CAP Laboratory Improvement Programs
Laboratory Critical Values Policies and Procedures
A College of American Pathologists Q-Probes Study in 623 Institutions
Peter J. Howanitz, MD; Steven J. Steindel, PhD; Nan V. Heard, MD
● Context.—Critical values lists have been used for many
years to decide when to notify physicians and other care-
givers of potentially life-threatening situations; however,
these lists have not been studied widely.
Objectives.—To investigate critical values lists in insti-
tutions participating in the College of American Patholo-
gists Q-Probes program and to provide suggestions for im-
provement.
Setting.—A total of 623 institutions voluntarily partici-
pating in the Q-Probes program.
Design.—A multipart study in which participants re-
sponded to information from preprinted lists, collected in-
formation about current practices, completed a question-
naire, monitored critical values calls, reviewed patients’
medical records, and surveyed nursing supervisors and
physicians about critical values.
Main Outcome Measures.—Defining critical values sys-
tems, including lists, personnel, costs, processes, useful-
ness, and related medical outcomes.
Results.—Critical values lists were determined for routine
chemistry and hematology analytes and were found to vary
widely among participants. In contrast, more than 95% of
participants reported positive blood cultures, cerebrospinal
fluid cultures, and toxic therapeutic drug levels as critical
values. Based on more than 13 000 critical values, partici-
R ecently, laboratorians have begun to evaluate clinicians’
requirements for rapid availability of test results and
have developed and implemented strategies to improve lab-
oratory test turnaround times.1 Although most clinical lab-
oratory results have diagnostic and therapeutic implications
that do not require urgent physician attention, occasionally
laboratory results are so extreme that they may indicate a
potentially lethal situation requiring immediate notification
and action by the physician. Lundberg2,3 was the first author
to identify the value of rapid reporting. He developed a
system in which clinical laboratory personnel reported
these values immediately to appropriate individuals, so that
Accepted for publication January 10, 2002.
From the Department of Pathology, State University of New York,
Downstate Medical Center, Brooklyn, NY (Dr Howanitz); the Public
Health Practice Program Office, Division of Laboratory Systems, Cen-
ters for Disease Control and Prevention, Atlanta, Ga (Dr Steindel); and
the Department of Laboratory Medicine, West Los Angeles Veterans
Administration Medical Center, West Los Angeles, Calif (Dr Heard).
Reprints: Peter J. Howanitz, MD, Department of Pathology, Box 25,
State University of New York, Downstate Medical Center, 450 Clarkson
Ave, Brooklyn, NY 11203 (e-mail: phowanitz@netmail.hscbklyn.edu).
Arch Pathol Lab Med—Vol 126, June 2002
pants’ data showed that most critical values reports (92.8%)
were made by the person who performed the test, and that
65% of reports for inpatients were received by nurses. For
outpatients, physicians’ office staff received the largest per-
centage (40%) of reports. The majority of participants
(71.4%) had no policy on how repeat critical calls should
be handled. On average, completion of notification required
about 6 minutes for inpatients and 14 minutes for outpa-
tients. Slightly greater than 5% of critical value telephone
calls were abandoned, with the largest percentage aban-
doned for outpatients. More than 45% of critical values
were unexpected, and 65% resulted in a change in therapy.
Although only 20.8% of 2301 nursing supervisors thought
critical values lists were helpful, 94.9% of 514 physicians
found critical values lists valuable.
Conclusions.—Critical values systems were medically
important, highly variable, but also costly practices for par-
ticipants. We propose a number of recommendations for
improvement, including that the critical values list should
be approved by the medical staff, each laboratory should
develop a written policy for handling initial and repeat crit-
ical values reports, a foolproof policy should be established
to report results from calls abandoned, and efforts at au-
tomating the process should become widespread.
