Original Article
Predicting Inpatient Stay Lasting 2 Midnights or Longer After Robotic Surgery for Endometrial Cancer Margaret I. Liang, MD, Maggie A. Rosen, MD, Kellie S. Rath, MD, Erinn M. Hade, PhD, Aine E. Clements, MD, Floor J. Backes, MD, Eric L. Eisenhauer, MD, Ritu Salani, MD, David M. O’Malley, MD, Jeffrey M. Fowler, MD, and David E. Cohn, MD* From the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology (Liang, Rosen, Rath, Clements, Backes, Eisenhauer, Salani, O’Malley, Fowler, Cohn), and Center for Biostatistics (Hade), The Ohio State University College of Medicine, Columbus, Ohio.
ABSTRACT Objective: To estimate the rate of inpatient stay and the factors predicting inpatient status after robotic surgery for endometrial cancer following the change in the Medicare definition of ‘‘inpatient’’ to include hospitalization spanning 2 midnights. Design: Retrospective chart review (Canadian Task Force classification II-1). Setting: Academic hospital. Patients: All patients (n 5 395) with endometrial cancer who underwent robotic surgical management between 2006 and 2010. Intervention: The outpatient stay group with hospitalization spanning 1 midnight was compared with the inpatient stay group with hospitalization spanning 2 midnights or longer through estimation of the adjusted relative risk (aRR) for various characteristics of interest. Results: Ninety-six of 395 patients (24.3%) stayed at least 2 midnights and thus were deemed inpatients. Clinical factors associated with inpatient stay were increasing age, history of myocardial infarction (aRR, 2.0; 95% confidence interval [CI], 1.0–3.7), surgery start time at or after 12 noon (aRR, 1.7; 95% CI, 1.2–2.4), perioperative blood transfusion (aRR, 3.2; 95% CI, 2.3–4.5), and surgery performed in the year 2010 (aRR, 0.5; 95% CI, 0.3–0.7). Age R60 years was associated with at least a 2-fold adjusted risk of prolonged hospitalization. Body mass index, other medical comorbidities, operative duration, estimated blood loss, and performance of lymphadenectomy or additional surgical procedures were not identified as significant risk factors. Conclusion: Approximately 75% of the patients undergoing robotic surgery for endometrial cancer were discharged as outpatients. Recognition of factors predicting inpatient stay can improve hospital resource allocation and throughput in women undergoing robotic surgery for endometrial cancer. Journal of Minimally Invasive Gynecology (2015) 22, 583–589 Ó 2015 AAGL. All rights reserved. Keywords:
DISCUSS
Endometrial cancer; Inpatient stay; Prolonged hospitalization; Robotic surgery
You can discuss this article with its authors and with other AAGL members at http://www.AAGL.org/jmig-22-3-JMIG-D-14-00586
Use your Smartphone to scan this QR code and connect to the discussion forum for this article now* * Download a free QR Code scanner by searching for ‘‘QR scanner’’ in your smartphone’s app store or app marketplace.
The authors declare no conflicts of interest. This paper was presented as an oral presentation at the Western Association of Gynecologic Oncologists Annual Meeting, Seattle, Washington, June 2013. Corresponding author: David E. Cohn, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, The Ohio State University College of Medicine, M210 Starling Loving, 320 West 10th Ave, Columbus, OH 43210. E-mail: [email protected] Submitted November 25, 2014. Accepted for publication December 30, 2014. Available at www.sciencedirect.com and www.jmig.org 1553-4650/$ - see front matter Ó 2015 AAGL. All rights reserved. http://dx.doi.org/10.1016/j.jmig.2014.12.167
Increasing attention is being given to improving the delivery of cost-effective health care. In the field of gynecologic oncology, robotic surgery has become an accepted and popular approach for the surgical management of endometrial cancer [1,2]. It is estimated that more than 60% of minimally invasive hysterectomies performed for endometrial cancer were completed robotically in 2010 [3]. Shorter hospital stay is among the many advantages of robotic surgery compared with laparotomy [4–6]. The high technology costs associated with the robotic platform remain a barrier to its widespread use, however [1].
