J Am Board Fam Pract: first published as 10.3122/jabfm.17.5.353 on 8 September 2004. Downloaded from http://www.jabfm.org/ on 1 January 2025 by guest.

Protected by copyright.
Colonoscopy in Rural Communities: Can Family
Physicians Perform the Procedure with Safe and
Efficacious Results?
Jeffrey K. Edwards, MD, and Thomas E. Norris, MD

Background: Colonoscopy is becoming increasingly necessary for many patients in screening, diagnos-
ing, and treating colorectal problems. Because the majority of rural doctors are family physicians, pro-
viding colonoscopy for the enlarging group of patients with valid indications in rural areas is difficult,
unless rural family physicians perform the procedure. Subspecialists in academic settings have been
responsible for most of the previously reported studies regarding colonoscopy. We have studied the
safety and efficacy of the procedure when performed by rural family physicians.
Methods: A total of 200 sequential colonoscopies performed by family physicians in a rural setting
were prospectively collected. Outcomes were measured based on current recommendations and bench-
marks, including rate of reaching the cecum, time to reach the cecum, time to completion of the study,
pathologic lesions found, and complications.
Results: The rate of reaching the cecum was 96.5%, and the average time to the cecum was 15.9 min-
utes. The average time to study completion was 34.4 minutes. The rate of neoplastic polyps and cancer
found was 22.5% and 2.5%, respectively. There were no serious complications.
Conclusions: Adequately trained family physicians can provide safe and technically competent
colonoscopy in a rural setting. Their results compare favorably to the currently reported comparative
benchmarks from other endoscopists. (J Am Board Fam Pract 2004;17:353– 8.)

Colon cancer is a preventable but potentially fatal We present 200 prospectively collected sequen-
disease. In the last 2 decades, research has estab- tial colonoscopy cases performed by rural family
lished that most colon cancers arise from neoplastic physicians, and we provide a comparison with re-
polyps within the colon. If these polyps are found gard to the quality and safety of the procedure as
early and removed, colon cancer can be prevented. found by a current review of the literature.1–23 It
There is currently a growing national movement to has been reported that a competent colonoscopist
screen patients at risk for colon cancer, with a can reach the cecum more than 90% of the
consequential growing need for surveillance of pa- time,1,3– 6,10 –12,15–18,21,22 in a reasonable amount
tients with a history of polyps or colon cancer. of time.1,5,6,8,16,18,20 –22 Colonoscopy can be com-
Many physicians and patients prefer colonoscopy as pleted with a minimal amount of risk to the
patient3,6,7,10 –12,16,18,20,23 and with a satisfactory
the “reference standard” screening study for colon
rate of detection of pathologic lesions.3,8,10,11,12,16,24
polyps and cancer as well as polyp surveillance.
Unfortunately, there is a lack of qualified colonos-
copists in rural areas. The majority of rural doctors Methods
are family physicians. If colonoscopies are to be Data were collected at the time of the procedure on
offered widely in rural areas, family physicians will 200 consecutive colonoscopies performed at St.
need to perform them. Mary’s Hospital in Cottonwood, Idaho. St. Mary’s
Hospital is a rural, 14-bed health care facility in
north central Idaho that serves the largest county,
in terms of area, in the state. This service area
Submitted, revised, 6 April 2004. includes a patient population of 12,000 to 15,000
From the St. Mary’s Hospital, Cottonwood, ID (JKE);
and the School of Medicine, University of Washington, people. All colonoscopies completed by family phy-
Seattle (TEN). Address correspondence to Thomas E. Nor- sicians at St. Mary’s Hospital over a 2-year period
ris, MD, School of Medicine, University of Washington,
Seattle, WA 98195-6340 (e-mail: tnorris@u.washington. from December 2000 through January 2002 were
edu). included in the study. These cases included both

