Telephone
Triage
by Primary
Care
Physicians
Susan Zelitch Yanovski, MD*; Jack A. Yanovski, MD, PhD; James D. Malley,
PhD; Richard L. Brown, MD, MPH*; and Donald
J.
Balaban, MD, MPH*ABSTRACT. To determine if experienced primary care
physicians are more likely to reach correct decisions on
the telephone than their less experienced colleagues, we asked 31 first-year and 29 third-year residents, 21 faculty, and 36 private practitioners in pediatrics and family
practice to evaluate three pediatric patients via a
tele-phone interview with a simulated mother and to decide
whether each patient needed to be seen that evening. Compared with first-year residents, the third-year residents, faculty and private practitioners decided less frequently to see children who were not severely ill (P
< .05) or injured (P < .01); however, less than half
ob-tamed histories considered adequate to rule out potential
serious illnesses. Faculty did better than either residents or private practitioners in managing a severely
dehy-drated child 100% of the faculty, but less than 60% of
the residents or private practitioners, chose to see the patient promptly (P < .001). More than one third of all
residents and private practitioners reached inappropriate
management decisions despite obtaining information that should have altered their decisions.
In these simulations, experience in private practice was not associated with improved telephone manage-ment of very sick children. Faculty physicians appeared to be better able to identify severely ill children without inappropriately evaluating those who were less ill. In all
three simulations, attainment of the correct decision ap-peared to be determined not by the number or type of questions asked, but rather by the physician’s
interpre-tation of the information collected. Pediatrics
1992;89:701-706; Telephone, primary care, pediatrics,
fam-ily practice, management, triage, decision making.
The telephone call is a major means by which
physicians provide care for their patients. Twelve
percent to 28 percent of primary medical care is
delivered by telephone,”2 and up to half of all
phy-sician-patient telephone contacts result in further
medical evaluation.3 Nevertheless, how physicians
arrive at such decisions remains unclear.
Several studies of telephone management have
evaluated the effects of pediatric experience upon the
clinical judgment of physicians. Pediatric house
offi-cers have been reported to perform either as well as
or better than practicing pediatricians when scored
on history taking, interviewing skills, and
disposi-tion.4 Other investigators have found that serious
From the 6Greenfield Research Center, Department of Family Medicine,
Thomas Jefferson University, Philadelphia, PA; the Children’s Hospital of
Philadelphia, Philadelphia, PA; and the §Division of Computer Research and Technology, National Institutes of Health, Bethesda, MD.
Received for publication May 2, 1991; accepted Oct 18, 1991.
Reprint requests to (S.z.Y.) 5000 Battery Lane, #801, Bethesda, MD 20814. PEDIATRICS (ISSN 0031 4005). Copyright © 1992 by the American
Acad-emy of Pediatrics.
deficiencies in the questions physicians ask on the
telephone are not rectified by physician training or
experience.58 The consistent findings of unchanged
or even worsening physician performance with
in-creasing experience seem counterintuitive, and may
reflect artifacts of study design. For instance, adequate
history taking has been defined as not missing more
than one of a list of questions suggested by a
consen-sus panel.5 Whereas Perrin and Goodman attempted
to address this issue by weighting the importance of
historical data,4 their approach was still dependent
on the absolute number of “most critical items’
in-cluded, not all of which may be necessary in a given
clinical situation. Additionally, most studies of
phy-sician telephone decision making were carried out
more than a decade ago, and were limited by their
small sample size.
The purpose of this study was to examine telephone
decisions made by primary care physicians as a
func-tion of experience using their actual decisions as the
primary outcome measure. Our hypothesis was that
experienced physicians would more frequently make
the correct management decision when performing
telephone triage. We further hypothesized that
ex-perienced physicians would ask fewer, but more ap-propriate, questions in reaching their decisions.
Subjects
SUBJECTS AND METhODS
The sample comprised all first- and third-year residents in family
medicine at a university- and community-based residency program,
all first- and third-year residents in pediatrics at two
university-based residency programs, all faculty in family medicine and gen-eral pediatrics at the same programs, and the family physicians
and general pediatricians in full-time practice who admifted
pa-tients to these institutions. Subjects were excluded if they reported
that less than 10% of their patients were younger than 14 years of age, or if they participated in the development of the study.