584
As with any new scientific advancement, capital costs likely will decrease over time. Minimizing the length of hospital stay after robotic surgery has the potential to help offset the overall higher supply costs. Several studies have compared the costs associated with various surgical approaches for the management of endometrial cancer, including laparotomy, laparoscopy, and robotic surgery [3–6]. The majority of the cost of laparotomy has been attributed to longer average hospital stay, whereas that of robotic surgery is related to disposable equipment. For all minimally invasive surgical techniques, shorter hospital stays offered significant cost savings compared with laparotomy [4]. Recent statutory changes implemented to satisfy provisions contained in the Patient Protection and Affordable Care Act and the Health Care and Education Reconciliation Act of 2010 (collectively known as the Affordable Care Act) will have an effect on the care of robotic surgery patients. On August 19, 2013, the Centers for Medicare & Medicaid Services (CMS) revised the Medicare hospital inpatient prospective payment systems (IPPS) applying to the designation of appropriate level of care and subsequent costs in acute care hospitals [7]. As part of the IPPS revision, hospital discharges occurring on or after October 1, 2013, are classified as inpatient if the hospital stay crosses 2 midnights (known as the ‘‘two-midnight rule’’) and thus are paid under Medicare Part A. In contrast, if the stay is unanticipated to or does not span 2 midnights, then the hospitalization is classified as outpatient, with appropriate payment under Medicare Part B [7]. The rationale for this IPPS revision includes the fact that outpatient and/or observation hospital encounters were not being appropriately converted to inpatient status, leading to higher out-of-pocket expenses for patients. With the more clearly defined classification for inpatient stay, hospitals also will have a further incentive to expedite safe medical discharges owing to differences in reimbursement. The CMS estimated that under the new guidelines for determining inpatient status, there would be a net shift of 40,000 additional inpatient admissions in fiscal year 2014, at a cost of $220 million [7]. With these changes, identifying patients at increased risk for hospitalization lasting 2 midnights or longer after robotic surgery for endometrial cancer is imperative for hospitals to achieve optimal resource allocation. Although the relevance of the ‘‘two-midnight rule’’ to other health systems is uncertain, it is likely that utilization of healthcare resources will continue to be an important subject internationally as well. Therefore, in response to the new CMS revision of the IPPS using the ‘‘two-midnight rule’’ to define inpatient status, we set out to evaluate the factors associated with a hospital stay spanning at least 2 midnights following robotic surgery for endometrial cancer. Materials and Methods The Ohio State University Institutional Review Board approved this study. All patients who underwent robotic sur-
Journal of Minimally Invasive Gynecology, Vol 22, No 4, May/June 2015
gical management of endometrial cancer between January 1, 2006, and December 31, 2010, at The Ohio State University Wexner Medical Center in Columbus, Ohio were identified. All patients underwent hysterectomy and bilateral salpingooophorectomy, with lymph node staging performed in 86% of cases based on the surgeon’s clinical judgment. Five robotically trained gynecologic oncologists at Ohio State were included in the analysis. Exclusion criteria included conversion to laparotomy or mini-laparotomy for delivery of the specimen, so that only patients who underwent robotically completed surgery were included in the analysis. Hospital and office charts were reviewed for patient characteristics, including age, height, weight, body mass index (BMI), and presence of prespecified medical comorbidities. The major comorbidities examined were hypertension, diabetes, hyperlipidemia, congestive heart failure, history of myocardial infarction, history of stroke, chronic renal disease, asthma, and obstructive sleep apnea. Operative data, including surgical start time, operative duration, estimated blood loss, perioperative blood transfusion (including transfusions administered during and after surgery), performance of pelvic lymphadenectomy with or without para-aortic lymphadenectomy, performance of other surgical procedures, and intraoperative complications, were collected. Procedure start times were stratified as before 12 noon or at 12 noon and after. Operative duration was defined as the time of skin incision to incision closure. Procedures performed concurrently included ureterolysis, enterolysis, omentectomy, hernia repair, cholecystectomy, and cystoscopy. Intraoperative complications were categorized as vessel injury, genitourinary injury, gastrointestinal injury, or vaginal laceration. Information on postoperative complications was gathered, including postoperative morbidity and readmission within 90 days. The postoperative morbidities were classified as gastrointestinal, genitourinary, infectious, cardiopulmonary, thromboembolism, and wound complications. Standard postoperative discharge criteria were used. Before discharge, patients were expected to be tolerating a regular diet, achieving adequate pain control with a combination of oral narcotics and nonsteroidal anti-inflammatory drugs, ambulating without difficulty, and voiding spontaneously with adequate urine output. Although the ‘‘twomidnight rule’’ was not in effect during study years 2006 to 2010, we applied the definition to our cohort of robotic surgical patients based on length of stay. As a result, patients who were discharged following a hospitalization spanning a single midnight (i.e., discharged on postoperative day 1) were designated as the outpatient stay group, and those discharged after a hospitalization spanning at least 2 midnights (i.e., discharged on postoperative day 2 or later) were classified as the inpatient stay group. Demographic and clinical preoperative and operative characteristics were summarized through descriptive statistics. Categorization of these variables was determined by clinical relevance, historically used cutpoints, and natural breaks in the data owing to collection methodology. For
Liang et al.
