 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 105 No. 4 April 2000, pp. 767-773
; Kimberly A. Freudigman, PhD*,
,, and
From the * Department of Pediatrics and Center for Medical
Informatics, Yale School of Medicine, New Haven, Connecticut.
 |
ABSTRACT |
|---|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Objective. To evaluate effects on the process and outcomes of care brought about by use of a handheld, computer-based system that implements the American Academy of Pediatrics guideline on office management of asthma exacerbations.
Design. A before-after trial with randomly selected, office-based Connecticut pediatricians. In both the control and intervention phases, physicians collected data from 10 patient encounters for acute asthma exacerbations. During the intervention phase, the computer provided for structured encounter documentation and offered recommendations based on the guideline of the American Academy of Pediatrics. Patients were contacted by telephone 7 to 14 days after the visit to assess outcomes.
Results. Nine study-physicians enrolled 91 patients in the
control phase and 74 in the intervention phase. Follow-up information
was available for 93% of encounters. Use of the intervention was
associated with increased mean frequency/visit of: 1) measurements of
peak expiratory flow rate (2.18 vs 1.57) and oxygen saturation
(1.12 vs .42), and 2) administration of nebulized 
2-agonists (1.25 vs .71). Visits in the intervention phase lasted longer and fees were
higher ($145.61 vs $103.11). There were no significant differences in
immediate disposition or subsequent emergency department visits, hospitalizations, missed school, or caretaker's missed work during the
7 days post visit.
Conclusion. Use of handheld computers that provide guideline-based decision support was associated with increased physician adherence to guideline recommendations; however, visits were prolonged, fees were higher, and no improvement could be demonstrated with regard to the observed intermediate-term patient outcomes. Guideline implementers (and users) should be cautious about putting unvalidated recommendations into practice. Key words: guideline adherence, asthma, computer-based decision support.
Practice guidelines have been devised for a large number of
medical conditions to diminish inappropriate variations in clinical practice, to control costs, and to improve patient outcomes. However, several factors have limited the success of guideline initiatives. Some
guidelines are difficult to apply because they are intrinsically unclear and incomplete.1 Others are based on unvalidated opinion rather than a cogent evidence base; in such cases, the consequences of imposition of the policies are unknown.2 Many physicians are resistant to the very concept of guidelines. Accustomed to functioning autonomously, they see the imposition of
practice policies as a threat to the traditional doctor-patient relationship.3 Finally, unless an effective implementation strategy is chosen, the expensive, resource-intensive guideline development process often fails to influence clinicians'
behavior.4-6
In a meta-analysis of studies selected for methodologic rigor, Grimshaw
and Russell7 reported improvements in the process of care
in 55 of 59 evaluations of guideline implementation. They found that
the best predictor of successful implementation was providing
guideline-based recommendations that were patient-specific at the time
and place of the consultation. Computer-mediated decision support
systems can process clinical data from individual patients and can
offer tailored advice at the point-of-care. A structured review of
evaluations of such decision support systems found improved clinician
performance in 43 of 65 trials.8
In 1994, the American Academy of Pediatrics (AAP) published a practice
parameter for management of asthma exacerbations in children who
present to an office setting.9 Asthma is the most common
chronic illness in childhood accounting for 25% of school absences;
the prevalence and rate of outpatient visits for asthma are rising, as
is the mortality rate.10 Total costs of illness related to
asthma exceeded $5.8 billion in 1987.11 Although effective
therapy exists, there is evidence of inconsistent and suboptimal care
of asthma in both children and adults.12-14
This report describes an investigation that evaluated the effectiveness
of a handheld, computer-based decision support device in implementing
the AAP asthma guideline. Specifically, we examined whether use of the
device led to improved adherence to the guideline regarding: 1)
measurement of peak expiratory flow rate (PEFR) and oxygen saturation,
2) prescription of corticosteroid, and 3) administration of oxygen.
Studies that investigate only process measures, eg, adherence to
guideline recommendations, may fail to identify policies with adverse
effects on health outcomes.15 Despite the best intent of
guideline developers, untested policies may have unanticipated side
effects. Because the AAP guideline recommendations had not been
previously validated, this study also explored whether adherence to the
guideline was associated with changes in immediate outcomes, eg,
improvement in the severity of the exacerbation or disposition from the
office, and other indicators of functional status in the week after the
office visit, including repeat office visits, delayed emergency
department (ED) visits and hospitalizations, missed school days, and
missed caretaker work days.