Colonoscopy in Rural Communities 353

 J Am Board Fam Pract: first published as 10.3122/jabfm.17.5.353 on 8 September 2004. Downloaded from http://www.jabfm.org/ on 1 January 2025 by guest. Protected by copyright.
data. Analysis of the data included only the calcu-
lation of totals (sums) and percentages, using all the
cases included in the study. No statistical calcula-
tions were required or performed. Data were ana-
lyzed by category, including demographics, proce-
dure indication, time to cecum, time to completion
of colonoscopy, complications, and findings. After
data collection, outcome results from this descrip-
tive study were compared with others reported in
the literature.
All procedures were completed by 1 of 4 rural
family physicians. The 3 younger physicians had
been trained to do colonoscopy in their residencies.
A younger physician proctored the oldest physician
(who had been doing flexible sigmoidoscopy) until
he was proficient with colonoscopy.
The colonoscopies were all performed in a ded-
icated hospital endoscopy room at St. Mary’s Hos-
pital. In all cases presented in this report, Certified
Registered Nurse Anesthetists provided intrave-
nous conscious sedation. It is noted that in other
settings, other trained personnel could provide this
service. Intravenous sedation used included a com-
bination of midazolam, fentanyl, and propofol.
Continuous cardiorespiratory monitoring was done
during each procedure. An Olympus video colono-
Figure 1. Standardized data collection form.
scope, model CFQ140L, was used for all proce-
dures.
inpatient (urgent) and outpatient (nonurgent) study
participants. Results
The data were collected prospectively during Two hundred consecutive sequential colonoscopy
the study period. A standardized data collection procedures were performed over a 2-year period at
form was completed at the time of each procedure. St. Mary’s Hospital from December 2000 through
Figure 1 shows the standardized form used to col- January 2002 by 4 family physicians. Of the 200
lect colonoscopy data. The physician completed patients, 91 were women and 109 were men. The
this form immediately after the colonoscopy (be- age range of patients was from 16 to 90 years, with
fore pathology reports were available), and endos- an average patient age of 62 years. The number of
copy room staff recorded procedure times. Study colonoscopies completed per physician varied from
patients were then called 1 to 3 days after the 23 to 108. Table 1 shows the estimated number of
procedure to review for any problems unless they
were inpatients already.
Pathology reports were collected for all proce- Table 1. Estimated Number of Colonoscopies and the
dures in which tissue specimens were obtained. Total Years Physicians Had Been Performing
Only in situations in which the pathologist con- Colonoscopy before Beginning the Study
firmed biopsy results were polyps of neoplastic or-
Number Years
igin reported (adenomatous, tubular, or villous) or Physician of Cases Performing Procedure
the case designated as colon cancer. Pathology re-
ports of hyperplastic polyps and other nonmalig- A ⬍50 3
B 50 to 100 8
nant findings were excluded from the analysis. Mi-
C 100 to 200 12
crosoft Access (Microsoft Corp., Redmond, WA)
D ⬎500 15
was the database used for storage and analysis of the

354 JABFP September–October 2004 Vol. 17 No. 5

 J Am Board Fam Pract: first published as 10.3122/jabfm.17.5.353 on 8 September 2004. Downloaded from http://www.jabfm.org/ on 1 January 2025 by guest. Protected by copyright.
Table 2. Most Common Indications for 200 Table 4. Average Time to Cecum and Procedure
Colonoscopies (Some Cases Had Multiple Indications) Completion by Physician
Number Percentage Average Time Average Time to
Indication of Cases of Cases to Reach Cecum Complete Procedure
Physician (minutes) (minutes)
Rectal bleeding 52 26
Hx of polyps 49 25 A 23.8 51.1
FHx of colon cancer 44 22 B 16.6 36.5
Stools positive for FOBT 35 18 C 16.9 33.7
D 6.5 16.3
Hx, history; FHx, family history; FOBT, fecal occult blood Average 15.9 34.4
testing.