We sent 141 physicians letters inviting each to participate in the study; nine (five pediatricians and four family physicians, all in
private practice) were unable to be contacted by telephone after
several attempts. Of the 132 physicians who were contacted by telephone, five (family practice faculty) saw less than 10% pediatric
patients and were excluded, leaving 127 physicians eligible for
inclusion in the study. Ten subjects refused to participate (one first-year family practice resident, one first-year pediatrics resident, three third-year pediatrics residents, and five private pediatricians). Thus, of the 127 eligible physicians who could be contacted by telephone, I 17 (93%) agreed to participate, and all completed the study. Some of the tape-recorded data for six cases involving four subjects could not be analyzed because of technical problems and were deleted for the purpose of analysis. However, the subjects’ triage decisions were available in all cases.
Characteristics of the study physicians are shown in Table 1. More pediatricians were studied than family physicians (P < .05).
practi-TABLE 1. Characteristics of Study Physicians: Description of Subject’s Age, Qualifications, and Practice Patterns
Gender, Specialty
First-year
Residents
Third-year
Residents
Private Practice
Faculty Total
Numberofsubjects 31 29 36 21 117
Specialty
Family practice, % 26 38 47 19 34
Pediatrics, % 74 62 53 81 66
Mean age, y: 27 30 43 38 35
Gender
Male,% 42 48 83 81 63
Female, % 58 52 17 19 37
Any training in telephone medicine, 23 59 50 58 46
%
Board certification, % 89 100 93
Years in practice
<5y 33 38 35
>5y 67 62 65
Mean number of pediatric phone 45 1 1 32
contacts per weeks
Mean number of pediatric office via- 65 36 54
its per week9
*P< .05.
P< .001. §P< .005.
tioners reported seeing significantly more pediatric patients weekly
(P < .05) and handled significantly more pediatric phone calls (P
< .001) than did faculty physicians. Of the subjects, 46% reported training in telephone decision making, primarily by lecture. Interns
were significantly less likely than all others to have received any
training in telephone decision making (P < .005). All but four nonresident physicians (two family physicians and two pediatri-cians) were board-certified.
Case Development
Three cases were developed by the investigators with ‘correct’
management decisions based upon the most recent American
Acad-emy of Pediatrics guidelines for telephone management9 (employed previously4’#{176}), and upon the recommendations of texts on pedi-attic telephone decision making.’113 Taken together, these sources provided guidance on whether the patient needed to be examined
immediately or could wait until office hours the following day.
The cases described three pediatric problems commonly
man-aged by telephone: an 1 1-month-old with severe diarrhea and
dehydration who required immediate medical attention, an
18-month-old with minor head trauma, and a 2#{189}-year-oldwith fever,
neither of whom were felt to have problems of an urgent nature
(Table 2).
Administration and Scoring
The simulations were administered by one of the investigators (S. Z. Y.) by telephone at the subjects’ convenience. All interviews
were tape-recorded with the subjects’ consent. Subjects were
pro-vided identical written and verbal information describing the sim-ulated conditions to insure that subjects were working from similar
premises. This information included the time of day of the patient’s call, the physician’s previous knowledge of the patient, the physi-cian’s location, and the physician’s goals for the encounter. Other
information was given only in response to questions asked by the
physician-subject. The exact wording of the responses to individual questions was developed before the start of the study and was
standardized to minimize possible effects of investigator bias. If a
subject asked a question which did not have a predetermined response, it was answered by the simulated mother in a manner consistent with the overall content of the case, and this exact
response was used for all subsequent administrations. The three cases were administered in the same order during one telephone call for each subject, to minimize performance variability.’4 The order of the cases was: (1) 1 1-month-old with severe diarrhea and
dehydration, (2) 18-month-old with minor head trauma, and (3)
2#{189}-year-old with fever.
