Introduction
Colonoscopy is considered the “gold standard” for colorectal cancer screening [1]. A diagnostic colonoscopy can be performed safely without prior consultation with
a gastroenterologist, which increases efficiency and reduces delays and costs [2]
[3]
[4]. This practice of “open access endoscopy” (OAE) has gained popularity, particularly
in the Western world. Despite these benefits, concerns exist about improper referrals
or patients referred for colonoscopy in whom the possible risks outweigh the benefits
such as patients with serious comorbidities, those who are frail, and those with a
limited life expectancy. In addition, improper referrals for colonoscopy may increase
complications, average waiting time for services, and health care costs [5].
In many centers, standardized referral letters (SRL) are screened by gastroenterologists
to evaluate whether patients should be referred for colonoscopy, medications, and
comorbidities. Unfortunately, SRLs do not always provide accurate information; therefore,
we introduced a 7.5 min outpatient consultation (OC), performed by a gastroenterologist,
to ascertain key features about patients to determine if colonoscopy is warranted
before authorizing the procedure. In this study, we prospectively evaluated the SRL
and the OC to define the differences in outcomes regarding regarding recommendations
for or against colonoscopy and patient priority scheduling. In addition, the patients’
use of medications (in particular, insulin and coumarin) and comorbidities were recorded.
Patients and methods
This prospective, observational, single-center study was performed at the Department
of Gastroenterology, Deventer Hospital, The Netherlands. Formal ethics committee review
or approval was not required. Between May 2012 and July 2012, all consecutive patients
referred for surveillance and diagnostic colonoscopy were de-identified and included
in the analysis. There were no exclusion criteria. Two methods of identifying patients
for colonoscopy were evaluated: the standardized referral letter (SRL) and the outpatient
consultation (OC).
Standardized referral letter: The SRL is a predefined list of approved indications
and symptoms that warrant colonoscopy. This includes patients’ use of anticoagulants,
a diagnosis of diabetes, with or without insulin therapy, and contagious diseases.
Current medications and serious comorbidities were also included. All SRLs were scrutinized
by an independent gastroenterologist who was instructed to use only the SRL for authorization
of colonoscopies. In instances where authorization for a colonoscopy was approved,
the gastroenterologist determined the patient’s priority for undergoing the colonsocopy:
A, within one week; B, within two weeks; and C, first regular opportunity (usually
within 3 – 4 weeks). Gastroenterologists were selected using a day-by-day rotation
system, ensuring that they had no previous contact with the study patients or their
medical records.
Outpatient consultation: Every outpatient referred for colonoscopy had a standard
7.5 min face-to-face consultation with a gastroenterologist who verified the data
in the patient’s SRL. The outpatient consultation occurred before and in addition
to colonoscopy, but without a physical examination. Additional information regarding
patient symptoms, comorbidities, and use of medications was obtained from the patient
and the electronic medical record system. The gastroenterologist subsequently approved
or did not approve the patient for colonoscopy. When a colonoscopy was authorized,
a priority level was given for that patient, and the patient met with a nurse for
additional information about the procedure. All patients gave informed consent prior
to the procedure. When colonoscopies were not authorized for a patient, both the patient
and the referring physician were informed by the gastroenterologist that the patient
did not meet the criteria for the procedure. Colonoscopies were performed by experienced
endoscopists, including five gastroenterologists and three nurse endoscopists. All
procedures were performed under conscious sedation using midazolam. The outcomes of
colonoscopies were retrieved from electronic hospital records.
Outcome parameters
The primary outcome was the number of patients who were not authorized for colonoscopy.
Secondary outcomes were the differences in the patient’s priority for colonoscopy,
the number of malignancies identified, the number of advanced neoplastic lesions identified,
and the number of new diagnoses of inflammatory bowel disease (IBD). In addition,
the completeness of the SRL with respect to the patient’s use of medications (insulin
and coumarins) and their important comorbidities were recorded. Important comorbidities
were defined as previous abdominal surgery, cardiac disease, chronic obstructive pulmonary
disease, malignancies, and coagulation disorders.
Statistical analysis
The number of patients in this study (255) was determined by the following calculation
(because no data were available in the literature): power of 90 %, α < 0.05, and an
estimation of 3 % of patients not authorized for colonoscopy using SRL and 10 % of
patients not authorized for colonoscopy using OC. Most of the data in both groups
were descriptive. Repeated measurements on a single sample were performed (SRL and
OC). McNemar’s test was used for nominal data, which is suitable for matched data
pairs. For categorical data, the Wilcoxon signed-rank test was used, a P-value less than 0.05 was considered to be statistically significant. Statistical
analyses were conducted using SPSS 20.0 software.