(Arch Pathol Lab Med. 2002;126:663–669)
suitable treatment could occur quickly. As part of his pro-
cedure, Lundberg coined the term critical (panic) values for
these results, defining them as representing a pathophysi-
ological state at such variance with normal as to be life-
threatening unless something is done promptly, and for
which some corrective action could be taken.2,3
Soon after Lundberg’s classic observations in 1972,
many laboratories adopted his critical value concept and
its use became widespread throughout the world.4,5 Ulti-
mately, accrediting agencies incorporated the critical val-
ues list into their requirements, so that at least in the Unit-
ed States, almost every laboratory uses critical values pol-
icies and procedures daily for results indicating poten-
tially life-threatening situations.6,7
Despite widespread and long-standing use of critical
values, the scientific literature describing these policies
and procedures has been limited. A number of publica-
tions have described critical values lists in a single insti-
tution,4,5,8 but more recently, publications have described
experiences from a number of institutions. In 1990, Kost9
published summary data from 92 medical centers in the
United States, identifying 32 different critical value limits
Laboratory Critical Values—Howanitz et al 663
 Table 1. Analytes From Supplied List Used on Participants’ Critical Values Lists*
Low Critical Value Percentile High Critical Value Percentile
50th 50th % Appearing
Analytes 10th (Median) 90th 10th (Median) 90th on Stat List (N)
Chemistry
Ammonia, mmol/L 0.0 4.4 22.2 19.4 43.3 110.8 46.4 (289)
Arterial pH 7.2 7.2 7.3 7.5 7.6 7.6 45.9 (286)
Arterial PCO2, kPa 2.5 2.7 4 6.7 9.3 9.3 45.3 (282)
Arterial PO2, kPa 5.3 5.3 8.0 7.3 14.8 33.3 44.4 (279)
Bilirubin (neonatal), mmol/L 0.0 0.0 17.1 206.2 256.5 307.8 66.1 (412)
Calcium (ionized), mmol/L 0.22 0.75 1.50 0.35 1.58 1.75 17.3 (108)
Calcium (total), mmol/L 1.50 1.50 1.75 3.0 3.25 3.50 82.5 (514)
Carbon dioxide, mmol/L 10 10 15 40 40 45 82.5 (514)
Chloride, mmol/L 70 80 90 115 120 130 73.4 (457)
Creatinine, mmol/L 0 18 44 265 442 884 60.5 (377)
Glucose (cerebrospinal fluid), mmol/L 1.10 2.15 2.31 4.13 11.00 27.50 63.7 (397)
Glucose (serum), mmol/L 2.20 2.20 2.75 16.50 24.75 38.50 57.6 (359)
Lactic acid, mmol/L 0.00 0.06 0.33 0.27 0.44 3.40 86.4 (538)
Lecithin/sphingomyelin (L/S) ratio 1.0 1.5 2.0 2.0 2.0 3.0 35.5 (221)
Magnesium, mmol/L 0.39 0.41 0.57 1.23 1.91 2.50 9.3 (58)
Phosphorus, mmol/L 0.32 0.32 0.65 1.78 2.58 3.23 57.9 (361)
Potassium, mmol/L 2.5 2.8 3.0 6.0 6.2 6.5 49.3 (307)
Osmolality, mOsmol/kg 219 250 270 300 323 350 86.2 (537)
Sodium, mmol/L 110 120 125 150 160 170 43.0 (268)
Urea nitrogen, mmol/L
Uric acid, mmol/L 0.0 1.1 2.1 17.9 28.6 35.7 85.9 (535)
0.0000 0.0590 0.1180 0.5990 0.7611 0.8850 65.2 (406)
Hematology
Activated partial thromboplastin time, s 0 19 23 40 78 118 81.7 (509)
Fibrinogen, mmol/L 0.50 1.00 1.50 3.75 8.00 8.00 65.8 (410)
Hemoglobin, g/L 50 70 80 180 200 210 83.0 (517)
Hematocrit, % 0.15 0.20 0.25 0.55 0.60 0.65 77.4 (482)
Platelet count, 3109/L 20 40 60 600 999 1000 83.0 (517)
Prothrombin time, s 0 9 15 17 30 40 83.0 (517)
White blood cell count, 3109/L† 1.0 2.0 3.0 20.0 30.0 60.0 78.8 (491)
* Values set in boldface type were reported by fewer than 100 participants; values set in italics were reported by 100 to 199 participants.
Conversion table for calculating conventional units follows references.
† Three hundred twenty-two (51.7%) of the participants included the differential on their stat lists.

used in many of his survey participants’ laboratories. Dur-
ing the next few years, surveys of 39 pediatric hospital
laboratories and 9 regional Veterans Administration hos-
pital laboratories found that critical limits are similar be-
tween institutions, but also specific for the clinical needs
of each institution.10,11 A few years ago, Emancipator12 au-
thored a practice parameter for the American Society of
Clinical Pathologists that provided guidelines for selection
of critical value limits, as well as policies and procedures
for their implementation. Others have studied critical val-
ues for a specific analyte, such as calcium,13 ionized cal-
cium,14,15 or potassium,16,17 but we are unaware of a com-
prehensive description and demonstrated value of critical
values systems.
Recently, measurement of clinical outcomes of patients
has come into vogue in American medicine. For some ar-
eas of medicine, such as clinical drug trials, patient out-
come measures have become second nature and are sub-
ject to routine regulatory review by the Food and Drug
Administration and other groups. However, for laboratory
medicine, only a handful of patient outcome studies have
been published, most of which were summarized recently
by Bissell.18 Because critical values occur frequently and
represent life-threatening situations, the therapeutic man-
agement physicians use in response to critical values could
serve as a potentially important laboratory outcome mea-
sure that could be implemented widely.
Since 1989, the College of American Pathologists’ (CAP’s)
664 Arch Pathol Lab Med—Vol 126, June 2002
voluntary quality improvement program, Q-Probes, has
studied more than 100 indicators of quality for pathology
and laboratory medicine, describing benchmarks of perfor-
mance and providing suggestions for improvement. In this
article, we describe policies, procedures, practices, and out-
comes for laboratory critical values systems and provide
suggestions for their improvement. We believe this is the
largest and most comprehensive report to date, studying
not only limits used as critical values, but also policies and
procedures, costs, and perceived values of these laboratory
critical values lists.