Predicting Inpatient Stay After Robotic Surgery for Endometrial Cancer
instance, the majority of measurements for estimated blood loss were recorded in 50-mL intervals. Crude relative risk (RR) and adjusted RR (aRR), along with the associated 95% confidence intervals (CIs), of inpatient stay for these factors were estimated via a modified Poisson regression approach [8]. The development of a multivariable model to determine predictive factors of inpatient stay proceeded in 2 stages. First, a model that included clinically important preoperative factors was established. Then operative factors of interest were added to the model, and the model was refined based on the relative information provided by the addition of each variable. Potential interactions between history of myocardial infarction and either age or transfusion status were explored and found to be noninfluential. All reported CIs are 2-sided. Analyses were performed with Stata version 13 (StataCorp, College Station, TX). Results A total of 395 patients met the inclusion criteria for this study (Table 1). The outpatient stay group comprised 299 patients (75.7%), and the inpatient stay group comprised 96 patients (24.3%). The average length of stay was 1.5 6 2.4 days, with a median length of stay of 1 day (range, 1– 46 days) across all patients. One patient with a prolonged hospitalization of 46 days underwent reoperation on postoperative day 2 for ileal perforation, requiring ileal resection with gastrostomy tube placement. This was further complicated by prolonged respiratory failure, bacteremia, and postoperative ileus. This patient was ultimately discharged to a rehabilitation facility and experienced complete recovery. In the inpatient stay group, the average length of stay was 2.9 6 4.6 days, with median length of stay of 2 days (range, 2–46 days). Patient demographic data and clinical preoperative and operative characteristics are summarized in Table 1. In univariable analyses, an increased risk of inpatient stay was associated with older age, history of renal disease, history of myocardial infarction, surgery starting at or after noon, need for perioperative transfusion, and surgery performed during earlier study years (2006 to 2009 vs 2010) (Table 1). Although the estimated unadjusted risk of inpatient stay was not associated with operative duration or estimated blood loss, patients with very long surgeries (R5 hours) or greater blood loss (R150 mL) were found to be at increased risk of longer hospitalization. The overall rate of intraoperative complications was low (2%), and the occurrence of complications was not associated with inpatient stay (Table 1). Adjusted multivariable models identified the demographic and preoperative characteristics of age, history of myocardial infarction, timing of surgery, and year of surgery as important predictors of inpatient stay when included together (Table 2). Patients aged R60 years had at least a twofold increased risk of inpatient stay on average compared
585
with patients aged ,55 years (age 60–64: aRR, 2.7; 95% CI, 1.5–4.8; age R65: aRR, 2.1; 95% CI, 1.2–3.7). Increased risk was also consistently found in patients with a history of myocardial infarction (aRR, 1.9; 95% CI, 1.0–3.5) and those whose surgery began at or after 12 noon (aRR, 1.6; 95% CI, 1.2–2.3). When further considering operative characteristics in this model, perioperative transfusion was found to be the most influential factor of all operative variables in predicting inpatient stay (aRR, 3.2; 95% CI, 2.3–4.5). Inclusion of this covariate with those previously included preoperative covariates had little effect on the previously estimated effects. Patients who experienced any postoperative complication (overall rate, 18.2%) were more likely to be hospitalized for at least 2 midnights and thus be discharged as an inpatient (RR, 1.9; 95% CI, 1.4–2.8). The 2 most common postoperative complications in the inpatient stay group, ileus and postoperative fever, were also independently associated with inpatient status (ileus: RR, 3.7; 95% CI, 2.6–5.3; postoperative fever: RR, 3.0; 95% CI, 1.9–4.7). Patients who required inpatient hospitalization after robotic surgery for endometrial cancer were more likely to be readmitted within 90 days of surgery (RR, 2.1; 95% CI, 1.0–4.2). A total of 30 patients, including 18 (6.0%) from the outpatient group and 12 (12.5%) from the inpatient group, were readmitted within 90 days. Adjustment for age and year of surgery attenuated this effect downward (aRR, 1.7; 95% CI, 0.9–3.4). Discussion Following robotic surgery for endometrial cancer, hospital stays for the majority of patients (approximately 75%) span a single midnight and thus are classified by the revised Medicare hospital IPPS as outpatient hospitalizations. Increasing age, history of myocardial infarction, surgery start time at or after 12 noon, and perioperative blood transfusion were independently associated with stay lasting 2 midnights or longer. Elucidating factors that predict inpatient status following robotic surgery for endometrial cancer is of interest to patients, their caregivers, providers, and hospital systems. Preoperative identification of patients with risk factors for inpatient stay (such as older age or history of myocardial infarction) can help surgeons set appropriate expectations regarding discharge date and course. Alternatively, these data are important to aid in counseling obese patients and patients with multiple comorbidities that the expected postoperative stay following robotic surgery for endometrial cancer is typically similar to that of patients who are of normal weight and with no comorbid conditions. The recognition that operative factors may place a patient at risk for inpatient stay is also important for planning surgery. Surgeries at the extreme of operative duration (R5 hours) or with greater blood loss (R150 mL) were associated with longer hospital stays. Although increasing estimated blood loss was not identified as a very strong predictor, the need for perioperative blood transfusion, a more
586
Journal of Minimally Invasive Gynecology, Vol 22, No 4, May/June 2015
Table 1 Potential clinical factors associated with inpatient stay, defined as length of stay spanning 2 midnights or longer, after robotic surgery for endometrial cancer Clinical factor Preoperative characteristics Age, yr, mean 6 SD Age ,55 yr, n (%) Age 55–59 yr, n (%) Age 60–64 yr, n (%) Age R65 yr, n (%) BMI, kg/m2, mean 6 SD Normal weight (18.5–24.9), n (%) Overweight (25–29.9), n (%) Obese (R30), n (%) Medical comorbidities, n (%) 0 1 R2 History of renal disease, n (%) No Yes History of myocardial infarction, n (%) No Yes Timing of surgery, n (%) Start before 12 noon Start at or after 12 noon Year of surgery, n (%) 2006–2009 2010 Operative characteristics Operative duration, hr, mean 6 SD Duration of ,3 hr, n (%) Duration of 3 hr to 3 hr and 59 min, n (%) Duration of 4 hr to 4 hr and 59 min, n (%) Duration of R5 hr, n (%) Estimated blood loss, mL, median (range) Loss of 0–49 mL, n (%) Loss of 50–99 mL, n (%) Loss of 100–149 mL, n (%) Loss of R150 mL, n (%) Perioperative transfusion, n (%) No Yes Pelvic lymphadenectomy, n (%) No Yes Para-aortic lymphadenectomy, n (%) No Yes Additional surgical procedure, n (%) No Yes Intraoperative complication, n (%) No Yes Postoperative characteristics
Outpatient stay (n 5 299)
Inpatient stay (n 5 96)
59 6 11 88 (88) 58 (76) 55 (66) 98 (72) 35 6 10 39 (76) 55 (67) 205 (78)
