State of the Art in Clinical and Anatomic Pathology (Informatics)

Analysis of On-Line Clinical Laboratory Manuals and

Practical Recommendations
Bruce Beckwith, MD; Robert Schwartz, MD; Liron Pantanowitz, MD

● Context.—On-line clinical laboratory manuals are a valu- 8.9 (56%) elements were present (range, 4–14). Basic sam-
able resource for medical professionals. To our knowledge, ple requirements (specimen and volume needed) were the
no recommendations currently exist for their content or elements most commonly present (98% of manuals). The
design. frequency of the remaining data elements varied from 10%
Objective.—To analyze publicly accessible on-line clin- to 90%.
ical laboratory manuals and to propose guidelines for their Conclusions.—On-line clinical laboratory manuals orig-
content. inate from both hospital and commercial laboratories.
Design.—We conducted an Internet search for clinical While most manuals were user-friendly and contained ad-
laboratory manuals written in English with individual test equate specimen-collection information, other important
listings. Four individual test listings in each manual were elements, such as reference ranges, were frequently ab-
evaluated for 16 data elements, including sample require-
sent. To ensure that clinical laboratory manuals are of max-
ments, test methodology, units of measure, reference
range, and critical values. Web sites were also evaluated imal utility, we propose the following 13 data elements be
for supplementary information and search functions. included in individual test listings: test name, synonyms,
Results.—We identified 48 on-line laboratory manuals, test description, test methodology, sample requirements,
including 24 academic or community hospital laboratories volume requirements, collection guidelines, transport
and 24 commercial or reference laboratories. All manuals guidelines, units of measure, reference range, critical val-
had search engines and/or test indices. No single manual ues, test availability, and date of latest revision.
contained all 16 data elements evaluated. An average of (Arch Pathol Lab Med. 2004;128:476–479)

L aboratory medicine is a broad and complex discipline,

involving the measurement of hundreds of different
analytes using many different techniques and instru-
ical laboratory facilities. To this end, on-line laboratory
manuals, which are typically Web-based, have begun to
complement, and even replace, traditional printed ver-
ments. Laboratories have long recognized the need to pro- sions. This evolution allows for more frequent content up-
vide easily accessible collections of information regarding dates, increased information about the laboratory and in-
the testing that they offer.1 This information typically has dividual tests to be included, and greater accessibility.5–7
been published as hardcopy clinical laboratory manuals.2 The College of American Pathologists Laboratory Accred-
These paper manuals have proven to be a valuable re- itation Program mandates that each laboratory create and
source for medical and laboratory professionals.3 Howev- distribute a specimen collection manual.8 However, the
er, their use has been hampered by practical issues, in- program limits its attention to data elements required for
cluding cost, limited space in order to keep the manual proper collection and transportation of samples. While
small enough to be easily used and printed, the long time each on-line manual should be designed to meet the re-
between revisions,4 inadequate distribution, and the in- quirements of each laboratory’s specific circumstances, a
herent limitations of physical items, such as loss or dam- generalized consensus on the data elements that should
age. The move toward centralized testing facilities, on-line be present for each test would enhance their utility as a
order entry, and networked laboratory information sys- clinical reference. This study aimed to evaluate publicly
tems necessitates easily accessible and up-to-date testing accessible on-line clinical laboratory manuals and to pro-
information for both hospital-based and commercial clin- pose guidelines for their content.
MATERIALS AND METHODS
Accepted for publication December 8, 2003.
We conducted an Internet search for publicly accessible clinical
From the Departments of Pathology, Beth Israel Deaconess Medical
Center and Harvard Medical School, Boston, Mass.