After all three cases had been completed, the interviewer
ob-tamed descriptive data and questioned subjects about their
percep-TABLE 2. Synopsis of Case Descriptions: Examples of Pertinent
Information Physicians Could Elicit from Simulated Mother for
Each Case
Case I : Severe Diarrhea and Dehydration
1 1-mo-old with watery stools for more than 24 h More than 10 stools last 12 h, each filling diaper
Drinking less than 12 oz last 24 h, won’t take more Won’t smile or play; whimpering last few hours Less alert than earlier, listless; has never been this sick
Is not making tears, eyes are sunken, mouth is dry
Case 2: Minor Head Trauma Without Loss of Consciousness 18-month-old, fell off changing table 3 ft onto a carpet
Cried immediately for 5 mm, then went to sleep
Able to wake up without difficulty, no vomiting
Now acting normally; moving arms/legs well, talking normally
Has lump on back of head which is tender, no other injuries
Case 3: Viral Syndrome with Fever
2’/2-y-old with temperature 102.6#{176}F(rectal) for 12 h Fever did not respond to acetaminophen
Dose of acetaminophen only 80 mg every 4 h (5 mg/kg) Emesis once, had two loose stools today
Minor pain at umbilicus, eating/drinking normally now
Otherwise acting well; has often been this ill in the past No significant past medical history
tions of the realism of the cases. Subjects were also asked to explain the reasons for their decision in each instance. Tapes of all
inter-views were reviewed to establish the decisions made by each
physician and the particular questions asked. Fifty percent of the
tapes were reviewed by two investigators. Inter-rater reliability for
the number of questions asked was 92% and for the specific
questions asked was 94%.
At the conclusion of the study, a complete list of the questions
asked by subjects and the responses offered by the simulated
mother were reviewed by five experienced physicians. This panel
consisted of two family physicians and three pediatricians of whom three were in private practice (one family practitioner and two
pediatricians). None of these physicians otherwise participated in
the study. Their task was to determine if, given all the information that subjects could elicit about each case, they would choose to see the child urgently. All five physicians agreed with the published guidelines that both the 18-month-old with head trauma and the 2#{189}-year-old with fever could wait until the next day to be exam-ined; all were similarly in agreement with the published guidelines that the 1 1-month-old with severe diarrhea and dehydration
*
+
I
Diarrhea & Dehydrationill
F
Head Trauma
Fever TABLE 3. Comparison of Performance Between Family Physicians and Pediatricians on Diarrhea and Dehydration Simulation’
Group Percent Correct
Decisions
Time to Decision in Seconds
Total No. of Questions
No. of Critical Areas
Ad-dressed (of
nine
possi-ble)
Pediatric private practitioners, n = 19
Family practice private practitioners, n = 17
47 (24-71)
41 (18-67)
141 ± 10
143 ± 10
18.4 ± 5.5 17.5 ± 1.1
7.0 ± 0.3 7.2 ± 0.3 Pediatric faculty, n = 17
Family practice faculty, n = 4
100 (84-100)
100 (47-100)
138 ± 13 183 ± 40
18.1 ± 1.6 21.0 ± 2.1
7.6 ± 0.3
6.8 ± 0.4 Pediatric third-year residents, n = 18
Family practice third-year residents, n = 11
61 (36-82) 55 (23-83)
155 ± 15 178 ± 22
19.1 ± 1.4 20.5 ± 2.2
7.7 ± 0.2 7.5 ± 0.4 Pediatric first year residents, n = 23
Family practice first-year residents, n = 8
52 (31-73) 50 (16-84)
123 ± 6
189 ± 36
16.6 ± 0.7 21.5 ± 2.8
7.2 ± 0.2 7.8 ± 0.2
9Within each group (ie, faculty, private practitioners, first-year and third-year residents), family physicians and pediatricians performed
equally on each of the outcome measures.
: For the Percent Correct Decision, the standard exact asymmetric 95% confidence intervals are given in parentheses.
§Mean and standard error of the mean.
