
The age and gender of users, as provided by users when they registered with Bing, was available for a subset (23%) of the users. The queries were filtered to include those queries which contained the word “colonoscopy” and any query made by people who searched for this word. Each query comprised of an anonymized user identifier, the time and date of the query, and its text. We extracted all queries made to Bing from people in the USA between October 1st, 2017 and September 30th, 2018 (one year). In this study we aimed to determine the prevalence and timing of post-colonoscopy complications using search engine queries. Such a strategy has been used to estimate the prevalence of symptoms that did rise to the threshold of seeking care, but that preceded a diagnosis of adenocarcinoma of the pancreas. These events are difficult to measure using traditional measures, but may be estimated using search engine queries. Such methods likely underestimate more mild, but still clinically significant events that do not meet the threshold for acute care, such as a febrile respiratory illness with cough. Prior studies of colonoscopy complications focused on the severe end of the spectrum, and mostly relied on post-colonoscopy emergency department visits or hospitalizations to identify complications. As rates of anesthesia assistance during colonoscopy have increased markedly in recent years, there is concern that respiratory complications may be an increasingly common event. However, respiratory complications, such as aspiration pneumonia, may be an underappreciated sequela of colonoscopy, particularly when deep sedation with anesthesia assistance is provided. Numerous studies have confirmed that the rate of serious complications after colonoscopy is low, and that the most severe complications of the procedure (perforation, bleeding, and mortality) have declined in recent years. Given its widespread use in the general population, the effectiveness of colonoscopy is dependent on its safety profile, particularly when it is employed as a primary screening modality in asymptomatic individuals. Among individuals undergoing alternative modes of screening, such as fecal immunochemical testing, colonoscopy is mandated for those who have a positive screening test, and colonoscopy is the effector arm of all colorectal cancer screening tests. Given the widespread use of colonoscopy as a screening modality and the rise of anesthesia-assisted colonoscopy in the United States in recent years, this signal is of potential public health concern.Ĭolonoscopy is the most commonly used method for colorectal cancer screening in the United States, and nearly 60% of individuals eligible for colorectal cancer screening have undergone this procedure. Internet search queries for respiratory symptoms rose approximately one week after queries relating to colonoscopy, raising the possibility that such symptoms are an under-reported late adverse effect of the procedure. Respiratory symptoms rose significantly at days 7–10 after the colonoscopy. 28% of colonoscopy-related queries were made afterwards, and included queries about diarrhea and cramps, with 2.6% of users querying respiratory symptoms after the procedure, including cough (1.2%) and pneumonia (0.6%). The majority of queries about colonoscopy preceded the procedure, and concerned diet. Of 641,223 users mentioning colonoscopy, 7013 (1.1%) had a query that enabled identification of their colonoscopy date. We then identified queries from those same users for medical symptoms during the time span from 5 days before through 30 days after the colonoscopy date. We identified queries made to Bing originating from users in the United States containing the word “colonoscopy” during a 12-month period and identified those queries in which the timing of colonoscopy could be estimated. We aimed to identify the rate and timing of adverse effects as reported by users querying symptoms on an internet search engine. The rate of more mild adverse effects is difficult to estimate, particularly when such effects do not result in hospital admission. Colonoscopy as a screening and diagnostic tool is generally safe and well-tolerated, and significant complications are rare.
