Safely reducing length of stay after open radical retropubic prostatectomy under the guidance of a clinical care pathway
Collaborative care pathways have proven to be a safe and effective method of decreasing length of hospital stay (LOS) and costs after radical retropubic prostatectomy (RRP). In the current study, the authors evaluated the safety and efficacy of a pathway transitioning from a 3-day to a 2-day LOS.
The authors performed a retrospective chart review of 994 patients who underwent RRP at the study institution between July 1994 and December 2001. A total of 561 patients were managed on a 3-day LOS pathway, 172 were managed during the transition period from 3 to 2 days, and 261 were managed on a 2-day LOS pathway. Statistical analysis was performed comparing preoperative variables and complications among the three groups.
No statistically significant differences were found in comparisons of preoperative and demographic variables including age, race, medical comorbidities, preoperative prostate-specific antigen level, clinical stage of disease at presentation, and biopsy Gleason score. Forty-nine of the 561 patients (8.7%) remained longer than their targeted 3-day LOS, whereas 14 of 261 patients (5.4%) were hospitalized for longer than their targeted 2-day LOS, a difference that approached statistical significance (P = 0.058). During the transition period, 99 of 172 patients (57.6%) were discharged on postoperative Day 2 and 73 patients (42.4%) were discharged on postoperative Day 3. Complication rates were found to be significantly lower (P = 0.013) in the 2-day LOS group (2.3%) compared with the 3-day LOS group (7.0%) and the transition group (8.1%). The rate of readmission remained constant at 3% during this time period.
Overall, greater than 90% of patients were discharged within 2–3 days of surgery with acceptable complication rates, suggesting that in a consistent patient population, the transition from a 3-day LOS to a 2-day LOS can be done successfully without compromising patient safety. Evaluation currently is ongoing with regard to a 1-day LOS. Cancer 2005. © 2005 American Cancer Society.
The U.S. health care system has witnessed much change in the past decade with the increasing impact of managed care. Hospital financial pressures have resulted in the development of strategies directed toward decreasing costs without sacrificing the quality of patient care. The treatment of prostate carcinoma has been shown to be a significant economic burden on our health care system. Greater than 200,000 patients are diagnosed with prostate carcinoma each year, a significant number of whom will undergo radical prostatectomy. With rising numbers of patients and the continued employment of radical prostatectomy as a popular treatment option, various methods for decreasing incurred costs have been explored.1
Length of hospital stay (LOS) has been identified as an important factor influencing both hospital charges and patient costs for radical retropubic prostatectomy (RRP).1, 2 Several studies have shown that the introduction of a standardized collaborative care pathway can significantly reduce both LOS and costs without adversely affecting patient satisfaction.3–8
In 1994, Koch et al.9 reported on our institution's introduction of a standardized pathway targeted toward discharge on postoperative day (POD) 3. This report demonstrated a 44% reduction per patient in the LOS and overall hospital charges.9 Klein et al. reported similar results, noting a reduction in the LOS with a 43% reduction in the mean cost per patient.7 Experiences such as these led to the wide acceptance of clinical pathways limiting postprostatectomy hospitalization. Unfortunately, in the current economic climate of increasing costs and dwindling reimbursements, continued pressure has been applied to reduce costs even further.
In addition, with the advent of minimally invasive surgery, laparoscopic/robotic prostatectomy has become increasingly viable as an option and is often compared with open radical prostatectomy.10, 11 In response to many pressures and our own desire to improve care, we introduced changes into our standardized post-RRP pathway and evaluated the safety and efficacy of the modified pathway. In the current study, we reviewed a series of 994 patients and evaluated the safety and efficacy of our pathway, transitioning from a 3-day LOS to a 2-day LOS.
MATERIALS AND METHODS
We examined the records of 994 consecutive patients who underwent RRP at the study institution between January 1994 and December 2001. We collapsed these patients into three groups; the first group was managed with a target discharge day of POD 3, the second group was managed during a transitional period, and the third group was managed with a target discharge day of POD 2.
A total of 561 patients were managed with a clinical care pathway targeting discharge on POD 3. This pathway outlined patients arriving on the day of surgery after home-administered preoperative mechanical bowel preparation. After surgery, patients were transferred from the postanesthesia unit directly to the regular urology floor. Ambulation was initiated on the evening of surgery. Postoperative pain was managed with scheduled intravenous ketorolac and parenteral narcotics via patient-controlled anesthesia (PCA). A full liquid diet was given on POD 1 and patients were advanced to a regular diet on POD 2. Bisacodyl suppositories frequently were administered on PODs 1 and/or 2 to encourage the return of normal bowel function. A single closed drain was placed at the time of surgery and generally was removed on the day of discharge. Each patient received individual care instructions from a clinical nurse specialist and was discharged home when they were able to tolerate a regular diet and oral analgesics without difficulty. There were no routine home health care arrangements included in the pathway. If home health care was needed, it was arranged on an individual basis.