During the study, 45 of 200 (22.5%) cases were

colonoscopies and years of experience performing
found to have neoplastic polyps that were con-
colonoscopy by a physician before beginning the
firmed by pathological examination. Table 5 dem-
study. The most frequent indication for colonos-
onstrates the frequency of neoplastic polyps found
copy was rectal bleeding. Some patients had mul-
based on the most frequent preprocedure indica-
tiple indications for colonoscopy, whereas others
tions. Five of 200 (2.5%) patients had confirmed
were for colon cancer screening only. Twelve pa-
colon cancer. The most common preprocedure di-
tients had procedures completed as inpatients for
agnosis, for patients subsequently found to have
urgent reasons (ie, acute gastrointestinal bleeding).
colon cancer, was hemoccult-positive stools [3 of
Table 2 shows the most common indications for
35 (8.5%)]. The 2 other patients with colon cancer
colonoscopy.
had a preprocedure diagnosis of anemia or rectal
The success rate for reaching the cecum was
bleeding.
found to vary among physicians from 91 to 100%.
Complications were considered to be adverse
The average rate for reaching the cecum overall
events that required intervention and occurred dur-
was 96.5% (193 of 200) as shown in Table 3. Of the
ing the procedure or after, if related to the proce-
cases in which the cecum was not reached, 2 of 7
dure. Complications included use of reversal agents
(29%) were the result of an obstructing colon can-
with sedation, cardiorespiratory problems with se-
cer, whereas 4 of 7 (57%) were caused by a tortuous
dation, bowel perforation, hospital admission,
colon, and 1 was secondary to an anastomosis ste-
emergency department visits, and bleeding requir-
nosis.
ing transfusion that may have been related to the
The average time to reach the cecum per phy-
procedure. A review of the patients’ charts 2 years
sician varied from 6.5 to 23.8 minutes and is shown
after completion of the study was completed to
in Table 4. The overall average time to reach the
look for any possible delayed complications related
cecum was 15.9 minutes. The physician with the
to the procedure or missed colon cancers. There
most experience had the shortest average time to
was 1 case of sedation-related bradycardia that re-
the cecum, whereas the physician with the least
quired a single dose of atropine. One patient was
experience had the longest. The overall average
admitted to the hospital for abdominal pain for
time for completion of the procedure was 34.4
minutes, including biopsies and polypectomies.
Table 5. Frequency of Neoplastic Polyps Based on Most
Common Procedure Indications
Table 3. Rate to Reach the Cecum Number Percentage
Number of Times Cecum Indication of Cases (Total) of Cases
Physician Colonoscopies Reached Percentage
Stools positive for FOBT 9 (35) 26
A 23 21 91 Hx of polyps 13 (49) 27
B 45 43 96 FHx of colon cancer 8 (44) 18
C 24 24 100 Rectal bleeding 11 (52) 21
D 108 105 97
Total 200 193 96.5 Hx, history; FHx, family history; FOBT, fecal occult blood
testing.

