The second day of the second session of the Leadership Strategies for Information Technology in Health Care started with another cold rainy day in Boston.
The morning session was led by Dr Ashish Jha, and unfortunately Dr David Bates had to go to Paris, so Dr Jha covered his stuff as well. As you are probably aware David Bates is THE EXPERT in quality and safety informatics, and I have had the pleasure of hearing him speak, and meeting with him a number of times in the past. I have to admit Ashish Jha did a great job driving the discussion amongst the 50 plus attendees from around the world, so while I would have loved to hear David Bates again, I was just as impressed by Dr Jha.
The first half of the morning we discussed Meaningful Use and its relevance not only to the USA but also the rest of the world. There was some concern by a few members of the group that the seemingly unreasonable time frame and demands. The opposite argument was raised of it setting a benchmark for action, and otherwise we would still be trying to define the problem. A very lively discussion limited only the amount of time we had ensued. It was daunting listening o folk from some of the world’s leading institutions talking about how it has affected them and their facility.
The classic safety papers were discussed in the next session. The 2005 paper by Han et al on the increased mortality on introduction of an EMR in a paediatric hospital triggered much discussion about how poorly we actually understand and document current workflow, so when we introduce big changes to it is when we first realise how important the current workflow is. The introduction of an EMR which required physicians to leave the bedside to go to the computer may have been the reason mortality rose in this paper. Pretty good case for making the technology fit the workflow rather than the other way around.
After morning tea the folks from CRICO RMF, who are essentially Harvard’s malpractice insurers, but do a lot more than just legal stuff. The RMF stands for Risk Management Foundation which as a fascinating discussion in itself. They are very active in developing how systems need to be developed and used to drive high quality safe practice, because that is less likely to get you sued. They brought up some very interesting points about what the new record is, how things like the ability to copy and paste information has led to a whole range of new issues about the authorship and validity of the EMR. A great example of how copy and pasting a troponin result across multiple days was in direct conflict with the actual results. The notes said it was high, the path results showed it wasn’t, but the clinical care was driven by what was in the notes, which was copied and pasted from the admission examination. Even so called smart paste has issues, because things like abdominal tenderness captured by the system was not highlighted because the smart paste focussed on the EKG so lack of chest pain was the decisive information. The patient ended up needing an emergency hernia repair, the signs were all there, but the cardiologist using copy and paste made everyone else miss it.
They then went on to talk about the effect of workarounds. Partners has an enterprise allergy tracking system, but discovered that the value of the system was challenged by the fact that in a 15 month period over 93,000 allergies entries were entered as free text, even though there was a system in place to capture coded allergies. About 5 percent of these were anaphylactic reactions, not just mild itching, but not captured ion the system designed to prevent this. When they looked at why this was so, it was basically a design issue, the box for selecting the coded entry was blank, so people just typed into the other box below it. However in the allergy list there was no way of identifying free text entries versus coded ones, even though the free text ones would not trigger an alert. With some simple User Interface (UI) changes they have dropped the non coded allergies from 20% to 8% over 12 months.
An interesting discussion arose about how copy and paste notes have led to some clinicians no longer using the notes as they are just a collection of copied data. The point was raised that documentation is driven by billing requirements rather than clinical requirements, so that may be what causes all the problems. Glad that t=in Australia I don’t have to document for the billing system, yet!
In the afternoon we looked at Clinical Decision Support. A good question was raised, what is a good or acceptable rate of over riding alerts. No answer but an interesting statement, poor decision support is worse than no decision support at all! Another key learning, all decision support is local, so listen to your users. The other great rule is that less is more, i.e. simpler is better! Another interesting statistic was the NNTR, Number Needed To Remind for an alert. Suggested that for patients with Coronary Artery Disease, the clinician needed to be reminded 4 times to start aspirin. Question arose that in actual fact response is better than that, but the information is not captured.
An interesting concept of tiered alerts was discussed, and an interesting experiment was shown of how a tiered alert had 2-3 times greater effect than non tiered alerts. It was an interesting concept, some alerts cannot be over ruled, but you have to make sure that their use is managed, and they should be rare, others can be over ruled but you need to do something, whilst level 3 was a little red note on the screen, which was probably ignored most of the time.
The golden rule for CDS is to make desired behaviours easy, undesired behaviours difficult.
The final session was a panel by the quality and safety managers from three of Boston’s major healthcare facilities, Massachusetts General, Brigham and Women’s and Harvard Vanguard. They are all doctors, one of them is also a lawyer, and they all have a strong background in policy. Costs management came up, meaningful use came up, and the discussion was enthralling. Personal Health Records came up and their adoption is an issue. Harvard Vanguard has 30% of their patients actually using the PHR, which is an astronomical adoption rate.
End of the day, shopping and downtown and dinner with my old friend Dr Ron Ribitzky who has been travelling the world driving ehealth adoption.