ClinicalRadiology (1991)44, 317-321
Barium Meal or Endoscopy? A Prospective Randomized Study of Patient Preference and Physician Decision Making G. W. STEVENSON, G. N O R M A N * , R. F R O S T and S. SOMERS
Departments of Radiology and *Epidemiology, McMaster University, and Chedoke-McMaster Hospitals, Hamilton, Ontario, Canada This study examines the preference of 64 out-patients for either a barium meal or an upper gastrointestinal endoscopy. The sequence of the examinations was randomized to avoid order bias. An initial preference for a barium meal of almost two-to-one, with 53% having no preference, was changed after the investigations to a preference for endoscopy of two-to-one with 5% having no preference. The use of mild sedation and the skill of the endoscopist had a major impact on the patients' opinion of the endoscopic procedure, and on their choice of examination for any repeat study. There was little agreement between clinical diagnosis and the result of investigation, but clinicians tended to accept the reported result of the investigation, especially if the result was abnormal. Despite a change in diagnosis in 34 of 49 patients there was little change in management as a result of the investigations, supporting the view that young patients with dyspepsia may be managed with symptomatic treatment initially, and without investigation. This study lends further support to the view that endoscopy should be the investigation of choice in patients with persistent dyspepsia, especially those whose age or infirmity may make barium examination suboptimal. Stevenson, G.W., Norman, G., Frost, R. & Somers, S. (1991). Clinical Radiology 44, 317 321. Barium Meal or Endoscopy? A Prospective Randomized Study of Patient Preference and Physician Decision Making
Double-contrast barium meal and endoscopy are both used in the investigation of patients presenting for the first time with dyspepsia. The choice of procedure may be determined by sensitivity, costand safety, but when these are similar patient preference should be taken into account. There was for many years a belief that patients preferred barium meals to endoscopy (Editorials, 1980, 1982). A few studies have found patient acceptance of endoscopy to be high (Hoare and Hawkins, 1976; Walker and Smith, 1978; Thompson et al., 1980; Dooley et al., 1986; Hacker et al., 1987), but these have for the most part been retrospective or open to order bias. Many of our patients have expressed a preference for endoscopy, particularly elderly and immobile patients, for whom a barium meal is more arduous and technically less satisfactory. A prospective study was therefore designed to examine the preference of a group of out-patients referred for investigation by their family physicians, using randomization of the sequence of investigation so as to avoid order bias.
P A T I E N T S AND M E T H O D S Patients included in the study were those referred routinely for the investigation of dyspepsia from the Department of Family Medicine, with whose members the study plan was discussed, The family physician explained the study, obtained consent and referred the patient to the Radiology Department for a barium meal and endoscopy. The order of testing was randomized, and the patients were interviewed before the first test to Correspondenceto: Dr G. W. Stevenson,Department of Radiology, Chedoke-McMaster Hospitals, McMaster Division, 1200 Main St. West, Hamilton, Ontario LSN 3Z5, Canada.
confirm that they understood the nature of the study. Patients were excluded if they had dysphagia, or if they had previously been investigated for dyspepsia. Patients completed a questionnaire before and after each study, to determine their level of anxiety about, and preference for each procedure. Anxiety, discomfort and recall were assessed on 7-point Likert scales. A research assistant distributed the questionnaires and telephoned the patients 7 days after the second procedure to obtain answers to a summary questionnaire. All patients completed a signed consent form, and the study was approved by the hospital's ethics committee. The barium meal was performed by one of two staff gastrointestinal (GI) radiologists or by one of three G! radiology fellows. The standard double-contrast technique of the department was used, using a high density barium suspension, gas granules and compression. After demonstration of the oesophagus and stomach, a 20 mg intravenous injection of hyoscine butylbromide (Buscopan, Boehringer) was given for improved duodenal visualization. A limited explanation of the findings was given at the end of the examination. Endoscopy was performed using a small calibre forward viewing endoscope (GIFXP10, Olympus Corporation). The examination was performed by one of two staff radiologists or one fellow in GI radiology, all of whom were experienced endoscopists. The pharynx was anaesthetized with xylocaine spray, and all patients were given the opportunity to receive intravenous sedation using fentanyl 0.1 mg (Sublimaze, Janssen) but some preferred to do without. The standard endoscopic technique of the department was used, with patients being encouraged to watch the procedure through the teaching attachment. If they did so, they were given a running discussion of the appearances as the examination progressed. A limited explanation of the findings was given,