MATERIALS AND METHODS
Participants registered and participated in the CAP’s volun-
tary quality improvement program, Q-Probes, according to a
previously described study design.19 For this study, a critical
value was described as a pathophysiological state at such vari-
ance with normal as to be life-threatening unless something is
done promptly and for which some corrective action could be
taken.2,3 The day shift, evening shift, and night shift were de-
fined in terms of time of day, and the physician was defined as
the primary caregiver.
The study required participants to (1) select from a supplied
list of 28 tests those laboratory tests that were included on their
critical values list and provide the limits they used for those tests;
(2) provide information on up to 3 other chemistry or hematology
critical values used in their laboratory that were not on the pro-
vided list; (3) select from supplied lists critical values for their
microbiology and drug assays; (4) answer general questions con-
Laboratory Critical Values—Howanitz et al
cerning the derivation and reporting of critical values; (5) review
the medical records of 10 patients with recent critical values to
determine the effect of critical values on medical care delivery;
(6) survey 3 to 10 nursing supervisors from different nursing
stations regarding the nursing staff’s awareness and use of crit-
ical values; and (7) survey up to 10 physicians regarding their
awareness and use of critical values.
Participants also collected data during a 30-day period on 24
critical value calls; 6 critical values were selected from the middle
of each of the 3 shifts for inpatients, as well as 6 critical values
for outpatients. These data included the length of time required
to find a person to take each critical value call, the job title of the
person making and taking the critical value call, the length of
time before each critical value call was abandoned, the inpatient
or outpatient status of the patient, and the shift during which the
call was made.
The 13 general questions concerning the derivation and re-
porting of critical values were multiple-choice type with direc-
tions to choose the best answer or all answers that applied, de-
pending on the question. For chart reviews, a preprinted single
sheet was provided so that each of 4 questions was answered by
making a mark in either the yes or no column. The nursing su-
pervisors were interviewed by telephone and were asked 3 ques-
tions, each requiring either a yes or a no answer, and their re-
sponses were indicated on a preprinted interview form. The nurs-
ing supervisors selected were to represent the emergency de-
partment, an intensive care or critical care unit, and at least 1
medical/surgical unit. Immediately after receiving a critical val-
ue, the physicians were mailed a form with 6 questions about the
critical value and were asked to circle either yes or no for each
question. An accompanying letter signed by the laboratory di-
rector requested the physicians to immediately complete and re-
turn the short questionnaire.
Demographic information was obtained from a yearly partici-
pant enrollment form. Based on the data submitted, participants’
results were ranked according to the usual Q-Probes percentile
rankings, with the 10th percentile corresponding to the lowest
critical values and the 90th percentile corresponding to the high-
est critical values. When participants did not complete all re-
quired information, they were excluded from the database only
for the missing information. Some laboratories did not complete
the yearly demographic form, and for these, their bed size was
listed as unknown. Laboratories reported data using the units of
the test results they commonly used, and these units were con-
verted to SI units at the CAP computer center by applying widely
used conversion factors.
RESULTS
Data were submitted by 623 institutions, of which 31%
were teaching and 69% were nonteaching. Most institu-
tions were nongovernmental (76.1%); the remainder were
governmental (5.5% federal and 18.4 nonfederal). In rela-
tionship to size, 23.2% had fewer than 150 beds, 34.6%
had 151 to 300 beds, 21.7% had 301 to 450 beds, 11.0%
had 451 to 600 beds, and 9.5% had more than 600 beds.
The majority of participants (62.8%) were located in cities,
19.3% were suburban, 16.0% were rural, and 1.9% were
classified as ‘‘other.’’ Occurrences of critical values ranged
from 1 in 120 to 1 in 957 tests, with more frequent values
occurring in larger hospitals.
Table 1 shows participants’ low and high critical limits
in terms of the number of institutions using various limits.
Also in this table is the distribution of values for each of
the low or high critical limits expressed as the 10th, 50th
(median), and 90th percentile. The choice of the limit for
either the 28 low or 28 high critical values was highly
variable, with no value being the same in 80% or more of
the laboratories. For some chemistry analytes, such as glu-
cose, potassium, sodium, and calcium, and for some he-
Arch Pathol Lab Med—Vol 126, June 2002
Table 2. Participants (%) Reporting Microbiology
Results as Critical Values
Microbiology Result Participants, %
Positive blood cultures 95.0
Positive CSF cultures 91.2
Positive AFB smear or culture 71.9
Positive Gram stains of sterile body fluids 66.8
Initial stool isolates of Salmonella, Shigella,
Campylobacter, and Yersinia 59.7
Positive latex agglutination and/or antigen
detection test 50.7
Positive CSF VDRL 25.7
Other microbiology critical values 44.8
* CSF indicates cerebrospinal fluid; AFB, acid-fast bacilli; and VDRL,
Venereal Disease Research Laboratory test.
