Development and implementation of a clinical pathway. See "Clinical Pathway Development" subsection of the "Patients, Materials, and Methods" section for detailed explanation.
Median comparative length of hospital stay (LOS) for the historical control group (prepathway group, n = 96 patients), contemporaneous nonpathway control group (n = 64 patients), and clinical pathway enrollees (n = 30 patients). Asterisks indicate the outliers; error bars, the distribution.
Total comparative costs for the historical control group (prepathway group, n = 96 patients), contemporaneous nonpathway control group (n = 64 patients), and clinical pathway enrollees (n = 30 patients). Asterisks indicate outliers; error bars, the distribution.
Amy Y. Chen, David Callender, Carol Mansyur, Kim M. Reyna, Ellen Limitone, Helmuth Goepfert. The Impact of Clinical Pathways on the Practice of Head and Neck Oncologic SurgeryThe University of Texas M. D. Anderson Cancer Center Experience. Arch Otolaryngol Head Neck Surg. 2000;126(3):322–326. doi:10.1001/archotol.126.3.322
To assess the impact of clinical pathways on the practice of head and neck oncologic surgery in an academic center.
Cancer treatment center.
The study population consisted of 3 groups of patients who underwent unilateral neck dissection and were treated in the Department of Head and Neck Surgery of the University of Texas M. D. Anderson Cancer Center, Houston. Additional procedures which may have been performed were direct laryngoscopy, rigid esophagoscopy, and/or dental extractions. Ninety-six patients treated during 1993-1994 prior to the implementation of the clinical pathway (historical control group) were compared with 94 patients treated during 1996-1998, 64 who were not (contemporaneous nonpathway group) and 30 who were managed on the clinical pathway (pathway group). Patients from 1995 were excluded since the pathway was in the planning stages then.
Main Outcome Measures
Median length of stay; median total costs of care.
The median length of hospital stay of the historical control, contemporaneous nonpathway, and pathway groups decreased from 4.0 to 2.0 days (P<.001). The total median costs of care were less in the pathway group as compared with the historical contol group ($6227 and $8459, respectively, P<.001) and also less in the contemporaneous nonpathway group compared with the historical control group ($6885 and $8459, respectively, P<.001). Mean and median length of hospital stay and costs were lower in the pathway group as compared with the nonpathway group but not significantly (P = .11 and P = .07, respectively) The contemporaneous nonpathway and pathway groups did not differ in complications or readmissions.
Development and implementation of this clinical pathway played a statistically significant role in decreasing length of hospital stay and total costs of care associated with neck dissection between nonpathway and pathway patients. Thus, a more cost-effective practice environment has resulted for all of our patients.
THE CHALLENGE of providing consistently high-quality care and of being cost-efficient is a major emphasis of the medical community especially in this age of managed care. In particular, tertiary care centers have been under increasing pressure to reduce costs. Clinical pathways were first implemented in engineering fields in an effort to maximize efficiency in resource allocation and to provide guidelines for timely job completions. As defined by Coffey et al, a clinical pathway is an "optimal sequencing and timing of interventions by physicians, nurses, and other staff for a particular diagnosis or procedure."1(p45) Ideally, as guided by a clinical pathway, patients receive necessary and essential aspects of care at the appropriate sequence of time, resulting in maximum efficiency of the delivery system.
The University of Texas M. D. Anderson Cancer Center, Houston, is a tertiary referral center for patients with cancer worldwide. More than 1200 new patients are evaluated annually in the Department of Head and Neck Surgery and 70% receive surgical treatment as part of their overall treatment plan. Because of the volume of patients and the desire to standardize care as well as provide cost-efficient management strategies, clinical pathways were developed and implemented in an effort to define critical aspects of care and provide them in the most effective and efficient manner.
The study population consisted of 3 groups of patients who underwent unilateral neck dissection and who were treated in the Department of Head and Neck Surgery of the University of Texas M. D. Anderson Cancer Center. Additional procedures which may have been performed were direct laryngoscopy (DL), rigid esophagoscopy (RE), and/ or dental extractions.