Guideline Recommendations
The AAP practice parameter on office management of asthma
exacerbations in the office setting was published in
Pediatrics in both algorithmic and tabular
formats.9 The developers suggested that the practice
parameter differed from common practice in recommending: 1) the use of
physiologic measures This was the first AAP guideline constructed with an evidence-based
development process. That process proceeds systematically through the
following sequence of activities: problem definition, comparison of
potential interventions, identification of health outcomes,
comprehensive literature review with structured abstraction of evidence
from the identified literature, development of evidence tables,
meta-analysis, assessment of benefits and harms, and development of
parameter recommendations (C. Herrerias, AAP, personal communication, 1997).
Selection and Randomization of Participants
Physician-subjects were drawn from a pool of 375 pediatrician
listings in the 1996 Fellowship Directory of the AAP, which categorizes members by state and city. The authors assigned a numeric
identifier to each pediatrician listed in Connecticut cities and towns
within a 20-mile radius of New Haven. In a rolling recruitment process,
the RAND random number tables were used to select candidates, who were
screened for eligibility and invited to participate until 11 participants had been recruited.
Eligibility criteria included pediatricians who: 1) actively practiced
primary care pediatrics within a 20-mile radius of New Haven,
Connecticut, 2) anticipated seeing 20 patients older than 5 years of
age with acute asthma exacerbations within the following year, and 3)
had equipment available in their offices for measurement of PEFR and
for providing supplemental oxygen if needed. Pediatricians in academic
practices, physicians-in-training, and subspecialists in allergy and
pulmonology were excluded. To achieve a randomization by practice unit
and to avoid potential contamination, only 1 physician from any group
practice was eligible for participation.
Eligible patients were children between 5 and 18 years old, who
presented to a nonhospital setting with acute exacerbations of asthma.
Physician-subjects and patients were masked with respect to specific
study hypotheses. They were told we were investigating "asthma care
for Connecticut children in which we would evaluate a computer-based,
decision support device and the costs of care." Data collection forms
were designed to resemble superbills and included checkboxes for
recording a number of items unrelated to the study hypotheses that
might be examined in any comprehensive study of asthma care, eg, blood
counts and chest radiographs ordered, epinephrine administered, and
theophylline prescribed. The study was approved by the Yale School of
Medicine Human Investigations Committee.
Study Design
The study was a randomized, prospective, before-after trial.
Physician-subjects served as their own controls, thus improving statistical power. Sample size calculations with the physician as the
unit of analysis indicated that a paired analysis of a sample of 10 physicians The intervention (known as AsthMonitor) consisted of a Newton
MessagePad 130 (Apple Computer Co, Cupertino, CA) outfitted with
custom-designed software and an Apple StyleWriter 1200 inkjet printer.
The system was designed for use at the point-of-care. It provided: 1)
structured documentation of the clinical encounter using a pen-stylus,
2) dynamically-generated recommendations based on the AAP practice
parameter, 3) assistance with calculation of predicted PEFR and
medication dosages, and 4) printed encounter summaries and
prescriptions. Study personnel provided in-office training in the use
of the intervention.
The system provided guideline recommendations in 2 formats. Reminders
for PEFR and oxygen saturation measurement were listed at the top of a
window for documentation of physical findings (which also listed
respiratory rate, alertness, dyspnea, retractions, color, chest sounds,
subcutaneous air, and pulsus paradoxus). Recommendations for
administration of oxygen, inhaled albuterol, and corticosteroids were
provided on screens that appeared when appropriate combinations of
findings triggered them. Null entries for PEFR and oxygen saturation
measurements were permitted. Likewise, users were not obliged to follow
any recommendation; physicians could deselect any recommendations and
make alternate choices. Users could retrieve explanatory information
that supported each recommendation by tapping an "Information
Button" next to each recommendation.