Colonoscopy in Rural Communities 355

 J Am Board Fam Pract: first published as 10.3122/jabfm.17.5.353 on 8 September 2004. Downloaded from http://www.jabfm.org/ on 1 January 2025 by guest. Protected by copyright.
Table 6. Review of Reported Rates of Reaching the Table 7. Review of Reported Time to Reach the Cecum
Cecum More Than 90% of the Time Mean Time
to Cecum
Rate to
Author Year n (minutes)
Cecum
Author Year n (%)
Marshall5 1995 34 9*
4
Marshall et al 1993 418 96 Chak et al1 1996 297 10.5*
Church15 1994 2,907 93.6 Kim et al21 2000 909 6.9
Marshall and Barthel5 1995 423 93.4 Anderson et al22 2001 755 5.6*
Hopper et al12 1996 713 93 Nelson et al6 2002 3196 10.5
Chak et al1 1996 315 94.3 Current study 2004 200 15.9
Pierzchajlo et al11 1997 751 91.5
* Median rather than mean.
Wexner et al16 1998 2,069 96.5
Tassios et al17 1999 430 91
Kim et al21 2000 909 96.4
lished studies shown in Table 6. The majority of
Wexner et al18 2001 13,580 92
studies reported in the last 10 years confirm that
Anderson et al22 2001 755 91.6
skilled endoscopists can reach the cecum more than
Thomas-Gibson et al10 2002 505 93
Nelson et al6 2002 3,196 97
90% of the time. However, when considering en-
Current study 2004 200 96.5 doscopists in training, the rate of successful cecal
intubation is usually lower than 90%.1,3,5,8,14,17,19
The 4 physicians in our study all had cecal intuba-
tion rates greater than 90%, suggesting that they
observation overnight but required no interven- meet the most widely studied standard in achieving
tion. One patient was admitted to the hospital after technical competence in colonoscopy.
colonoscopy for observation for several hours be- Confirmation of reaching the cecum can some-
cause of other medical problems, the length of times be a difficult task. All colonoscopies in our
procedure, and the number of polyps removed. study were either videotaped or photographs were
This patient required no intervention. One patient taken of cecal anatomy for confirmation, but these
required placement of a rectal tube to relieve re- cases were not reviewed by outside staff. One study
tained air within the colon, and symptoms resolved. has considered the photograph documentation of
There were no bowel perforations. No cases re- cecal landmarks as confirmation of a complete
quired sedation reversal medications. colonoscopy.25 They found that it was difficult for
reviewers to consistently agree on whether ade-
Discussion quate visual documentation had been obtained.
This study demonstrates that the quality of Since completion of our study, there have been no
colonoscopy performed by adequately trained rural missed colon cancers reported in the study group.
family physicians compares favorably with bench- The procedure time was divided into 2 primary
marks reported in the literature.1– 4 Although measurements: (1) time to reach the cecum and (2)
benchmarks have been poorly defined in the past, time to procedure completion. We found that our
standards are now becoming more apparent. The average time to the cecum was 15.9 minutes. How-
literature suggests that the colonoscopist should be ever, there were cases in the study group in which
able to reach the cecum more than 90% of the time
(Table 6). They should be able to perform a
Table 8. Review of Reported Time to Complete
colonoscopy in a reasonable amount of time (Ta-
Colonoscopy
bles 7 and 8). They are required to find and diag-
nose all significant pathologic lesions. Finally, they Mean Time to
Completion
must be able to complete the procedure with min- Author Year n (minutes)
imal risk of complications and patient discomfort.
The reported success rate for reaching the ce- Wexner et al16 1998 1,023 34.7
Wexner et al18 2001 13,580 22.7
cum has, in the past, varied from 57% to more than
Nelson et al6 2002 3,196 30.6
95%.3 Our success rate to reach the cecum was
Current study 2004 200 34.4
96.5%. Our rate is similar to that of other pub-