A total of 172 patients were managed during a transitional period during which discharge on POD 2 or POD 3 was deemed acceptable.
Two hundred sixty-one patients were managed on a pathway with targeted discharge on POD 2. The plan of care was similar to that of the other groups; however, a full liquid diet was begun in the afternoon of POD 1 and was advanced to a regular diet that evening. Patients were not required to demonstrate full recovery of bowel function prior to discharge. Discharge instructions included a bowel regimen of milk of magnesia and glycerin suppositories to be administered as needed.
The charts of all the patients were reviewed retrospectively and preoperative variables, transfusion and complication rates, and LOS were recorded. Preoperative characteristics included patient age, American Society of Anesthesiologists (ASA) score, prostate-specific antigen (PSA) level at presentation, clinical stage of disease at presentation, biopsy Gleason score, and the existence of comorbid conditions. Postoperative complications were recorded if they occurred within a 30-day postoperative follow-up period. Major complications were defined as myocardial infarction, cerebrovascular accident, pulmonary embolus, sepsis, respiratory failure, or return to the operating room. LOS was recorded and those patients who did not meet their targeted discharge day were identified. Reason(s) for increased hospital stay were recorded. All data were available in 604 patients.
The three data sets were analyzed using commercially available statistical software. Predictor variables were collapsed into binary (true/false) format and included comorbid condition(s), major complications, minor complications, and complications after hospital discharge. Univariate analysis was performed comparing the preoperative patient demographic and clinical values among the three groups. Missing values were not replaced. A chi square test was performed for categoric values and the Student t test was performed for continuous variables. The correlations among variables were examined using the chi-square test. Variables found to be significant (P ≤ 0.05) then were included in a logistic regression equation to determine the relative contribution of variables for predicting increased LOS.
Data were available on 994 patients ages 39–78 years (mean age, 60.7 years) who underwent RRP between January 1994 and December 2001. Patient demographics and preoperative clinical information are summarized in Table 1. Overall, the mean ASA score was 2.26 (median score, 2) and the median preoperative PSA level was 7.5 ng/mL (range, 0.6–76.8 ng/mL). At the time of initial presentation, 52% (n = 315) of the patients had clinical T1c disease, 38% (n = 231) had clinical T2 disease, and 10% (n = 58) had T3 disease. Of the sample of 604 patients with all available data, 433 (72%) were otherwise healthy and reported no preoperative comorbid conditions, 91 patients (15%) had only 1 comorbid condition, and the remaining 80 patients (13%) reported > 1 preexisting condition.
Table 1. Patient Characteristics
|Mean age (yrs) (range)||61.6 (39–78)|
|Median ASA category||2|
|Median PSA level (ng/mL) (range)||7.5 (0.6–76.8)|
|Clinical disease classification (no. of patients)|| |
| T1c||315 (52%)|
| T2||231 (38%)|
| T3||58 (10%)|
|No. of comorbidities (no. of patients)|| |
| 0||433 (72%)|
| 1||91 (15%)|
| >1||80 (13%)|
|Median pre-operative HCT (range)||43 (33–55)|
|Median operative time (mins) (range)||190 (98–408)|
|Median EBL (mL) (range)||400 (50–3200)|
|Median post-operative HCT (range)||34 (26–50)|
Various major complications developed in 4 patients (0.66%) (perforated duodenal ulcer, pulmonary embolus, retained drain, and a cardiovascular event in 1 patient each). There were 40 minor complications reported (6.6% of the total cohort without data), with postoperative ileus reported to occur most often, accounting for 18 of these 40 reported complications (45%). Other minor complications included postoperative fever, wound infection, persistent JP output, and hematuria (Table 2). The median LOS was 3 days (range, 2–22 days). Of the study population with 3-day target, 512 of 561 patients met the targeted hospital discharge of POD 3. Of the 49 patients with an increased LOS (8%), 36 patients (73.5%) remained hospitalized only 1 day longer than expected. It is important to note that the 30-day readmission rate during this time period remained constant at 3%. Similarly, 247 of 261 patients (94.6%) met discharge goal of POD 2.
Table 2. Complications
|Major||4 (0.66% overall)|
| Wound infection||4|
| Persistent JP output||3|
| Total minor||40 (6.6% overall)|
Univariate analysis using the chi-square and Student t test to compare the demographic and clinical variables of the two datasets (target discharge vs. increased LOS) was performed. This analysis indicated no statistically significant difference with regard to LOS based on age, ASA score, presenting PSA level, clinical stage of disease, comorbid condition(s), surgical time, estimated blood loss, or preoperative or postoperative HCT. The only variables found to affect the LOS significantly were those that presumably caused the deviation from the pathway: the occurrence of a complication or the existence of a social problem with discharge. On further chart review, “social” reasons were found to be either logistic (e.g., no ride, a lack of assistance at home, etc.) or emotional (e.g., “not ready”). This finding will be discussed in greater detail later in this article.