318
CLINICAL RADIOLOGY
Table 1 - Answers to preference questionnaires
No. of patients (%)
Prefer endoscopy Definirely Slightly No preference Prefer barium meal Slightly Definitely
Before meal
Before endoscopy
AJ?er 7 days
WouM choose again
11 (17) 5 6 34 (53) 19 (30) 10 9
9 (14) 6 3 29 (45) 26 (41) 14 12
37 (58) 29 8 6 (9) 21 (33) 12 9
41 (64) 3 (5) 20 (31)
meal. Patients receiving no sedation were few in number, but their preference was split evenly between endoscopy and radiology in the case of physicians A and C_ The preference for endoscopy over barium meal for patients who had sedation with fentanyl and were examined by physician A or C was in a ratio of 9/1. Both the endoscopist's technique, and the use of even a mild amount of sedation had a considerable effect on patient preference. The effect of sedation was not due to amnesia, as the score for recall of the endoscopy was 1.5 (with 1 representing total recall and 7 being complete amnesia), and was similar to recall after barium meal. Management
Table 2 - Effect of physician and sedation on preference
No. of patients
Prefer endoscopy No preference Prefer meal
Physician A
Physician B
Physician C
S
NS
S
NS
S
NS
9 1 1
4 1 5
4 1 10
0 0 2
18 1 2
1 2 2
s, sedation; NS, no sedation. to satisfy the patient without encroaching on the prerogatives of the referring physician. In all cases, endoscopy and the barium meal were performed by different radiologists, the radiologist being unaware of the result of the other investigation. The procedures usually took place on consecutive days, and always within 5 days of each other. The referring physicians completed a form giving their provisional diagnosis and intended treatment, and completed further forms repeating this information after each test result. The forms were collected regularly from the physicians' offices by the research assistants.
RESULTS Preference Data
Sixty-four patients completed the study. Their answers to the preference questionnaires are given in Table 1. Before both barium meal and endoscopy, many patients expressed no preference, but of those that had a preference, the choice was 45/20 in favour of barium meal. However, 7 days after completing the second procedure there had been a significant swing in favour of endoscopy, with 58% preferring endoscopy and 33% preferring the meal (P
Barium Meal or Endoscopy? A Prospective Randomized Study of Patient Preference and Physician Decision Making G. W. STEVENSON, G. N O R M A N * , R. F R O S T and S. SOMERS
Departments of Radiology and *Epidemiology, McMaster University, and Chedoke-McMaster Hospitals, Hamilton, Ontario, Canada This study examines the preference of 64 out-patients for either a barium meal or an upper gastrointestinal endoscopy. The sequence of the examinations was randomized to avoid order bias. An initial preference for a barium meal of almost two-to-one, with 53% having no preference, was changed after the investigations to a preference for endoscopy of two-to-one with 5% having no preference. The use of mild sedation and the skill of the endoscopist had a major impact on the patients' opinion of the endoscopic procedure, and on their choice of examination for any repeat study. There was little agreement between clinical diagnosis and the result of investigation, but clinicians tended to accept the reported result of the investigation, especially if the result was abnormal. Despite a change in diagnosis in 34 of 49 patients there was little change in management as a result of the investigations, supporting the view that young patients with dyspepsia may be managed with symptomatic treatment initially, and without investigation. This study lends further support to the view that endoscopy should be the investigation of choice in patients with persistent dyspepsia, especially those whose age or infirmity may make barium examination suboptimal. Stevenson, G.W., Norman, G., Frost, R. & Somers, S. (1991). Clinical Radiology 44, 317 321. Barium Meal or Endoscopy? A Prospective Randomized Study of Patient Preference and Physician Decision Making
Double-contrast barium meal and endoscopy are both used in the investigation of patients presenting for the first time with dyspepsia. The choice of procedure may be determined by sensitivity, costand safety, but when these are similar patient preference should be taken into account. There was for many years a belief that patients preferred barium meals to endoscopy (Editorials, 1980, 1982). A few studies have found patient acceptance of endoscopy to be high (Hoare and Hawkins, 1976; Walker and Smith, 1978; Thompson et al., 1980; Dooley et al., 1986; Hacker et al., 1987), but these have for the most part been retrospective or open to order bias. Many of our patients have expressed a preference for endoscopy, particularly elderly and immobile patients, for whom a barium meal is more arduous and technically less satisfactory. A prospective study was therefore designed to examine the preference of a group of out-patients referred for investigation by their family physicians, using randomization of the sequence of investigation so as to avoid order bias.