356 JABFP September–October 2004 Vol. 17 No. 5

 J Am Board Fam Pract: first published as 10.3122/jabfm.17.5.353 on 8 September 2004. Downloaded from http://www.jabfm.org/ on 1 January 2025 by guest. Protected by copyright.
polyps were removed before reaching the cecum; of patients who were not offered colonoscopy by
this would obviously lead to an increase in the time the family physicians was not followed. Despite
to reach the cecum. Further analysis was not per- this, there were no serious complications in the
formed on these cases. study group. One adverse cardiopulmonary event
Table 7 shows the time reported to reach the (1 of 200, 0.5%) required intervention. There were
cecum in recent studies. It has been stated that a no colon perforations. Our rate of procedure-
“reasonable standard” would be cecal intubation related morbidity was consistent with morbidity
within 15 minutes or less.1 The 4 physicians in the reported by others.6,7 The reported rate of mortal-
current study showed a large amount of variation in ity as a result of colonoscopy is greater than 1 in
their average time to reach the cecum (6.5 to 23.8 5000, and major morbidity is approximately
minutes). This seemed to be related to the number 0.4%.6,7 The rate of colon perforation is reported
of previous procedures completed before beginning to vary from 0.14 to 0.65% for diagnostic proce-
the study. A similar variation is seen when one dures and 0.15 to 3.0% for therapeutic colonoscopy
compares the time to procedure completion among (including polypectomy).23
these same physicians (16.3 to 51.1 minutes). Table The primary weakness of the study is the rela-
8 shows recently reported studies that measured tively small sample size; however, we continue to
total procedure times for colonoscopy. The re- collect sample data prospectively and look forward
ported studies also show a considerable amount of to reporting this in the future. It is possible that
variation in reported average times (22.7 to 34.7 patient complications that occurred after the day of
minutes). the procedure could have been missed if the patient
Various factors affect the amount of time re- presented to another hospital for evaluation and
quired to reach the cecum and complete the treatment. However, this is unlikely because of the
colonoscopy procedure (including biopsies and rural locality of the study hospital and the distance
polypectomies). Factors affecting these outcome to the next largest medical facility. In addition, this
measurements include physician experience, pa- patient study group is part of a rural primary care
tient anatomy, quality of bowel preparation, pa- 5-clinic system with ongoing longitudinal fol-
thology encountered, and reporting differences. low-up that uses central data management.
For example, the study with the shortest reported
time to reach the cecum subtracted the time con- Conclusion
sumed with polyp removal before reaching the ce- As the number of indications for the use of colonos-
cum.22 At this time, more research is required to copy increase, rural patients face mounting geo-
clarify these variables comparing colonoscopists graphic and distance-related obstacles to obtaining
and their procedure time. the endoscopy services that they need. This study
Our study group included a heterogeneous col- provides an appraisal of the quality of 200 consec-
lection of subjects. Patients were included regard- utive colonoscopies provided by 4 rural family phy-
less of their indication for colonoscopy. The indi- sicians with various degrees of experience for a
cations ranged from asymptomatic screening to heterogeneous group of patients. Outcomes were
surveillance follow-up for polyps or prior colon measured and compared with other published re-
cancer. Among this diverse patient population, we sults. We found that well-trained rural family phy-
found a prevalence of neoplastic polyps of 22.5% sicians could safely provide diagnostic and thera-
and a prevalence of colon cancer of 2.5%. These peutic colonoscopy for their patients.
findings are consistent with what has been reported Further research regarding quality measures for
previously.24 colonoscopy and setting “standards” regarding
The study group included a wide variety of pa- compliance with these measures is needed to allow
tients based on age, indications (both diagnostic objective comparison between colonoscopists. Re-
and therapeutic) and complicating medical condi- search in this area will advance the quality of
tions. The group also included both an inpatient colonoscopy provided to all our patients, both ur-
(urgent) and outpatient (nonurgent) population. ban and rural.
Some patients were believed to be too “high risk”
and were referred to a larger medical center with We thank Cheri Holthaus and Pat Forsman, of St. Mary’s
subspecialist management. However, the number Hospital, who assisted greatly with data input and management.

