| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Article |
From the Institute of Oncology Ljubljana (IE), Ljubljana, Slovenia; University of Calgary/Tom Baker Cancer Centre (WJT), Calgary, Alberta, Canada; Cross Cancer Institute (CJdeG), Edmonton, Alberta, Canada; and Faculty of Medicine (GCES), University of British Columbia, Vancouver, British Columbia, Canada.
ABSTRACT
Background: Quality of surgery is a proven prognostic factor in many tumors. It is critical to ensure that an effective method is in place to evaluate surgery accurately.
Material and Methods: A provincial Cancer Surgery Working Group designed and piloted a computerized synoptic operative report template (WebSMR) in rectal cancer surgery, to replace the standard narrative operative record (NR). This included a precise description of the procedure, data on demographics, diagnostic evaluation, staging, and functional measures. A total of 70 items for anterior resection (AR) and 63 items for abdominoperinal excision (APR) were included. The WebSMR was assessed for comparison with 40 NR randomly selected from seven hospitals in Southern Alberta from 2001 to 2003.
Results: The NR contained 45.9% of the specified data elements and the WebSMR captured 99%. The most complete NR data (68.8% to 97%) concerned hospital and patient data, anesthetist and surgeon information, approach, and closure details. The important details of laparotomy and tumor resection were the next most complete data (33.5% to 47.5%) and the least complete (0 to 25%) concerned preoperative treatment, comorbidity, and metastatic and local assessment. All differences among these groups were statistically different (P < .001). No statistically significant differences were seen in the completeness of the NR according to the type of surgery (AR vs. APR; P = .1) or the dictating surgeon (colorectal vs. general vs. resident; P = .175). The time needed to complete the WebSMR test was only 6 minutes.
Conclusion: The science of surgical technique can be better measured by this unique instrument and will create accountability in surgery.
Key Words: Rectal cancer • Quality control • Computer • Operative report • Surgery • Synoptic • Narrative
Quality of medical and surgical treatment is an issue that is becoming a critical concern for patients, health care physicians, and institutes.1–3 This is particularly relevant for patients with cancer, where the characteristics of both surgeons and hospitals are significant prognostic variables in both short- and long-term outcomes for the patient.4,5 In 1999, the Alberta Cancer Board established the Cancer Surgery Working Group (CSWG) to address the quality of cancer surgery in the province of Alberta (Canada). The CSWG, consisting of community surgeons, colorectal surgeons, surgical oncologists, and oncology program leaders as information systems specialists, and Alberta Health and Wellness representatives convened to examine the issue.
A survey of the literature identified multiple guidelines6 developed to improve the quality of the medical records and notes, but no indication of its application to operative procedures. The initial project was to examine the feasibility of a web-based computerized operative report template that would collect the specified data necessary to measure the quality of the rectal surgery and to replace the narrative operative record currently used. The rationale of this approach is that the template, which describes the surgery, can be used as ‘‘real-time’’ data to derive outcomes. Additionally, a report can be sent immediately to all appropriate health care providers and to the pathologist in synoptic format. Furthermore, as a sequela, a synoptic report serves as a reminder of the essential steps of the procedure and, therefore, is a powerful educational tool.
The development of the tool began after two province-wide surgical workshops were convened on rectal cancer surgery. These included demonstrations of total mesorectal excision by Mr. Bill Heald, a recognized world expert and originator of the technique. These were followed by 10 sessions to define a template describing an operative procedure in synoptic fashion, which would include most of the possible surgical scenarios. This template included the details of surgery and those items essential to the decision-making process for the procedure (e.g., colonoscopy and endorectal-ultrasound) and functional outcome parameters (e.g., continence). These data elements were identified based on consensus among the CSWG members. To retain a reasonable length for the template, the ancillary items that many surgeons dictate in a narrative report (e.g., bowel preparation and deep vein thrombosis prophylaxis) were excluded. Most of the variables identified, however, were based on what was believed to be current practice (i.e., among surgeons who perform rectal cancer surgery and other oncology consultants).
As the use of the synoptic operative report is different from standard surgical practice, it is critical to investigate and demonstrate the value of this format. The objective of this article will identify the degree of congruency in detail between these two operative reports systems.
MATERIALS AND METHODS
The operative records of 40 patients with a diagnosis of rectal cancer during the period from 2001 to 2003 were randomly selected from the Health Records departments of hospitals from the southern part of Alberta (Canada). The University of Calgary ethics committee approved this study.
