Background

There is increasing pressure to perform traditional inpatient surgical procedures in an outpatient setting. The aim of the current trial was to determine the safety and cost savings of performing laparoscopic cholecystectomy in an outpatient setting using a "mock" outpatient setting.

Methods

Patients who were scheduled for laparoscopic cholecystectomy by four attending surgeons and for whom operating time was available in the outpatient center were studied. All patients received a standardized anesthetic, including ondansetron, and were discharged from the outpatient postanesthesia care unit if appropriate. At discharge, all patients were admitted to a clinical research center where they were observed in a "mock home" setting and monitored for complications that would have necessitated readmission. A decision analysis was created assuming all patients underwent outpatient surgery with either direct admission or discharge to home and readmission if complications developed.

Results

Of 99 patients who were enrolled in this study, 96 patients would have met the discharge criteria for home. No major complications were observed in these 96 patients. Eleven patients experienced postoperative nausea and vomiting, 3 of whom required an additional 24 h of hospital observation. In the decision model, the optimal strategy would be to perform the procedure on an outpatient basis and readmit patients only for complications, with an average baseline cost savings of $742/patient. Conclusions The results show that outpatient laparoscopic cholecystectomy is safe and cost-effective in selected patients, and that the mock home setting provides a means of studying the safety of transition of care. Key words: Ambulatory; complications; decision analysis; economics; nausea. RECENT trends in surgery reported the increasing incidence of traditional inpatient surgical procedures in the outpatient setting. It has been estimated that more than 60% of all surgical procedures in the United States will be performed on an outpatient basis by the year 2000. [1]The impetus for this development has been the desire to reduce healthcare expenditures, particularly by reducing hospital duration of stay. Frequently, the decision to perform a surgical procedure in the outpatient setting is based on anecdotal evidence and insurance company guidelines involving healthy patients. The evidence supporting the safety of ambulatory surgery is predominantly composed of case series of selected patients who would tolerate an outpatient procedure. However, the practice of immediate postoperative discharge (i.e., outpatient surgery) of selected patients frequently leads to generalization to higher-risk patients. As these changes in practice patterns are diffused, the safety of early discharge needs to be established. [2]For example, despite the widespread practice of outpatient tonsillectomy, there have been several recent reports that have suggested that there are subsets of patients in whom such a practice would lead to unacceptable outcomes. [3,4]There are currently no objective methods of prospectively determining the safety of early discharge before adoption of this approach for a given surgical procedure, particularly in patients with comorbidities. Laparoscopic cholecystectomy has largely replaced open cholecystectomy for the treatment of gallstone disease in the United States, and is an example of a procedure that is increasingly performed on an outpatient basis. [5]Importantly, a substantial portion of patients undergoing this procedure may require postoperative admission or readmission after discharge to home. [6-9]Many surgeons prefer longer observation periods, e.g., for 24 h or overnight, to observe for potential complications. Considering the potential increased morbidity of early discharge, a rigorous evaluation of the safety of this approach and appropriate patient selection is necessary. The literature supporting the routine practice of sending patients home within hours after laparoscopic cholecystectomy includes studies that support and refute the safety of this practice, many without rigorous patient selection or postoperative surveillance. [6,8-18]Specifically, with respect to laparoscopic cholecystectomy, observation for and treatment of postoperative nausea and vomiting (PONV) is critical because it represents one of the greatest potential obstacles to same-day discharge, with a reported incidence of 12-52%. [6,7]Therefore, one goal of any anesthetic regimen must be to reduce the incidence of PONV, which must be incorporated into the perioperative plan. The primary aim of this study was to develop a model to study the transition of surgical care from an inpatient to outpatient setting, using laparoscopic cholecystectomy as an example. We hypothesized that laparoscopic cholecystectomy could be performed on an outpatient basis, i.e., discharged to home on the day of surgery, with a low incidence of complications and that it would prove to be cost-effective from a provider perspective. To determine the safety of performing laparoscopic cholecystectomy on an outpatient basis before routinely adopting such a practice, we designed a "mock" outpatient setting for overnight observation after discharge from a outpatient surgery unit. We used admission to a clinical research center as a surrogate for home, because at the time of initiation of the study we believed that a medical "safety net" was necessary, and this scenario represented the least-intensive care setting. As a means of assessing the potential cost savings of an outpatient laparoscopic cholecystectomy, a decision-analysis model was developed. After obtaining approval from the Institutional Review Board, all patients older than 18 years scheduled for laparoscopic cholecystectomy at the Johns Hopkins Hospital Outpatient Center were evaluated for inclusion in this open-label observational trial between July 1994 and June 1996. The number of patients studied was determined by the predefined duration of the pilot study (2 yr) and was not related to a predetermined sample size. Four faculty surgeons experienced in laparoscopic cholecystectomy performed all procedures with resident assistant. The decision to perform the surgery on an inpatient or outpatient basis was determined by the available operating room time. If morning time was available in the Johns Hopkins Hospital Outpatient Center then surgery was performed on an outpatient basis. Time was designated for outpatient laparoscopic cholecystectomies at least twice a week, and additional time was available if other services did not use their designated time. Eligibility criteria included all patients with an American Society of Anesthesiologists classification of I, II, or III. Patients who had an ASA classification of IV were excluded from the study because the current standard of care would preclude them from undergoing laparoscopic cholecystectomy in an outpatient surgery setting. Informed