More Thoughts on the Rx Issue

Fred_Nickols1 · August 19, 2001, 10:59am

[From Fred Nickols (2001.08.19.0640)] --

More thoughts on the Rx issue that Rick raised won't leave me alone so I'm
going to dump them here...

Like Kenny, I'm a consultant who makes his living improving productivity
and performance. The Rx error rate is the kind of issue I deal with on a
regular basis. That the person you're working with would be interested in
root causes reflects not just that particular concern but probably a larger
concern with processes as well. There are many, many people working in the
TQM, reengineering and performance areas who would immediately zero in on
the larger system or process in which those errors are occurring and not
focus on the people first.

One of the first things I'd do is set the boundaries for the process or
system under consideration. Five seconds' quick thought suggests to me
that the process begins with the physician's decision to prescribe and ends
with the patient having the prescription in hand (probably at a
pharmacy). What lies between those two boundaries? Writing the
prescription for one. Giving it to the patient or calling it in to the
pharmacy. Filling it (at the pharmacy). Transcribing what the physician
wrote to the label that will be placed on the
bottle/box/tube/container. Labeling the container. Getting the prescribed
medicine from stock and placing it in the container, which often involves
counting things like pills. Placing the container containing the
prescription in a larger container (usually a pick up envelope). Giving
the prescription to the patient (usually in exchange for some money at a
cash register/counter).

The process just described is fraught with the potential for error all
along the way. The physician could make an error in writing the
prescription. As Bill Powers already pointed out, the physician's
handwriting could be an issue. Transcription errors happen all the
time. The wrong label can get on the wrong bottle and, believe it or not,
people do sometimes wind up getting the wrong medicine, wrong dosages, and
even someone else's medicine.

Electronic transmission of prescriptions could no doubt reduce some errors
but it will not touch many of them. One of the first things I would
undertake is to compile an analysis of the Rx errors. A look at the
frequency distribution of the various kinds of errors would tell me where
to look and for what. Absent the data, I would focus on getting agreement
regarding the process, the likely cause(s) of errors (including the
obligatory homage to "root" causes), and some reasonable preventive steps.

As a final note (at least on my part), it is certainly conceivable that the
entire prescription-filling process could be automated. It is certainly
technologically feasible to do so; whether it is economically and
operationally feasible to do so remains to be seen.

Regards,

Fred Nickols
The Distance Consulting Company
"Assistance at A Distance"
http://home.att.net/~nickols/distance.htm
nickols@att.net
(609) 490-0095

Kenneth_Kitzke_Value · August 19, 2001, 2:46pm

[From Kenny Kitzke (2001.08.19)]

Probably like you Fred, I have found executives very willing to spend time
and money on improving business results. It is easy to demonstrate how
eliminating defects in outputs, or productivity in work processes, improves
business results.

It has not been so easy to demonstrate to CEO's how improving ones
understanding of human behavior improves business results. When it comes to
the management systems we teach to improve quality, or improve business
strategy/competitiveness/and productivity, they like it, want it and reap
tangible reward quite quickly. Sometimes they are appreciative and grateful
for your help, but you can't count on it being expressed (don't establish a
reference for that).

I have still not found one CEO who really wants to understand psychology and
is willing to devote personal and corporate time and money to learn a "new"
theory known as PCT to improve the way they manage and lead people. Even
though I personally believe that doing so will provide a new paradigm of
performance improvement (just as powerful as the quality or strategy
management systems I promote quite successfully), I can't seem to find a way
for them to want it.

Sometimes it seems like the old chicken and egg routine. IOW, if I could
show them irrefutable dramatic and valuable application results (like from
the other systems), they could imagine doing something similar themselves and
say, "Kenny, what do we need to do to achieve these results?" But, since I
can't get anyone to try PCT as a basis for a radically different people
management system, I can't show any actual results; only theoretical results.
It looks like basic research to them and they have more practical tasks to
pursue that seem more important to their livelihoods.

I have come very close with a few, especially those CEOs who share my
Christian belief system. They generally have a bigger heart for their
employees and are more receptive to learning something new about why people
do what they do so they can serve them better. But, time after time, their
good intentions and attention get sidelined by more pressing economic issues
perceived in their environment.

But, I shall not give up as long as I am still breathing.

Your comments about the entire process (the better ones having a feedback
loop not just a linear chain) of getting correct prescriptions to the
intended user is right on. We first collect enormous amounts of data and
analyze it. Here we use Pareto Analysis to find out what to work on that
will produce the biggest result.

The work at Rand in electronic prescription writing/communication mistakes
strikes me as working on the tail of the dog, a tiny piece of the Rx defect
puzzle. It is like pursuing a topic for intellectual interest (basic
research) that might have a useful application someday, somewhere. What Rick
has already discovered in his project, which he perceives to be profound,
about human errors in PCT lingo, may well be as uninteresting to those
managers or executives who could apply it as PCT itself is (next to nothing),
at least for the time being. This in no way suggests it should not be done.