Sampling of Records and Data Collection
Physicians completed a data form for each patient during both
the control and intervention phases, which documented the severity of
the exacerbation at presentation and discharge (mild, moderate, severe, and resolved); tests performed in the office; medications and
treatment performed in the office; prescribed medications; the duration
of the visit, defined as time from initial contact with a clinician
until discharge from the office; immediate patient disposition; and
total fee charged. Pediatricians determined the severity of the
exacerbation based on a table of physical signs and physiologic
measurements from the published guideline (reproduced on the data
collection form). PEFR was considered to have been measured if either
an actual measurement was obtained or the physician recorded that an
effort was made but the child was unable to cooperate. Data forms,
office notes, and (during the intervention phase) AsthMonitor encounter
summaries were faxed to the study office after each encounter. Parents
and/or patients were contacted by telephone 7 to 14 days after the
office visit by 1 of the authors to ascertain outcomes in the 7 days
after the initial office visit. A structured set of questions was read
to the patient, eg, "In the 7 days after your (child's) visit to
Dr____'s office, did you miss any school? If so, how
many days?" A study assistant (masked to study hypotheses) entered
all data into a database and all entries were rechecked for accuracy.
Each physician supplied data on customary charges for asthma-related
services before beginning patient enrollment. A questionnaire was
administered at the conclusion of the study to assess practitioners'
reactions to AsthMonitor.
Analysis
Adherence rate was defined as the proportion of visits at which
a physician performed an intervention in compliance with a guideline
recommendation (eg, measured PEFR or prescribed corticosteroids) compared with the number of visits at which the intervention was appropriate. The AAP guideline recommended: 1) measurement of PEFR
(and/or oxygen saturation) in all patients who presented with acute
asthma exacerbations; 2) administration of supplemental oxygen to all
patients who presented with moderate or severe exacerbations, and 3)
consideration of prescription of steroids for any patient who presented
with an asthma exacerbation. We designated the prescription rate of
corticosteroids Because guideline adherence is strongly dependent on each physician's
knowledge, abilities, and acceptance of guideline recommendations, the
computation of adherence rates was based on the individual physician as
the unit of analysis.16 This permitted a paired analysis
of adherence data for each physician during the control and
intervention phases. Continuous variables (eg, adherence rates, number
of nebulization treatments, and fees) were analyzed using t
tests and analysis of variance. Categorical variables (eg, nebulization
treatment General
Data collection began on September 30, 1996 and the study was
terminated on October 1, 1998. The average time interval for each
physician to complete phase I was 228 days (range: 65-361 days); the
average time interval in phase II (before completion or study
termination) was 192 days (range: 60-336 days). A schematic summary of
the study is shown in Fig 1.
METHODS
Top
Abstract
Methods
Results
Discussion
References
PEFR and oxygen saturationto better assess the
severity of asthma exacerbations, 2) increased frequency and dosage of
2-agonists, and 3) increased use of corticosteroids.9
each of whom enrolled 10 patients in the control and
intervention phaseswould have a power of .80 to detect a 20%
difference in adherence rates. On enrollment, physician-participants provided informed consent and were instructed in the selection of
patients and completion of study forms. Each pediatrician completed a
questionnaire that supplied information on personal demographics, computer experience, and fees. In the control phase, each physician enrolled 10 consecutive eligible patients and managed them in the
conventional manner. Next, in the intervention phase, physicians were
asked to enroll 10 additional patients using the handheld, computer-based system (see below) to assist care. Physicians received $200 at the conclusion of the study in partial compensation for time
spent.
ie, the number of encounters for which the physician
prescribed steroids divided by the total number of encounters for that
physicianas an adherence rate. We also tabulated the mean number of
times per visit that a repeated intervention (eg, PEFR measurement or
nebulization treatment) was performed.
yes or no) were analyzed using 
2
analysis. Analysis of covariance was used to control for the effects of
severity on outcome data. Patient outcomes were calculated as the
proportion of patients who experienced that outcome compared with all
patients in that phase of the study. All statistical calculations were
2-tailed and performed with SPSS 8.0 (Chicago, IL) and EpiInfo 5 (Centers for Disease Control and Prevention, Atlanta, GA).
RESULTS
Top
Abstract
Methods
Results
Discussion
References
View larger version (45K):
[in a new window]
Fig. 1.
Schematic summary of study.
Of the 375 listings of pediatricians within a 20-mile radius of New
Haven, 138 were eliminated because of categorization as residents or
fellows in-training (n = 80) or an address at a medical center or medical school (n = 58). The remaining 237 were subjected to a random, rolling recruitment process. Of those
selected, 28 were not in active primary care pediatric practice22
were retired, working in administration, or practicing part-timeand 6 had moved away; 4 did not anticipate seeing 20 asthma patients in the
coming year; 7 did not have required equipment; and 5 were partners of participants who had been enrolled previously. Eighteen potentially eligible physicians declined to participate in the study, 7 of whom
belonged to a single multisite group practice whose administration proscribed participation despite the willingness of individual group
members to take part. Of the 11 physicians who enrolled in the study, 2 dropped out during the control data collection phase1 because of a
move out-of-state, and the other because of excessive workload. Data on
the 11 control phase patients they submitted were eliminated from the
analysis because of our inability to make comparisons with intervention
data.