written consent was obtained. ### Anesthetic Regimen All patients received 4 mg ondansetron intravenously before induction of anesthesia. General tracheal intubation was performed using propofol and mivacurium. The maintenance anesthetic consisted of a continuous propofol infusion, nitrous oxide (70%) and fentanyl (up to 250 [micro sign]g). Neuromuscular blockade was maintained with an infusion of mivacurium. No reversal of neuromuscular blockade was administered. Ketorolac, 15 or 30 mg intravenously, was given intraoperatively. Bupivacaine, 0.5%, was sprayed over the gallbladder bed, between the diaphragm and the liver, and infiltrated into operative wounds for a maximum of 20 ml. At the conclusion of the case, the surgeons subjectively categorized the difficulty of the procedure as "easy," "moderate," or "difficult." Postoperatively, the patients were admitted to the outpatient Post Anesthesia Care Unit (PACU) where they were monitored per established protocols. Assessments of pain and nausea-vomiting were performed by the nurses according to the standard of care in the PACU. Rescue antienemic therapy was administered at any time at the request of the patient or physician. The choice of the antiemetic was left to the discretion of the physician. The time from arrival to the PACU until the patient was able to achieve standard discharge criteria from the PACU using a postanesthesia discharge scoring system (i.e., able to sit, stand, ambulate, void, tolerate fluids) was noted. Patients were initially admitted to a phase I PACU and transferred to a phase II PACU when the modified Aldrete score was met. When discharge criteria were met, the patient was given discharge instructions in the PACU as if they were going home. If all of these criteria were met and the patient had no continuing medical problems, an attending anesthesiologist determined that the patients met criteria for discharge to home. If discharge criteria were met within 6 h, or at 6 h if they were not met, the patient was transported by wheelchair (through a connecting tunnel) and admitted to the Clinical Research Center (CRC) of The Johns Hopkins Hospital. Six hours is defined for all patients as the maximum duration of stay in the outpatient PACU before admission to the hospital. Patients whose surgery was converted from laparoscopic to open cholecystectomy or in whom major perioperative morbidity developed before discharge from the outpatient unit were admitted to a surgery inpatient floor and were excluded from the study. ### Postoperative Care (Clinical Research Center) For postoperative care, the patients were assigned to low- (ASA I) or high-risk groups (ASA II or III). For the low-risk group, vital signs were evaluated at arrival and before discharge. To maintain the medical safety net, the high-risk group had vital signs recorded every 4 h. The patients were given instructions regarding the need to notify medical personnel as if they had gone home (e.g., severe pain, nausea, vomiting, lightheadedness). Oral intake and activity were ad libitum. Any nursing interventions had to be initiated by the patient (i.e., request for pain medications). The patients were monitored for study endpoints that would indicate the need to use hospital resources (based on patient-initiated problems, and/or abnormal vital signs or laboratory tests), and recorded as the time initially reported. These included intractable pain necessitating intravenous medications and intractable nausea or vomiting, or both, requiring intravenous fluids or medications; respiratory distress; angina; arrhythmias; or hemodynamic instability. Patients were discharged to home on the morning of the first postoperative day (24-30 h) if they were adequately tolerating liquids and had stable vital signs, or they were admitted to an inpatient ward if complications developed. During their stay in the CRC, the nurse-to-patient ratio varied between 1:3 during the day to 1:9 during the night. Both the patient and the nursing staff of the CRC were queried at the patient discharge regarding whether the patient's postoperative course was sufficiently uncomplicated that the patient could have been discharged to home on the day of surgery. The patient also completed a survey at discharge to determine the factors influencing same-day discharge. Specific questions included, "If you were to go home yesterday after having surgery without staying overnight in the hospital, would you have had someone to help you at home, needed someone to help take care of you, been able to walk to the bathroom, walk up stairs and get food on you own?" ### Assessment of Charges A cost-identification analysis was performed from the perspective of the provider using charge data for calendar year 1995 from the case-mix files of The Johns Hopkins Hospital. Data were obtained for the subset of patients who were either enrolled in the prospective trial during 1995 (50 patients) or admitted on the day of surgery for a 24-h stay and underwent surgery performed by one of the four surgeons during 1995 (56 patients). Professional fees were not included in the analysis. The Medicare general inpatient routine service charge per diem was used to estimate the charges of an uncomplicated overnight admission. Hospital charges were obtained from the case-mix files of The Johns Hopkins Hospital. Total charges and charges by charge category (e.g., operating room, supplies, admission, and laboratory testing) were evaluated. For the purposes of developing a decision model, charges related to a theoretical home care and readmission were determined. Charges related to home health care were determined based on schedules of the Hopkins Home Health Care Group. Charges related to the emergency room were based on an average charge for an intermediate visit to the emergency room. ### Statistical Analysis and Cost-effectiveness Model Construction Demographics, intraoperative, and PACU times are expressed as the mean +/- standard deviation. Predictors of postoperative complications and willingness to be discharged to home was determined using the Student t test or Fisher exact test, where appropriate, with P < 0.05 considered statistically significant. Multiple logistic regression was used to determine independent predictors of the patient's belief that they were ready to go home on the day of surgery. A backward stepwise logistic regression analysis was performed using JMP v3.0 software (SAS Institute, Cary, NC) for the Macintosh (Apple Computers, Cupertino, CA). Variables were entered into the model based on the magnitude of their effect and statistical significance, with P < 0.25 used as statistical significant for this part of the process. Ability to walk to the bathroom, walk up stairs, and get food on one's own were entered into the model separately because they are not independent variables. Those variables that were independently and significantly (P < 0.05) related to readiness to go home on the day of surgery were