Results
Consecutive patients (255; 51 % men; mean age, 61 years; range, 19 – 88 years) were
included in the study. Eighty-three percent of the patients were referred by a general
practitioner, the others were referred by internists and surgeons. After OC, 226 patients
(89 %) were authorized for colonoscopy; two patients cancelled their colonoscopies,
and 224 colonoscopies were performed (87 %). In 211 (94 %) patients, the colonoscopy
extended to the cecum. After SRL, 239 (93.7 %) consecutive patients were authorized
for colonoscopy. The results of the SRL and OC authorizations are shown in [Table 1]. Fewer colonoscopies were authorized by OC (11.4 % of patients were not authorized
for a colonoscopy) than by SRL (6.3 % of patients were not authorized for a colonoscopy;
P = 0.02).
Table 1
Results of SRL and OC authorizations for colonoscopy.
|
OC-authorized
|
OC-not authorized
|
Total
|
|
SRL-authorized
SRL-not authorized
|
219
7
|
20[1]
9[1]
|
239
16 (6.3 %)
|
|
Total
|
226
|
29 (11.4 %)
|
255
|
Abbreviations: OC = 7.5 min outpatient consultation; SRL = standard referral letter.
1 These patients did not have a colonoscopy
As shown in [Table 1], seven of sixteen patients that were not authorized for colonscopy using the SRL
, were authorized for colonoscopy using OC, because of incomplete information on the
SRL (eg., missing indications). For example, one patient underwent a colonoscopy a
year previously, which was sufficient reason to decline the colonoscopy using the
information in the SRL. Using OC, however, it was noticed that the prior colonoscopy
of this patient was incomplete because of fecal contamination, and the colonoscopy
for this patient was then authorized. In five of the seven patients, lesions were
found during colonoscopy. These lesions were symptom-related and included hemorrhoids,
colonic diverticular disease, and adenomatous polyps.
[Table 2] shows the reasons why colonoscopies were not authorized, which included symptoms
related to irritable bowel syndrome (using the Rome-III-criteria), recently performed
colonoscopy, incorrect surveillance (early referral or no reason for colonoscopy according
to the guidelines), or a patient’s choice for a second consultation or diagnostic
tests (upper endoscopy, sigmoidoscopy, laboratory tests, or imaging) instead of colonoscopy.
Other reasons why colonoscopies were not authorized included changes in patients’
medications and referrals to other specialists. More colonoscopies were prioritized
when gastroenterologists use the SRL (30 % level A, 14 % level B) than when they used
the OC (18 % level A, 8 % level B). Wilcoxon signed-rank test revealed a significant
difference in priority assignments (P < 0.001). Colorectal cancer was detected in 5.8 % (13/224) of patients who underwent
a colonoscopy. These malignancies were confirmed by histopathology of the biopsies
that were taken during colonoscopy. [Fig. 1] shows that although fewer colonoscopies were prioritized using OC (level A + B),
more patients were diagnosed with malignancies (10.8 % versus 6.3 %, P < 0.001). When colonoscopies were not prioritized, malignancies were detected in
4.7 % and 3.7 % of patients using SRL and OC, respectively. The accuracy of prioritization
was demonstrated in that the number of patients in level A and B decreased while the
number of patients with colorectal cancer in those levels increased.
Table 2
Reasons for de-authorization of colonoscopy.
|
SRL
n (%)
|
OC
n (%)
|
P-value
|
|
Total number of patients who were not referred for colonoscopy:
|
16 (6.3 %)
|
29 (11.4 %)
|
0.02[1]
|
|
Reasons colonoscopy was not recommended:
Symptoms of IBS (Rome III)
Recently performed colonoscopy
Incorrect surveillance
Second consultation or diagnostic tests
Other
|
2 (12.5 %)
5 (31.3 %)
2 (12.5 %)
8 (50.0 %)
1 (6.3 %)
|
5 (17.2 %)
2 (6.9 %)
7 (24.1 %)
16 (55.2 %)
3 (10.3 %)
|
|
Reasons for colonoscopy was not recommended. Data from 16 (SRL) and 29 (OC) subjects
who were not recommended for colonsocopy. Some of the patients were not recommended
for more than one reason, causing the total number of reasons that patients were not
recommended for colonoscopy (SRL, n = 18 and OC, n = 30) to be higher than the total
number of patients who were not recommended for colonoscopy (SRL, n = 16 and OC, n = 29).Abbreviations:
SRL = standard referral letter screening; OC = 7,5 min outpatient consultation; IBD = inflammatory
bowel disease.
1 significance (P < 0.05)
Fig. 1 The assessment of the priority of outpatient colonoscopy based on SRL and OC plotted
against the number of malignancies found at colonoscopy. Abbreviations: SRL = standard
referral letter; OC = 7.5 min outpatient consultation.