During the control phase of the study, physician-participants enrolled
91 patientseach of the 9 physicians enrolled 10 patients (1 enrolled
11). In the intervention phase, 74 patients were enrolled6 physicians
each enrolled 10 patients and 3 physicians enrolled 8, 5, and 1 patient(s), respectively. In calculating changes in adherence rates, we
eliminated data from the single physician, who enrolled only 1 patient
in the intervention phase. For each of the remaining 8 physicians, we
calculated the adherence rate in each phase and performed a paired
analysis. Follow-up data were collected from 92% of control patients
(84/91) and 93% of intervention patients (69/74).
Characteristics of the 9 study physicians are summarized in Table 1. The distribution of practice settings seemed to fairly reflect the practice milieu in this part of New England. The majority of physicians were relatively computer-naïve by self-description. The average age of the patients enrolled during the control phase was 10.3 years (range: 5.0-17.4) and during the intervention phase was 10.8 (range: 5.0-17.8).
|
Guideline Adherence
Adherence to guideline recommendations for assessment of PEFR and
oxygen saturation, office administration of -agonists and oxygen, and prescription of corticosteroids increased during the intervention phase of the study (Table
2). Mean adherence rate for measurement
of PEFR was .86 at baseline and rose to .94 with the intervention
(P = .32). Adherence rate for measurement of oxygen
saturation nearly doubled (.29-.56; P = .007). The
difference between the practitioners' rates of adherence in the
control and intervention phases approached statistical significance for
-agonist nebulization treatments (.73-.91; P = .064) and corticosteroid prescription (.43-.57; P = .055). As shown in Table 3, the number of
measurements per visit of PEFR and oxygen saturation and the number of
nebulization treatments administered per visit also rose significantly
during the intervention phase.
|
|
Clinicians recorded the duration of office visits as <30 minutes, 30 to 60 minutes, and >1 hour. During the control phase, 44% of the
visits lasted <30 minutes and 46% lasted 30 to 60 minutes. Only 10%
lasted more than an hour. During the intervention phases, visit
duration was prolonged. Only 15% of visits lasted <30 minutes, while
67% lasted 30 to 60 minutes and almost twice as many lasted >1 hour.
The difference in visit duration was statistically significant (2 = 15.62; P < .0004).
Use of the intervention was also associated with increased charges. The average fee charged by the study physicians during the control phase was $103.11 and rose to $145.61 during the intervention phase. This may be accounted for by charges for the increased number of interventions described above or by upcoding attributable to the more extended nature of the visits. Among the 8 physicians (from whom fee schedules were obtained at enrollment), the most commonly used Current Procedural Terminology code for asthma visits was 99213 for which the median fee was $65.00. The median fee for the next level of Current Procedural Terminology code 99214 was $77.50. The median charge for PEFR measurement was $12.50 and for oxygen saturation measurement was $30.00 (although some physicians did not bill repeatedly for physiologic measures), and for nebulization treatment was $30.00.
At presentation in the control phase, 78% (n = 71) of
the patients were assessed to have mild exacerbations and 22%
(n = 20) moderate exacerbations. During the
intervention phase, 59% (n = 44) were judged to have
mild exacerbations, 36% (n = 27) had moderate
exacerbations, and 4% (n = 3) had severe exacerbations at presentation. The increased severity of illness during the intervention phase was statistically significant
(2 = 8.72; P = .0128) and
introduced a potential confounder into the analysis. No other
confounding variables were identified with univariate analysis.
To control for the potential confounding effect of the presenting severity of the exacerbation, we performed analysis of covariance and examined the effect of the intervention. The analysis showed a statistically significant effect of the intervention on the number of PEFR measurements (F1,158 = 8.64; P < .01), the number of oxygen saturation measurements (F1,159 = 8.5; P < .01), the number of nebulization treatments (F1,153 = 12.3; P < .001), and fees (F1,156 = 14.14; P = .0002).