determined. A model was constructed to determine cost savings using DATA 3.0 software (Treeage software, Boston, MA). Because there are substantial cost savings related to performing the procedure in an outpatient facility, a strategy using outpatient laparoscopic cholecystectomy with routine admission to an inpatient unit was compared to outpatient surgery with an emergency room or home healthcare visit only if complications developed. For each complication, it was assumed that one third would require an emergency room visit and the remaining patients could be treated at home with a home healthcare nursing visit. These estimates were derived from discussion with the attending surgeons and patients. Data for the model were based on the probability of complications and hospital charges, as described previously. Sensitivity analysis was performed to determine the influence of the probability of complications on the cost savings, as well as the influence of the rate of emergency room admission versus treatment at home. A total of 99 patients were enrolled in the study. Of all the eligible patients, only three declined to participate in the study. Of the 99 who were enrolled, two of the patient procedures were converted from laparoscopy to open cholecystectomy when good surgical exposure could not be obtained. These patients were excluded from the study. The laparoscopic procedure in a third patient required placement of a closed suction drain in the subhepatic space at the end of the procedure. This patient was not eligible for discharge to home on the day of surgery and was also excluded from the study. Demographics on the remaining 96 patients are shown in Table 1. All had an uneventful intraoperative course. The average anesthesia time was 1.4 +/- 0.3 h (range, 0.9-2.4 h). The average operating time was 1.1 +/- 0.3 h (range, 0.5-2.2 h). Of all the laparoscopic cholecystectomies, 54 of 96 (56%) were rated as easy, 33 of 96 (34%) were rated as moderate, and 9 of 96 (10%) were rated as difficult. Table 1. Demographics of Patients (N = 96) Who Underwent Outpatient Laparoscopic Cholecystectomy ### Postanesthesia Care Unit While in the outpatient PACU, none of the patients experienced any cardiovascular or pulmonary complications. Postoperative pain was managed with intravenous fentanyl, which was supplemented with intravenous morphine if the pain was severe. The average amount of fentanyl administered was 94 +/- 81 [micro sign]g (range, 0-400 [micro sign]g). Seven patients required morphine supplementation (range, 1-20 mg). Only 6% of the patients experienced nausea necessitating pharmacologic intervention (Compazine [SmithKline-Beecham, King of Prussia, PA] or metoclopramide), and there was no emesis in the PACU. All 96 patients met discharge-to-home criteria within 6 h and were therefore transferred to the CRC. The average time from arrival to the PACU until time ready to discharge was 3.1 +/- 1.2 h (range, 1.3-5.5 h). ### Clinical Research Center While in the CRC, none of the patients experienced any cardiovascular or pulmonary complications. Ninety of 96 patients (94%) did not require any additional pharmacologic intervention for postoperative pain other than routine oral medications (acetaminophen with oxycodone or meperidine). Six patients required intramuscular analgesics (morphine or meperidine), which would not have been available at home. Four of these six patients also had symptoms of PONV, necessitating pharmacologic therapy. Eleven (12%) of 96 patients experienced PONV necessitating intravenous or intramuscular pharmacologic therapy. The time between discharge from the PACU and onset of treatment is shown in Figure 1. Of the 11 patients with PONV requiring treatment, only 3 required an additional day in the hospital because of the inability to tolerate oral intake, whereas symptoms in the other patients resolved after a single treatment. One of the three patients who required a second day in the hospital complained of lightheadedness and demonstrated orthostatic hypotension, probably caused by decreased oral intake. The symptoms resolved with intravenous fluid hydration. A history of PONV was associated with an increased risk of PONV, although this trend did not reach statistical significance. Total fentanyl dosage in the PACU was significantly higher in the group in which PONV requiring treatment subsequently developed (130 +/- 23 [micro sign]g in the group with PONV versus 77 +/- 8 [micro sign]g in the no complications group; P < 0.05), although there was significant overlap in total fentanyl usage between the two groups. No other demographic factor predicted complications. Figure 1. Cumulative probability of being complication free after surgery. Episodes of postoperative nausea and vomiting requiring treatment occurred throughout the first 24-h period. Figure 1. Cumulative probability of being complication free after surgery. Episodes of postoperative nausea and vomiting requiring treatment occurred throughout the first 24-h period. A discharge survey completed by the nursing staff indicated that 77 of 96 patients (79%) using the current regimen could have been discharged to home on the day of surgery. Ninety of 93 patients (97%) who were discharged to home from the CRC completed the discharge survey. Three patients who were discharged to home did not complete a survey. Only 26 of 90 (29%) patients who responded stated that they were ready to be discharged to home on the day of surgery, the remainder of the patients stated they would have required an overnight admission (71%). The percentage of patients who stated they were ready to be discharge to home directly after surgery was significantly lower (29%) than the nurses thought were acceptable for discharge after surgery (79%, P < 0.001). Univariate predictors of readiness to go home on the day of surgery according to the patient are shown in Table 2. Independent predictors of readiness to go home on the day of surgery included age (odds ratio, 1.07/yr [1.01-1.22, 95% confidence interval]) and ability to get food on one's own (odds ratio, 3.24 [1.62-6.50, 95% confidence interval]). Table 2. Predictors of Readiness for Immediate Discharge after Surgery on Exit Interview ### Costs Analysis of charge data for outpatient and inpatient stays at The Johns Hopkins Hospital is shown in Table 3. Outpatient laparoscopic cholecystectomy was associated with an average$2,127 less compared to an inpatient procedure associated with a one-day hospital stay. Some of the lesser charges relate to a less-expensive charge/min of operating room usage, which reflects less-intensive resources necessary to maintain the outpatient center operating rooms. The operating room time was also shorter in the outpatient center because of a lesser extent of resident teaching in this setting. Charges are also less for supplies in the outpatient setting. Pharmacy charges reflect operating room nursing and ward charges, but do not include costs of anesthetic drugs.