Table 3. Participants’ Reporting Practices
Drug Results Participants, %
Drugs of abuse
Telephone all positive results 25.2
Do not telephone all positive results 52.8
Not applicable 22.0
Therapeutic drugs
Telephone all toxic values 96.3
Do not telephone all toxic values 2.9
Not applicable 0.8
matology analytes, such as hemoglobin, hematocrit, plate-
let count, and white blood cell count, almost all of labo-
ratories had both high and low critical values. In contrast,
fewer than 100 participants had both high and low critical
values for ammonia, ionized calcium, lactic acid, and the
lecithin/sphingomyelin (L/S) ratio. For some analytes,
such as arterial PO2, the majority of laboratories had only
low critical values, whereas for other analytes, such as cre-
atinine and prothrombin time, the majority of participants
had only high critical values. Sixty-five other chemistry or
hematology analytes appeared on at least 1 laboratory’s
critical values list, and of these, amylase was the most
common. Amylase was found on 66 (10.7%) participants’
lists. Also listed in this table is the percentage of analytes
with critical limits that were on participants’ stat test lists.
Table 2 shows the percentage of participants with re-
sults of common microbiology procedures on their critical
values lists. Almost all laboratories provided notification
for a positive blood or cerebrospinal fluid culture. Table 3
shows participants’ reporting practices for positive find-
ings for drugs of abuse and toxic levels for therapeutic
drugs. The majority (52.8%) of participants did not im-
mediately report all positive drugs of abuse results,
whereas most (96.3%) immediately reported all toxic val-
ues for therapeutic drugs.
Of the methods used to report critical values, 99.2% re-
ported they used telephone contact, 29.5% used a fax ma-
chine or similar transmission device, 10.0% used a com-
puter report as the primary means, 42.2% used a computer
report as a secondary means, and 6.9% used an answering
machine or voice mail system.
Participants made a total of 275 872 critical value calls
during the 30-day study period and recorded specific in-
formation on 12 930 (4.7%) calls. Table 4 shows that the
mean time participants spent to complete a critical value
call for inpatients was 6.1 minutes, whereas for outpatients
Laboratory Critical Values—Howanitz et al 665
 Table 4. Time Spent to Complete or Abandon a Critical Value Call
Mean Time to Complete Call, min (No. of Participants) Mean Time to Abandon Call, min (No. of Participants)
Inpatient Inpatient
Outpatient, Outpatient,
No. of Beds Night Evening Day All Shifts Night Evening Day All Shifts
Unknown 6.8 (8) 12.6 (9) 6.9 (9) 16.5 (8) ... 5.0 (1) ... ...
0 ... ... ... 11.2 (2) ... ... ... ...
1–150 7.0 (110) 10.1 (127) 8.3 (136) 13.2 (123) 9.8 (4) 46.6 (8) 54.8 (9) 23.4 (14)
151–300 5.7 (190) 5.8 (196) 5.9 (203) 11.4 (185) 3.5 (8) 7.3 (8) 7.5 (8) 27.2 (26)
301–450 5.4 (124) 6.0 (128) 7.5 (131) 21.2 (123) 15.7 (4) 36.4 (5) 6.4 (4) 108.6 (15)
451–600 2.9 (63) 3.4 (65) 4.0 (66) 9.5 (60) 0.8 (1) 0.5 (1) 6.1 (3) 28.3 (3)
.600 3.1 (51) 4.7 (53) 3.3 (54) 9.3 (49) 15.0 (2) 13.3 (3) 0.0 (1) 34.2 (5)
All 5.3 (546) 6.5 (578) 6.4 (599) 13.7 (550) 8.5 (19) 25.3 (26) 23.9 (25) 46.3 (63)

Table 5. Personnel Involved in Critical Value Calls

Inpatients, No. (% of Total) Outpatients, No. (% of Total)
Reporting personnel
Person performing test 11 587 (91.0) 2308 (77.3)
Section supervisor 330 (2.6) 117 (3.9)
Laboratory clerk 718 (5.6) 515 (17.2)
Other management personnel 96 (0.8) 47 (1.6)
Receiving personnel
Registered nurse 4775 (37.8) 622 (21.2)
Any staff nurse 2307 (18.3) 393 (13.3)
Unit clerk/office staff 3985 (32.5) 1237 (42.1)
Medical student 21 (0.2) 3 (0.1)
Any physician on call 455 (3.6) 196 (6.7)
Physician ordering test 1090 (8.6) 490 (16.7)

Table 6. Repeat Critical Value Calls Table 7. Use of Critical Values as Indicated by Chart
Review and Physician Surveys
Policy Participants, %
5426 4237
No policy on how handled 71.4 Charts Physicians
Repeat values not called 11.6 Finding Reviewed Surveyed
Values not called on physician request 6.8
Seed physicians’ permission to not call 1.9 Critical value anticipated, % 54.3 59.3
Have preset policy on no. of calls 1.8 Critical value influenced
Seek physicians’ permission not to call therapy, % 64.9 62.9
after preset no. of values 0.3 Test reordered, % 66.3 43.3
Use another policy 6.1 Found in nursing or progress
notes, % 75.5 ...

it took 13.7 minutes to report a value. In contrast, on av-

erage, critical value calls were abandoned in 20.2 minutes
for inpatients and in 46.3 minutes for outpatients. The time
spent in reporting a critical value call was related to the
shift, the bed size of the institution, and the patient’s status
as an outpatient.
