To date, no single therapy for IBS has proved satisfactory [8-12]. A critical review of randomized controlled trials in patients with IBS concluded that no therapy has been shown to be superior to placebo [8]. Therefore, a nonpharmacologic approach to the patient with IBS is important [13]. Drossman and Thompson suggested that the foundation of therapy for IBS should be confidence in the diagnosis and a strong physician-patient relationship [13]. However, whether physicians are effectively meeting the physical and emotional needs of patients is unclear [14]; in some studies, as many as one third of patients have sought alternative medical treatment [15, 16].

Our purpose was to evaluate the long-term course and prognosis associated with IBS by following a cohort of patients from a community using their medical records, after they were initially diagnosed with IBS according to current criteria. We also sought to objectively assess, using the written record, whether the physician-patient interaction was associated with prognosis as indicated by the number of return visits for IBS-related symptoms.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Our study was approved by the Institutional Review Board of the Mayo Clinic. In the medical record system of the Mayo Clinic, a single dossier is maintained for all visits of each patient; this dossier includes notes from the patient's primary physician, notes from specialists seen in consultation, and surgical and hospital records [17]. Since its inception, the Mayo Clinic has provided primary medical care for most persons living in Olmsted County, Minnesota, and has maintained complete histories on these persons from birth through death [17]. The organizing principles guiding the indexing of records have remained constant since 1907, and past and present records are readily accessible. All records are permanently maintained. The consistency of design, reliability over time, and organization of the Mayo record system allow for epidemiologic research on patterns and causes of disease [17]. Epidemiologists have recognized that some of the problems of prospective design can be avoided by following a large population through the medical record rather than by following patients "in life" [17]. The Olmsted County population is sociodemographically similar to the U.S. white population and findings are likely to be generalizable to that population [2, 17].

We reviewed the medical records of all residents of Olmsted County who were 20 to 64 years of age and who were diagnosed with IBS at Mayo Clinic in the years 1961-1963; 276 patients were screened. Criteria for inclusion in the cohort were the presence of abdominal pain associated with either disturbed defecation (diarrhea or constipation or both) or abdominal distension (with or without gaseousness) and the absence of organic bowel disease [11, 12]. Diarrhea was defined as loose watery stools, soft stools, or increased stool frequency. Constipation was defined as hard, scybalous, or pebble-like stools or decreased stool frequency. In contrast to the study by Chaudhary and Truelove [6] but in concert with that of Harvey and colleagues [4] and the Rome criteria [11, 12], patients with painless diarrhea (n = 29), painless constipation (n = 7), painless distension (n = 7), or abdominal pain in the absence of colonic symptoms (n = 48) were excluded.

Eleven patients were excluded because their symptoms were inconsistent with IBS but were consistent with another functional gastrointestinal disorder or because they had a parasitic or bacterial (n = 7) or other gastrointestinal disease (n = 3). Patients were also excluded if they had concurrent diagnoses of neurologic disturbances such as multiple sclerosis or the postpolio syndrome (n = 7), if their medical records were incomplete (n = 8), if their symptoms disappeared after cholecystectomy (n = 5), if they had no symptoms listed in the history (n = 9), or if they did not have symptoms consistent with a diagnosis of IBS (n = 10). One patient was excluded because of coexistent radiation treatments for abdominal cancer. Three patients were considered to have been misdiagnosed and were not included in subsequent analyses: One died of acute colonic infarction 4 months after diagnosis of IBS, and two were excluded because the diagnosis of IBS was superseded within 4 months by diagnoses of colon cancer. Patients whose first IBS diagnosis was made during their last visit to the Mayo Clinic were excluded (n = 9). One hundred sixty-four patients were excluded; 112 patients with a firm diagnosis of IBS according to current criteria [12] remained in the cohort.

Review of Medical Records

The complete medical record (inpatient and outpatient) of each of the 112 patients was reviewed twice. During the first review, the accuracy of the diagnosis, the number of follow-up visits for IBS-related symptoms, and the numbers of hospitalizations, surgeries, and general clinic visits were determined. General clinic visits included eye and dental examinations, general medical examinations, psychiatric evaluations, and other specialty clinic visits. The purpose of the first review was to evaluate the long-term clinical course of each patient.

