Keywords
anesthesiology - education - educational needs - obstetrics/gynecology - surgical
and anesthesia information systems
Background and Significance
Background and Significance
The Accreditation Council for Graduate Medical Education (ACGME) establishes minimum
case experience requirements for trainees in core residency programs and subspecialty
fellowship programs.[1] Programs are required to ensure that graduates are “able to competently perform
all medical, diagnostic, and surgical procedures considered essential for the area
of practice,” in part, through demonstrating compliance with these case minimums.
In the subspecialty of obstetric anesthesiology, most case minimums are easily met
through routine presence on the labor and delivery suite. This includes vaginal and
cesarean deliveries involving high-risk maternal and fetal conditions and obstetric
antenatal procedures including cervical cerclage, external cephalic version, and fetal
surgery ([Table 1]). In contrast, when pregnant women require nonobstetric antenatal procedures, their
surgery and anesthesia care is often delivered in perioperative areas remote from
maternal care units and not uncommonly by nonobstetric anesthesiologists.[2]
[3] This presents a unique challenge to the fellowship program, as the scheduling and
planning of these nonobstetric surgeries must be somehow communicated to fellows to
facilitate their participation in and learning from these cases. Historically, programs
may have utilized unreliable methods for this communication, including phone calls
from schedulers or anesthesia in-room providers to facilitate obstetric anesthesiology
fellow presence for these procedures; however, these methods are fraught with multiple
opportunities for communication breakdown and failure. Prior work suggests that communication
challenges within the health care system can be successfully overcome through the
use of informatics-based solutions.[4]
[5]
[6]
Table 1
Accreditation Council for Graduate Medical Education case minimums for obstetric anesthesiology
fellowship
|
Case types
|
Minimum required number of cases
|
|
High-risk maternal comorbidity
|
|
|
Vaginal delivery
|
30
|
|
Cesarean delivery
|
30
|
|
High-risk fetal condition
|
|
|
Vaginal delivery
|
30
|
|
Cesarean delivery
|
20
|
|
Antenatal procedures
|
10[a]
|
a Limited to no more than five cases accrued from cervical cerclage placement or removal.
Objectives
Thus, we designed and implemented a quality improvement project which created an informatics-based
electronic notification system, with the aim of increasing fellow participation in
nonobstetric antenatal surgeries performed during pregnancy.
Methods
This quality improvement study was determined to be exempt by the Human Research Protections
Program at Vanderbilt University Medical Center. In December 2014 an automated email
notification system to inform obstetric anesthesiology fellows of scheduled nonobstetric
surgeries in pregnant patients was implemented. Prior to December 2014, cases were
identified by nurse practitioners in the anesthesiology preoperative evaluation clinic.
If the patient was noted to be pregnant, the nurse practitioner was instructed to
notify the obstetric anesthesiology attending on service, who communicated by word-of-mouth
to the fellow the patient's information and date and time of surgery. As a backup
means of identifying these cases, labor and delivery charge nurses were also requested
to notify fellows whenever requests for preoperative fetal monitoring were received
from the main operating room.
Following implementation of this quality initiative, nonobstetric cases in pregnant
patients were identified via daily automated query of the preoperative evaluation
database looking for structured documentation of current pregnancy ([Supplementary Material A], available in the online version). A functionality which identified positive point-of-care
urine pregnancy test results and assessed laboratory values for plasma human chorionic
gonadotropin indicating pregnancy (value > 2 mIU/mL) was later added in June 2016.
Information on flagged cases including patient medical record number, operating room
location, and date and time of procedure were communicated to fellows via automated
email daily at 7:00 a.m. ([Supplementary Material B], available in the online version).
ACGME case logs for obstetric anesthesiology fellows from July 2013 to July 2018 were
accessed, and case numbers for antenatal procedures, including both obstetric and
nonobstetric surgeries, were recorded. Median nonobstetric antenatal procedures per
quarter completed by fellows before and after implementation of the automated email
notification system were compared using an exact Wilcoxon-Mann-Whitney test. The fraction
of fellow-logged antenatal cases per year representing nonobstetric procedures was
compared using a Fisher's exact test. To assess for possible changes in the overall
number of nonobstetric antenatal procedures performed at our institution during the
study period, the structured query language was retrospectively applied to records
in the pre- and postimplementation years to identify the total number of flagged cases
annually.