Table 5 describes the personnel involved in critical value
calls. For inpatients, the call usually was made by the per-
son performing the test (91.0%) to either a registered
nurse (37.8%), a clerk (31.5%), or staff nurse (18.3%). In
contrast, for outpatients the call was made by the person
performing the test (77.3%) or a laboratory clerk (17.2%)
to either a clerk (42.1%), a registered nurse (21.2%), or the
physician ordering the test (16.7%).
Participants’ policies on repeat critical values are shown
in Table 6. Most (71.4%) did not have a policy on repeat
critical values for the same analyte on the same patient,
although in some institutions (11.6%), the policy was that
repeat values were not reported.
Table 7 demonstrates results of 5426 chart reviews (8.7
per institution) and 4237 surveys of physicians (6.8 per
666 Arch Pathol Lab Med—Vol 126, June 2002
institution) about recently reported values. A survey of
physicians and review of the medical records found that
the majority of critical values were expected. For those
unanticipated critical values, both the majority of inter-
views and review of the medical records indicated therapy
was influenced by the result. Although a minority of phy-
sicians contacted stated they reordered the test result
(43.3%), the majority (66.3%) of medical records contained
documentation that the test was reordered.
Table 8 shows hospital personnel’s perceptions of criti-
cal values as surveyed by clinical laboratory personnel. An
average of 7.4 nursing stations and 6.8 physicians recently
receiving a critical value report were queried. The majority
of nursing personnel surveyed were not aware of the crit-
ical values list, did not think the list is a valid indicator of
the patient’s condition, nor did they feel calls were helpful.
In contrast, most physicians surveyed were aware of the
critical values list, thought it was a valid indicator, and felt
calls were helpful.
Table 9 contains descriptions of the derivation of lab-
Laboratory Critical Values—Howanitz et al
 Table 8. Perception of Critical Values
Location* Respondents
Emergency department 575 nursing supervisors
ICU/CCU 576 nursing supervisors
Medicine/surgery 576 nursing supervisors
Other station 574 nursing supervisors
Hospital 514 physicians
* ICU indicates intensive care unit; CCU, critical care unit.
Table 9. Source of Laboratory Critical Value Policy
Literature review only (includes manufacturer’s
help), % 18.60
Literature review and laboratory meetings
(includes manufacturer’s help), % 36.00
Literature review modified by hospital committees
(includes in-house studies), % 16.80
Literature review modified by in-house studies
and medical staff consultations, % 72.70
oratory critical values policies. Most institutions based
their choices on data from several sources, including
manufacturers, hospital committees, and medical staff
members.
COMMENT
We used the definition of critical (panic) values origi-
nally described by Lundberg as ‘‘life-threatening unless
something is done promptly and for which some correc-
tive action could be undertaken.’’2,3 We chose this defi-
nition because we thought that all participants would be
able to identify their institutional lists in these terms.
Even though all participants were able to provide the
required information about their critical values lists, it
was clear that most included critical limits for analytes
that were not life-threatening or for which some correc-
tive action could not be undertaken. Many would argue
that all 28 analytes listed in Table 1 would fulfill Lund-
berg’s definition; however, among the list of the 65 other
analytes submitted were some that might be considered
as not fulfilling this definition, such as mean corpuscular
volume, urine crystals, haptoglobin, prealbumin, and li-
docaine. The use of critical limits for non–life-threatening
situations has led others to conclude that these lists
should be now be entitled ‘‘Vital Laboratory Values’’ or
‘‘Alert Values,’’ rather than critical values.5,8,20 Hortin and
Csako20 found that alert values frequently serve not to
alert clinicians of a medical emergency, but rather iden-
tify for laboratorians a valuable quality improvement op-
portunity. This distinction arises because a substantial
proportion of critical values occur from preanalytic prob-
lems or analytical interferences and offer opportunities
for improvement of processes or methodologies that re-
side within the clinical laboratory.
Our survey led to presentation of critical values in the
usual percentile ranking used in Q-Probes studies; such a
presentation can be used to provide guidance to labora-
torians. By listing values in terms of the number of par-
ticipants and critical limits medians, 10th, and 90th per-
centile critical limits, interested laboratorians can compare
their own values with this list and adjust their values to
be similar to the median, the most lenient, or the most
stringent policy. Laboratories also had critical values lists
Arch Pathol Lab Med—Vol 126, June 2002
% Think List Is a % Feel Calls Are
% Aware of List Valid Indicator Helpful
17.3 16.7 19.1
23.6 23.3 27.6
33.5 33.4 40.8
16.0 16.2 20.6
76.9 78.6 94.9
for microbiology. Almost all laboratories reported results
of a positive blood culture and had critical values lists for
therapeutic drugs, and most reported all toxic values;
however, the majority did not include drugs of abuse on
their lists. Perhaps one reason that so many laboratories
had similar practices for critical values for blood and ce-
rebrospinal fluid cultures was that these results are qual-
itative values. In contrast, critical value limits differed
among participants for hematology and chemistry because
these limits are quantitative. Although we did not ask par-
ticipants to list critical values used in their transfusion
medicine laboratories, others have indicated that there are
some tests that should be considered.4
One of the ways of defining expert opinion is to ascer-
tain the practice patterns of the majority and use these
patterns as the standard of practice. A number of critical
values lists have been published for a group of laborato-
ries, and there are major similarities among these lists.