laboratory manuals written in English with individual test list-
Presented as an electronic poster at the Advancing Pathology Infor- ings. Our goal was to find on-line laboratory manuals from each
matics, Imaging, and the Internet (APIII 2002) conference, Pittsburgh, of the following laboratory subtypes: academic/teaching insti-
Pa, October 2–4, 2002. tutions, commercial/reference laboratories, and community hos-
The authors have no relevant financial interest in the products or pitals. Searching was done by a combination of Internet search
companies described in this article. engines, such as google.com, and also targeted manual searching
Reprints: Bruce Beckwith, MD, Department of Pathology, Beth Israel of Web sites that were considered likely to host a publicly acces-
Deaconess Medical Center, 330 Brookline Ave, Boston, MA 02215 sible laboratory manual (eg, labcorp.com). Laboratory manuals
(e-mail: brucepbeckwith@bidmc.harvard.edu). that were clearly intended to be used only for research purposes
476 Arch Pathol Lab Med—Vol 128, April 2004 On-Line Clinical Laboratory Manuals—Beckwith et al
and manuals that were produced by institutions not providing Table 1. General Features Available in On-Line
clinical laboratory services for human patients were excluded. We Clinical Laboratory Manuals
did not include sites targeted only at the general public, such as
labtestsonline.org. In addition, manuals that were clearly identi- No. (%) of Manuals
fied as being incomplete or ‘‘under construction’’ were excluded. Feature Containing Feature
Each laboratory manual Web site was analyzed for 5 general Laboratory contact information 47 (98)
features, including the ability to perform a text search for specific Alphabetic listing of tests 41 (85)
tests, presence of an alphabetic list or index of available tests, a Free text search for tests 36 (75)
listing of the laboratory’s accreditations, contact information for Basic specimen-collection
the laboratory, and basic-specimen collection information and information 28 (58)
guidelines (eg, how to obtain a sputum sample). In addition, 4 Accreditation information 26 (54)
specific tests (serum sodium, glycated hemoglobin, prolactin, and
human immunodeficiency virus [HIV] antibody) in each manual
were evaluated for the presence of individual data elements. So-
dium was chosen because it is a commonly ordered test, and Table 2. Presence of Data Elements in Individual Test
therefore most laboratories should have an assay to measure so- Listings
dium. In addition, it is a test that commonly has critical values, No. (%)
that is, levels above and below which the laboratory may contact of Manuals
the ordering provider so that a potentially life-threatening value Containing
may be acted on in a timely fashion. Glycated hemoglobin was Data Element Element*
chosen because there are a variety of synonyms for this test, such
as hemoglobin A1c and HbA1c. Prolactin was selected to repre- Type of sample needed 47 (98)
sent a less common assay and one for which additional clinical Specimen volume required 47 (98)
Test availability or turnaround time 43 (90)
and analytic information might be present in the laboratory man-
Test name synonyms (hemoglobin
ual (eg, reasons for ordering this test or an explanation of ma- A1c only) 39 (81)
croprolactinemia). Lastly, HIV antibody was chosen since there Reference range 38 (79)
are a number of ethical, logistic, and patient confidentiality issues Units of measure 37 (77)
associated with this test (eg, specific HIV consent forms may Current Procedural Terminology code 31 (65)
need to be signed by the patient prior to performing this test). Test methodology 29 (60)
As such, the HIV antibody entry was anticipated to have addi- Transportation information 29 (60)
tional information and possibly links to other Web pages regard- Specimen collection information 18 (38)
ing ordering and result interpretation. Critical values (sodium only) 18 (38)
The 4 specific test listings selected were each evaluated for the Hyperlinks (any test) 15 (31)
presence of the following 16 data elements: test description, syn- Date of latest revision 14 (29)
onyms, test methodology, sample type, specimen volume require- Test description or explanation 9 (19)
ments, collection guidelines, transport guidelines, units of mea- Contact information for individual test 7 (15)
sure, reference range (including age and sex differences), critical Literature references (any test) 5 (10)
values, test availability or turnaround time, interpretative infor- * Unless otherwise indicated, data elements had to be found in 3 of
mation (including literature references), hyperlinks to additional the 4 test listings evaluated in each manual in order to be scored as
information, Current Procedural Terminology codes, contact infor- present.
mation for the individual test, and the date that the specific test
information was last updated. Each of these 16 data elements was
scored as present or absent in the laboratory manual according had functional search engines, and 41 (85%) had an al-
to the following criteria. Most data elements were scored as pre- phabetic listing of available tests. Forty-seven (98%) in-
sent if at least 3 of the 4 specific test listings for each laboratory cluded contact information, 28 (58%) provided general
manual included them. However, the synonym category was specimen-collection information, and 26 (54%) indicated
scored as present if there were any synonyms listed for ‘‘glycated
hemoglobin’’ (HbA1c, A1c, etc), since we felt that this was the
their accreditations.