Stop wondering. If you are ever in Pittsburgh, and would like to see robots
filling prescriptions at incredibly high quality and productivity levels and
at lower cost, give me a ring. By the way, it is so economical that this
unit has regularly collected the highest management bonuses of any unit in
the entire J. C. Penny corporation.

If you would guess that the rest of the organization hierarchy would want to
learn why this unit does so well year in and year out (now over a decade),
you would guess wrong. They are controlling for other variables; certainly
not for being stupid leaders needing to learn better management systems.

Respectfully,

Kenny

Richard_Marken2 · August 19, 2001, 7:17pm

[From Rick Marken (2001.08.19.1120)]

Kenny Kitzke (2001.08.19)--

The work at Rand in electronic prescription writing/communication
mistakes strikes me as working on the tail of the dog, a tiny piece
of the Rx defect puzzle. It is like pursuing a topic for
intellectual interest (basic research) that might have a useful
application someday, somewhere.

This strikes me as a somewhat presumptuous conclusion considering the
fact that I have not described the RAND electronic prescribing
project in any detail at all. If you would like to read a brief
overview of the whole project go to http://www.rand.org/health/eRx/.
The project has just begun. I have been doing an analysis of the
prescribing process as it occurs in various circumstances (in-patient,
out-patient, new prescription, refill, etc) and collecting prescribing
error data from the literature. My question to the net about "root
causes" was related to how we could do our future evaluation of
features that should (or shouldn't) be standard for e-prescribing
prescribing products.

What Rick has already discovered in his project, which he perceives
to be profound, about human errors in PCT lingo, may well be as
uninteresting to those managers or executives who could apply it
as PCT itself is

What I discovered was a _principle_ for evaluating features that
e- prescribing products should have in order to reduce or eliminate
known types of prescribing errors (OEs). The principle is simple: the
e-prescribing system should have features that make potential OEs
into SEs for the user. The principle is not just based on PCT lingo;
it is based on an understanding of how the PCT model works and how
the workings of this model relate to actual behavior (such as the OEs
that have been observed in studies of the prescribing process).

This principle has some surprising implications from the point of
view of conventional human factors principles (as I noted in a reply
to one of Fred's posts). Based on his understanding of PCT, Fred
suggested that a major cause of prescription errors (OEs) is
disturbances. My principle, also based on PCT, leads to exactly the
opposite conclusion: disturbances should have _no_ effect on the
occurrance of OEs if there is no SE when they occur. The absense of
SE -- the physician has written the correct prescription from his/her
point of view -- suggests successful _resistance_ to any disturbance
to the controlled variable. Not only are disturbances _not_ the cause
of OEs but, looking at the process from an HPCT perspective, they
may actually _reduce_ the occurance of OEs. For example, if a
disturbance (such as a flakey pen) makes it difficult for the
physician to write "Take 1 per day", by the time the physician
reaches for a new pen a higher order system may realize that
the lower level intention should be to write a "2" rather than a
"1". _Appropriate_ disturbances are more likely a _feature_ than
a flaw in the prescribing process.

As a final note (at least on my part), it is certainly conceivable
that the entire prescription-filling process could be automated.
It is certainly technologically feasible to do so; whether it is
economically and operationally feasible to do so remains to be seen.

It is definitely economically feasible. The e-prescribing project I
am working on addresses how to make it operationally feasible.

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: marken@mindreadings.com
mindreadings.com

Kenneth_Kitzke_Value · August 20, 2001, 12:59am

[From Kenny Kitzke (2001.08.19)]

[From Rick Marken (2001.08.19.1120)]

It was my presumptuous opinion only. That is what strikes me means in my
perception. Why do you infer it was some kind of studied "conclusion" on my
part?

Thanks Rick. I did not know this was possible.

I sent you a response about the concept of "root causes" which I have worked
with for 18 years in doing problem solving, including eliminating or
overcoming human errors. I did not get any response from you.

I am thrilled you are discovering things. Others do too, even those without
"an understanding of how the PCT model works."

What is the conventional human factor principle to which you refer?

I read Fred's post. I think he is always cautious to explain he does not
perceive himself as a PCT scientist who uses PCT terms in a technical
glossary sense when he posts.

I don't disagree with your principle. How have you determined what the CV is
for physicians when they write prescriptions? Or are you telling us how it
strikes you?

If that is what people want you to tell them, I guess that would be helpful
to their reference for knowledge they desire. I will try to follow the
progress.

May you find great insights for them.

Richard_Marken2 · August 20, 2001, 3:31pm

[From Rick Marken (2001.08.20.0830)]

This principle has some surprising implications from the point of
view of conventional human factors principles (as I noted in a reply
to one of Fred's posts).

Kenny Kitzke (2001.08.19)--

What is the conventional human factor principle to which you refer?

