Concerns Emerge

Colonoscopy is a reasonably effective way of finding and removing colon polyps, the benign lesions from which most colon cancers develop. The United States Preventive Services Task Force has recommended initiating colonoscopy screening at 50 years of age for men and women who are at average risk for colorectal cancer. As a general rule, they have said, people over the age of 50 should undergo colonoscopy once every five years. If a doctor has previously found and removed a polyp, that schedule is moved up to a three-year interval for the next procedure.

However, concerns have emerged over the accuracy of many colonoscopies. The ability of colonoscopy to detect abnormalities ultimately relies on the vigilance and experience of the operator. The rate of false negatives (i.e., missed polyps or colorectal cancers) after colonoscopy is influenced by a number of factors. For example, who performs the colonoscopy and where it is carried out can have a major influence on the reliability of the test. University of Western Ontario researchers found that colonoscopy is far more likely to result in a false negative (i.e. the cancer is missed) when an internist or family physician performs the test. This is true for when it is done in an office setting as well. Colonoscopy is far more accurate when done by a gastroenterologist in a hospital (Bressler 2007).

Higher Cancer Miss Rates

The senior author of this paper, Linda Rabeneck, MD, of the Dalla Lana School of Public Health, University of Toronto, said:

  • “There is something different about the practice of colonoscopy in these settings [i.e., an office setting] that gives rise to higher cancer miss rates, a worrisome finding” (Douglas 2007).

Of 12,487 patients included in the study, 430 (or 3.4%) had a new or missed cancer within six months to three years of having a colonoscopy. Reducing this interval to just two years yielded a 2.4% failure rate. An increase to five years gave a figure of 4.6%. Thus, to be really safe, patients may need more frequent colonoscopies for accurate results.

Hospital vs. Doctor’s Office

Compared to a colonoscopy performed in a hospital, having the procedure done in a doctor’s office yielded an odds ratio of new or missed colorectal cancer of 3.07 in men and 1.95 in women. In plain language, this means that you have a two-to-three times greater risk of a dangerous growth being missed just by having your colonoscopy done in an office, rather than a hospital! Also, if you have it done by a family doctor, you nearly double the missed cancer risk. This, compared to undergoing a gastroenterologist-performed colonoscopy.

Time of Day

Oddly enough, another study found that the colonoscopy failure rate was higher in the afternoon than when the procedure was performed in the morning (Sanaka 2006). The reason was partly because of the all-too-human factor of fatigue among endoscopists. Some doctors were apparently weary of doing one after another of these somewhat tedious procedures. (People tend to miss things when they get tired.) This particular study involved 2,087 colonoscopies, roughly half of which were performed in the morning and half in the afternoon. The colonoscopy failure rate was 6.5% in the afternoon compared to 4.1% in the morning, a difference of 2.1 percent, which was statistically significant.

Ideally, performing all colonoscopies in the morning might reduce the number of repeat procedures. However, this is not feasible given the huge number of patients undergoing the procedure. The study’s authors suggested that one way to counteract the increased afternoon failure rate. This would be to ensure that any patients who are known to be at higher risk for colon cancer be tested in the morning rather than in the afternoon. A word to the wise: schedule your own colonoscopy, if possible, for the morning (and preferably early in the week.)

The Rate of Discovering Tumors Varies Greatly

A third study found that even among experienced gastroenterologists, the rate of discovering tumors varies greatly. The time devoted to examining the mucous lining of each patient’s colon—which is generally performed during the withdrawal of the instrument—appears to be crucial to the successful detection of abnormalities. This is called the “withdrawal time.” Researchers monitored outcomes among 12 board-certified gastroenterologists and data from a total of over 2,000 colonoscopies. These gastroenterologists were all fully qualified, but they varied greatly in how long they took to remove the instrument. Some took as little as 3.1 minutes, while others took as much as 16.8 minutes–more than five times as long.

The authors saw what they called a “striking, seemingly linear relationship” between the withdrawal time and the rates of polyps and cancers that were detected. The overall rate of detection of polyps among operators who had relatively slow withdrawal times was nearly four times as great as the rate among those who had relatively fast withdrawal times.

Speed Matters

Slow workers were also about three times more likely to find an abnormality than fast workers. The author of this study (which was published in the top-drawer New England Journal of Medicine) concluded that “a minimum adequate amount of time for colonoscopic withdrawal can be equated with quality of colonoscopy” (Barclay 2006).

Yet in our personal experience, sometimes doctors are pressured to perform as many colonoscopies as possible in a limited amount of time. This can only lead to worse outcomes for the patients.

Recommendations for More Accurate Results

The takeaway message is this: if you can, choose your endoscopist carefully! As a general rule, pick a board-certified gastroenterologist rather than an internist or general practitioner. Make sure to have the procedure done in a hospital, not in a doctor’s office. And, (very importantly) insist on having the procedure done in the morning, not in the afternoon. (I would also avoid Fridays.)

As to getting your doctor to slow down and take his or her time in examining each patient, it is hard to know how lay people can exert much influence. This is especially true since they are usually under deep sedation while the procedure is being performed! But the next time I go for a colonoscopy, I intend to tell my gastroenterologist that I have read these these papers, particularly the New England Journal of Medicine article showing that slower procedures yield more accurate outcomes. I shall ask him or her (politely, of course) to not rush things and to do the most thorough job possible even if it takes more time. Hopefully, David Lieberman’s heartfelt call to action (in an accompanying New England Journal of Medicine editorial) for endoscopists to take their time will have gotten through to them (and their department heads) by then.

Read Choosing a Hospital for guidance on choosing a practitioner or facility.

See also,  A Visit to the Hospital – Patient Experience


  • Barclay RL, Vicari JJ, Doughty AS, Johnanson JF, Greenlaw RL. Colonoscopic withdrawal times and adenoma detection during screening colonoscopy. N Engl J Med. 2006;355:2533-2541,2588-2589.
  • Bressler B, Paszat LF, Chen Z, Rothwell DM, Vinden C, Rabeneck L. Rates of new or missed colorectal cancers after colonoscopy and their risk factors: a population-based analysis. Gastroenterology. 2007;132:96-102.
  • Douglas D. Colon cancers missed more often in office setting. Reuters Health. February 23, 2007. Available at:
  • Kaitin KI, ed. Cost to develop new biotech products is estimated to average $1.2 billion. Tufts Center for the Study of Drug Development Impact Report. 2006;Nov/Dec;8(6).
  • Lieberman D. A call to action–measuring the quality of colonoscopy. N Engl J Med. 2006 Dec 14;355(24):2588-2589.
  • Sanaka MR, Shah N, Mullen KD, Ferguson Dr, Thomas C, McCullough AJ. Afternoon colonoscopies have higher failure rates than morning colonoscopies. Am J Gastroenterol. 2006;101:2726-2730.
  • Wazana A. Physicians and the pharmaceutical industry: is a gift ever just a gift? JAMA. 2000;283(3):373-80.

Original Publication 2007