The computerized template developed by the CSWG, consisting of 13 consensus grouped sets of variables (Table 1), was used for the evaluation. The total number of variables for abdominoperineal excision (APR) and low anterior resection (AR) was 63 and 70, respectively. All variables were coded as 1, 0, or NA. If the required information could be retrieved from the report, it was coded as 1 (positive) and if the required information could not be retrieved, it was coded it as 0 (negative). If the variable was not applicable for the particular patient, it was coded as NA (non-applicable). This review was conducted to check if the required information existed, regardless of the quality of the surgery. For example, information about mesorectal fascia was coded as positive if it was described in the report, regardless of whether it was surgically penetrated or not. Contrary, the same information was coded as negative if mesorectal fascia was not mentioned in the report. For each group of variables in each report, the total number of positive, negative, and non-applicable codes was counted. The non-applicable variables were excluded and the percentage of positive codes was calculated for each group of variables. Mean value of all groups of variables presented final rate for each report. Mean values of the same groups of variables in all reports were also calculated.
|
|
The relative completeness of reports by colorectal surgeons, general surgeons, and general surgery residents was examined separately.
The time necessary to complete the synoptic report was measured. Three surgeons were asked to complete four separate computerized synoptic reports, each based on the information from a narrative report, which they had read previously. Their first two completion times were excluded as learning curve and the mean value of the third and fourth completion time was calculated.
A database was created and all calculations were performed using Microsoft Excel 97 and SPSS SigmaStat 3.0 for Windows.
RESULTS
Forty randomly selected operative reports describing procedures performed on a cohort of patients with a diagnosis of rectal cancer from seven hospitals (two university, three community, and two rural) were used. The AR was described in 29 operative reports and the APR was described in 11 reports. Dictating these reports were 12 surgeons (4 colorectal surgeons, 5 general surgeons and 3 residents).
The mean rate of retrieved information from narrative (dictated) operative report was only 45.9% (Table 2) from all reports taken together (47.4% for AR and 41.9% for APR). Difference between the AR and APR groups was not statistically significant (two-tail t test, P = .1). According to the rate of retrieved information, the groups of variables fell into three categories. The most frequently (68.8% to 97%) retrieved data were those elements concerning the hospital, patient, surgeon, and anesthetist; incision and indication for the surgical procedure; and details describing the closure. Less frequently (33.5% to 47.5%) retrieved data were those elements concerning the laparotomy and details of the procedure. The least frequent (0 to 25%) retrieved data were related to preoperative comorbidity, local and metastatic assessment, carcinoembryonic antigen (CEA) levels, and preoperative treatment. Differences between these three groups are statistically significant (Kruskal-Wallis 1-way analysis of variance [ANOVA] on ranks, Dunn method: P < .001)
|
Colorectal surgeons, general surgeons, and general surgery residents dictated 29, 5, and 6 reports with mean rate of retrieved information 44.2%, 48.9%, and 51.6%, respectively (Table 3). No statistically significant differences were noted in the rates of retrieved information between these three data groups of surgeons (Kruskal-Wallis 1-way ANOVA on ranks, Dunn method: P = .175). The most common noted items, which were described in narrative but not included in the synoptic report, are found in Table 4.
|
|
On the synoptic reports, the total number of items not included on the 29 AR and 11 APR operative reports was 113 (80 and 33, respectively). If one excludes items deliberately omitted by CSWG to keep the form to a reasonable length, then the total number of items missing in all reports is 27, which is only 1% compared with the total of 2723 items in all 40 synoptic reports.
Mean times three surgeons needed to complete the four synoptic reports were: 10'46'', 7'42, 6'27'', and 5'32''. Assuming the first two attempts were learning ones, mean time for completing the computerized operative report was 5'59''.
DISCUSSION
This is the first systematic study of the value and completeness of the operative record in rectal cancer surgery. The strength of this study is that the template was developed by a cross section of surgeons from different backgrounds and practice settings after attendance at an intensive workshop in evidence-based techniques in rectal cancer surgery.
This study shows that the rate of the retrieved selected information from the narrative operative reports was only 45.9%, which was surprisingly low. Furthermore, the most completed parts of operative reports were those concerning the less important information (patient and hospital information, indication for the procedure, and closure technique). For the most important part of the report (details of the resection procedure), the rate of retrieved information was from 33% to <50%. It was clearly shown that a dictated operative report failed to document adequately the performed procedure. Although it is not likely that these missing details correlate with surgery of less quality in most cases, without these details it is not possible to assess the quality of the surgery. Additionally, it is not possible for a surgeon to consider personal results to improve his or her surgical technique.
Also shown in this study, was a trend that residents dictate more complete reports than consultants, although difference was not shown to be statistically significant. This is not surprising in view of the small numbers (n = 3). Baigrie et al.7 studied whether 264 operative reports followed the Guidelines6 of The Royal College of Surgeons of England for the clinicians on medical records and notes, requiring very basic details (e.g., procedure performed, sutures used). They showed that trainees produced better operative reports and that 
70% of reports written by consultants were not useful. This latter finding was not the case in this report. Resident staff can focus on operative reports as a learning experience.
