#29 LEAN in The Lab 2

THE CLINICAL LABORATORY AND OUTCOMES
Your laboratory can play an important role in improving

efficiency, lowering costs, increasing patient and physician
satisfaction as you participate in outcome studies. For our
discussion an outcome is the end result from an action (or lack of
action). For example, the statement, “If treated, a cold will go away
in 7 days,” is based on an outcome study. Every outcome study has
variables. This statement has four variables – treated, cold, go
away and 7 days.
Treated – with chicken soup? with aspirin? with tender loving

care and bed rest? are four possible criteria for “treated.”
Cold – by whom is ‘cold’ defined? what criteria are used to

define ‘cold?’
Go away—meaning the person can return to work full time?

part time? feels ‘better?’
7 days – not 6? or 8. But 168 hrs!? no wiggle room?
Before an outcome study can begin, the variables must be listed

and defined. It is easier and more efficient to consider all possible
variables and discard some rather than finish the study and wish
that a variable that would have been of value had it been included
was not.
In our work there are several outcomes

1. Physician sends patient home
2. Physician admits patient
a. To observation bed
b. To medical ward or similar
c. To ICU or CCU, etc.
3. Diagnostic tests ordered
4. Treatment ordered
a. More
b. Less
5. Biopsy performed
6. Radiographic scan performed
7. Ultrasound performed
The laboratory can assist the clinical staff in studying the outcomes
from the use of (or not) of diagnostic tests; in studying the
outcomes of draw times and turnaround times, as well as length of
stay, triage of patients, to name a few.
Let’s step through a generic outcome study to see the sequence and
what are some of the pitfalls to try to avoid. In designing the
outcome study, it is probably wise to include representatives from
not only the laboratory (including a pathologist) but clinicians,
nurses, pharmacy (when drugs are involved) and at least at one
stage someone from the financial side and even a patient
ombudsman and perhaps clergy. Each of these need not be present
at all meetings if more than one meeting is needed. The goal is to
cover all the aspects of the question before the study starts. What
follows is a list of parameters to be discussed. After the list, each is
discussed in a bit more detail.
1. Determine the measure(s) of interest
2. Define the patient population to include
3. Determine the parameters needed for the measurement of

interest.
4. Collect the data
5. Analyze the data
6. Interpret the data
7. Consider the impact of the data.
8. Make the changes needed to improve the outcome(s)
9. Evaluate the effect of the change(s)
Now let’s discuss each of these in some detail:
1. Determine the measure(s) of interest

Some outcome studies in which the laboratory could assist
include; 1)measure turnaround time for cardiac markers; 2)
the utility of d-dimer in the patient with chest pain, 3) which
test(s) are ‘best’ for detecting DIC, 4) is white count better
than body temperature to detect an infection, 5) what cut-off
to use for say troponin or PSA or d-dimer. These are only a
few of many. The outcome to be studied must be precisely
defined – what are the criteria for the outcome. For example Is
it length of stay? Or is this test more accurate in aiding in the
diagnosis of ___ than that test?
2. Define the patient population to measure
The population might be patients in the ED, or more narrowly,

patients in the ED older than 45; or persons at a health fair.
3. Determine the parameters needed for the measurement of

interest.
In this case it might be better to have more than to get to data

analysis and say “Oh, I forgot to get the age of the patients.”
To help in determining the data you will want to use, ask what
are the possible variables. In thinking about testing for
anemia for example would you include a CBC, a serum iron, an
IBC or TIBC or ferritin and sex and age? Would you include
certain medications (e.g. Vitamin B-12)? Stop and ask have you
left out something that you might later wish you had chosen to
collect.
4. Collect the data. Organize it appropriately. Back it up.