Comparisons between the list we developed and the lists
used for both adults and children as reported by Kost in-
dicate our median values approximate those of Kost.9,10
Our list is the most complete to date, but will need to be
updated as new tests are developed and added to critical
values lists. An example of the changing list of critical
values is that now almost all laboratories have toxic levels
of therapeutic drugs on their critical values list. However,
when the critical values concept first was developed, ther-
apeutic drug monitoring was not widespread, consequent-
ly very few laboratories had critical values for these drug
levels. Since the time this study was conducted, troponin
measurements have become widespread, and we predict
that a number of laboratories already have added troponin
to their critical values lists.
Tate and Gardner21 found that fewer than 10% of critical
values were reported in their institution. They identified
weaknesses in the reporting process as (1) critical values
were not always reported by the clinical laboratory; (2)
when critical values were reported, it was often to someone
not directly involved in the patient’s care; (3) documenta-
tion of critical value reporting by the laboratory was incom-
plete; (4) clinicians’ awareness of critical values was not al-
ways documented; (5) clinicians’ decisions on corrective ac-
tions were not documented adequately; and (6) the time
interval between the availability of critical test results and
the institution of corrective measures was sometimes un-
acceptably long. Improvement in the reporting of critical
values requires that the process is well defined and widely
supported by users of laboratory services.
In our study, technologists performing the procedure
made the critical value calls for inpatients in more than 90%
of cases; for outpatients, the persons performing the proce-
dure still made most of the calls, although more than 17%
of calls were made by laboratory clerks. In most laboratories,
Laboratory Critical Values—Howanitz et al 667
personnel performing chemistry and hematology procedures
work in teams. We recommend that all critical values be re-
ported by one of the team members performing the proce-
dure, thereby providing the most direct system for the action.
Critical value policies require notification of the abnormal
result to someone who can take appropriate action, and in
our study, a variety of personnel received these calls. For
inpatients, registered nurses received almost 40% of the
calls, compared with the clerk or other office staff, who
received slightly more than 30%. For outpatients, office staff
and clerks received more than 40% of the calls, registered
nurses slightly greater than 20%, and physicians 17%. Crit-
ical value calls are costly for laboratories in that extensive
time is required to handle reporting. For inpatients, these
calls took an average of 6.1 minutes, whereas for outpa-
tients they took 13.7 minutes. One of the reasons outpatient
calls took more time was that approximately 17% of out-
patient reports were received by physicians, and it took an
average of 19 minutes to reach the physician on call.
At what point do laboratorians give up on reaching a
person with notification of a critical value? We suspect that
this determination is related to many variables, including
the availability of personnel who make these calls. For out-
patients, calls were abandoned, on average, after 46 min-
utes. This time expenditure was far longer than for inpa-
tients when the result occurred during the night shift (9
minutes), the evening shift (25 minutes), or the day shift
(24 minutes). For some outpatients, laboratorians tried to
reach an appropriate physician for almost 2 hours. Unfor-
tunately, a small percentage of calls (5.4%) were aban-
doned. However, from our own practices we have realized
the futility of continuing to try to find someone for an
extended period of time to assume responsibility for the
critical value, even when time may be important in treat-
ing a patient who has a potentially lethal condition.
It is unfortunate that any calls are abandoned; hence,
we recommend that a fail-safe mechanism be implement-
ed to assure that all life-threatening values are fully eval-
uated, and if in fact critical, then proper care can be start-
ed for patients. The following policy was established with
great success in the laboratories of one of the authors, and
we recommend it as simple, effective, and practical. If no
one can be found by the laboratory personnel to accept
the critical value after a short period of time, then a phy-
sician representing the laboratory should determine the
significance of the result and determine whether further
effort is required. When further evaluation of the patient
is needed, we believe it should be performed by a physi-
cian able to provide immediate care for the patient. In our
experience, this procedure works best if the patient’s avail-
able medical record is brought to the physicians in the
emergency department, who, after receiving the patient’s
available record, may contact the patient and discuss the
results with the patient. If the patient requires immediate
attention, it can be provided either as a patient in the
emergency department or as an inpatient. If the patient
cannot be reached by telephone, we have found that local
police are responsive to requests to hand-carry a notice
with the desired actions that should occur to the patient’s
residence. Of course, such a policy requires agreement of
the medical staff, as well as the review of each occurrence
when a physician cannot be reached as part of the labo-
ratory’s and institution’s quality improvement program.
Although critical values lists are widespread and similar
among institutions, our data indicate that policies on how
668 Arch Pathol Lab Med—Vol 126, June 2002
a second critical value for the same analyte on the same
patient should be handled have not been defined clearly in
most institutions. It is our belief that with the limited in-
formation available to the technologist and the short period
of time a technologist has to make a critical decision on a
patient’s results, it is difficult to develop a policy that is
tailored uniquely to the suggestions or requirements of each
physician. Rather, a generic policy that covers all patients
is what we favor, as it is easiest for technologists to imple-
ment and causes no errors in judgment by the technologists.