Table 3. Average Charges Associated with the Subset of Laparoscopic Cholecystectomy Performed at the Johns Hopkins Hospital during 1995

### Decision Analysis

A decision model comparing routine overnight admission versus treatment for complications after outpatient surgery was constructed and is shown in Figure 2. The probabilities of being admitted after surgery, requiring readmission and a home healthcare visit were based on those probabilities observed in the current trial. Specifically, the probability of complications in the base-case scenario was set at 0.11 (11 of 96); the probability of direct admission for a surgical reason was 0.03, based on the two patients who required an open procedure and on the one patient who had a drain placed; and a probability of readmission was 0.03, based on the three patients who required an additional day of admission in the current trial. Charges associated with outpatient surgery were based on those obtained from the case-mix files of The Johns Hopkins Hospital during 1995. An emergency room visit that includes intravenous fluids and medication would be associated with a $350 charge, whereas a 1-h home healthcare visit that included a dose of ondansetron would be associated with a$145 charge. Considering the complication rate observed in the current trial, the optimal decision would be to perform the procedure on an outpatient basis, with an average baseline cost savings of $742. The probability of development of a complication of PONV varied from 0% to 50% in a one-way sensitivity analysis and is shown in Figure 3. Even if there was a 50% incidence of complications, outpatient laparoscopic cholecystectomy would be associated with an average$647 savings for each patient. Charge savings would increase if the percentage of patients who required treatment of PONV decreased.

