Dear Dr. McGing:

On several occasions, I have done such cost-per-test calculations, and my
conclusion is that there are numerous ways to approach the problem. Each
method has its particular strengths and weaknesses, and none is without some
degree of ambiguity related to essential estimations. Let me address your
specific questions, and then I will comment on general approaches to cost
analysis.

> 1) are labs using specific computer packages?

I'm not familiar with any of these, but they probably exist. I would be very
careful, though, about using any program if you do not understand explicitly
how the calculations are being made. Like statistics, cost-per-test
estimates can be revealing, helpful, or utterly useless and misleading. The
key to making that distinction is understanding exactly how the numbers are
calculated. Off-the-shelf software rarely reveals such detail.

> 2) are there any recent publications on this matter which members have
> seen?

Once again, I don't know of any, but they probably exist.

> 3) in respect of charging for tests do labs base this entirely on
> calculated cost or are 'historical' factors and/or commercial lab
> charges taken into account?

Yes, yes, and yes. For contract work, we base our charges on both cost
analysis and market price (charges by commercial labs). Like commercial
labs, pricing may include some "loss leaders," in order to attract more
lucrative tests. In general, our charges are determined by (1) compliance
with federal and state regulations, (2) reimbursement, (3) competetion, and
(4) direct costs, in that order.

As for calculating costs per test, I have always considered two types of
expense associated with producing a laboratory result: Incremental cost and
non-incremental (fixed) cost. The former increases with each test added,
whereas the latter remains the same, at least within certain limits of
workload. The "devil in the details" emerges when you try to assign line
items in your budget to one or the other category. Incremental costs clearly
include reagents, but a certain component of reagent cost is not
incremental--QC and calibration frequency do not ordinarily increase with
higher test volume. Many disposables are non-incremental, as well, such as
gloves and wipes.

Staffing costs are the largest part of the laboratory budget, so how they
get assigned will make a huge difference in the cost-per-test estimate. Most
accountants with whom I've debated this poing insist that five minutes (or
whatever) of technologist time required to perform a laboratory test is an
incremental expense corresponding to five minutes' worth of salary and
benefits. I have always argued that, unless you need to hire an additional
technologist to accomodate extra workload, staffing expenses are fixed, and
should not be included as an incremental expense. It does not cost the
laboratory five minutes' worth of a technologist's salary to perform one
more glucose test. Whether or not you include staffing as an incremental
cost depends, mostly, on the intended use of the cost-per-test estimate. If
you are justifying your budget based on workload volume, it usually favors
you to include staffing as an incremental cost. If you are trying to
establish a competitive price for a test, it is better to regard staffing as
a fixed cost, as long as your workforce is sufficient to meet needs.

Once you decide on a formula for assigning incremental and fixed costs, you
may need to allocate fixed costs by a formula that takes into account the
fact that some tests are more resource-intensive than others, although this
is not always necessary. For example, if your hospital is merging with
another, similar, hospital and you want to assess the impact on your
laboratory budget, you may be able to assume that the additional work will
include essentially the same mixture of laboratory tests that you are
already performing, so an overall average cost-per-test is sufficient to
predict the budgetary impact. However, if you are adding workload that
disproportionately includes more labor-intensive tests, you need to know the
cost of the individual tests. Several schemes exist for assigning relative
"weights" to individual tests: (1) You can do your own process analysis to
assess the relative drain on resources that each individual test requires,
(2) The College of American Pathologists used to publish "Workload Units"
that provided a measure of the time required to perform individual tests, or
(3) we often use "Relative Value Units," upon which Medicare (the federal
health insurance program) reimbursement is based.

I've found the process of estimating costs per test a tedious, but
ultimately rewarding, exercise. I wish you luck.

Kind regards,

Roger

Roger L. Bertholf, Ph.D.
Associate Professor of Pathology
Chief of Clinical Chemistry & Toxicology
University of Florida Health Science Center/Jacksonville