Recognizing that obtaining an adequate history does not neces-sarily require asking all of a predetermined list of questions, we developed the concept of critical areas. A critical area was defined as information that should materially influence the physician’s decision whether to see the patient promptly. For example, the state of hydration in the diarrhea case constituted a critical area that might be addressed adequately by one or more questions. Critical areas were reviewed for content validity by the five-physician panel. Only items about which there was 100% agree-ment were included (see the Appendix).
Data Analysis
Statistical analysis was performed using Statview II and KaleidaGraph on a Macintosh IIcx, and also using mainframe SAS. One-way analyses of variance were followed by the post hoc tests: Fisher LSD, Bonferroni and Scheffe multiple comparisons. All outcomes agreed for all applications. Across groups analyses were also checked diagnostically using the Kruskal-Wallis nonparametric
test. Statistical tests for counted data were done using chi-square,
contingency table analyses. Tests for trend in tables with ordered categories were examined using the Goodman-Kruskal Gamma and the Kendall taub statistic.
RESULTS
No significant differences were found between
spe-cialty groups (family practice vs pediatrics) in the
number of questions asked, the number of correct
decisions made, or in the number or type of critical
areas addressed for any of the cases (see Table 3 for
example). Because of the lack of difference in outcome
measures as well as the small sample size of some
groups (ie, family practice faculty), specialty groups
are combined and-data are shown based on level of
training and type of practice. However, all of the
findings presented below remained unchanged when
only the data from pediatricians were analyzed.
When evaluating a child who was severely ill with
diarrhea and dehydration, faculty physicians were
significantly more likely than all others to reach the
correct triage decision (Figure). All the faculty, but
only 44% of the private practitioners, made the correct
decision to evaluate the child urgently (P < .001).
First-year residents were more likely than all others
to see a child urgently when the case described a
patient with minor head trauma (P = .014) or fever
without serious illness (P = .005).
Experience was measured by number of years in
practice, number of pediatric patient visits per week,
and number of pediatric phone contacts per week.
100 90 80 C 0 U) #{149}3 70 U) U 0 U C
I:
10 0Figure. Correct decision by level of training. Nominal P values
are reported. ‘ P < .0001, faculty physicians vs all other groups;
t P = .014, interns vs all other groups; # P .005, interns vs all
other groups; U, first-year residents (n = 31); 0, third-year residents
(n = 29);
#{149}
private practice (n = 36);,
faculty (n = 21).These measures of experience were examined both
individually and jointly to determine if correct
deci-sions or number of critical areas addressed were
re-lated to experience. No significant correlation was
found between these measures of experience and
physician performance for private practitioners or
faculty in any of the cases.
Training in telephone decision making was not
correlated with any measure of performance; nor did
triage decisions differ significantly according to the
sex of the physician for any simulation. There was no
association between board-certification and any
out-come measure.
When evaluating the children with problems that
did not require urgent evaluation, many physicians
failed to obtain adequate histories to rule out serious
illness (Table 4). Sixty percent of physicians did not
determine circumstances surrounding the injury in a
child with head trauma sufficiently to rule out child
abuse. More than half of the subjects failed to
condition or was taking any medications other than acetaminophen.
While interns were less likely to inquire about the
presence of concomitant underlying ifiness in a child
(P < .05), there were no significant across-group
differences either in the total number of questions
asked or in the number of critical areas assessed.
Furthermore, neither the number of critical areas
ad-dressed nor the total number of questions asked was
associated significantly with correct decision making
for any of the groups in any of the three case
simu-lations. Of the four critical areas in the diarrhea and
dehydration simulation that indicated the child had a
significant medical ifiness (see Appendix A), at least
one was addressed by all study physicians. The
fre-quency with which any of the four critical areas was
addressed did not correlate with correct decision
mak-ing. Thus, more than one third of all residents and
private practitioners reached inappropriate
manage-ment decisions despite obtaining information that,
according to published guidelines,9’”3 should have
altered their decisions. Upon receiving the same
an-swers to their questions, all faculty physicians chose
the correct decision.