Information about current medications was incomplete in 31.5 % of the SRLs. The most
dangerous medications (insulin, coumarins, and platelet inhibitors) were reported
in 1.6 % of the SRLs. Medications that patients take are less relevant for colonoscopy
than for other procedures, but antihypertensives, beta-blockers, and proton pump inhibitors
were missing on the SRLs. Of all patients, 63.5 % had a past medical history, which
included abdominal surgery (35.8 %) cardiac disease (15.4 %), COPD (3.1 %), prior
malignancy (3.7 %), diabetes mellitus (3.0 %), and coagulation disorder (0.6 %). Relevant
past medical history was incomplete for 54.9 % of all patients.
Discussion
In the Netherlands, 191,339 colonoscopies were performed in 2009 [6]. In the current study, SRL, which is current practice, revealed that 16 patients
(6.3 %) did not require colonoscopy whereas OC revealed that 29 patients (11.4 %)
did not require colonoscopy. In the Netherlands, 191,339 colonoscopies were performed
in 2009 [6]. Extrapolating the results from this study to the 2009 data revealed that SLR and
OC would result in 12,058 and 21,812 fewer annual colonoscopies nationwide, respectively.
One could argue that in real practice SRL will result in more unwarranted colonoscopies.
After the introduction of the screening program for colorectal cancer (CRC) in the
Netherlands, the benefits of the OC will increase, because the number of patients
authorized for colonoscopies will increase by 66,000 annually in the long term [7]. Reducing the number of non-screening colonoscopies by using OC reduces the workload
and the financial consequences.
Nearly half of of those who were not recommended for colonoscopy by gastroenterologists
using SRL alone were authorized for colonoscopy using OC, and many of these patients
had lesions that were identified during colonoscopy. Thus, there are two problems
with SRL; first, fewer patients were authorized to undergo colonoscopy, and, second,
patients who were not originally authorized by SRL were ultimately diagnosed with
a lesion when the procedure was authorized after OC. This supports the importance
of OC for maintaining high quality and safety for patients. Relying on SRL could cause
delays in diagnoses of diseases and increase health care costs.
We found OC is better than SRL for establishing patient priority for colonoscopy and
that fewer referrals were assessed as urgent (level A or B). Although the number of
urgent patient referrals was lower for OC, the number of diagnoses of malignancies
was comparable to SRL. This is supported by Ramsay et al., who reported high numbers
of urgent patient referrals by general practioners and that in 92.2 % of these no
malignancies were found [8]. Determining patient priority is important because of its logistical consequences.
A long patient waiting period for a colonoscopy is common for most hospitals. When
the number of referrals assessed as urgent is too high, it could cause problems scheduling
colonoscopies that actually are urgent.
We found incomplete referral information regarding the patients’ past medical history
and current medications. In particular, previous abdominal surgery was not reported
and it is a factor known to contribute to the difficulty of a colonoscopy [9]
[10]. In 2006, an American study reported inaccurate relevant information in 8.8 % of
patient referrals in an assessment of the SRL [11]. In the current study, the incomplete information on the SRLs was revealed upon
OCs and the colonoscopies were subsequently approved.
We performed a prospective study, with blinded evaluation of SRL and OC to determine
if patients would be recommended for colonoscopy or not. Patients who were not recommended
for colonoscopy by OC, did not receive colonoscopy, which is a limitation of this
study. It is impossible to evaluate whether these subjects needed a colonoscopy or
not. However, that determination was not the purpose of this study, and the number
of patients in this group was small relative to the total number of patients. Furthermore,
these patients were evaluated by an experienced gastroenterologist who concurred with
the decision made by OC.
Prioritization of colonoscopies at the Deventer Hospital may be relatively easy because
of a short waiting period and this could have influenced the percentages of prioritized
patients in our study. However, both SRL and OC were evaluated using the same conditions
so the differences noted remain statistically valid. Patients recommended for a colonoscopy
using OC met with a nurse for additional procedural information.This has a positive
impact on the quality of bowel preparation, however, the impact of this meeting was
not evaluated in this study.
The study shows that the implementation of OC to authorize pateint referrals for colonoscopy
reduced the number of authorizations for patients with improper indications. Furthermore,
it reduced the number of procedures that were not authorized from the SRL, provided
a better prioritization of patients for colonoscopies, and provided important additional
information that was lacking on the SRL. These findings show that OC provides important
health care benefits for patients referred for colonoscopies, a reduced of risk of
colonoscopies that are not justified or among patients that are too frail, and reducted
costs and patient waiting times.
Abbreviations
IBD:
Inflammatory bowel disease
OAE:
Open access endoscopy
SRL:
Standard referral letter
CRC:
colorectal cancer
OC:
outpatient consultation