According to the guideline, 50 patients should have received oxygen
treatment because they presented with moderate or severe exacerbations20 in the control phase and 30 in the intervention phase. Although no physicians administered oxygen to their patients during the control phase, 3 of the 30 recommendations to administer oxygen from the AsthMonitor system were heeded. In their responses to
the exit questionnaire, clinicians made it clear that they often
disagreed with the recommendation to prescribe oxygen: "Did not feel
the need was acute enough to treat with oxygen," "I ... feel oxygen recommendation (was) overly given," "Recommendations with regard to giving oxygen and oral steroids more conservative than what I
felt necessary," "I felt oxygen prescribed too often by AsthMonitor," and "Too cumbersome to bring out oxygen tank when I
know pt will improve readily with nebulization therapy."
Patient Outcomes
Comparing the physicians' assessments of each patient's asthma
severity at presentation and discharge, during the control phase 43.3%
of patients (n = 39) improved over the course of the office visit, whereas 57.7% (n = 41) improved during
the intervention phase. This relative advantage for short-term
improvement approached statistical significance
(2 = 3.30; P = .069).
There was no difference in immediate patient disposition from the office between control and intervention phases. Almost all patients were discharged from the hospital with only 2 in 88 control patients and 1 in 73 intervention patients requiring transfer to the ED or hospitalization directly from the office setting (Table 4).
|
During the week after the visit, there were no significant differences between the control and intervention phases in the number of children who missed school (or the number of days they missed), the number of their caretakers who missed work (and the number of missed workdays), and the number of children who revisited their pediatrician's office (Table 4). When we controlled for discharge severity as a potentially confounding variable, analysis of covariance showed no significant difference in missed school days (F = .63; P = .42) or missed caretaker work days (F = .25; P = .62) attributable to the intervention.
There was a small difference (not statistically significant) in the percentage of children who required ED visits and hospitalization in the week after discharge but the absolute number of patients was quite small.
Because we found no significant difference in outcomes between the
control and intervention phases, we examined the statistical power of
this study to detect an effect. The null hypothesis is that the
outcomes are the same in both the control and intervention phases.
Using the number of patients actually enrolled in the study and the
observed frequencies of intermediate term outcomes during the control
and intervention phases, we calculated (the probability of a type
II error) and power. We can be >90% confident that we could have
detected a true difference (if it existed) between frequencies of
missed school, missed work, and office revisit in the control and
intervention with the sample size that we used. Power was considerably
lower for avoiding a type II error with respect to ED visits (.55) and
hospitalization (.62).
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Use of a handheld, computer-based guideline implementation system
was associated with increased adherence to the AAP guideline on office
management of acute asthma exacerbations. Physicians measured PEFR and
oxygen saturation more frequently, administered nebulizations more
often, and there was a tendency toward more frequent prescription of
systemic corticosteroids and nebulization of -agonists when
pediatricians were prompted by the system. Physicians resisted
recommendations to administer oxygen to patients, however.
Not surprisingly, providing more services tended to prolong visits and
was associated with higher fees. Although there was a tendency toward
greater immediate clinical improvement in the children whose physicians
used AsthMonitor, the additional procedures performed in compliance
with the guideline recommendations did not seem to meaningfully improve
intermediate-term outcomesmissed school, missed work by caretakers,
office revisits, delayed ED visits, and hospitalizations in the week
after the visit.
Other investigators have found similar, unanticipated side effects of successful guideline implementation. Gleason et al17 noted that implementation of American Thoracic Society guidelines for management of community-acquired pneumonia led to 3- to 10-fold higher antimicrobial costs without any difference in patient outcomes. Suarez-Almazor et al18 evaluated the potential impact of Agency for Health Care Policy and Research guidelines for management of acute low-back pain and found that adherence would increase the use of radiographs by 238%. Using computer technologies, Safran et al19 showed that although electronic medical record alerts dramatically improved physicians' response times to clinical events for patients with human immunodeficiency virus, no change in admission rates, ED visits, admissions for pneumocystis, or survival could be demonstrated. A recent review of the effectiveness of guidelines in improving patient outcomes in primary care concluded that there is need for more research on guidelines implementation that includes outcome assessment.15
A more careful look at the specific interventions that were advised by
this guideline may help to explain the apparent paradox of improved
physician adherence and unchanged patient outcomes. Although
measurement of arterial oxygen saturation has been shown to predict
hospital admission in children who present to the ED with
wheezing,20 the value of measurement of PEFR during acute
exacerbations has not been studied systematically.21 No
evidence was presented in the AAP guideline to support its recommendation for repeated assessment of physiologic functions. Likewise, it is not common practice for pediatricians to administer oxygen when children present with exacerbations of moderate severity in
the office setting22 and the guideline offered no evidence
to suggest the value of oxygen therapy. Increased numbers of
nebulization treatments with -agonists during the intervention phase
may have resulted in the tendency toward greater immediate clinical
improvement that was observed. An increase in the frequency of
prescription of corticosteroids might be expected to have an effect on
intermediate term outcomes. Unfortunately, this study did not
demonstrate a substantial increase in steroid prescription rate, so the
potential impact of wider use of antiinflammatory agents on outcomes
cannot be determined.