The computerized synoptic operative report provides a template with a mandatory list of items that have to be identified; otherwise, the report cannot be completed. This means that the compliance would be 100% once this format has been accepted as part of the medical record by an institution. It provides a powerful tool to ensure that important details of a surgery have been described. Few operative techniques are found in randomized studies that have proved superior, although certain maneuvers are determined essential (e.g., mesorectal excision) to achieve maximal local regional control. In the absence of randomized trials, the synoptic reporting system will provide a database to identify critical elements by providing complete prospective data compared with that found in the narrative report. In a narrative operative report, frequently found were descriptions such as "patient was draped in the usual way" or "we reached the bottom of the pelvis"); however, in the same report, was no documentation of whether the right ureter or mesorectal fascia had been identified. The impression is that the surgeons cannot explicitly describe routine items, which later on can become important (e.g., not perforating the rectum or dissecting in or out of the mesorectal plane).
Using the computer template is effective and simple. A software engine termed the ‘‘Survey Administration Module’’ is generic survey tool that customizes the rectal template into a WebSMR. Its smart navigational features include the following:
Having these features means:
A template cannot be used to predict every possible event, so a text box is included, which, in the view of the narrative report, is sufficient for most of the infrequently encountered situations.
The quality of rectal cancer surgery is currently defined by outcomes (e.g., locoregional recurrence, sphincter-saving surgery, function, and lower operative mortality and morbidity). The real challenge, however, is to define what are critical components of a surgical operation that do contribute to this quality. One such example, of course, is the use of mesorectal excision. It is recognized that the conduct of the surgery (i.e., the surgeon) is one of the most important prognostic factors in the rectal cancers.2,4,5 The direct entering of the data into the computer template provides a real-time accurate description of the procedure, which can be used immediately to assure quality control for patients with cancer by monitoring the completeness of surgery. The CSWG has plans to implement this format of operative reports in all hospitals in Alberta. Access to the database is web based, which means that it can be accessed throughout the province in any hospital setting with any Internet connection. Prospective data collection is of great value, which this computerized operative report provides.8
The only other published attempt at computerizing the operative records was described by De Orio,9 who created a word processor template for standard orthopedic reports. Using this template, the surgeon had to dictate only key words, which reduced time necessary to complete the report and saved up to 1 hour of the surgeon’s time daily. Furthermore, assuming 48 operative reports created daily in the department, they calculated savings of up to $98,755/year, mainly because of dramatically reduced typing time. With the computerized operative record, the savings are equally substantial, because no typing is required and the surgeon can verify the report immediately. A formal cost-benefit analysis of the surgeon’s time is yet to be completed. Because this system is online, the report can be accessed by any caregiver and sent immediately to all necessary locations (e.g., recovery room, in-patient units, the referring physician, the cancer center, and the pathologist). This will enhance the quantity and quality of information provided to the pathologist to process the specimen appropriately.
Acceptance of the synoptic report by physicians and surgeons may be problematic because a study from Lissauer et al.10 described the narrative form of medical documentation is more readable than computer-generated ones. If one analyzes the synoptic report, however, it is far more efficient in providing important details and is much faster to extract critical information. A physician, familiar with such forms, can summarize data in these reports in seconds, whereas reading the two or more pages of the same pathology report takes much longer. In other medical specialties, the use of computerized templates or preprinted forms has positive echoes. Sleszynski et al.11 compared the accuracy and efficacy of information obtained by using standardized outpatient osteopathic note forms and those obtained using physician’s progress notes. They found that the greater content of information was almost always found in standardized note forms. Van Walraven et al.12 compared discharge summaries created from computer databases and those created by voice dictation. They found that computer-created discharge was faster and also preferred by hospital staff.
To be objective in assessing the value of the narrative record, this study compared information only found in that format. Of the four items frequently missed in synoptic report, three of them (type of anesthesia, preparation of the skin and draping, and deep vein thrombosis prophylaxis) were not felt to be essential and, therefore, were deliberately omitted to shorten the form. The clinical status of the gallbladder, as it has significance in subsequent care, and the length of the rectal stump were felt to be useful additions to the synoptic report and will be added. The other missing items can be easily included in the comments section provided.
Developing a synoptic report was not without challenge. It required a great deal of compulsion to create a usable form and it also required consensus on what constituted a minimal standard for rectal cancer surgery. The consequence is that it represents a standard for rectal cancer surgery and an effective educational tool reminding the surgeon of essential steps and details of the surgery.
CONCLUSIONS
Quality of surgery, as with all other scientific disciplines, can be accurately measured using a web-based synoptic record to create the operative record.
This study showed that narrative operative record is not as complete and that the routine use of a computer template can significantly improve quality measurements of rectal cancer surgery. Unless a surgeon documents accurately personal judgement of a procedure, the description of surgery will remain more as an art than a science.
CSWG Current and Past Members
FOOTNOTES
Received December 30, 2003; accepted June 17, 2004.
Address correspondence and reprint requests to: Ibrahim Edhemovic, MD, MSc, Institute of Oncology Ljubljana, Zaloska 2, 1000 Ljubljana, Slovenia; Fax: +386-1-587-9407; E-mail: ibrahim.edhemovic{at}mf.uni-lj.si.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