only test of the 4 that was likely to have synonyms listed. Also, All 48 manuals had listings for the 4 tests we specifi-
listings for sodium were used to score critical values, as the other cally targeted for evaluation and were suitable for inclu-
tests were deemed unlikely to have critical values defined. Web sion in our study. Analysis of the specific test listings (Ta-
links or literature references were scored as present if they were ble 2) showed that no single laboratory manual contained
present in any specific test listing. Any additional or unique fea- all 16 data elements that we evaluated. A mean of 8.9
tures specific to the various manuals were also recorded. Anal- (56%) data elements per manual was present (range, 4–
ysis of the data elements in each laboratory manual was per- 14). There was only a minimal difference in the number
formed using Microsoft Excel (Microsoft Corporation, Redmond, of data elements present when the laboratory manuals
Wash).
were divided according to the type of laboratory. Among
the 24 academic or community hospital laboratories
RESULTS
found, the average number of data elements present was
We identified 48 publicly available on-line clinical lab- 9.2, and among the 24 commercial or reference laborato-
oratory manuals written in English, of which 24 originated ries, the average number was 8.5. The 2 data elements re-
from hospital-based laboratories (21 were academic insti- lated to sample requirements (type of specimen and vol-
tutions and 3 were community hospitals) and 24 from ume required) were the most common, with 47 (98%)
commercial or reference laboratories that appear to pro- manuals containing them. The turnaround times for spe-
vide primarily outpatient testing services to a wide variety cific tests were present in 43 manuals (90%), while almost
of hospitals and clinics. Forty-two (88%) of the 48 manuals 80% of manuals provided reference ranges, units of mea-
were from institutions located in the United States. Two sure, or test name synonyms. Data elements with partic-
were from Canadian institutions, 3 from Australia or New ularly low representation included critical values for so-
Zealand, and 1 from Hong Kong. Our analysis of general dium (18 manuals [38%]), links to related information (15
features (Table 1) revealed that 36 (75%) of the 48 manuals manuals [31%]), an explanation or medical significance of
Arch Pathol Lab Med—Vol 128, April 2004 On-Line Clinical Laboratory Manuals—Beckwith et al 477
a specific test or analyte (9 manuals [19%]), and the date complete on-line laboratory manual. Ideally, an on-line
the Web page was last updated (14 manuals [29%]). The laboratory manual should draw whatever information is
frequency of the remaining data elements varied from 10% possible from the LIS in real time, so that a secondary
to 65%. database will not need to be created and kept synchro-
Finally, several notable features were found in individ- nized. Multiple laboratory manuals have been implement-
ual laboratory manuals, including sample-collection infor- ed using this design strategy of drawing information di-
mation with supporting images or graphics, color photo- rectly from the LIS6 and other networked resources.7 Re-
graphs of sample collection tubes, sample rejection crite- cently, an extensible markup language (XML) syntax for
ria, information regarding the effect of physiologic states clinical laboratory procedure manuals was proposed,9
(eg, pregnancy) on reference ranges of various analytes, which, in addition to test method details, includes infor-
printable HIV consent forms, and downloadable Adobe mation relevant to physicians and patients, and might
portable document format (PDF) and Palm personal dig- prove useful in implementing laboratory manuals de-
ital assistant (PDA) versions of the laboratory manual. signed for health care providers outside the laboratory.
In 2000, we implemented a Web-based on-line labora-
COMMENT tory manual at our institution (http://home.caregroup.