It's the one that Fred Nichlols mentioned. It's no a principle so much
as a rule of thumb. It's the idea that errors are caused by disturbances
(properties of the "system" in which the error is made). I guess the
human factors "principle" is that errors are caused by system
characteristics. An example of this "principle" is the idea that one
cause of prescription error is a noisy environment. I think most human
factors engineers would take it for granted that noise can cause error
(by disturbing the process of producing the result). My analysis
suggests that this is not the case.

I don't disagree with your principle. How have you determined what
the CV is for physicians when they write prescriptions? Or are you
telling us how it strikes you?

I have not determined the CV and I don't believe my analysis requires
it. All I have observed is that the physician has produced result X. And
I know from medical informants that the physician was supposed to
produce result X'. The difference between X and X' is what I call an
objective error (OE). If the physician _intended_ to produce result X',
then PCT tells me that, when the result was X the physician would have
acted to change X to X'. The fact that the physician did _not_ change X
to X' suggests either that the physician 1) was _unable_ to change X
into X' or 2) intended to produce X rather than X', and thus experienced
no subjective error (SE) when X was produced. I reject possibility 1)
because physicians are skilled at writing prescriptions and are readily
able to change mistaken prescriptions when given the chance. So I
conclude that when we see an OE (prescription X rather than X') the
physician who wrote the prescription experienced no SE; X was actually
the intended result.

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
MindReadings.com
10459 Holman Ave
Los Angeles, CA 90024
Tel: 310-474-0313
E-mail: marken@mindreadings.com

Bill_Powers1 · August 20, 2001, 8:20pm

[From Bill Powers (2001.08.20.1348 MDT)]

All I have observed is that the physician has produced result X. And
I know from medical informants that the physician was supposed to
produce result X'. The difference between X and X' is what I call an
objective error (OE). If the physician _intended_ to produce result X',
then PCT tells me that, when the result was X the physician would have
acted to change X to X'. The fact that the physician did _not_ change X
to X' suggests either that the physician 1) was _unable_ to change X
into X' or 2) intended to produce X rather than X', and thus experienced
no subjective error (SE) when X was produced. I reject possibility 1)
because physicians are skilled at writing prescriptions and are readily
able to change mistaken prescriptions when given the chance. So I
conclude that when we see an OE (prescription X rather than X') the
physician who wrote the prescription experienced no SE; X was actually
the intended result.

I think you also have to consider the case in which the physician thought
he was prescribing X', but as far as the nurse, pharmacist, or practitioner
was concerned, he was prescribing X. The patient got X not because the
doctor prescribed it, but because the doctor's prescription or order was
misheard, misread, misinterpreted, or misremembered by the person who
actually gave the treatment or provided the drug.

All this fits your principle that the mistake is not corrected because the
doctor does not experience any subjective error, SE. But the case I'm
putting before you doesn't involve the doctor's intending the erroneous
order or prescription. If the doctor knew how his order or prescription was
being interpreted, he would correct the error, the OE. But you can't say
that because X was actually administered, X is what the doctor intended to
be administered. Doctors, in my experience, do not usually apply treatments
(excepting surgeons), fill prescriptions, or give pills to patients. They
write prescriptions and treatment orders, and it is up to someone else to
act as the output function. It is, unfortunately, a common occurrance for
doctors to give order or prescription X', only to have order or
prescription X actually reach the patient. I seem to recall news reports
that there is an incredible number of deaths in hospitals every year --
tens of thousands -- from just such "mixups".

If the loop were closed, such mixups would happen far less often.

Best,

Bill P.

Richard_Marken2 · August 20, 2001, 10:16pm

[From Rick Marken (2001.08.20.1515)]

Bill Powers (2001.08.20.1348 MDT)--

I think you also have to consider the case in which the physician thought
he was prescribing X', but as far as the nurse, pharmacist, or practitioner
was concerned, he was prescribing X.

Yes. Of course. And I am considering it in my analysis. But the fact of
the matter is that most prescribing errors are _not_ transcription
errors. The nurse or pharmacist -- the one filling the order -- can read
the order just fine. It's just that the wrong thing was written on the
order. For example, 0.5 mg instead of .05 mg, erythromycin ordered for
patient on cisapride (a drug the interacts badly with erythromycin),
atrovent ordered instead of alupent (presumably because they sound
alike). These were not handwriting ambiguities; these were cases where
the physician wrote X (intentionally) when what _should_ have been
written was X'.

If the loop were closed, such mixups would happen far less often.

Yes. In the case of handwriting "errors", "closing the loop", by telling
the doctor how the pharmacy actually interpreted his order, is the way
to create an SE for the doctor when there is an OE.

Best

Rick

···

---
Richard S. Marken, Ph.D.
MindReadings.com
10459 Holman Ave
Los Angeles, CA 90024
Tel: 310-474-0313
E-mail: marken@mindreadings.com