking in Implementing Healthcare IT

There was a great article published in the Health Affairs Journal by Carol Diamond from the Markle foundation - "Health Information technology: a few years of magical thinking" - abstract here the full text requires a subscription

The concept of magical thinking in this context was that implementers must resist the concept where this notion that
..isolated work on technology will transform our broken system...Another tempting and related notion suggests that a lack of technical standards is the main barrier to health IT adoption..
Given that as they state
...the literature on computerization, stretching back to the 1980s, is unambiguously clear on this point: computers are amplifiers. If you computerize an inefficient system, you will simply make it inefficient, faster. IT can contribute to improving care only when underlying system processes are transformed at the same time.
To be successful instead of joining the stampede of standard creations from the likes of Health Information Technology Standards Panel (HITSP) and the Certification Commission for Health Information Technology (CCHIT) which are increasing awareness in the public and helathcare industry but have not according to recent testimony by Sam Karp of the California HealthCare Foundation stated
"Not a single data element has been exchanged in real world health care systems using standards this process has developed or deployed."
The point being that standards are adopted and the process of standardization is incremental. The internet being a great case in point that was developed over severalyears and floated to users unfinished and lacking consistency to allow usr interaction and use to help refine and develop a standard that we all use on a daily basis

So taking this concept to the next level and looking at the 60% of healthcare notes that are simply free-form notes - providing simple easy access to essential parts of this information would be invaluable. And there lies the beauty of the of the CDA4CDT system allowign for the capture and sharing of this infomation wihtout the impostion of structure, coding or limitations of choices. Sections can be easily identified and shared in a meaningful way. At a high level in the first instance but in more detail as people explore and adopt this standard and actually use it. THis is a standard to use without the imposed and artifical limitations that normally accompany the typical narrow "ideal" standard that does not meet the test for real use and genuinely useful interoperability

Wednesday, November 5, 2008

User Design - Basic Principles

The British National Health Service (NHS) has one one of the largest healthcare IT projects (National Programme for IT NPfIT) in the world rivaling size and complexity in almost any industry. It's no wonder given the size and complexity that there are problems but in some cases these are just plain simple mistakes and getting basic simple things wrongs does nothing to engender the support and confidence with the embattled clinical users

A Recent post from "Phil Hammond 'Medicine Balls' " Confusing for Health highlights the reasons why the NHS Care Record Service (CRS) implementation systems has been "indefinitely postponed". (I've copied the full post below in case this link fails or ceases to work) but it was item 9 that caught my attention.....
Users who select 'died in department' by mistake (which is default outcome if D is used to navigate drop down), do not receive a warning resulting in appropriate letters sent to the GP. For genuinely deceased patients, the code is mapped incorrectly and does not bring up fields to complete, to register patient dead on the spine.
(The spine in this context is the NHS network backbone)

In a cock-up worthy of the Colemenballs moniker the default value for patient outcome is "Died in Department"!! Someone, somewhere is not firing on all cylinders and there there is a chain of people who let this through into a live systems that are diagonally parked in a a parallel universe.

As Phil Hammond states
...you can only succeed by starting as simply as possible and then getting those who have to use the new system to help build it into something fit for purpose
Sage advice!

>>>>> Confusing for Health >>>>>
Original posting should be available here
The NHS Care Records Service (CRS) is a secure service that links patient information from different parts of the NHS electronically, so that authorised NHS staff and patients have the information they need to make care decisions.' So says the Connecting for Health website, but there isn't much sign of it up and running in the West Country. A computer literate consultant tells me that CRS has been 'indefinitely postponed' where he works because 'it cannot migrate data across the system into the correct coding slots which means there is no way for a hospital to record what work it actually carries out and be paid appropriately.'

Our new market driven NHS depends for its survival on the right money following the right patients. And then there's the extra cost not just of installing the new system but paying extra staff to run it properly. A document sent to me for the Royal Free hospital in London shows not just how many bugs still need ironing out in the CRS, but how we all need to learn a whole new language to communicate with each other.

'1. There are problems associated with the use of smart cards to log on, which takes 7 key strokes, resulting in a very long log-in time, as much as 10% of each hour. This discourages use of the system, and encourages staff to leave the card in place, which then prevents the identification of the user when requesting X-rays, bloods etc.