As mentioned previously, there were 4 major complications and 40 minor complications reported. It is interesting to note that all patients who developed a major complication also developed a minor complication, resulting in a total of 40 patients with complications. All these patients deviated from the pathway and stayed longer than the predicted 2–3 days, an occurrence that was statistically significant (P < 0.001). In addition, the nine patients with social problems with discharge also all deviated from the targeted discharge date (P < 0.001). When a statistical analysis was performed using the variables already mentioned, there were no identifiable risk factors that determined that certain patients were more likely to develop a complication and/or have social problems with discharge.
One of the current challenges in health care is to develop effective methods of cost containment while continuing to improve both the quality of patient care and patient satisfaction. In the current study, we present long-term data regarding a large cohort of patients supporting the safety and efficacy of our post-RRP collaborative care pathway. The purpose of our investigation was to evaluate the long-term effectiveness of our postprostatectomy pathway in meeting the targeted LOS and to attempt to identify any controllable risk factors contributing to failure to meet this target. The results of the current study reveal that greater than 90% of our patients were successfully discharged on or before their target dates while demonstrating no increased incidence of complications.
In 1993, Litwin et al. performed a charge-based analysis of patients undergoing radical prostatectomy at two large teaching hospitals and found that LOS was the primary determinant of higher cost for both healthy and sick patients.5 With increasing pressures from hospitals and insurance companies to cut costs, information such as this led to the development of patient care pathways designed to decrease and standardize LOS. In 1994, Koch et al. reported on our institution's prospective development of a cost-efficient collaborative care pathway for prostatectomy patients. In an initial experience with a small number of patients during the implementation of the program, they reported a reduction in the mean LOS from 5.7 days to 3.6 days and a total cost reduction of 44% without any apparent increase in the incidence of complications.9
Despite the diversity of the current study cohort, 92% of our postprostatectomy patients were discharged by the target day of POD 3 and, of those who did not meet the targeted date, 73.5% (36 of 49 patients) stayed only 1 day longer than expected. The reason(s) for the increased LOS included complication(s) in 40 patients and social reasons in 9 patients.
The occurrence of a complication was a statistically significant predictive variable for increased LOS. Our overall perioperative complication rate was quite low (0.66% for major complications and 6.6% for minor complications), and is comparable to that published by other high-volume centers.12–14
Paralytic ileus was responsible for 45% of the complication-related increases in LOS. In an attempt to address this occurrence, we limited the amount of intravenous narcotics used by each patient by discontinuing the routine postoperative use of PCA machines and making intravenous pain medications available on an as-needed basis. However, patient satisfaction surveys after this change indicated a decrease in satisfaction with pain control and, as a result, the use of postoperative PCA was reinstituted.
An interesting and potentially controllable variable believed to be responsible for an increase in LOS was a “social” difficulty in nine patients. On further chart review, “social” reasons were found to be either logistic (e.g., no ride, lack of assistance at home, etc.) or emotional (e.g., “not ready”). It was believed that these variables could be controlled easily by intensifying preoperative counseling, identifying any need for and arranging home assistance preoperatively, and increasing the amount of one-on-one postoperative teaching time.
There are several important caveats of the current study that bear mentioning. Clearly, multiple other variables that we did not examine may impact hospitalization, such as surgical history, previous neoadjuvant therapy, patient body habitus, patient socioeconomic status, and others. We were not able to assess all variables due to the nature of this study. In addition, the pathway currently in use at the study institution, which is a high-volume center, may not be applicable in all clinical scenarios. Assuming that performing a significant number of these procedures regularly is the key to controlling the LOS, however, underestimates the impact of the multidisciplinary approach, which requires careful cooperation, resources, and a constant monitoring with the ability to adapt. Preoperative written and verbal counseling, regular follow-up phone conversations, prospective data collection and analysis, and intensive in-patient teaching, among other endeavors, require manpower. In return, however, are continued patient satisfaction and safe and efficient care.
However, the results of the current study do offer a point of comparison for patients undergoing robotic or laparoscopic prostatectomy. With the intensification of preoperative readiness and postoperative teaching and the success of these modifications, we have continued to modify and refine our care guidelines and currently are prospectively evaluating the safety and efficacy of a pathway targeting a 1-night hospital stay.
In the current study, we presented what to our knowledge is one of the largest series published to date evaluating the efficiency and safety of a collaborative care pathway for RRP patients over an extended period of time. The validity of altering our pathway was demonstrated through greater than 90% of the study patients meeting the targeted day of discharge with acceptably low rates of complication(s). The occurrence of major and/or minor complications was found to correlate with delayed discharge, with the occurrence of postoperative ileus being the most frequent reason for increased LOS. No preoperative or intraoperative variables were identified that appeared to impact the LOS significantly or accurately predicted the occurrence of a complication. The success of our postprostatectomy clinical pathway has led us to explore ways to reduce our LOS further to 1 day.