The Vermont Colorectal Cancer Project
Abstract & Commentary
Synopsis: This study reports results of a statewide quality improvement project that was designed to evaluate prospectively gathered surgical and adjuvant treatment data for colorectal cancer patients. Surgeon compliance in reporting was estimated at 78% statewide. Among 364 elective cases, 82% were identified after presentation with symptoms, compared with only 18% identified by screening. Eighty-five percent of patients underwent resection for cure, and 100% of patients with stage III colon cancer and stage II and III rectal cancer were offered referral for adjuvant therapy. The most important finding of the study was the feasibility of statewide surgeon participation in the development of a database devoted to quality-of-care improvement for colorectal cancer patients.
Source: Hyman N, et al. Arch Surg. 2002;137:413-416.The medical establishment, the lay press, and legislators have focused increasingly on methods for assessing and improving the quality of health care in recent years, the latter 2 perhaps partially spurred by the Institute of Medicine’s high-profile report on medical error released in 2000.1 One response to the impetus for accurate assessment of quality of care has been development of a colorectal cancer patient database by the Vermont Chapter of the American College of Surgeons and the Vermont Program for Quality Health Care. The database contains patient-, tumor-, and procedure-associated information submitted by the operating surgeon at the time of the index procedure and pathology and short-term outcome data from a 30-day follow-up questionnaire. The primary goal of this study was to evaluate the feasibility of performing a quality-of-care study, specifically addressing the surgical care of colorectal cancer patients. A secondary goal was to describe cancer demographics and consider their relationship to current screening methods.
Between April 1999 and March 2001, 33 Vermont surgeons submitted initial operative data and 30-day follow-up questionnaires with pathology and perioperative morbidity/mortality data. Emergent cases were excluded, leaving 364 cases for analysis. According to Vermont Tumor Registry 2000 data, 670 colorectal cancer cases were expected over 2 years. After excluding the expected 30% of cases presenting emergently, 469 expected cases remain. Thus, surgeon compliance was estimated at 364/469 = 78%.
A substantial majority of patients (82%) were identified by presentation with symptoms, such as obstruction, bleeding per rectum, or change in bowel habits. The remaining patients were identified at screening. No information regarding whether the symptomatic patients had been previously screened or by what method was given.
Screening method accuracy is clearly linked to tumor site in the case of sigmoidoscopy vs. colonoscopy, and some evidence indicates that tumor site may have an effect on the accuracy of fecal occult blood studies.2 Hyman and colleagues reported the frequency of tumor location in the right (36.6%), transverse (9.6%), descending (6.0%), and sigmoid colon (21.2%), and the rectum (26.6%). Again, no information regarding tumor site and previous screening or screening method was given.
Hyman et al reported an impressively high rate of appropriate treatment offered; 100% of patients underwent operation, 85% for cure, and 100% of patients were at least offered referral for adjuvant therapy. This contrasts a recent study reviewed by Clinical Oncology Alert in which 88% of rectal cancer patients underwent operation and 50-60% of those patients with stage II and III rectal cancer received adjuvant therapy,3,4 the standard of care per 1990 NIH Consensus Conference recommendations.5 Perioperative morbidity and mortality were 1.9% and 12.3%, both at the minimum of expected limits, indicating good short-term outcomes for patients undergoing care in this setting.
Comment by Arden Morris, MD
By its broadest definition, studying "quality of care" is the goal of virtually every published investigation. In 1988, Donabedian described a framework for this concept consisting of 1) structure, the attributes of the setting in which care occurs; 2) process, what is actually done in giving and receiving care; and 3) outcome, the effects of care on the health status of patients and populations.6 He argued that a meaningful quality-of-care study mandated examination of each of these major components and the links between them. Surgical care for oncologic disease is an especially worthy arena for such study, given that surgical resection forms the foundation of care for most solid tumors, hard data regarding short-term outcomes is readily available, and the incidence of solid tumors remains high.
Hyman et al have begun this process by evaluating the surgical care of colorectal cancer in patients in Vermont. They are to be commended for developing one of the first surgeon-driven, population-based, quality-of-care projects. The primary aim was to examine the feasibility of this endeavor. They achieved a 78% estimated compliance rate among surgeons and demonstrated impressive short-term results. Still, there are substantial gaps in the information reported.
Although the secondary aim of this study was to assess the implication of cancer demographics for screening, no information was given regarding screening resources available, or even patient resources such as insurance type. They appropriately draw the only conclusion allowed by the data, which is that colorectal tumor sites appear to be migrating more proximally, furthering the case for screening by colonoscopy. This may be an arising biological imperative, but implementation of recommendations won’t occur without resources (or, as implied by Donabedian’s framework, Process depends on Structure).
Process, the second construct of quality of care, is addressed by describing the types of operations performed. However, because the point of entry for this study was performance of elective operation, no denominator is available to determine the frequency of appropriate or even any operation performed. Furthermore, "use of adjuvant therapy" is listed as a main outcome measure but the only information given is that all patients were offered referral for adjuvant therapy. Ideally, information about adequacy of adjuvant therapy should be supplied, although most database studies simply report a dichotomous variable for any adjuvant therapy received. In this study, the reported 100% "use of adjuvant therapy" is misleading.
Outcome, Donabedian’s third construct, is the most thoroughly explored. Hyman et al describe short-term surgical morbidity (for specific complications) and mortality, with excellent results compared to the published literature. They acknowledge the absence of more meaningful long-term health status outcomes but clearly plan to continue gathering applicable data on ascertained cases. Cost information and patient utility information could round out this section of the overall study, and would be interesting to follow in the future.
Dr. Morris is a Robert Wood Johnson Clinical Scholar, University of Washington, Seattle, WA.
References
1. Kohn LT, et al. To Err is Human. Building a Safer Health Care System. 1st ed. Washington, DC: National Academy Press; 2000:287.
2. Nakama H, et al. Can J Gastroenterol. 2001;15(4): 227-230.
3. Schroen AT, Cress RD. Ann Surg. 2001;234(5): 641-651.
4. Morris A. Clinical Oncology Alert. 2002;17(3):17-19.
5. NIH consensus conference. JAMA. 1990;264(11): 1444-1450.
6. Donabedian E. JAMA. 1988;260(12):1743-1748.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content