DISCUSSION
This study was designed to determine if
experi-enced physicians demonstrate superior telephone
triage skills as determined by the frequency with
which they reach the correct management decision
and by the nature of their questions. The design was
developed specifically to explore adequacy of clinical
judgment in an explicit manner and to avoid reliance
on a ‘cookbook’ approach or checklist to demonstrate
good management skills. Because of differences in
on-call responsibilities of faculty, private
practition-ers, and residents, we believe that biases among
groups were minimized by administering the cases to
all subjects under optimal conditions, with specific
instructions given as to conditions of the simulation.
For example, most interns and faculty physicians in
this sample did not take patients’ telephone calls at
night and thus could only be studied through the use
of case simulations.
For the two children without serious illness,
expe-rienced physicians were more likely to reach the
correct management decision than were first-year
res-idents. Thus, interns recommended that more
chil-then with benign, self-limited ifinesses be seen
im-mediately. Although such actions may increase the
cost of health care and may inconvenience patients
and their families, children who are examined are
unlikely to experience an adverse medical outcome.
More importantly, both residents and physicians in
private practice were likely to underestimate the
de-gree of illness in a child with severe diarrhea and
dehydration regardless of the information they
elic-ited through their telephone evaluations. Faculty
phy-sicians were much more likely than private
practition-ers to see the ill child promptly. This may be because
faculty physicians have more contact with seriously
ill children than those in private practice. Faculty
physicians might be more familiar or more in
agree-ment with practice guidelines such as those developed
by the American Academy of Pediatrics and these
results conceivably may reflect the differences in
de-cision making of policymakers/teachers and
practi-tioners. Despite instruction prior to the simulation
that the physician would be required to assess the
patient at a local emergency department in the event
the child warranted urgent evaluation, this finding
may reflect the fact that the task of examining a
patient is usually delegated by faculty to their house
staff. Under such circumstances, faculty might be
TABLE 4. Percent of Subjects Addressing Specific Critical Areas’
First-year Residents
Third-year Residents
Private Practice
Faculty Total
Case 1 : Diarrhea
Age 90 100 100 100 97
Fever 100 100 86 100 96
Stool character 90 93 75 95 87
Stool frequency 87 93 83 86 87
Vomiting 71 79 83 62 75
State of hydration 42 62 36 62 49
Urine output 77 70 64 71 72
Fluidintake 97 97 95 95 96
Severity of illness 74 73 81 71 75
Case 2: Head Trauma
Age 90 100 92 86 92
Loss of consciousness 65 66 72 71 68
Neurologic status 97 97 89 90 93
Otherinjury 48 31 44 42 42
Circumstances of in- 38 52 39 29 40
jury Case 3: Fever
Age 94 100 97 90 96
Heightoffever 100 93 94 95 96
Severity of illness 39 34 36 38 37
Other medical condi- 16 59 33 43 37
tions
S Percentage of subjects who asked questions about the areas considered critical by a consensus panel
for each case.
expected to request examination of more children
with self-limited illnesses; however, faculty
per-formed as well as private practitioners in detecting
children who did not require urgent evaluation.
Rather than reflecting a bias on the part of faculty
that children are likely to have serious illnesses, a
“wellness bias,”5 in which the most favorable
diag-nostic possibility is embraced, may be present in
experienced general physicians for whom the vast
majority of contacts are with patients who have mild
and self-limited illnesses. This bias may have
contrib-uted to the observed tendency of private practitioners
to minimize the severity of symptoms reported.
Al-ternatively, faculty may be more adept in listening
skills or may interpret the information they gather
differently from their colleagues in private practice.
Faculty physicians also may be less likely to have
trust in a family’s ability to continue close
communi-cation with the physician, and thus insist on earlier
evaluation.
Physicians who knew they were being recorded
and evaluated might be expected to be at least as
thorough as in an actual after-hours call situation.
Thus, our overall results should represent optimal
physician decision making. It was surprising,
there-fore, to find that even when physicians were not
under time constraints and knew that their skills were
being studied, their assessments were often
inade-quate and the consequent decisions incorrect.