The historical control group (prepathway) included 96 patients who were treated between September 1, 1993, and December 31, 1994. Because discussion concerning development of clinical pathways commenced in 1995, 1993 and 1994 were chosen as the prepathway control group to minimize any bias the medical staff may have had. Thus, patients treated during this period were not subject to any preconceived notions concerning the outcome of clinical pathway implementation. Four hundred sixty-three patients during this period who had neck dissection Current Procedural Terminology codes of 38700, 38720, and 38724 were extracted from the patient care database provided by the Department of Medical Informatics, University of Texas, Houston. Each patient's surgical record was then reviewed and only those patients who had unilateral neck dissection; unilateral neck dissection and DL; unilateral neck dissection, DL, and RE; or a unilateral neck dissection with DL, RE, and dental extractions were included in this study's analysis.
The second group was the contemporaneous nonpathway control group and included 64 patients who underwent the same surgical procedure during the time of implementation of the clinical pathway (September 1, 1996, to August 31, 1998) but were not managed based on the pathway. Each patient's surgical record was reviewed and only those patients who had unilateral neck dissection; unilateral neck dissection and DL; unilateral neck dissection, DL, and RE; or unilateral neck dissection with DL, RE, and dental extractions were included in this study's analysis. Although this group of patients was eligible for the clinical pathway, this group of patients was not enrolled in the clinical pathway because of treating faculty preference or omission. The decision to place patients on the clinical pathway was made solely by the treating physician. Enrollment was not mandatory.
The third group included 30 clinical pathway-managed patients who underwent neck dissection using the pathway between September 1, 1996, and August 31, 1998. Unilateral neck dissections may have included DL, RE, and dental extractions.
Patients from 1995 were excluded since the pathway was in the planning stages then.
The clinical pathway was developed systematically (Figure 1). First a core group determined which pathways were to be developed. A disease site working group was then organized to draft the pathway. The proposed pathway was then reviewed by all members of the core team. Any revisions were supported by peer-reviewed literature. A final draft was then approved by the disease site working group and presented to the Guideline Review Committee for approval. After the Guideline Review Committee granted approval, the pathway was then implemented. Implementation of the pathway was performed only after the care team involved in processing the patient on the pathway was educated. Several conferences and seminars were conducted for outpatient and inpatient health care teams working with the disease site working group. Physicians were also briefed on the pathway. Enrollment commenced after education of the health care workers was complete. Patients who were enrolled on the pathway between September 1, 1996, and August 31, 1998, were included in this study's analysis. The details of the neck dissection clinical pathway are demonstrated in Table 1.
Clinical and financial data were reviewed for each patient. Main outcome measures that were evaluated were length of hospital stay and total costs, including hospital and professional fees. Costs were also compared within categories as designated on the pathways based on type of resource use. These categories included surgery-related costs,treatment-related costs, medications, consultations, and assessment and diagnostic tests. Diagnostic tests were further broken down into pathology, laboratory test, and diagnostic imaging.
The hospital-related costs were calculated using the institution's cost-accounting software (Hospital Cost Consultants, Pleasanton, Calif. The costs for professional fees were calculated using average cost per procedure. Hospital and professional costs were then combined into one model that has been developed to set them constant over time. Charges reflect the amount that is billed to the consumer. Costs reflect the indirect and direct resources used by the hospital. Costs, rather than charges, were compared because costs were set constant over time. Thus, with a single established cost model, comparisons among groups can be made.
Data were analyzed using Minitab 11.12 (Minitab Inc, State College, Pa). Nonparametric tests were used because of the large numbers of outliers relative to the small sample size. Differences in median length of stay and total median costs among the 3 groups were tested for significance using the Kruskal-Wallis test. The Mann-Whitney test was used to test median differences separately between historical control group (prepathway group) and each of the contemporaneous groups (nonpathway and pathway).
The median ages on hospital admission for the historical control group and the contemporaneous nonpathway and pathway groups were 58 years, 59 years, and 60 years, respectively. The percentage of female and male patients for the historical control group was 24% (23 women) and 76% (73 men); nonpathway group, 36% (23 women) and 64% (41 men); and the pathway group, 27% (8 women) and 73% (22 men).
The median length of hospital stay for the historical control group was 4.0 days. Implementation of the clinical pathway significantly decreased the median length of stay to 2.0 days for the contemporaneous nonpathway and pathway groups (Figure 2; P<.001, Kruskal-Wallis). Total costs also fell significantly after implementation. Median total costs of the historical control group was $8459. After implementation of the clinical pathway, median total costs decreased to $6885 for the contemporaneous nonpathway group and to $6227 for the pathway group (Figure 3; P<.001, Kruskal-Wallis).
