It was a frustrating situation. From 1979 to 1982, the cost of
reference laboratory service to our 340-bed acute-care hospital rose
more than 123 per cent. In each of those years, we had negotiated
successively deeper discounts with our principal reference lab. Yet our
reference costs were spiraling despite an actual drop in cost per test,
as shown in Figure I.
With New York State’s own complex version of prospective
payment on the horizon, we knew we had to do something. Reference work
was growing at a much faster rate than in-house testing, which rose some
12 per cent a year. Send-outs constituted less than 1 per cent of the
laboratory’s total volume, while eating up roughly 10 per cent of
its gross costs. Furthermore, these fees were almost never fully
reimbursed, dollar for dollar, by third-party payers.
We began to look for strategies to cut reference testing costs and
improve utilization. Our first step was to switch to a reference
laboratory that offered us a better deal. At the time, the hospital lab
was actively involved with the Joint Purchasing Corporation of New York,
an intermediary organization for hospital service providers, in
developing a group buying contract with major reference laboratories.
Under the JPC’s guidance, a committee of physicians, lab
administrators, and clinical chemists developed criteria, terms and
conditions for a proposed contract.
The effort paid off in early 1983, when a number of area hospitals
signed a highly economical group purchasing contract with one major
regional reference lab. This contract alone, with its 60. per cent
discount off list price, would have reduced our reference test costs for
the year, even without a cut in send-out volume.
Utilization control, our second tactic, began with a thorough
review of ordered reference tests and studies. It was soon clear that
our log books were not organized to provide easy access to the data we
If our busy pathologist/laboratory director was to participate
fully in the review process, we would need a mechanism for retrieving
data rapidly in a variety of formats. Under our existing system, one
had to flip through a sheaf of test request slips to assemble meaningful
information on individual physician ordering patterns, including
redundant, excessive, or inconsistent orders, and other possible
A microcomputer appeared the most likely solution to storing,
organizing, and printing our reference test data. In late 1982, we
purchased a Radio Shack Model II, and began recording data in the
beginning of 1983. Once the system was implemented, it involved no
additional clerical work, since the computer took the place of our
regular manual log of reference tests. As for software, any good data
base management system or flexible filing program that allows the user
to sort data by various parameters is adequate for this purpose.
During the first month of computer monitoring, we did nothing out
of the ordinary to influence the ordering behavior of attending,
consulting, or house staff physicians. When the month ended, we
prepared a cumulative printout. To identify ordering patterns or heavy
reference lab users, the report arranged the send-out list by
physicians’ last names, and separated the house staff orders from
those of attending and consulting physicians. Figure II shows an
abbreviated version of one of these reports.
In one stroke, the neatly labeled printout solved two of our main
hurdles: It clearly showed the volume and test mix of each physician
using reference lab services, and capsulized the extent and costs of
such testing in an easily readable form.
We presented a copy of the report at the next meeting of the
medical board. Attendees received it with interest and recognized the
need to cut these costs. We deliberately refrained, however, from using
the meeting as a forum to establish ordering policies or guidelines.
Instead, we opted for an approach that we believe is the key to the
program’s success. Rather than make a unilateral attempt to change
ordering habits, we focused on offering the laboratory’s services
for test ordering consultation. After the meeting, we discussed the
data with the director of medicine and various attending phsicians,
individually or in small groups. We suspected from a review of the
literature that this one-on-one method would prove more effective than
any mandated change. So far, experience has proven us right.
Over the next two months, all reference tests requests–some 10 per
day on average–were entered into the computer, a task that takes about
10 minutes daily. Each day’s activities were printed in accession log format and reviewed by the laboratory manager, who alerted the
director to any unusual or unlikely-looking requests. The director then
called physicians immediately to check out such orders.
These calls, which usually took about two minutes, stressed the
following points: the clinical need for the test ordered; less costly
alternatives, where appropriate, that yielded equal or better
information than the test requested; correction of occasional clerical
or other errors; and cost containment, a topic discussed frequently with
clinicians to reinforce their sense of responsibility in this area.
In some 90 per cent of all cases where the laboratory director
intervened, the ordering physician voluntarily canceled or changed the
test in question. The other 10 per cent justified their choices or
simply remained adamant. Almost every cancellation due to erroneous ordering resulted from one of the following: improper test selection by
inexperienced house staff; improper clerical transcription of
physicians’ orders; or simple lack of knowledge of a test’s
medical value in relation to its cost-effectiveness.
In the following months, referred laboratory fees began to drop.
Part of the savings arose from the new reduced fee schedule, of course.
But the actual number of tests ordered in the survey months also
declined slightly from the same period in the previous year. In
addition, test mix appeared to be shifting somewhat from very costly
procedures to more moderately priced tests.
The monthly summary serves as a guide for the laboratory
director’s review and follow-up, while the laboratory manager uses
it for statistical and record-keeping purposes. These concise reports,
which present data in several formats, allow selective intervention when
The computer can produce reports to monitor compliance with
laboratory guidelines or identify specific procedures for review. We
soon learned, for example, that physicians were routinely ordering
several different panels, containing some expensive and often redundant
tests, to rule out pheochromocytoma on hypertensive patients. The
director discussed alternative diagnostic approaches with clinicians
most likely to use the procedures. As a result of these discussions, we
selected VMA and total metanephrines as our panel of choice for this
screening, and the hospital’s physicians readily accepted it. We
discovered and stopped other irregularities as well–such as extensive
esoteric tests ordered by a staff physician out of intellectual
The laboratory director’s close involvement is the most
important element in our utilization program. His status as a physician
and colleague enables him to present alternate choices to the medical
staff without arousing resistance or resentment; he has the best
information at hand, and the patient’s welfare in mind. Most
clinicians, in fact, have been grateful for “free
consultations” to learn the latest consensus on cost-effective
The computer-generated reports, in turn, have become the
laboratory’s most valuable utilization tool. They provide a vital
element of objectivity to discussions on a highly sensitive subject.
These reports make clinicians instantly aware of their reference
laboratory ordering activities for the entire month, and remind them
that an efficient monitoring method is in ongoing use. Since the
computer performs all the searches, sorts, and collations, the
pathologist can afford the time for regular participation, which usually
requires no more than five to 10 minutes a day.
The results have been impressive. In the first year, the
combination of our computerized audit and group buying contract reduced
reference testing costs by 31 per cent, or $27,000 less than our 1982
The laboratory and the hospital have done more than save money. We
have improved service and enhanced diagnostic efficiency. Our review
system stimulates consultation between the pathologist and the ordering
physician, resulting in performance of tests that gives us the most for
our money. Everyone gains–the hospital, the physician, and the