Such a policy should require that all values exceeding the
critical value limits require notification of the appropriate
caregiver. For example, although nephrologists know their
patients on dialysis have elevated serum urea nitrogen and
creatinine values, it may be difficult for a technologist to
ascertain that the elevated renal function results just veri-
fied are those of a patient on dialysis because of missing,
incorrect, or hard-to-find demographic information. At the
same time, we believe laboratories must develop policies
not to annoy physicians needlessly with values they know
previously are elevated, and we suggest that only the first
of a series of values within 24 hours should require notifi-
cation. Although we chose 24 hours for the first series of
values, others may feel more comfortable with a longer time
frame, such as only the first of a series of values within 48
hours as requiring notification.
Laboratory productivity has increased remarkably dur-
ing the past 30 years because of the use of automated
equipment. The primary means of critical value notifica-
tion in most institutions is the use of telephone calls to the
physician and nursing personnel responsible for the pa-
tient’s care. The use of electronic devices automatically di-
aling the physician’s beeper offers opportunities to reduce
laboratorians’ time on the telephone with clinicians, and
at the same time improves the likelihood that the physi-
cian or another person charged with the patient’s care re-
ceived the result. Some instruments have been introduced
that provide automatic notification of preprogrammed
critical value results,22,23 and others have discussed the po-
tential value of laboratory information systems automati-
cally delivering results to physician pagers.12 Automated
e-mails or faxes to physicians require other newer tech-
nologies that can be used, although mechanisms are es-
sential to assure receipt of this notification.
Laboratorians generally believe critical values lists are
important and have worked with others to help design
the lists. The perceived lack of importance of a critical
value list to nursing supervisors was surprising. Not only
were most of these supervisors in the emergency de-
partment, intensive care unit, and other nursing units
unaware of the critical values lists, they did not think
they were a valid indicator of the patients’ conditions and
reported they did not think the calls were helpful. In
contrast, the overwhelming majority of physicians sur-
veyed were aware of the lists, thought they were a valid
indicator, and almost all thought the calls were helpful.
Perhaps the supervising nurses in their administrative
role had become out of touch with the medical care that
was being provided by their staff and physicians. Anoth-
er interpretation of these findings is that physicians usu-
ally see their hospitalized patients only once or twice a
day and are further removed from emergency care than
the nurses. Nurses are communicating among themselves
a number of times during the day about the patients and
the patients’ critical values, hence these critical value re-
Laboratory Critical Values—Howanitz et al
sults represent less concerning information. However, References
1. Howanitz JH, Howanitz PJ. Laboratory results: timeliness as a quality attri-
these discrepancies suggest that laboratorians must dis- bute and strategy. Am J Clin Pathol. 2001;116:311–315.
cuss with nursing staff the value of critical limits, as a 2. Lundberg GD. When to panic over abnormal values. Med Lab Observer.
better understanding of the value of these calls should 1972;4:47–54.
3. Lundberg GD. Managing the Patient-Focused Laboratory. Oradell, NJ: Med-
reduce the time required to give the critical value report ical Economics Books; 1975.
to someone on the nursing stations. 4. Murphy J, Henry JB. Effective utilization of clinical laboratories. Hum Pathol.
1978;9:625–633.
The frequency of critical values of our participants var- 5. Lundberg GD. Critical (panic) value notification: an established laboratory
ied between 1 in 120 and 1 in 957 tests, well within the practice policy (parameter). JAMA. 1990;263:709.
range found by others of between 1 in 100 and 1 in 2000 6. Clinical Laboratory Improvement Amendments of 1988: final rule (42 CFR
Part 405, et al), 57 Federal Register 7001–7186 (1992).
tests.2,9 If one chooses the average time our participants 7. Rabinovitch A. The College of American Pathologists (CAP) Laboratory Ac-
used for notification, and a frequency of panic values of 1 creditation Checklists. Northfield, Ill: College of American Pathologists; 2000.
in 1000 tests, then in a large hospital laboratory that per- 8. Hobbs GA, Jortani SA, Valdes R. Implementation of a successful on-call
system in clinical chemistry. Clin Chem. 1997;108:556–563.
forms 5 million tests, slightly less than one full-time 9. Kost GJ. Critical limits for urgent clinician notification at US medical cen-
equivalent would be required for notification of critical ters. JAMA. 1990;263:704–707.
10. Kost GJ. Critical limits for emergency clinician notification at US children’s
values. Although this may not seem like a large resource hospitals. Pediatrics. 1991;88:597–603.
requirement for a large hospital laboratory, in a small lab- 11. Lum G. Critical limits (alert values) for physician notification: universal or
oratory where only 1 technologist may be working during medical center specific limits? Ann Clin Lab Sci. 1998;28:261–271.