Results
Prior to implementation of the automated email notification system in December 2014,
fellows participated in 1 and 3 nonobstetric antenatal case, respectively, in academic
years 2013 to 2014 and 2014 to 2015. Following implementation, fellows participated
in 11, 25, and 16 nonobstetric antenatal cases in academic years 2015 to 2016, 2016
to 2017, and 2017 to 2018 ([Fig. 1]). This represents a significant difference in quarterly nonobstetric cases, with
fellows performing a median of 0 (interquartile range [IQR] 0,1) cases/quarter (n = 6 quarters) prior to implementation and 3 (IQR 1,6) cases/quarter (n = 14 quarters) following implementation (p = 0.007). Additionally, yearly nonobstetric antenatal cases completed by fellows
as a percentage of total antenatal cases completed increased from 14% in preimplementation
years to 52% in postimplementation years (p < 0.001). There was no significant difference in total nonobstetric antenatal cases
performed at our institution throughout the study period (average 37 cases per year
preimplementation vs. 31 cases per year postimplementation, p = 0.34).
Fig. 1 Total number of obstetric and nonobstetric antenatal surgery cases logged by obstetric
anesthesiology fellows in each academic year from July 2013 through June 2018. Arrow
indicates timing of intervention, with automated email system initiated in December
2014, halfway through academic year 2014 to 2015. Obstetric antenatal cases include
cervical cerclage, external cephalic version, and fetal surgery cases.
Discussion
This report describes the successful implementation of an informatics-based solution
to a clinical education problem, namely augmenting identification of nonobstetric
antenatal surgery cases with the goal of greater obstetric anesthesiology fellow participation
in such cases. We observed a considerable increase in the number of nonobstetric antenatal
cases logged by fellows in the 3 years following implementation.
Our results indicate that informatics-based communication solutions may be superior
to traditional means of communication when approaching obstacles within the health
care system. The original case notification system for obstetric anesthesia fellows
at our institution had numerous vulnerabilities which likely contributed to its failure,
including (1) multiple personnel serving in the various roles involved in communicating
cases, (2) infrequent occurrence of cases preventing reinforcement of the communication
procedures, and (3) multiple points of potential communication breakdown. By leveraging
technology, all of these vulnerabilities were overcome to ensure proper communication
of cases to fellows. Prior work has similarly demonstrated improved and more efficient
communication between health care entities using automated notification systems. In
a study of automated pager notifications following computerized provider order entry,
Jacobs et al were able to demonstrate shorter times to order implementation of both
radiographic studies and respiratory therapy treatments.[7] In another example, Weingart et al assessed whether automated electronic notifications
to providers regarding incomplete imaging tests could improve test completion rates.
They found higher rates of completion in the intervention group, although some baseline
differences in patient characteristics may account for their findings.[5]
There are also numerous ways in which an informatics solution similar to the one proposed
here can be applied to improve processes in graduate medical education. First, the
process of identifying educationally high-yield cases or achieving case minimum requirements
for trainees can be similarly applied in other specialties. By automating the identification
of such cases, program directors can ensure that educational opportunities are recognized
in a timely fashion and appropriately communicated to trainees. Second, this type
of informatics solution can be applied to other areas of graduate medical education
in which electronic capture of data may be more timely or accurate than current practices.
For example, automated operative case reporting systems have been shown to increase
the capture of operative cases for general surgery interns compared with manual case
entry by residents.[8] Similar results have been shown for cardiology fellows, where procedure logs automatically
generated from the clinical record increased the number of documented procedures and
reduced administrative time required for logging of cases.[9] Patient attribution, or attributing patient-level metrics to providers, can be modeled
using electronic health record data to improve the capture of trainee provider–patient
interactions, which is vital in providing appropriate feedback and evaluations to
trainees.[10] In another anesthesia-specific example, Starnes et al describe the development of
an automated case cancellation review system for anesthesiology residents to better
understand the causes and consequences of day-of-surgery case cancellations, which
is important for their education pertaining to systems-based practice as required
by the ACGME.[11] Understanding the value and benefits of informatics-based educational solutions
in graduate medical education, other institutions can certainly adapt the methods
we present here to a variety of clinical situations using the structured query language
included in supplementary material as a possible starting point.