Think of a table in a spread sheet (e.g. Excel or Access).
5. Analyze the data. How will you analyze the data? In other
words, what statistics will you use? It is not appropriate to
analyze the data with one statistic only to find that that
statistic did not let you state what you wanted as your
outcome and then go back and redo the analysis with another
statistic. Not only should the statistic be chosen before
analysis but the level of confidence (e.g. 95%) must be chosen.
6. Interpret the data

This can be straight forward or a real disappointment if you
look at the data and say “OMG where are the ____ (e.g. ages,
weight, X-rays, etc, etc)?! The data must be interpreted in the
light of the criteria set for the outcome and the statistic and
confidence limit you chose.
7. Consider the impact of the data.
It is possible that your study showed that dropping CK-MB

from the rule out MI algorithm had no effect on the outcome.
The next question is how best to eliminate this test. What will
you use as your argument? Time? Time and money? Turn
around time? How are the doctors to be informed of this
suggestion? Then ask, How shall it be dropped – today or on
some day in the future? Shall there be a discussion of the data
amongst the staff?
8. Make the changes needed to improve the outcome(s)
9. Evaluate the effect of the change(s)
Does the change have to be modified? Revoked? Can this

change lead to other changes that will further improve
outcomes? If so start a new study.
Let me present three sketches of published outcome studies in

which the laboratory played a significant role.
A. In a study of 1285 patients presenting to the ED at the VA

Hospital in Davis, CA, three cardiac markers were measured:
CK-MB, cTnI and myoglobin at 0, 45 and 90 min after
presentation to the ED. The study found that 90% of the
patients of those with negative sets of markers and a non-
diagnostic ECG could be discharged home within 90 minutes of
presentation. In addition to the outcome measure of discharge
time, the study also found that there was s substantial cost
saving and increased patient satisfaction.
B. This second outcome study was based a questionnaire given to

a group of some 700 member of an HMO. Of those who
responded, 33% (n = 63) reported that they were diabetics
who had been tested less than twice for their A1c level in the
last year. Additionally 49% were unaware that A1c levels are
recommended; 38% were not informed of the need for the test
by their physician, 33% had never heard of the A1C test , and
19% were not seen regularly by their physician. These
outcomes should certainly lead to several changes in patient
education and even physician awareness.
C. In a study of over 900 consecutive patients presenting to the

ED with suspected venous thromboembolism (DVT/PE), the
availability of a quantitative d-dimer assay was found to
reduce the number of pulmonary angiographies by 31%. In this
case the outcome (reduction of angiographies resulted in
saving of both time and money for the hospital, length of stay
for the patient with increased patient satisfaction.
These three examples are only to give you ideas of how the
laboratory can interface with other departments in improving
many aspects of patient care.

#29 LEAN in The Lab 2

Uploaded by

Copyright:

Available Formats

#29 LEAN in The Lab 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

#29 LEAN in The Lab 2

Uploaded by

Copyright:

Available Formats

THE CLINICAL LABORATORY AND OUTCOMES

Your laboratory can play an important role in improving

Treated – with chicken soup? with aspirin? with tender loving

Cold – by whom is ‘cold’ defined? what criteria are used to

Go away—meaning the person can return to work full time?

7 days – not 6? or 8. But 168 hrs!? no wiggle room?

Before an outcome study can begin, the variables must be listed

In our work there are several outcomes

2. Physician admits patient

b. To medical ward or similar

c. To ICU or CCU, etc.

3. Diagnostic tests ordered

6. Radiographic scan performed

1. Determine the measure(s) of interest

2. Define the patient population to include

3. Determine the parameters needed for the measurement of

4. Collect the data

5. Analyze the data

6. Interpret the data

7. Consider the impact of the data.

8. Make the changes needed to improve the outcome(s)

9. Evaluate the effect of the change(s)

Now let’s discuss each of these in some detail:

1. Determine the measure(s) of interest

2. Define the patient population to measure

The population might be patients in the ED, or more narrowly,

3. Determine the parameters needed for the measurement of

In this case it might be better to have more than to get to data

4. Collect the data. Organize it appropriately. Back it up.

6. Interpret the data

7. Consider the impact of the data.

It is possible that your study showed that dropping CK-MB

8. Make the changes needed to improve the outcome(s)

9. Evaluate the effect of the change(s)

Does the change have to be modified? Revoked? Can this

Let me present three sketches of published outcome studies in

A. In a study of 1285 patients presenting to the ED at the VA

B. This second outcome study was based a questionnaire given to

C. In a study of over 900 consecutive patients presenting to the

You might also like