Contrary to our hypothesis, neither the quantity
nor the quality of questions asked was associated
with any of the measures of experience examined.
We acknowledge that experience measured as the
number of years a physician has been in practice may
be confounded with the calendar time when the
physician completed training, but addressing this
is-sue was beyond the scope of this study.
Physicians who addressed a greater number of
critical areas were not more likely to reach the correct
triage decision. It has been suggested that the
ade-quacy of assessment and plan may not correlate with
the amount of information gathered or the number
of hypotheses generated.6’’17 Thus, factors other
than the information obtained by asking questions
recommended by experts in telephone decision
mak-ing would seem to be of greater importance in guiding
the telephone decisions of the physicians in our
sam-ple.
Although it was impossible to determine the exact
point during a simulation at which the physician
determined the severity of the problem and was
pre-pared to prescribe therapy, some physicians appeared
to make a triage decision early in the course of the
conversion, and then ‘shut off,’ ignoring subsequent
data which should have led them to reconsider their
decision.4 The following is an excerpt from an
inter-view with a third-year pediatrics resident who has
not yet adequately assessed hydration or the severity
of the child’s illness:
Resident: It sounds like your baby has a viral
gas-troenteritis but is well-hydrated and should do just
fine at home. What we worry about with these
little ones is that they can get dehydrated which
can make them very ill. We can tell this if the baby
isn’t urinating, if the mouth is dry ...
Mother: But doctor, I think her mouth is dry. Resident: Well, it’s often hard to tell that. Just keep
doing what you’re doing, and if she’s not better in
the morning, bring her into the office.
In combination with the ‘wellness bias,’ which can
be even more accentuated when the patient is not
physically present for the physician to evaluate, a
shutting out of information could have serious
con-sequences.
In postsimulation discussions, private practitioners
often commented that they relied on their knowledge
of most of their patients (and their mothers) when
making the decision about the seriousness of a child’s
ifiness. Unfortunately, today many practitioners do
not know all of the patients with whom they have
telephone contact.18 Even small groups often share
coverage, and large, staff-model health maintenance
organizations are becoming more prevalent. Many
primary care physicians no longer have the luxury of
relaying on a previously established relationship with
the patient to guide telephone decision making. Thus,
expertise in telephone management becomes
essen-tial.
As assessed by these case simulations, clinical
de-cisions by primary care physicians improved in some
areas with experience. However, many physicians in
private practice did not demonstrate the skills needed
to perform adequate telephone triage despite their
great number of patient contacts. We have found, as
have others using blinded simulations,4’7’8”9 that
phy-sicians performing telephone triage may not always
ask all the right questions. However, most physicians
in this study did obtain information sufficient to
determine the severity of a child’s illness. Thus,
fac-tors other than the type or amount of information
obtained may account for deficiencies in telephone
management. Physicians must not only ask good
questions, but must carefully listen to and evaluate
the answers if they are to manage patients effectively on the telephone.
ACKNOWLEDGMENTS
We are indebted to Dr Robert E. Merrill for his critical review of
our work and valuable suggestions. We would also like to thank
APPENDIX: CRITICAL AREAS
Simulation 1: Severe Diarrhea with Dehydration
1. Age of patient 2. Presence of fever 3. Description of stools
Defined by asking at least 2of the following:
Consistency
________Color
________Amount ________Blood ________Duration 4. Frequency of stools’ 5. Presence of vomiting
6. Urine output(Note: In the present simulation, the mother cannot estimate urine output, as every diaper is fified with watery
stool.) 7. Hydration’:
Defined by asking at least 1of the following: Presence of tears
Dry mucous membranes
Sunken fontanelle
Skin turgor Sunken eyes
8. Intake’
9. Presence of severe illness’:
Defined by asking at least 1of the following:
_________Will she smile
________Will she play
Question eliciting that child is whimpering
Is child easily comforted Is mother worried about her ________Does mother think she is very sick ________Does mother think she is getting worse
Questions on child’s activity Questions on child’s alertness
‘indicates areas which, in this simulation, the answer given to a single
question may have been sufficient to lead to urgent evaluation of the child.