We identified 48 laboratory manuals that met our cri- org/departments/pathology/). The design was guided by
teria for inclusion in the study, including the laboratory our desire to make each test listing brief, yet still provide
manual at our own institution. We suspect that there are sufficient information for clinicians to guide their test se-
many more on-line laboratory manuals that were not lection, sample collection, and test result interpretation.
available to us owing to restrictions on outside access or We use a homegrown LIS that has no provision for storing
to our choice of search phrases and search engines. How- the information needed for a laboratory manual, nor for
ever, we believe that our sample was sufficient for our pur- real-time links to the data tables containing test informa-
poses. About half of the on-line clinical laboratory man- tion. As a result, we designed and implemented a Micro-
uals originated from laboratories affiliated with academic soft SQL Server database to serve as a repository of this
medical institutions. information. We have entries for approximately 500 tests
All except 1 of the manuals listed information necessary in our database. We then created a Web site that pulls the
for adequate specimen collection. However, interpretive test-specific information from the database using active
and educational information about the tests offered was server pages. A large number of static Web pages were
not a prominent feature of most Web sites. Specifically, also created, which contain general information about the
data elements of potential medical importance, such as ap- laboratory, methods of specimen collection (ranging from
propriate reference ranges and critical values (for sodium), venipuncture to eye cultures), and expanded information
were found to be absent in several manuals. Educational regarding specific tests. This format has served us well
data elements, including Web links to additional infor- and provides the flexibility to continually and easily tailor
mation or resources and literature references were absent the content and format to meet our needs. However, it
in the majority of manuals. The lack of such important does have the drawback of having to maintain some test
data elements in the manuals we evaluated may be related information (eg, units and reference ranges) in 2 places—
to these manuals being designed for use prior to or during the LIS and the laboratory manual database.
sample collection, and not as a clinical resource that might Two other design elements merit discussion. First, we
help health care providers understand or interpret test re- decided to formulate all of the content of our manual in-
sults or select additional appropriate tests. While sample house. Another option would have been to engage a com-
collection is obviously an important function of a labora- mercial vendor, such as Lexi-Comp (Hudson, Ohio). For
tory manual, we believe that information supporting test many years, this company provided a service to create
selection and result interpretation is equally important. customized hardcopy and electronic laboratory manuals
This is likely to become more pertinent with the increasing that were based on the information in their Laboratory Test
use of Web-based order entry and laboratory result re- Handbook Concise.10 Although Lexi-Comp no longer active-
porting, both of which offer additional entry points into ly pursues custom laboratory database projects, our un-
the laboratory manual. Fewer than half of the manuals derstanding is that they may still accept a few new labo-
provided hyperlinks to additional information in the in- ratories for such development work. The second consid-
dividual test listings. One of the great strengths of the eration is whether the resulting laboratory manual will be
World Wide Web format is that voluminous supporting freely accessible via the Internet or restricted in some way
information can be made available simply by creating hy- (eg, password protected or limited to local intranet users).
perlinks. While laboratories may not have the resources Since many of our laboratory clients are located off-site
or desire to generate their own supporting material, many and may not have easy access to our intranet, having our
reliable sources of medical information are currently avail- laboratory manual freely available on the Internet made
able on the Internet. The use of appropriate hyperlinks sense for our situation. When making a decision about
would allow the main test listing to remain brief and un- access, laboratories will need to consider their local cir-
cluttered, and in our opinion, provides a means for inter- cumstances, including whether a secure intranet infra-
ested parties to further their understanding. structure is already in place and available to all of their
The inclusion of supporting information in a manual users.
brings up an important design consideration. Laboratory Currently, health care providers at our institution use a
information systems (LISs) contain much of the relevant secure, Web-based, on-line provider order entry system to
test information, such as units, reference ranges, and syn- enter almost all orders on inpatients, including laboratory
onyms for ordering. However, most of these systems do tests, radiologic examination, and pharmaceutical orders.