In interpreting these results, it is important to consider multiple limitations to
the study design. First, there are other possible reasons for the increase in fellow
participation in nonobstetric antenatal cases throughout the study period. In conjunction
with the automated email system, communications from the program director also emphasized
the importance and educational value of these cases, which could have influenced fellows
to seek out these opportunities more actively. Additionally, this increased emphasis
could have resulted in more reliable logging of cases by fellows into the ACGME system
during the postimplementation years. We also considered whether the observed increase
may have been due to an overall increase in the number of nonobstetric antenatal cases
performed at our institution during the study period. However, by retrospectively
applying the structured query language to records in the pre- and postimplementation
years, we found no change in the average annual number of positive query outputs.
Another limitation of this work is that we were unable to define a gold standard for
identification of a pregnant patient for nonobstetric surgery. This difficulty was
structural—performance of using structured data to document a pregnancy was presumably
not perfect, laboratory studies were not ordered on all patients, administrative billing
and coding data could not be relied on as not all surgical encounters took place within
the setting of an inpatient admission, and we did not collect ongoing accuracy data
from our fellows. This precluded us from calculating performance characteristics of
our automated script.
Post-2018 fellow case log data was not included in our analysis as this email-based
alert system was ultimately discontinued with the implementation of Epic (Epic Systems,
Madison, Wisconsin, United States) at our institution. As part of the Epic implementation,
similar functionality was created as part of the go-live build. Specifically, a status
board was developed that displayed cases in the main operating room where the patient
was marked as having a current pregnancy. This was achieved using a custom extension
which filtered cases with a rule-based approach using documentation of an active pregnancy.
Although the effort for this implementation was relatively small, being completed
by a physician builder with only several hours of effort, it was not well utilized
in the initial implementation due to lack of active alerting of fellows. The workflow
was subsequently revised to include inbasket messages sent within Epic that are triggered
by the scheduling and rescheduling of these cases via case tracking events that are
distributed to an inbasket pool, which includes our fellows. Additionally, efforts
were made to encourage structured documentation of active pregnancies for surgical
patients as part of the surgical and anesthesia preprocedure workflows. This reimplementation
was conducted in a multidisciplinary fashion in coordination with our education personnel
and took approximately 6 months to complete.
Conclusion
In conclusion, through the use of an easy-to-implement, low-cost, automated email
system to identify nonobstetric antenatal surgical procedures, we were able to demonstrate
an increase in the number of these cases completed by fellows during the 3 years following
implementation.
Clinical Relevance Statement
Clinical Relevance Statement
Modern training institutions are large and intricate, and informatics systems are
uniquely suited to accomplish complex educational initiatives which require the transfer
of information within those systems. Importantly, these automated informatics solutions
do not add, but rather reduce, human resource burdens, placing pertinent information
directly into the hands of stakeholders with the greatest interest—in this case, trainees
and program directors.
Multiple Choice Questions
Multiple Choice Questions
-
Which of the following is a benefit of an informatics-based notification system over
a personal communication-based system?
Correct Answer: The correct answer is option d.
Each of the answers a to c represent benefits of an informatics-based system.
-
Which of the following outcomes has been shown to result from implementation of an
informatics-based notification or reporting system?
-
Decreased operative case cancellations.
-
Decreased logging of ACGME-required cases.
-
Increased trainee involvement in educational cases.
-
None of the above.
Correct Answer: The correct answer is option c. The work presented here supports the use of an automated
query and email system to identify nonobstetric antenatal cases during pregnancy.
With the implementation of the system, trainee involvement in those cases rose appreciably.
Automated case reporting has been shown to increase logging of ACGME-required cases in general surgery interns. An automated case cancellation
review system was previously shown to improve resident education in systems-based
practice (one of the core ACGME competencies), but has not been shown to decrease
case cancellation rates.