The AAP practice parameteras published in
Pediatricslacked explicit statements regarding evidence
quality and strength of recommendations. It was supported by a
technical report that was made available from the AAP for a nominal
charge.23 The technical report noted a lack of specific
evidence in the medical literature to support many of the
recommendations in this practice parameter. As is the case with many
guideline development efforts, in which panel members struggle to
produce useful products amid deficits in the evidence base, this panel
relied on clinical experience and expert opinion for many of its
recommendations. The technical report stated that the systematic
evidence review performed during development of the guideline only
addressed issues of -agonist dosing and steroid toxicity. All other
recommendations were based on expert consensus, achieved through a
nominal group process.
The small number of participants in this study raises justifiable questions about the generalizability of the findings. We believe that the random sample of physicians was representative of community pediatric practice in Connecticut, but the extensive literature on practice variation suggests that any geographically limited sample may be unique. In addition, the baseline adherence rate to some recommendations was unexpectedly high suggesting the possibility of a Hawthorne effect. However, despite the small numbers, we were able to detect differences in guideline adherence that were statistically and clinically significant.
Our study describes associations between the use of the intervention and the observed behavior changes and outcomes, not cause-and-effect relationships. It is possible that the changes in behavior that were found reflect influences external to the study intervention. The time-series study design is subject to time-dependent confounding and this study did not control for secular trends in asthma management. The National Heart Lung and Blood Institute's Expert Panel Report II was published near the end of this study and may have influenced practice in the intervention phase.21 However, other investigators have shown that simply publishing recommendations is often ineffective in changing physician behavior.4,24 Moreover, the before-after approach is the most commonly applied design for evaluation of computer-based, clinical decision support systems.25,26 Because evaluations of computer-based interventions tend to be resource-intensive and have profound impacts on organizational dynamics, the before-after design leverages the contributions of smaller numbers of participants. Innate characteristics of the participants are not merely balanced, but they are actually eliminated as confounding variables with a paired before-after design.27
Successful use of computer-based decision support systems to improve
guideline adherence has been reported for a number of medical
conditions,8 but to our knowledge, this is the first
report that evaluates the use of these technologies in the management
of childhood asthma. The AsthMonitor system was designed to promote
integration of guideline recommendations into clinical workflow. It
provided a number of information management servicesincluding recommendation, documentation, explanation, registration, presentation, and calculationintended to increase its perceived value and offset the inconvenience associated with a change in work
habits.28 The vast majority of previous studies have been
performed in academic settings, whereas this work investigates
community-based care. In addition, this report describes the use of a
handheld, pen-input device, whereas other studies of decision support
tools have primarily investigated desktop computers as providers of
guideline recommendations.
Implications for Clinical Practice
Although possibly flawed in terms of its claim to being evidence-based, the AAP guideline clearly described recommendations for changes in asthma management that were widely perceived as best practices. Although an ideal scenario would provide for validation of guideline recommendations before publication, this is impractical for most guideline development efforts. The development process often takes several years from topic nomination to final publication. Additional delays for extensive testing would often be associated with the appearance of new evidence that might lead to guideline revision, resulting in a continuous development process. Moreover, a budget for testing would greatly augment the already high costs of guideline development.
