Comparison of the utilization of endoscopy units in selected teaching hospitals across Canada Utilisation comparative de l’endoscopie dans les hôpitaux universitaires au Canada

year), and it would be interesting to learn whether this represents an opportunity for cost-saving in some units.


RESULTS
The total number of gastroscopies, colonoscopies and endoscopic retrograde cholangiopancreatographies (ERCPs) performed at each of the eight units in 1992 varied from 1966 to 6392 (Table 1). In one unit only the total number of procedures was recorded. Gastroscopies represented the majority of the procedures, with 50.9% to 68.5%; colonoscopies represented 25.5% to 39.5% and ERCPs represented 3.0% to 6.3%. The absolute number of ERCPs performed each year varied widely, from 106 to 332. The survey did not distinguish between diagnostic and therapeutic ERCPs. A total of 60.7% to 86.0% of these three procedures were performed on out-patients and 14.0% to 39.3% were performed on inpatients ( Table 2).
A ratio of procedures per descriptor unit was derived. Because the total patient base was unknown, these endoscopic numbers (gastroscopies, colonoscopies and ERCPs) were expressed on the basis of total endoscopic procedures per in-patient bed or per endoscopist. The number of inpatient beds per hospital ranged from 217 to 804. The number of endoscopies per year per total number of in-patient beds ranged from 6.0 to 9.8 (Table 3). Two hospitals did not have designated gastrointestinal in-patient beds, while the remaining units had from five to 18 gastrointestinal beds. The number of endoscopic procedures performed per year per gastrointestinal bed varied from 109.2 to 1278.4.
The workload on the support staff, including receptionists and endoscope cleaners, in the endoscopy units was determined. The number of registered nurses (RNs) per unit varied from two to eight, and the number of other support staff varied from 1.5 to six (Table 4). Total number of endoscopies per RN varied from 453 to 1398, and the number of total endoscopies per total support staff including RNs varied from 323.7 to 1065.3.

DISCUSSION
There was a 10-fold difference in the number of endoscopic procedures performed on the basis of a designated gastrointestinal bed, and a 150% difference between centres in the number of endoscopies per in-patient bed. The workload per RN in the endoscopy unit varied by a factor of five and the workload for total number of support staff varied threefold. Thus, there was no consistent trend in these eight major teaching hospitals that allowed prediction of the number of appropriate procedures per descriptive designator. Furthermore, there does not appear to be any consistent level of support staff in these units.
Gastroscopies clearly represent the greater proportion of endoscopic procedures (between 50.9% and 68.5% of all endoscopic procedures were gastroscopies). From institution to institution the numbers varied from 1205 to 3980 per year. The total number of colonoscopies varied between 640 and    (Table 1), and there was no obvious relationship between the number or proportions of gastroscopies and colonoscopies. This variability likely represents local referral patterns and physician interest. The number of ERCPs performed per year depends on the interest of the teaching unit, the skill of attending physicians and the demands placed by affiliated laparoscopic and transplantation programs. In this study, the number of annual procedures ranged from 106 to 514. Minimum standards have been set for the number of ERCPs required before a gastroenterology training fellow may achieved a minimal standard of competence. Indeed, there is controversy whether diagnostic ERCPs should be taught to all gastrointestinal trainees. The results of this survey clearly raise concern because some centres may not perform sufficient ERCPs to provide a minimum basis for this important diagnostic and therapeutic skill.
With the steady reduction in the number of available in-patient beds, more gastrointestinal endoscopy procedures are being performed on out-patients. In each of the eight centres surveyed, at least two-thirds of total endoscopic procedures were performed on out-patients (Table 2). No relationship existed between the number of designated gastrointestinal in-patient beds and the percentage of out-patient endoscopies, and there was no relationship found between the number of total in-patients beds and the proportion of out-patient endoscopies.
The job description of gastroenterologists in teaching hospitals varies from a major emphasis on clinical practice and teaching to a major emphasis on research. Only the total number of endoscopists performing procedures is available, with no information on the proportion of their time used to care for patients or to perform endoscopies. In this study, it was not specified whether the endoscopist was a physician, surgeon, radiologist or pediatrician. Thus, the wide variation in the total number of endoscopies performed each year per endoscopist (Table 3) likely represents physician interest and job descriptions, rather than any difference in endoscopy indications or utilization.
Every endoscopy unit across Canada is under pressure because of economic downsizing and every effort must be made to identify possibilities for cost containment. The number of endoscopies performed in each unit per year per RN, or more importantly per total number of support staff, varied widely. Some units have proportionately more nurses, whereas other units may depend heavily on nursing assistants. In this survey, the nature of support staff other than RNs was not defined, and these persons may have been endoscopic assistants, receptionists or endoscope cleaners. In some units endoscopies may be performed without an RN in attendance and some units may make proportionately more use of endoscope washing machines versus cleaning by support staff. Nonetheless, the total number of endoscopies performed per support staff varied immensely -from 370.6 to 1065.3 per year (Table 4). This may represent an opportunity for some units to save on their annual budget allocations for support staff. It would be useful and interesting to learn how the high volume per personnel unit (such as hospital numbers 1, 3 and 7) were able to achieve this workload and to determine whether this was achieved without any loss of quality.
Endoscopic procedures are an important part of the diagnostic and therapeutic armamentarium of gastroenterologists. In some units endoscopic or medical quality improvement programs have been established. The Canadian Association of Gastroenterology has taken an active role in the review of quality standards in endoscopy (1). This survey was intentionally undertaken in teaching hospitals, and it is recognized that as such there was an inherent potential bias. It is likely that the appropriateness of endoscopic procedures needs to be based on accepted standards of practice (2) and accepted indications, and needs to be constantly monitored by a peer-review process to ensure quality maintenance. Basing 'appropriateness' of rates of one procedure on rates of other endoscopic procedures or of in-patient beds is unlikely to be valid.
This survey only addressed the issue of the quantity of endoscopies performed, and not the quality and outcome of procedures. The Alberta Endoscopy Project has been established to attempt to address issues of endoscopy outcome. It is hoped that the information obtained as a result of this study may be used to identify means of modifying endoscopic practices across Canada. We believe that the present survey clearly establishes that there is a very wide range of endoscopic utilization in these eight teaching hospitals, with approximately comparable rates of procedures of out-patient versus in-patient procedures and of gastroscopies versus colonoscopies, but no obvious linking of the ratios of inpatients:out-patients versus total number of designated gastrointestinal beds or total number of hospital beds. We also noted with interest that many units are staffed with 'other staff' rather than with RNs, and cost saving opportunities may be available to perform more endoscopies.

CONCLUSIONS
Endoscopic use varies widely in major teaching hospitals. We propose that the appropriateness of endoscopic procedures be based on adherence to standard guidelines with procedures in place at each institution to ensure guideline adherence. We suggest that appropriateness cannot be based on comparing one unit with another in terms of procedures performed per gastrointestinal bed, in-patient bed or endoscopist.