The second review was done independently of the initial review; its purpose was to evaluate and score the physician-patient interaction on the basis of the written record. Only the initial visit or composite initial visits leading to the first definitive diagnosis of IBS were scored; we chose to score only the first visit so that subsequent visits could serve as an indirect index of prognosis. We hypothesized a priori that fewer return visits for IBS-related symptoms would indicate a better clinical course.

We also developed a priori criteria to evaluate facets of the physician-patient interaction on the basis of the medical record. The eight variables selected to provide an index of the strength of the physician-patient interaction were 1) documentation of the patient's psychosocial history; 2) notation of reassurance about the diagnosis; 3) reference to the patient's name within the physician's notes; 4) use of pronouns and adjectives that suggest the collaboration of patient and physician; 5) referral for psychiatric counseling only if evidence of a definite psychiatric diagnosis existed; 6) no unnecessary invasive procedures ordered without firm indication; 7) notation of a factor or factors that precipitated the patient's seeking medical help; and 8) notation of discussion with the patient about test results and diagnosis. These criteria were selected because they were consistent with Drossman and Thompson's overall recommendations for establishing a therapeutic relationship with patients with IBS [13]. Our preliminary analysis showed that no invasive studies were ordered without firm suspicion of organic disease and that no psychiatric referrals were made in the absence of a psychiatric diagnosis. Thus, these two variables had no effect on the composite index of strength and were removed from the final composite score. This adjustment resulted in a maximum possible score of 6.

Statistical Analysis

Analysis of variance was used to assess the relation between the number of return visits for IBS-related symptoms and the strength of the physician-patient interaction. Two-group comparisons of the number of return visits for each of the individual strength criteria were evaluated using the Wilcoxon rank-sum test for nonparametric data. All P values calculated were two-tailed. Patient survival after initial diagnosis of IBS was estimated using the Kaplan-Meier method and compared with expected survival using the one-sample log-rank test; expected survival was based on that of the 1960 U.S. Caucasian North Central population.

Results

Top Methods Results Discussion Author & Article Info References

Demographic Characteristics

Of the 112 patients in the cohort, 36 were men and 76 were women. Men tended to have fewer return visits for IBS-related symptoms than did women (1.7 and 3.0 visits, respectively; P = 0.03). The median age at first diagnosis was 41 years (39 years for men, 43 years for women). All persons in the cohort were white.

Years of Follow-up

The median number of years between the first and last Mayo Clinic visits made for any reason was 40 years (range, 1 to 80 years). The number of years between the first visit and the first diagnosis of IBS ranged from 0 (if the first IBS diagnosis and the first Mayo Clinic visit coincided) to 50 years (median, 16 years). The median number of years from the first diagnosis of IBS to the last recorded visit was 29 years (range, 1 to 32 years). Over that period, patients made a median number of two return visits for IBS-related symptoms (range, 0 to 12 return visits). Thirty-two of 112 patients in the cohort made no return visits for IBS-related symptoms. Twenty-five patients (22%) were lost to follow-up a median of 11 years (range, 1 to 26 years) after their first IBS diagnosis: Twelve moved away from Olmsted County, and 13 were lost to follow-up for unknown reasons.

Physicians requested return visits from only eight patients. Five of the eight patients reported no IBS-related symptoms at their return visits. One patient was lost to follow-up after two return visits for IBS symptoms over 16 years, and the remaining two patients died (one patient made three return visits in the 29 years before death from pneumonia and the other made three return visits in the 26 years before death from a motor vehicle accident). Thus, almost all return visits were patient-initiated, and physician requests were not a major determinant of the number of return visits.

Symptom Profile

As per the definition, 100% of patients presented with abdominal pain coupled with disordered defecation, abdominal distention, or both. The predominant symptoms of our cohort are shown in Table 1. Data are summarized for the entire cohort and for subgroups on the basis of the frequency of return for IBS-related visits. Other than abdominal pain, diarrhea, reported by 50% of patients, was the predominant bowel symptom.