The largest decrease (−38%) in categories of costs was that of treatment costs, which included room/board and nursing costs. This decrease could be attributed to the decrease in the length of hospital stay. Other important decreases (−16% each) occurred in surgery-related costs and diagnostic test costs. There was no statistically significant difference between the contemporaneous nonpathway and pathway groups for costs or length of hospital stay.
Several methods have been used to decrease variance among physicians' practices and to decrease costs. Utilization review is one such method.2 This method, however, is performed from the perspective of the reviewers and thus may be very different from the perspective of physicans and the perspective of patients. Another disadvantage of utilization review is that it relies on claims data that are not based on the idea of properly allocating health care dollars. It also is not based on clinical outcomes.
Managing outcomes may be most effectively performed by the use of practice guidelines and clinical pathways. In the practice of surgical oncology, several studies have evaluated the effects of clinical pathways. Leibman et al3 reported on the impact of a clinical pathway for radical retropubic prostatectomy. Average hospital charges and average length of stay were lowered significantly with the implementation of the clinical pathway. Koch and Smith4 reported that their pathway resulted in a decrease in length of stay and charges for their patients undergoing radical prostatectomy as well. Morris et al5 determined that the use of a clinical pathway for total abdominal hysterectomy and oophorectomy with pelvic and para-aortic node sampling for endometrial cancer decreased costs and the length of stay. Incremental improvements were also measured during the planning stage of the clinical pathway.
The steps of development of a clinical pathway are crucial to the success of implementation as well as the successful enrollment of patients (Figure 1). At M. D. Anderson Cancer Center, the Practice Outcomes Program was implemented in 1996 to oversee the development and implementation of clinical pathways. This oversight assisted in continuity and consistency of the pathways that have been developed for use by the center.
We have demonstrated that the development and implementation of a clinical pathway has resulted in a decrease in median length of hospital stay and a decrease in median costs of treatment. Some limitations of our study exist, however. First, the number of pathway-managed patients is smaller than the other groups. This may lead to an increase in type II error in our study. Future studies are necessary that include more patients handled using the clinical pathway, so that we can increase the power of the study and decrease type II error. Second, the costs may have been trending downward before the implementation of the clinical pathway. Thus, the percentage of decrease in costs may be different than that documented in this article. This may also explain why the costs are not significantly different between the contemporaneous nonpathway and the pathway-managed patients. Third, selection bias may have affected our results although this bias was minimized by several methods. One way to minimize bias was choosing patients during 1993 and 1994 as the historical control group enrollees, thus assuring that any discussion regarding clinical pathways had not been formally begun. Another method was to use the same inclusion criteria for each group; ie, only patients who had unilateral neck dissection; unilateral neck dissection with DL and RE; or unilateral neck dissection with DL, RE, and dental extractions were included for costs analysis. This study is limited by the fact that the patients were not randomized to either the contemporaneous nonpathway group or the pathway-managed group, thus, selection bias may exist for these groups of patients.
Although it is difficult to distill the effects of the clinical pathway into direct and indirect effects, overall, the benefits have been worthwhile. Because cooperation and assistance of all aspects of patient care were sought, an improved atmosphere of camaraderie has developed and has resulted in improved efficiency and cost-minimization. We have demonstrated that the development and implementation of a clinical pathway has resulted in a decrease in median length of hospital stay as well as a decrease in median costs of treatment. Furthermore, the process of thinking about and developing clinical pathways benefits all patients, whether or not they were managed on the pathway. Although a decrease in costs and a decrease in length of hospital stay are indicators of efficiency, future research is necessary so that outcomes such as variance from the clinical pathway, complications of surgery, and patient satisfaction can be measured. Outcomes such as readmissions, deaths, complications of surgery, and patient satisfaction need to be measured as well so that the impact of clinical pathways on quality of life can be assessed.
Accepted for publication December 22, 1999.
Presented at the annual meeting of the American Head and Neck Society, Palm Desert, Calif, April 25, 1999.
Corresponding author: Amy Y. Chen, MD, MPH, University of Texas M. D. Anderson Cancer Center, Department of Head and Neck Surgery, 1515 Holcombe Blvd, Box 69, Houston, TX 77030.