Figure 2. A decision-analysis model to compare a strategy of outpatient surgery with routine overnight admission to a ward compared with outpatient surgery and discharge to home if patients meet established criteria. For the discharge-home strategy, the model assumes that two thirds of complications could be treated at home by a home healthcare nurse, whereas the remainder of the patients in whom a complication develops require an emergency room visit and possible admission (see text). The probabilities for each of the nodes are listed in the box, and are based on the rates observed in the current series. Using the baseline values listed in the tree (base-case scenario), the optimal strategy would be outpatient surgery with discharge to home. #Complimentary probability for a given node; i.e., 1 - probability of the complimentary node; cER = cost of an emergency room visit; CHHC = cost of a home healthcare visit; cInpatient = cost of an overnight admission based on the Medicare per diem rate; cOutpatient = cost of surgery on an outpatient basis; pAdmission = probability of direct admission after surgery; pAdmission2 = probability of readmission after discharge; pComplications = probability of development of complications at home.

Figure 2. A decision-analysis model to compare a strategy of outpatient surgery with routine overnight admission to a ward compared with outpatient surgery and discharge to home if patients meet established criteria. For the discharge-home strategy, the model assumes that two thirds of complications could be treated at home by a home healthcare nurse, whereas the remainder of the patients in whom a complication develops require an emergency room visit and possible admission (see text). The probabilities for each of the nodes are listed in the box, and are based on the rates observed in the current series. Using the baseline values listed in the tree (base-case scenario), the optimal strategy would be outpatient surgery with discharge to home. #Complimentary probability for a given node; i.e., 1 - probability of the complimentary node; cER = cost of an emergency room visit; CHHC = cost of a home healthcare visit; cInpatient = cost of an overnight admission based on the Medicare per diem rate; cOutpatient = cost of surgery on an outpatient basis; pAdmission = probability of direct admission after surgery; pAdmission2 = probability of readmission after discharge; pComplications = probability of development of complications at home.

Figure 3. A one-way sensitivity analysis to evaluate the effect of varying the complication rate from 0% to 50% on the potential savings from discharge to home immediately after outpatient laparoscopic cholecystectomy. The strategy of outpatient surgery with routine overnight admission would be associated with an average charge of $2,853. If there were no complications, an outpatient strategy would be associated with an average charge of$2,111, whereas a 50% complication rate would be associated with an average charge of $2,206. Therefore, outpatient surgery with immediate discharge would always provide substantial savings within the clinically relevant range using the assumptions based on the results of the current series. Figure 3. A one-way sensitivity analysis to evaluate the effect of varying the complication rate from 0% to 50% on the potential savings from discharge to home immediately after outpatient laparoscopic cholecystectomy. The strategy of outpatient surgery with routine overnight admission would be associated with an average charge of$2,853. If there were no complications, an outpatient strategy would be associated with an average charge of $2,111, whereas a 50% complication rate would be associated with an average charge of$2,206. Therefore, outpatient surgery with immediate discharge would always provide substantial savings within the clinically relevant range using the assumptions based on the results of the current series.