Thirteen patients (12%) identified an acute gastrointestinal illness as predating their first visit for IBS-related symptoms. Twelve patients (11%) reported that they had had symptoms for 1 to 4 weeks before seeking medical assistance; 30 patients (27%) had had symptoms for 1 to 12 months; and 60 patients (54%) had had symptoms for more than 1 year. For 10 patients (9%), the duration of IBS symptoms before the first clinic visit was not noted in the record.

Organic Disease and Concomitant Conditions

Before the first IBS diagnosis, 72 patients (65%) had proctoscopic examination to rule out organic disease and 88 (79%) had a radiograph of the colon. Ten patients (9%) developed other gastrointestinal disease 2 to 30 years (median, 15 years) after diagnosis of IBS. These diseases were chronic pancreatitis (n = 2; 17 and 30 years later), gastrointestinal cancer (n = 4; 13 to 30 years later), small-bowel obstruction (n = 2; 7 and 12 years later), and gastric ulcers (n = 2; 2 and 12 years later). These subsequent diseases were considered to be unrelated to preexisting IBS; with the possible exception of the gastric ulcer detected 2 years after IBS diagnosis, all occurred long enough after IBS diagnosis to exclude a contribution to IBS symptoms.

Forty-five patients (40%) had previous or concurrent back pain, 45 (40%) had nonspecific chest pain, 42 (38%) had tension headache, 18 (16%) had migraine headache, and 16 (14%) had fibromyalgia.

Hospitalization and Surgery

The number of hospitalizations after the first diagnosis of IBS ranged from 0 to 16 (median, 3), the number of general surgeries ranged from 0 to 22 (median, 3), and the number of abdominal surgeries ranged from 0 to 13 (median, 0). Surgery was done in 100 patients and included appendectomy (n = 40), hysterectomy (n = 36), cholecystectomy (n = 23), and laparotomy (n = 14).

Mortality Rate

Twenty-seven patients died before August 1993; age at death ranged from 47 to 84 years (median, 76 years). Causes of death included cancer (n = 8), ischemic heart disease (n = 7), cerebrovascular disease (n = 2), pneumonia (n = 3), ruptured aortic aneurysm (n = 1), motor vehicle accident (n = 1), carbon monoxide poisoning (n = 1), and asphyxiation (n = 1). Three patients died of unknown causes. None of the deaths was considered to be related to IBS; deaths occurred from 5 to 30 years after diagnosis of IBS (median, 20 years). Observed survival was not different from expected survival (Figure 1).

View larger version (14K): [in this window] [in a new window]   Figure 1. Survival of the IBS cohort did not differ from expected survival. (P = 0.59, one-sample log-rank test). Twenty-seven of 112 patients with IBS died during a 32-year follow-up.

Physician-Patient Interaction

The number of return visits made for IBS-related symptoms was inversely related to the strength of the physician-patient interaction Figure 2, top). Patients perceived by the physician to be improved at a return visit (n = 44) had fewer follow-up visits for IBS than did patients not perceived to be improved (1.6 and 3.5 visits, respectively; P < 0.02). We found no overall association between strength of the physician-patient interaction and number of general or abdominal surgeries (P > 0.1, analysis of variance), but comparison of the strongest and weakest interaction groups (1.8 and 4.9 hospitalizations, respectively; P < 0.05) indicated that a positive interaction was associated with fewer hospitalizations over time.

View larger version (34K): [in this window] [in a new window]   Figure 2. Effects of physician-patient interaction on number of return visits.Top. Number of return visits (mean ±SE) for IBS-related symptoms was inversely related (P < 0.01, analysis of variance) to the strength of the physician-patient interaction as reflected by the written medical record. Maximum possible score is 6; a higher score suggests a better interaction. Bottom. Data viewed with respect to absence (open bars) or presence (hatched bars) of individual components making up the composite index of strength. A notation of any given variable contributed to the overall scoring of the physician-patient interaction.

Of the six component variables making up the composite index of strength of the physician-patient interaction Figure 2, bottom), notation of the patient's psychosocial history (P < 0.01), notation of the precipitating factors causing the patient to seek medical help (P < 0.01), and notation of discussion with the patient (P < 0.02) were associated with fewer follow-up visits for IBS-related symptoms. The use of pronouns and adjectives suggesting the patient's collaboration with the physician in treatment options was of borderline significance (P = 0.05). A notation of reassurance was not significantly related to the number of return visits. Only 1 of 112 records contained a physician's reference to the patient's name.