P A T I E N T S AND M E T H O D S Patients included in the study were those referred routinely for the investigation of dyspepsia from the Department of Family Medicine, with whose members the study plan was discussed, The family physician explained the study, obtained consent and referred the patient to the Radiology Department for a barium meal and endoscopy. The order of testing was randomized, and the patients were interviewed before the first test to Correspondenceto: Dr G. W. Stevenson,Department of Radiology, Chedoke-McMaster Hospitals, McMaster Division, 1200 Main St. West, Hamilton, Ontario LSN 3Z5, Canada.
confirm that they understood the nature of the study. Patients were excluded if they had dysphagia, or if they had previously been investigated for dyspepsia. Patients completed a questionnaire before and after each study, to determine their level of anxiety about, and preference for each procedure. Anxiety, discomfort and recall were assessed on 7-point Likert scales. A research assistant distributed the questionnaires and telephoned the patients 7 days after the second procedure to obtain answers to a summary questionnaire. All patients completed a signed consent form, and the study was approved by the hospital's ethics committee. The barium meal was performed by one of two staff gastrointestinal (GI) radiologists or by one of three G! radiology fellows. The standard double-contrast technique of the department was used, using a high density barium suspension, gas granules and compression. After demonstration of the oesophagus and stomach, a 20 mg intravenous injection of hyoscine butylbromide (Buscopan, Boehringer) was given for improved duodenal visualization. A limited explanation of the findings was given at the end of the examination. Endoscopy was performed using a small calibre forward viewing endoscope (GIFXP10, Olympus Corporation). The examination was performed by one of two staff radiologists or one fellow in GI radiology, all of whom were experienced endoscopists. The pharynx was anaesthetized with xylocaine spray, and all patients were given the opportunity to receive intravenous sedation using fentanyl 0.1 mg (Sublimaze, Janssen) but some preferred to do without. The standard endoscopic technique of the department was used, with patients being encouraged to watch the procedure through the teaching attachment. If they did so, they were given a running discussion of the appearances as the examination progressed. A limited explanation of the findings was given,
318
CLINICAL RADIOLOGY
Table 1 - Answers to preference questionnaires
No. of patients (%)
Prefer endoscopy Definirely Slightly No preference Prefer barium meal Slightly Definitely
Before meal
Before endoscopy
AJ?er 7 days
WouM choose again
11 (17) 5 6 34 (53) 19 (30) 10 9
9 (14) 6 3 29 (45) 26 (41) 14 12
37 (58) 29 8 6 (9) 21 (33) 12 9
41 (64) 3 (5) 20 (31)
meal. Patients receiving no sedation were few in number, but their preference was split evenly between endoscopy and radiology in the case of physicians A and C_ The preference for endoscopy over barium meal for patients who had sedation with fentanyl and were examined by physician A or C was in a ratio of 9/1. Both the endoscopist's technique, and the use of even a mild amount of sedation had a considerable effect on patient preference. The effect of sedation was not due to amnesia, as the score for recall of the endoscopy was 1.5 (with 1 representing total recall and 7 being complete amnesia), and was similar to recall after barium meal. Management
Table 2 - Effect of physician and sedation on preference
No. of patients
Prefer endoscopy No preference Prefer meal
Physician A
Physician B
Physician C
S
NS
S
NS
S
NS
9 1 1
4 1 5
4 1 10
0 0 2
18 1 2
1 2 2
s, sedation; NS, no sedation. to satisfy the patient without encroaching on the prerogatives of the referring physician. In all cases, endoscopy and the barium meal were performed by different radiologists, the radiologist being unaware of the result of the other investigation. The procedures usually took place on consecutive days, and always within 5 days of each other. The referring physicians completed a form giving their provisional diagnosis and intended treatment, and completed further forms repeating this information after each test result. The forms were collected regularly from the physicians' offices by the research assistants.
RESULTS Preference Data
Sixty-four patients completed the study. Their answers to the preference questionnaires are given in Table 1. Before both barium meal and endoscopy, many patients expressed no preference, but of those that had a preference, the choice was 45/20 in favour of barium meal. However, 7 days after completing the second procedure there had been a significant swing in favour of endoscopy, with 58% preferring endoscopy and 33% preferring the meal (P