not have the capability of storing, formatting, and dis- When selecting laboratory tests, the providers at our med-
playing all of the various data elements required for a ical center have the ability to click on the name of a de-
478 Arch Pathol Lab Med—Vol 128, April 2004 On-Line Clinical Laboratory Manuals—Beckwith et al
sired test in order to open a new browser window, which Table 3. Minimum Recommended Data Elements for
then displays the relevant test listing from our on-line lab- Individual Test Listings
oratory manual. This ability to link to our laboratory man-
Test name
ual gives us an opportunity to provide clinically relevant Synonyms
information, such as expected turnaround time for a given Test description
test, at a point in the patient’s care when it may be par- Test methodology
ticularly helpful to the clinician. Moreover, within our se- Sample type
cure Web-based inquiry system for health care providers, Specimen volume requirements
Specimen collection guidelines
individual result names are again linked to their corre-
Transportation information
sponding listing in the laboratory manual. Thus, our cli- Units of measure
nicians are able to access specific test information on-line, Reference range
which includes a contact person, at their convenience, 24 Critical values (if established)
hours a day. Our institution also has a secure on-line sys- Test availability or turnaround time
tem designed for patients to use, which displays different Date of latest revision
types of information drawn from their electronic medical
record, including laboratory test results. As mentioned,
our laboratory manual is available on the Internet, but portation information, units of measure, reference range,
since our test-specific information is designed for use by critical values (if established), test availability or turn-
health care providers, we have not created links from our around time, and the date this information was last re-
patient-based system into our laboratory manual. How- vised.
ever, we do provide links to another Web site designed to Additional information that can be profitably included
educate the public regarding laboratory testing. For labo- in the test listings includes hyperlinks to further educa-
ratories with the resources and desire to customize their tional material, literature references, expanded explana-
information for use by patients, systems such as this might tions of the medical significance of test results, sample
represent another venue in which to provide accurate and rejection criteria, potential sources of interference, Current
informative information regarding laboratory testing. Procedural Terminology codes, and relative cost information.
While the addition of these or other elements may enhance
RECOMMENDATIONS a clinical laboratory manual, they may also increase the
length of each entry. To keep the manual user friendly, we
Our study indicates that the majority of publicly avail- recommend that the individual test listings remain brief
able on-line clinical laboratory manuals are useful for ob- and that hyperlinks be used if extensive supporting in-
taining information needed for the collection, transporta- formation is provided.
tion, and analysis of patient samples in a given laboratory. References
However, their potential value as a reference and educa- 1. Grams R. Clinical laboratory test reference (CLTR). J Med Syst. 1993;17:59–
tional resource for medical and laboratory professionals in 67.
2. Nordenson N. How to plan and produce your laboratory test catalog. MLO
selecting or interpreting laboratory tests requires the in- Med Lab Obs. 1998;30:23–26.
clusion of more comprehensive and standardized infor- 3. Hardy R, Grubbs F. Revised laboratory manual helps solve problems. Am J
mation. Med Technol. 1976;42:441–444.
4. Croft S, Koehler M. New concept: a lab manual that withstands changing
To be useful to health care providers, general features times. Am J Med Technol. 1978;44:1090–1092.
of a clinical laboratory manual should include an alpha- 5. Lazinger B, Steif J, Granit E. An online tests catalog for clinical laboratories.
betic listing of tests and a search engine to facilitate find- J Med Syst. 1989;13:187–192.
6. Bennett S, Kern D. Automated production of an on-line laboratory reference
ing individual test listings, as well as basic specimen col- manual from a laboratory information system. J Med Syst. 2002;26:145–149.
lection and transportation information pertinent to that 7. Fine J, Ching A, Schneider J, Pollum D, Astion M. BibleCard: network-based
laboratory. The core of any laboratory manual is the in- virtual database for laboratory information. Clin Chem. 1995;41:1349–1353.
8. Laboratory general checklist: July 2003 version. College of American Pa-
dividual test listings, and for these listings we propose thologists. Available at: http://www.cap.org/apps/docs/laboratorypaccreditation/
that the following 13 data elements (Table 3) be included, checklists/laboratorypgeneralpjuly2003.pdf. Accessed November 28, 2003.
as a minimum, for each test: test name, synonyms, test 9. Saadawi G, Harrison J. An XML syntax for clinical laboratory procedure
manuals [abstract]. Arch Pathol Lab Med. 2003;127:809.
description, test methodology, sample type, specimen vol- 10. Laboratory Test Handbook Concise. 2nd ed. Hudson, Ohio: Lexi-Comp
ume requirements, specimen collection guidelines, trans- Inc; 2002.

Arch Pathol Lab Med—Vol 128, April 2004 On-Line Clinical Laboratory Manuals—Beckwith et al 479