2. The GP letter is very poor quality and requests to change this via the RFC (Referral Facilitation Centre?) have not been implemented. The treatments and investigations form does not support the user in choosing clinically important tests or support HRG 4. The Inbox not emptying automatically leads to significant delays in the system.

3. Clinics must have DNA patients dealt with as DNA, otherwise 18 weeks reporting is a problem. Reporting for the four-hour target is clunky and can only be done in Explorer because of the delay in updating IM200. This still takes up to 2 hours to validate the performance each day, causing 40-50 breaches a day for first few weeks.

4. Data can be entered but may not be visible to other users, and is difficult to find the forms and summaries. Multiple issues raised with LPfIT and BT/Cerner since going live have not been resolved

5. Free text fields on the discharge letter only allow 750 characters, resulting in limited summaries and poor communication with GPs.

6. Excessive time taken for scanning, registering, creating pending admission and GP letter printing. Five extra A & C (administration and coding?) staff have needed to be employed.

7. Discharge time of patient is displayed rather than check-out time (i.e. time patient left department). This makes reviewing 4 hour breaches impossible, since discharge time may be as much as 24 hours after checkout time

8. Manchester Triage does not populate white board. (I've no idea what this means either).

9. Users who select 'died in department' by mistake (which is default outcome if D is used to navigate drop down), do not receive a warning resulting in appropriate letters sent to the GP. For genuinely deceased patients, the code is mapped incorrectly and does not bring up fields to complete, to register patient dead on the spine.

10. The system crashes 2 or 3 times a week.'

I think I only understand number 10 with absolute certainty, but the message is clear enough. Implementing large scale IT projects is unbelievably complex. And you can only succeed by starting as simply as possible and then getting those who have to use the new system to help build it into something fit for purpose (preferably without inventing a whole new geeky language). I'm generally an optimist but I'm not sure about this. How long before we have more coders than doctors? Or are we there already?

Monday, November 3, 2008

Healthcare CIO's Grappling with EMR Adoption

SearchCIO online magazine ran an article on EMR adoption that made for interesting reading:
When patients, physicians and payers embrace the electronic health record (EHR), life will be different in pretty amazing ways.....For the first time, patients will be treated by a personal team of clinicians. When a new drug for hypertension comes on the market, all patients (not just Nobel laureates like James Watson) will be able to map their genotypes and phenotypes to that medication to determine if it's right for them. Hospitals will be held to the "perfect care" standard -- the elimination of all medical errors in instances of preventable harm.
Wow! But the problem is we are nowhere near the level of adoption necessary to achieve these kinds of advances and the barriers to adoption remain frustratingly present and challenging. Privacy, interoperability, liability issues and physician reimbursement are all main stays of resistance to the move towards wide scale adoption of the EMR. As expected there are some frightening stories to hammer home the point from an emergency room physician who estimated he treated 80,000 patients "with my own hands
...the thing that stuck out as he looked back on his career was how many times he was put in a position of "guessing over and over," "flying solo," in an information vacuum. In situations where people "die right in front of you," he said he often felt he was "one data element away" from stopping a patient from dying.
Needless to say there continues to be the naysayers who are convinced that physicians " know what they are doing; why do you want to tell them what to do" but in all this seem oblivious to the tsunami of knowledge rushing down the luge of clinical practice that is impossible to keep up with.

I agree with John Halamka
that the lives of primary care physicians -- snowed under by paperwork that does not require an M.D. but is required nonetheless, frustrated by prescribing a medication only to find out it's denied by the insurance company and terrified of making a mistake -- is sheer misery. He predicted they will welcome the help, and patients will be better off for it. As the system stands now, "all the medical students are becoming dermatologists," he said.
And it's easy to see why with the information overload with "medical literature published every month that is is more than a doctor could read in a year". Not to mention declining reimbursements and shattered dreams that litter the halls of our hallowed medical facilities. We need EMR's and EMRs need data to provide the decision support that an automated and optimized medical technology infrastructure can provide physicians in their daily practices. But all of this should not turn clinicians into data entry or data capture clerks - they are not good at this task and technology is available to facilitate this issue and provide clinicians with the tools to ease the burden and provide them with the necessary clinical decision support they want and need.