Discussion

Top Methods Results Discussion Author & Article Info References

Our study provides information about the course and prognosis associated with IBS. Other studies have acknowledged the lack of data on the long-term course of IBS [7]; the major strength of our study was the length of prospective follow-up using the medical record. Previous studies have followed patients for up to 8 years [4, 6, 7, 18, 19]; we were able to follow patients for up to 32 years. In the context of our study, patients not being seen at Mayo Clinic (including those who had died) at the end of the 32-year period were considered "lost" even though their full records remained for our review. However, it should be noted that these patients were followed for up to 26 years after initial diagnosis (median, 11 years); thus, even those who were "lost" remained part of our followed cohort for longer than patients in any previous study of IBS.

Earlier investigators have reported that the prognosis for patients with IBS is variable [4, 6, 7, 18, 19]. We found that only three patients had been misdiagnosed as having IBS despite a prolonged follow-up, confirming the results of shorter-term studies from Europe that indicate IBS to be a "safe" diagnosis unlikely to be altered with time. Previous investigators have also shown that most patients with IBS continue to be symptomatic during follow-up, although as many as one third of patients may become symptom-free over time (Table 2). Harvey and colleagues [4] noted that men have a better prognosis for IBS than do women, and we found that men tended to have fewer return visits for IBS-related symptoms than did women. Survival over the 32-year period did not differ from expected survival.

We indirectly assessed the strength of the physician-patient relationship through the medical record. Our results suggest that a strong physician-patient interaction may be related to a reduced number of return visits for IBS-related symptoms. Waller and Misiewicz [7] stressed the benefit of a sympathetic follow-up for patients with IBS. The component variables contributing to our strength index were compatible with the recommendations of Drossman and Thompson [13], who suggested that physicians should obtain a careful history; conduct a careful, cost-efficient examination; identify the patient's concerns; discover the precipitating factors bringing the patient to the physician; explain the patient's disorder; set consistent limits; respond realistically to the patient's expectations; and involve the patient in treatment to optimize results.

To our knowledge, our study was the first to use the medical record of a large patient cohort as a historical document for analysis of the physician-patient interaction. In Great Britain, the medical record of a single patient was used to analyze a physician's attitude toward the patient through the physician's choice of words [20]. The use of personal records is an accepted method in many fields for the examination of human relationships [21]. Physicians are authors of medical records; their choice of words may intentionally or unintentionally yield information about the author's state of mind and is an attempt to depict the functioning and structure of another personality, that of the patient [20, 21]. The consistency in the Mayo Clinic record system that has aided epidemiologic research [17] also provided us with a stable base from which to assess the physician-patient interaction.

Previous studies of physician-patient interaction have measured patient and physician satisfaction by direct observation, using video and audio recordings. However, these have been costly and fraught with methodologic difficulties so that they have not been replicated easily [22-30]. Nevertheless, these studies have suggested that a positive physician-patient interaction leads to greater patient compliance and satisfaction, and they have shown that the physician's awareness of the patient's psychosocial history is important in creating a strong physician-patient interaction [28]. Our study supports the concept that attention by the physician to the emotional needs of the patient with IBS may affect outcome; notations in the medical record relating to psychosocial history, precipitating factors, and discussion of diagnosis and treatment were all associated with fewer return visits for IBS-related symptoms. However, the validity of the written record for assessing the physician-patient interaction is uncertain, and our approach is limited to the physician's point of view. Although we found a clear association between the strength of the physician-patient interaction and a reduction in the number of return visits for IBS-related symptoms, we cannot deduce cause and effect. Our assumptions, however, are supported by the fact that patients continued to return over the 32-year period for symptoms unrelated to IBS.

We show that an accurate diagnosis of IBS according to current criteria is associated with a good prognosis. In addition, our findings suggest that a positive physician-patient interaction is associated with a reduced number of follow-up visits for IBS-related symptoms.