PURPOSE:Physiologic tests such as manometry, colonic transit times, balloon compliance, defecography, pudendal nerve latency, and electromyography are used to evaluate patients with severe constipation. Patients referred because of severe constipation between 1991 and 1996 were studied to examine the role that physiologic testing played in making a diagnosis and directing treatment.METHODS:Of 139 patients referred for severe idiopathic constipation, physiologic testing was recommended in 127, and 104 patients underwent the studies. The pretesting impression was noted, and test results were evaluated to determine diagnostic accuracy. If a specific initial impression was documented, tests were classified as refuting it, confirming it or confirming and adding significant information. If there was no clear pretest impression, tests were evaluated for their ability to indicate a diagnosis. The patient's history also was evaluated to determine what information was most useful in making a diagnosis. Historical features including duration of constipation, symptoms consistent with outlet obstruction or dysmotility, age, associated urinary incontinence, and prior hysterectomy were analyzed. Data were collected prospectively, then reviewed by an independent observer.RESULTS:Ninety‐eight study patients remained after 29 were excluded who did not undergo the recommended studies (19) or because no initial impression was documented (10). In 43 patients (44 percent), testing did not provide additional useful information. In 8 patients, testing confirmed the initial impression and added information impacting the treatment plan. Test results clearly refuted the initial impression in only one patient. In 46 (47 percent) patients the initial impression was uncertain, and in 43 (94 percent) of these, testing aided in making the diagnosis. In three cases, the diagnosis remained uncertain after testing. Prior hysterectomy (P=0.003), urinary incontinence (P<0.001), and symptoms of pelvic outlet obstruction (P=0.03) were associated with a high incidence of rectocele. Defecography and transit times were the most useful tests. Surprisingly, symptoms of outlet obstruction or dysmotility did not show an overall correlation with transit times.CONCLUSIONS:In one‐half of these patients with severe constipation, physiologic testing added significant information, leading to a specific diagnosis. Pretesting history and symptoms did not predict which patients were most likely to benefit from these studies.