Our results show that outpatient laparoscopic cholecystectomy is safe and is associated with cost savings using protocols developed at our institution. We observed no major morbidity from early discharge in this series, with only minor complications observed within the first 24 h. The transfer from an inpatient to an outpatient procedure was associated with an average savings of $2,127 in total hospital charges, although the savings in hospital costs would be much smaller because actual medical costs represent only a fraction of charges. Using a decision-analysis model and assuming an efficiency and cost saving observed in an outpatient setting, outpatient laparoscopic cholecystectomy is cost-efficient compared to inpatient surgery, even assuming the rate of readmission suggested by the current study. Even hypothetically considering that all of our patients with PONV would require readmission, this outpatient strategy would result in cost savings from a payer perspective. The pressure to transition a particular surgery from an inpatient to an outpatient procedure is frequently based on reimbursement rather than medical issues. [19]Initially, highly selected patients are chosen for the outpatient procedure, and eventually the results are generalized to include patients with some degree of comorbidity. [2]The ability to evaluate such practices is dependent on the accuracy of follow-up after discharge, which may be poor or unreported. For example, outpatient mastectomy has been advocated as a safe and cost-effective procedure. [2]In a recent analysis of Medicare claims data, however, the incidence of readmission for complications, such as sepsis, urinary tract infection and pneumonia, were significantly higher in the women who underwent outpatient procedures than an overnight inpatient stay. The current trial is unique in that a mock outpatient environment was developed to evaluate the safety of routinely performing a procedure on an outpatient basis before fully adopting such practices at our institution. Such an approach could be used to study other surgical procedures before widespread transfer to an outpatient setting. With respect to outpatient laparoscopic cholecystectomy, there are several issues regarding safety and the readiness to discharge to home. Patients undergoing laparoscopic surgery may require conversion of surgery to an open procedure or may not meet discharge criteria. We showed a 3% incidence of direct admission to an inpatient ward (which is similar or lower than suggested by other reports using selected patients) after surgery related to surgical issues such as conversion to an open procedure and placement of a subhepatic drain. [16]Although all of our patients who underwent uncomplicated laparoscopic cholecystectomy met discharge criteria for home from the PACU, a large patient population may also include patients who do not meet additional discharge criteria. A larger population may have also demonstrated a higher rate of major complications after discharge to home, because the actual rate of complications may be as high as 3% based on the confidence intervals for the absence of complications in the current study. We had a diverse patient population and patients were not specifically selected for their ability to be discharged to home after surgery, as reported by our patient demographics. The acceptability of performing more-invasive surgery in an outpatient facility assumes that there is an acceptable rate of admission or readmission after ambulatory surgery, although this is currently debated. The direct and readmission rate must be accounted for in designing an optimal "critical pathway" for these patients. With respect to other surgeries, such as outpatient tonsillectomy, minimum times of observation in the PACU have been defined to detect the majority of early complications. [20]We were unable to define such a time period, with a continuous rate of complications occurring during the ensuing 24 h. Early discharge after laparoscopic cholecystectomy offers a safe option if the patient is willing to accept a defined rate of readmission for minor complications. Although the nurses in the CRC believed that the majority of the patients could have been discharged immediately after surgery, the patients expressed the opposite opinion. In attempting to understand the determinants of their response, the perception that they would need help at home and their ability to perform tasks were the best predictors of a feeling that they could have gone home immediately after surgery. Additionally, younger patients were less willing to go home immediately after surgery. Further research will be necessary to elucidate the other determinants of patient preferences and the potential role of preoperative education on patient satisfaction with outpatient laparoscopic cholecystectomy. Despite the low incidence of PONV in our study, it still remains a problem. All of the patients met discharge criteria from the PACU and would have been discharged to home. Yet in 11 patients problems with PONV developed in the ensuing 24-h period. All patients reported in their exit interviews that if they had been sent home, they would have had to seek medical attention. Although this medical attention would traditionally involve a visit to the emergency room, many of these patients may have remained at home if a home healthcare nursing visit was arranged. Three patients (3%) required hospital admission for an additional day for intractable nausea and vomiting necessitating intravenous fluids and antiemetic therapy. If these were truly outpatients, they would most likely have required hospital readmission. Although the overall incidence is low, it would depend on the individual practitioner and patient to determine whether the risk is acceptable. Because the majority of patients expressed a desire to remain overnight in the hospital, further education may be required to facilitate patient acceptance of outpatient laparoscopic cholecystectomy. Prevention of PONV in the immediate postoperative period is important, but it is imperative that PONV does not occur during the subsequent 24-h period if the goal is to avoid readmission or inpatient stays after laparoscopic cholecystectomy. Achieving this later goal is often the limiting factor. This problem is shown in a study by Michaloliakou et al. [21]in which patients were randomized to receive either preoperative multimodal analgesia (treatment group) or single-modality analgesia (control). Both groups received droperidol at induction. In the PACU, there was a significant decrease in the incidence of nausea in the treatment group (4.7%) versus the control group (29.7%). However, after the patients were on the ward, the incidence of nausea was high and similar to that of controls (treatment 23.8%, control 33.8%), as well as to previous reports. We designed the anesthetic to minimize the incidence of PONV, including the prophylactic use of ondansetron. The use of a total intravenous anesthetic, including ondansetron, would be associated with a greater overall cost than an inhalational anesthetic, but should still yield cost savings compared to an inpatient admission. Further research will be necessary to determine whether another, less costly, anesthetic regimen could achieve similar or superior results. The cost-effectiveness of outpatient compared to inpatient surgery has been well-established. There are several areas of cost savings by performing laparoscopic cholecystectomy in the outpatient setting. An overnight admission at our institution was associated with a charge of$720, which included an admission charge and a room charge. In our series of outpatients, there were no abnormal laboratory values that led to any change in treatment. Elimination of these postoperative laboratory tests would result in an incremental average saving of \$246. There are several differences in charges associated between performing the procedure in the inpatient versus the outpatient setting related to lower overhead and differences in efficiency costs between the two settings. Decision analysis is a methodology for evaluating the economic impact of the variability in admission rates, emergency room visits, readmission rates, and costs for the optimal strategy of care for a given condition. [22-24]Because ambulatory surgical center costs are usually lower than inpatients costs, it will always yield cost savings for the payers and providers to perform a procedure on an outpatient basis, assuming that the potential complications of such a practice are not severe. We therefore chose to use charges associated with an outpatient procedure for both the one-day stay and outpatient limbs of our model. In the base-case scenario, we included in our model an 11% complication rate observed in our clinical trial. In assessing charges for readmission we included the costs of an emergency room visit, a room fee, and an estimated charge for drugs and supplies. Considering that the majority of these problems (80%) are related to PONV without other associated complications, they could theoretically be resolved in the emergency room or by a home healthcare nurse. Alternatively, Carroll et al. [25]suggest that many of these episodes would have been self-limited and not have required treatment if the patient were home. The advantage of using decision analysis is that a sensitivity analysis can be performed, assuming lower rates of emergency room admission or readmission. In such situations, greater cost savings could be achieved.