Simulation 2: Minor Head Trauma, No Loss of
Consciousness
1. Child’s age 2. Rule out Abuse
Defined by asking at least 1of the following:
_______Mechanism of injury (What happened, Did you see it happen, how did he fall, etc.)
_______Past history of injuries, illnesses, hospitalizations 3. Loss of consciousness
4. Neurologic assessment
Current general appearance or at least 2 questions assessing
the following:
Walking normally
Talking normally
Normal behavior
Playing normally
Normal vision
Normal pupils/eye movements
________Moves all extremities Presence of headache Presence of seizures
Is child easily awakened
Presence of vomiting
Elicits that child is currently watching television ________Normal activity
________Presence of irritability 5. Other significant injury
Is he hurt anywhere? or
Defined by asking at least 2of the following: Bleeding
Lump Bruising ________Pain
Fluid from nose/ears
Simulation 3: Viral Syndrome with Fever 38.5#{176}C
1. Age of patient 2. Height of fever
3. Presence of severe illness
Defined by asking at least 2 of the following:
_______Will she smile
_______Will she play
Questions on child’s alertness
Questions on child’s responsiveness Talking normally
Tating normally
_______Assessment of fluid intake Is mother worried about her _______Does mother think she is very sick
_______Does mother think she is getting worse
Questions on child’s activity Is child easily comforted
4. History of underlying ilinesses/hospitalizations/medi-cations
REFERENCES
1. Mendenhall RC. Medical Practice in the United States. Princeton, NJ:
Robert Wood Johnson Foundation; 1981:32-47
2. Curtis, P. Talbot A. After hours calls: an aspect of primary care education.
IMed Educ. 1980;55:55-57
3. Radecki SE, Neville RE, Girard PA. Telephone management by primary care physicians. Med Care. 1989;27:817-822
4. Perrin E, Goodman H. Telephone management of acute pediatric illness.
N EnglJ Med. 1978;298:130-135
5. Brown S. Eberle B. Use of the telephone by pediatric house staff: a technique for pediatric care not taught. IPediatr. 1974;84:117-119 6. Curry T, Schwartz MW. Telephone assessment of illness: what is being
taught and learned? Pediatrics. 1978;62:603-605
7. Greitzer L Stapleton FB, Wright L, Wedgwood RJ. Telephone assessment of illness by practicing pediatricians. I Pediatr. 1976;88:880-882
8. Ott JE, Bellaire J, Machotka P. Moon JB. Patient management by tele-phone by child health associates and pediatric house officers. IMed Educ. 1974;49:596-600
9. Standards of Child Health Care. 3rd ed. Evanston, IL: American Academy of Pediatrics; 1977. Appendix K.
10. Strasser PH, Levy JC, Lam GA, Rosekraus J. Controlled clinical trial of
pediatric telephone protocols. Pediatrics. 1979;64:553-.557
1 1. Schmitt PD. Pediatric Telephone Advice. Boston: Little, Brown and
Corn-pany; 1980:39-94
12. Brown JL. Pediatric Telephone Medicine: Principles, Triage, and Advice.
Philadelphia: JB Lippincott; 1989:57-135
13. Katz HP. Telephone Manual ofPediatric Care. New York: John Wiley and Sons; 1982:33-155
14. Wood PR, LittlefieldJH, Foulds DM. Telephone management curriculum
for pediatric interns: acontrolled trial. Pediatrics. 1989;83:925-930 15. Nazarian LF. Odd thoughts on well-child care. PediatrRev.
1990;11:227-228
16. Fosarelli PD. The telephone in pediatric medicine. Clin Pediatr.
1983;22:293-296
17. Sloane PD, Egelhoff C, Curtis P, et al. Physician decision making over the telephone. I Fam Prac. 1985;21:279-284
18. Fosarelli P, Katz H Residents on the phone. Pediatrics. 1987;79:311-312 19. Isaacman DJ, Verdile VP, Kohen FP, Verdile LA. Pediatric telephone