Colonoscopy in Rural Communities 357

 J Am Board Fam Pract: first published as 10.3122/jabfm.17.5.353 on 8 September 2004. Downloaded from http://www.jabfm.org/ on 1 January 2025 by guest. Protected by copyright.
References 13. Carr KW, Worthington JM, Rodney WM, Gentry
1. Chak A, Cooper GS, Blades EW, Canto M, Sivak S, Sellers A, Sizemore J. Advancing from flexible
MV Jr. Prospective assessment of colonoscopic intu- sigmoidoscopy to colonoscopy in rural family prac-
bation skills in trainees. Gastrointest Endosc 1996; tice. Tenn Med 1998;91:21– 6.
44:54 –7. 14. Cass OW, Freeman ML, Peine CJ, Zera RT, Onstad
2. Quality and outcomes assessment in gastrointestinal GR. Objective evaluation of endoscopy skills during
endoscopy. American Society for Gastrointestinal training. Ann Intern Med 1993;118:40 – 4.
Endoscopy. Gastrointest Endosc 2000;52:827–30. 15. Church JM. Complete colonoscopy: how often? And
if not, why not? Am J Gastroenterology 1994;89:
3. Bond JH, Frakes JT. Who should perform colonos-
556 – 60.
copy? How much training is needed? Gastrointest
Endosc 1999;49:657–9. 16. Wexner SD, Forde KA, Sellers G, Geron N, Lopes
A, Weiss EG, et al. How well can surgeons perform
4. Marshall JB, Barthel JS. The frequency of total
colonoscopy? Surg Endosc 1998;12:1410 – 4.
colonoscopy and terminal ileal intubation in the
17. Tassios PS, Lada SD, Grammenos I, Demertzis K,
1990s. Gastrointest Endosc 1993;39:518 –20.
Raptis SA. Acquisition of competence in colonos-
5. Marshall JB. Technical proficiency of trainees per- copy: the learning curve of trainees. Endoscopy
forming colonoscopy: a learning curve. Gastrointest 1999;31:702– 6.
Endosc 1995;42:287–91.
18. Wexner SD, Garbus JE, Singh JJ. A prospective
6. Nelson DB, McQuaid KR, Bond JH, Lieberman analysis of 13,580 colonoscopies. Reevaluation of
DA, Weiss DG, Johnston TK. Procedural success credentialing guidelines. Surg Endosc 2001;15:251–
and complications of large-scale screening colonos- 61.
copy. Gastrointest Endosc 2002;55:307–14.
19. Mitchell RM, McCallion K, Gardiner KR, Watson
7. Dafnis G, Ekbom A, Pahlman L, Blomqvist P. Com- RG, Collins JS. Successful colonoscopy; completion
plications of diagnostic and therapeutic colonoscopy rates and reasons for incompletion. Ulster Med J
within a defined population in Sweden. Gastrointest 2002;71:34 –7.
Endosc 2001;54:302–9. 20. Cotton PB, Connor P, McGee D, et al. Colonos-
8. Harper M, Pope JB, Mayeaux EJ, Davis TJ, Myers A, copy: practice variation among 69 hospital based
Lirette A. Colonoscopy experience at a family prac- endoscopists. Gastrointest Endosc 2003;57:352–7.
tice residency: a comparison to a gastroenterology 21. Kim WH, Young JC, Park JY, Min PK, Kang JK,
and general surgery services. Fam Med 1997; Park IS. Factors affecting insertion time and patient
29:575–9. discomfort during colonoscopy. Gastrointest Endosc
9. Worthington DV. AAFP position paper. Colonos- 2000;52:600 –5.
copy: procedural skills. Am Fam Physician 2000;62: 22. Anderson JC, Messina CR, Cohn W, et al. Factors
1177– 82. predictive of a difficult colonoscopy. Gastrointest
10. Thomas-Gibson S, Thapar C, Shah SG, Saunders Endosc 2001;54:558 – 62.
BP. Colonoscopy at a combined district general hos- 23. Putcha RV, Burdick JS. Management of iatrogenic
pital and specialist endoscopy unit: lessons from 505 perforation. Gastroenterol Clin North Am 2003;32:
consecutive examinations. J R Soc Med 2002;95: 1289 –309.
194 –7. 24. Rex DK. Colonoscopy: a review of its yield for can-
11. Pierzchajlo RP, Ackermann RJ, Vogel RL. Colonos- cers and adenomas by indication. Am J Gastro 1995;
copy performed by a family physician. A case series 90:353– 63.
of 751 procedures. J Fam Pract 1997;44:473– 80. 25. Marshall JB, Brown DN. Photodocumentation of
12. Hopper W, Kyker KA, Rodney WM. Colonoscopy total colonoscopy: how successful are endoscopists?
by a family physician: a 9 year experience of 1048 Do reviewers agree? Gastroinstest Endosc 1996;44:
procedures. J Fam Pract 1996;43:561– 6. 243– 8.

358 JABFP September–October 2004 Vol. 17 No. 5