The current study involved analysis of a selected cohort of patients based on available operating room time, which may have introduced substantial selection bias to our analysis. Additionally, the presence of nursing personnel in the CRC could have led to over-reporting of complications. However, an analysis of a subsequent cohort of 130 patients who underwent "true" outpatient surgery at our institution had a similar incidence of complications and readmissions to the mock setting we designed. [26]In this study, a total of eight patients (6.2%) were admitted to the hospital directly from the PACU, with six being discharged on the first postoperative day. After PACU discharge, an additional six patients (4.6%) required hospital admission. Although we attempted to detect major morbidity, which would have resulted in readmission to the hospital, we did not monitor patients for subclinical evidence of dysfunction. Laparoscopic cholecystectomy has been associated with marked decrements in pulmonary function. [14]We did not specifically test pulmonary function after laparoscopic cholecystectomy and did not observe any overt pulmonary problems. Our analysis is also limited, similar to other economic analyses, in that we used charges rather than actual costs for our economic analysis other than drug acquisition costs. Charges in Maryland hospitals are regulated, and the cost:charge ratio at The Johns Hopkins Hospital varies between 74-76% for the services provided. Therefore, all of the current estimates could be adjusted by this ratio to calculate "costs," although there are also problems inherent in the use of such ratios. Primarily, fixed and marginal costs should be calculated, as demonstrated by Macario et al. [27]Additionally, operating room costs in the outpatient setting are frequently less than those of the inpatient setting because of greater efficiency and less-intense resource use related to case mix.

Although outpatient laparoscopic surgery is feasible, appears to be safe, and dramatically reduces hospital charges, patients frequently prefer the security and comfort of a hospital. Our analysis did not include potential patient (society) costs of early discharge because all patients were admitted to the hospital overnight, nor did it include assessment of patient satisfaction. Indirectly, we assessed patient preferences by querying patients of their desire to stay in the hospital, but the strength of these desires was not quantified. Future research will be necessary to determine the additive value of these services in decreasing readmission and increasing patient satisfaction and the generalizability of using a "low-intensity care" setting in predicting outcome when patients are discharged directly to home.

The authors thank Delilah Salvana and Victoria Williams for their assistance with data collection and administration of the survey and the nurses in the outpatient PACU and inpatient CRC for their superb assistance.