12. Emancipator K. Critical values: ASCP practice parameter. Am J Clin Pathol.
some shifts, reporting a critical value may delay the re- 1997;108:147–153.
porting of other results. In 1996, the laboratory testing 13. Lum G. Evaluation of a laboratory critical limit (alert value) policy for
hypercalcemia. Arch Pathol Lab Med. 1996;120:633–636.
volume was estimated to be approximately 7.25 million 14. Kost GJ. The challenges of ionized calcium cardiovascular management
tests; when expressed in terms of the cost of laboratory and critical limits. Arch Pathol Lab Med. 1987;111:932–934.
testing in the United States, the cost of critical values sys- 15. Kost GJ, Jammal MA, Ward RE, Safwat AM. Monitoring of ionized calcium
during human hepatic transplantation: critical values and their relevance to car-
tems seems large.24 diac and hemodynamic management. Am J Clin Pathol. 1986;86:61–70.
Patient outcomes data collected in this study demon- 16. Bigger JT Jr. Diuretic therapy, hypertension, and cardiac arrest. N Engl J
strated that about 65% of critical values influenced thera- Med. 1994;330:1899–1900.
17. Wong KC, Schafer PG, Schultz JR. Hypokalemia and anesthetic implica-
py, the test was reordered 66% of the time when a critical tions. Anesth Analg. 1993;77:1238–1260.
value occurred, and physicians documented these results 18. Bissell MG. Laboratory-Related Measures of Patient Outcomes: An Intro-
duction. Washington, DC: AACC Press; 2000.
in the patients’ medical record 66% of the time. We know 19. Howanitz PJ. Quality assurance measurements in departments of pathol-
of no previous studies of the effects of critical values on ogy and laboratory medicine. Arch Pathol Lab Med. 1990;114:112–115.
patient care. Because critical values occur in every hospital 20. Hortin GL, Csako G. Critical values, panic values or alert values? [letter].
Am J Clin Pathol. 1990;109:496–497.
and are found in high frequency, we would recommend 21. Tate KE, Gardner RM. Computers, quality and the clinical laboratory:
that those interested in patient outcomes consider docu- a look at critical value reporting. Ann Symp Comp Appl Med Care. 1993;
193–197.
menting the value of this system in their institutions. 22. Catrou PG. How critical are critical values? Am J Clin Pathol. 1997;108:
Our study was based on critical values originating from 245–246.
the central laboratory, but with the recent trend to decen- 23. Shabot MM, LoBue M, Leyerle BJ, Dublin SB. Decision support alerts
for clinical laboratory and blood gas data. Int Clin Monitor Comput. 1990;
tralize these laboratories to point-of-care testing, consid- 27–31.
erations should be developed on how critical values 24. Steindel SJ, Rauch WJ, Simon MK, Handsfield J. National Inventory of
should be handled at these remote sites. In our own ex- Clinical Laboratory Testing Services (NICLTS): development and test distribution
for 1996. Arch Pathol Lab Med. 2000;124:1201–1208.
periences, the policies and procedures that have been es-
tablished for the central laboratory should be extended to
testing at these sites, with the required action included in Conversion From SI Units to Conventional Units
the procedure. In some situations, protocols and proce- Conversion
dures undoubtedly have been developed for such analytes Factor Conventional
as glucose, for which physicians traditionally document Analyte SI Unit (Divide by) Unit
the required responses for nurses to a variety of glucose Ammonia mmol/L 0.714 mg/dL
levels. In other cases, these values may cause problems, in Bilirubin mmol/L 17.1 mg/dL
that they may be unexpected and algorithms have not yet Calcium, ionized mmol/L 0.25 mg/dL
been developed for markedly elevated or low values for a Calcium, total mmol/L 0.25 mg/dL
Carbon dioxide mmol/L 1.0 mEq/L
specific analyte, such as calcium. Chloride mmol/L 1.0 mEq/L
In conclusion, critical values lists are widely used and Creatinine mmol/L 88.4 mg/dL
are medically important, but they also are widely vari- Fibrinogen mmol/L 29.41 g/dL
able in their content and use. Laboratorians must develop Glucose,
practices that ensure that each result is transmitted to cerebrospinal fluid mmol/L 0.05551 mg/dL
Glucose, serum mmol/L 0.05551 mg/dL
appropriate caregivers rapidly and effectively, so that Hemoglobin g/L 10.0 g/dL
quality patient care can be rendered. Medical and nurs- Lactic acid mmol/L 0.1110 mg/dL
ing staffs should be involved effectively in developing Magnesium mmol/L 0.4114 mg/dL
the laboratory’s critical value policy, as well as in partic- Phosphorus mmol/L 0.3229 mg/dL
ipating and investigating appropriate automation. A ma- Platelet count 3109/L 1.0 3103/mL
Potassium mmol/L 1.0 mEq/L
jor reduction in the time spent in critical values notifi- Sodium mmol/L 1.0 mEq/L
cation may not occur until automated systems are better Urea nitrogen mmol/L 0.357 mg/dL
used and a better understanding of the importance of the Uric acid mmol/L 0.0595 mg/dL
critical value policy is achieved among those whose ac- White blood cell
count 310 9/L 0.001 /mL
tions have an impact on patient care.
Arch Pathol Lab Med—Vol 126, June 2002 Laboratory Critical Values—Howanitz et al 669