Computer-mediated reminders presented at the time and place of a
consultation provide powerful enabling tools to improve the fidelity
between physicians' intentions and actions.29 Because of
their capability to influence behavior, they must be applied wisely and
responsibly. As with all sources of medical knowledge (eg, journal
articles, textbooks, and consultants), guideline users should not
accept statements on faith even when the guideline is described as
evidence-based. Guideline developers would assist both policy
implementers and end-users in the proper assessment of the quality of
guideline knowledge if they annotated each recommendation with an
indicator of evidence quality and/or strength of expert opinion. The
AAP included such annotations in a subsequently published practice
parameter.30 Potential usersincluding managed care
organizations, hospital systems, and individual cliniciansshould
critically evaluate each recommendation before
implementation.31,32
Implementation of guideline recommendations using handheld computers seems to be an effective mechanism for influencing physicians' behavior. We believe that as the technology matures, we can look forward to the appearance of decision support applications that assist clinicians with management of a wide variety of disorders. We predict that ultimately the integration of flexible data entry, guideline-based decision support, and a comprehensive computer-based patient record will transform pediatric care.
| |
ACKNOWLEDGMENTS |
|---|
This work was supported by the Robert Wood Johnson Foundation and the National Library of Medicine through Grants 1-R29-LM 05552-01A1 and T-15-LM07056. Dr Shiffman is a Robert Wood Johnson Generalist Physician Faculty Scholar.
We gratefully acknowledge the efforts of the following physicians and their office staffs in support of this study: David Bonheim, MD; Marguerite Dillaway, MD; Greg Germain, MD; Michael Rokosky, MD; Craig Summers, MD; Juan Hernandez-Trujillo, MD; Ellen Wolfson, MD; and the other anonymous participants.
We also thank Donald Miller, MD, for his work on the initial design and programming of the AsthMonitor software.
| |
FOOTNOTES |
|---|
Received for publication Mar 2, 1999; accepted Jul 28, 1999.
Reprint requests to (R.N.S.) PO Box 208009, TMP-3, New Haven, CT 06520-8009. E-mail: richard.shiffman{at}yale.edu
| |
ABBREVIATIONS |
|---|
AAP, American Academy of Pediatrics; PEFR, peak expiratory flow rate; ED, emergency department.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
United States, 1980-1993.
MMWR CDC Surveill Summ
1996;
45:350-353 This article has been cited by other articles:
 |
|
 |
|
  S. M. Strayer, L. K. Rollins, and J. R. Martindale A handheld computer smoking intervention tool and its effects on physician smoking cessation counseling. J Am Board Fam Med, July 1, 2006; 19(4): 350 - 357. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Sanders and D. Aronsky Biomedical Informatics Applications for Asthma Care: A Systematic Review J. Am. Med. Inform. Assoc., July 1, 2006; 13(4): 418 - 427. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Halterman, K. M. McConnochie, K. M. Conn, H. L. Yoos, P. M. Callahan, T. L. Neely, and P. G. Szilagyi A Randomized Trial of Primary Care Provider Prompting to Enhance Preventive Asthma Therapy Arch Pediatr Adolesc Med, May 1, 2005; 159(5): 422 - 427. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. B. Johnson, J. R. Serwint, L. M. Fagan, R. E. Thompson, and M. H. Wilson Computer-Based Documentation: Effect on Parent and Physician Satisfaction During a Pediatric Health Maintenance Encounter Arch Pediatr Adolesc Med, March 1, 2005; 159(3): 250 - 254. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Fischer, T. E. Stewart, S. Mehta, R. Wax, and S. E. Lapinsky Handheld Computing in Medicine J. Am. Med. Inform. Assoc., March 1, 2003; 10(2): 139 - 149. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Ratib, J. M. McCoy, D. R. McGill, M. Li, and A. Brown Use of Personal Digital Assistants for Retrieval of Medical Images and Data on High-Resolution Flat Panel Displays RadioGraphics, January 1, 2003; 23(1): 267 - 272. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Rothschild, T. H. Lee, T. Bae, and D. W. Bates Clinician Use of a Palmtop Drug Reference Guide J. Am. Med. Inform. Assoc., May 1, 2002; 9(3): 223 - 229. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Mellman, A. L. Miller, E. M. Weissman, M. L. Crismon, S. M. Essock, and S. R. Marder Evidence-Based Pharmacologic Treatment for People With Severe Mental Illness:A Focus on Guidelines and Algorithms Psychiatr Serv, May 1, 2001; 52(5): 619 - 625. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. N. Shiffman, B. T. Karras, A. Agrawal, R. Chen, L. Marenco, and S. Nath GEM: A Proposal for a More Comprehensive Guideline Document Model Using XML J. Am. Med. Inform. Assoc., September 1, 2000; 7(5): 488 - 498. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2000 American Academy of Pediatrics. All rights reserved. |
||
|
||