1.
Kozak L, Hall M, Pokras R, Lawrence L: Ambulatory in the United States, 1994, Advance Data, 1997, National Center for Health Statistics, Centers for Disease Control, Washington, DC
2.
Warren JL, Riley GF, Potosky AL, Klabunde CN, Richter E, Ballard-Barbash R: Trends and outcomes of outpatient mastectomy in elderly women. J Natl Cancer Inst 1998; 90:833-40
3.
Schloss MD, Tan AK, Schloss B, Tewfik TL: Outpatient tonsillectomy and adenoidectomy: Complications and recommendations. Int J Pediatr Otorhinolaryngol 1994; 30:115-22
4.
Mitchell RB, Pereira KD, Friedman NR: Outpatient adenotonsillectomy. Is it safe in children younger than age 13? Arch Otolaryngol Head Neck Surg 1997; 123:681-3
5.
Legorreta AP, Silber JH, Constantino GN, Kobylinski RW, Zatz SL: Increased cholecystectomy rate after the introduction of laparoscopic cholecystectomy. JAMA 1993; 270:1429-32
6.
Arregui ME, Davis CJ, Arkush A, Nagan RF: In selected patients outpatient laparoscopic cholecystectomy is safe and significantly reduces hopitalization charges. Surg Laparosc Endosc 1991; 1:240-5
7.
Madsen MR, Jensen KE: Postoperative pain and nausea after laparoscopic cholecystectomy. Surg Laparosc Endosc 1992; 2:303-5
8.
Voitk AJ: Is outpatient cholecystectomy safe for the higher-risk elective patient? Surg Endosc 1997; 11:1147-9
9.
Voitk AJ: Establishing outpatient cholecystectomy as a hospital routine. Can J Surg 1997; 40:284-8
10.
Llorente J: Laparoscopic cholecystectomy in the ambulatory surgery setting. J Laparoendosc Surg 1992; 2:23-6
11.
Reddick EJ: Laparoscopic cholecystectomy in freestanding outpatient centers. J Laparoendosc Surg 1992; 2:65-7
12.
Farha GJ, Green BP, Beamer RL: Laparoscopic cholecystectomy in a freestanding outpatient surgery center. J Laparoendosc Surg 1994; 4:291-4
13.
Taylor E, Gaw F, Kennedy C: Outpatient laparoscopic cholecystectomy feasibility. J Laparoendosc Surg 1996; 6:73-7
14.
Saunders CJ, Leary BF, Wolfe BM: Is outpatient laparoscopic cholecystectomy wise? Surg Endosc 1995; 9:1263-8
15.
Lam D, Miranda R, Hom SJ: Laparoscopic cholecystectomy as an outpatient procedure. J Am Coll Surg 1997; 185:152-5
16.
Mjalan O, Raeder J, Aasboe V, Trondsen E, Buanes T: Outpatient laparoscopic cholecystectomy. Br J Surg 1997; 84:958-61
17.
Voitk AJ: Routine outpatient laparoscopic cholecystectomy. Can J Surg 1995; 38:262-5
18.
Narain PK, DeMaria EJ: Initial results of a prospective trial of outpatient laparoscopic cholecystectomy. Surg Endosc 1997; 11:1091-4
19.
Carithers JS, Gebhart DE, Williams JA: Postoperative risks of pediatric tonsilloadenoidectomy. Laryngoscope 1987; 97:422-9
20.
Rakover Y, Almog R, Rosen G: The risk of postoperative haemorrhage in tonsillectomy as an outpatient procedure in children. Int J Pediatr Otorhinolaryngol 1997; 41:29-36
21.
Michaloliakou C, Chung F, Sharma S: Preoperative multimodal analgesia facilitates recovery after laparoscopic cholecystectomy. Anesth Analg 1996; 82:44-51
22.
Kassirer JP, Moskowitz AJ, Lau J, Pauker SG: Decision analysis: A progress report (review). Ann Intern Med 1987; 106:275-91
23.
Pauker SG, Kassirer JP: Decision analysis (review). N Engl J Med 1987; 316:250-8
24.
Pauker SG, Pauker SP, Wong JB: The utility of thresholds and the threshold utility (editorial). Med Decis Making 1987; 7:201-2
25.
Carroll NV, Miederhoff PA, Cox FM, Hirsch JD: Costs incurred by outpatient surgical centers in managing postoperative nausea and vomiting. J Clin Anesth 1994; 6:364-9
26.
Lillemoe KD, Talamini MA, Yeo CJ, Snyder DS, Parker SD: Laparoscopic cholecystectomy as a "true" outpatient procedure: Initial experience in 130 consecutive patients. J Gastrointest Surg (in press)
27.
Macario A, Vitez TS, Dunn B, McDonald T: Where are the costs in perioperative care? Analysis of hospital costs and charges for inpatient surgical care. Anesthesiology 1995; 83:1138-44