What exactly is this beast?
Yes, medication errors have been reported (such as in this article) to affect 1.5 million people every year (according to the Archeives of Internal Medicine), costing between $77 billion and $177 billion a year.
How do these medication errors occur? Because little is actually known about the information flows for medicine reconciliation.
Here's the problem : Nobody has a good plan for organizing enough to answer the question, "What meds is the patient actually on?"
Here's the scenario : A 60-year-old male shows up in the Emergency Room needing urgent care.
You're the doc responsible for this patient. How are you going to figure out what meds the patient is actually taking?
Potential sources of data include :
- The patient - Works for some patients, but many don't know the details of their meds.
- The family - Works for some, but many don't know the details of the patient's meds.
- The PCP - Works sometimes, but may not know what the specialist prescribed last week.
- The Specialist - Works sometimes, but may not know what the PCP prescribed last week.
- The Pharmacy - Works sometimes, but many pharmacies close after 5pm, many aren't electronic, and a lot of patients go to many pharmacies (including mail-order)
- Electronic "Insurance databases" - Works sometimes, but mainly only with insured patients - Doesn't work if the patient pays out-of-pocket
- Herbal medications - Often missed as a data source
- The old hospital record - Sometimes helpful, but sometimes out-of-date. Also remember, except for the VA, virtually no hospital shares pharmacy data with the outside PCPs.
So which of these data sources are you going to use?
What you start to realize is - It's almost impossible to know with 100% accuracy what meds a patient is on.
(Well, maybe in a best-case scenario : A well-educated, non-intoxicated patient, only on one or two medications - Maybe then, you can achieve 100% accuracy. Outside of this scenario, you're generally working at 95% accuracy or less.)
So for the majority of patients, you will never achieve 100% accuracy.
So you have to ask yourself - What level of accuracy is acceptable?
I'm proposing a new standard, which I lovingly call : The Mother Standard.
The Mother Standard is "The amount of data you would collect to achieve a level of accuracy that you would find acceptable for your own mother, knowing that 100% is impossible." I figure, for most people, this is a level of accuracy of >95%.
And this is my proposal, for an acceptable way to perform Medicine Reconciliation to the degree of the Mother Standard (>95%) :
Step 1 : Ask the patient what meds they are on - Including herbal medications. If you, as a clinician, do not feel you've met the Mother Standard - Proceed to step 2.
Step 2 : Ask the family what meds the patient is on - If still not the Mother Standard, proceed to Step 3.
Step 3 : Ask the pharmacy(ies) what meds the patient is on - If still not the Mother standard, proceed to Step 4.
Step 4 : Ask the PCP what meds the patient is on - If still not the Mother Standard, proceed to Step 5.
Step 5 : Ask the specialist what meds the patient is on - If still not the Mother Standard, proceed to step 6.
Step 6 : Check the old hospital/office chart for what meds the patient is on - If still not the Mother Standard, proceed to step 7.
Step 7 : Check the electronic insurance report - If still not the Mother Standard, then at least you can say you made every attempt to achieve the Mother Standard, and were unsuccessful.
So why are there so many errors nationally, every year? Because patients who don't have their meds clearly tracked require an enormous amount of work, just to try to get to the Mother Standard - And in some cases, it's just impossible.
Looking at the coordination of care among multiple specialists, sometimes even multiple PCPs, and hospitals - If the patient, or their family, does not keep track of the meds - Then achieving the Mother Standard is virtually impossible.
Fortunately, doctors are trained to work with incomplete information, but when incomplete information isn't enough, we have to ask ourselves : Who can fix this?
I'm hoping that SpeakFlower (http://speakflower.org) helps our country move into this realm, gradually.
Anyway, I'm hoping to do some research into the time it takes to reach the Mother Standard. Will try to publish what I can shortly. Stay tuned.
The Mother Standard. Love it! Of course, there are some people whose Mother Standard might kill the old lady, so one might want to be sure they actually love their mother, first.
You left out the problem with recording samples dispensed, but I really wanted to point out that achieving this maternal standard will require different levels of effort depending on the situation and the medications in question.
You may not need to go beyond the patient when asking about their their statin when being seen for pneumonia, but it could be very important when seeing someone with acute muscle pain and elevated CPKs. So, the calculus needs to include:
-the current clinical situation (risk, urgency), and
-the risks involved if you get it wrong.
Steve - Good point about "What's not really important for this hospitalization" - But I've studied this to death, in a "retrospectoscope" it's easy to say which meds aren't important, but when it comes to the question :
- What process are you going to leave off the list? (e.g. which of my steps in the "Mother standard" can you safely leave out?) -
... This is where it gets tough. Hard to accept leaving any one of them out. And yet, you probably don't need all of them in every patient.
Hence... The mother standard. :)
(I do totally agree, though - The risk of "getting it wrong" should be weighed...)
I have a somewhat different definition of the Mother Standard. Do you have the level of accuracy your own mother would accept? For my mother that would get you to 99% or better. Nonetheless, I love the idea.
Post a Comment