64 6 10 12 (12) 18 (24) 28 (34) 38 (28) 35 6 10 12 (24) 27 (33) 57 (22)
1.0 1.4 (0.8–2.5) 0.9 (0.5–1.6)
120 (79) 99 (73) 80 (74)
31 (21) 37 (27) 28 (26)
1.0 1.3 (0.9–2.0) 1.3 (0.8–2.0)
297 (76) 2 (40)
93 (24) 3 (60)
1.0 2.5 (1.2–5.3)
297 (76) 2 (33)
92 (24) 4 (67)
1.0 2.8 (1.6–5.1)
185 (81) 114 (68)
43 (19) 53 (32)
1.0 1.7 (1.2–2.4)
139 (67) 160 (86)
69 (33) 27 (14)
1.0 0.4 (0.3–0.6)
200 6 43 99 (79) 149 (78) 45 (71) 6 (35) 100 (10–1400) 61 (80) 87 (77) 80 (80) 71 (67)
218 6 57 26 (21) 41 (22) 18 (29) 11 (65) 100 (10–500) 15 (20) 26 (23) 20 (20) 35 (33)
297 (77) 2 (20)
88 (23) 8 (80)
1.0 3.5 (2.4–5.0)
43 (80) 256 (75)
11 (20) 85 (25)
1.0 1.2 (0.7–2.1)
98 (76) 201 (76)
31 (24) 65 (24)
1.0 1.0 (0.7–1.5)
271 (77) 28 (67)
82 (23) 14 (33)
1.0 1.4 (0.9–2.3)
292 (75) 7 (88)
95 (25) 1 (13)
1.0 0.5 (0.1–3.2)
RR (95% CI)
1.0 2.0 (1.0–3.9) 2.8 (1.5–5.2) 2.3 (1.3–4.2)
1.0 1.0 (0.7–1.6) 1.4 (0.8–2.5) 3.1 (1.9–5.1) 1.0 1.2 (0.7–2.1) 1.0 (0.6–1.8) 1.7 (1.0–2.8)
(Continued )
Liang et al.
Predicting Inpatient Stay After Robotic Surgery for Endometrial Cancer
587
Table 1 Continued Clinical factor Postoperative complication, n (%) No Yes Ileus, n (%) No Yes Postoperative fever, n (%) No Yes
Outpatient stay (n 5 299)
Inpatient stay (n 5 96)
RR (95% CI)
256 (79) 43 (60)
67 (21) 29 (40)
1.0 1.9 (1.4–2.8)
298 (77) 1 (14)
90 (23) 6 (86)
1.0 3.7 (2.6–5.3)
296 (77) 3 (30)
89 (23) 7 (70)
1.0 3.0 (1.9–4.7)
objective measure of operative blood loss, was independently associated with inpatient stay in multivariate analysis. Perhaps not surprisingly, surgical start time at or after 12 noon was associated with a significantly greater risk of hospital stay of 2 midnights or longer. This information may lead surgeons to consider scheduling older patients or anticipated longer and more difficult cases as a ‘‘first start’’ to reduce the number of robotic endometrial cancer patients discharged as inpatients, thereby decreasing health care costs and improving health care utilization. At the time of surgery, a surgeon may consider the proposed risk factors for inpatient stay to enhance appropriate communication to the healthcare team. To promote surgical patient safety, the World Health Organization recommends surgeon involvement in a ‘‘time out’’ after induction of anesthesia and before the surgical incision, as well as a ‘‘sign out’’ after wound closure and before the patient is transferred from the operating room. The surgeon’s role during the surgical ‘‘time out’’ is to inform all team members of the expected operative duration and blood loss, as well as ‘‘any steps of the anticipated surgery that may put the patient at risk for rapid blood loss, injury, or other major morbidity’’ [8]. During the surgical ‘‘sign out,’’ the surgeon is urged to focus on particular intraoperative issues that might affect the patient, including ‘‘events that present a specific risk to the patient during recovery and that may not be evident to all involved’’ [9]. The aim of this step is the efficient and appropriate transfer of critical information to the entire team; in many institutions (including Ohio State), this step includes information regarding resource utilization (such as the anticipated need for postoperative telemetry), expected postoperative course, and early information regarding discharge planning. As expected, patients who experienced a postoperative complication related to surgery were more likely to be admitted as an inpatient regardless of the timing of the complication. However, this analysis was limited by the inability to reliably differentiate between complications that represented the reason for prolonged hospitalization and those that might have accounted for later medical eval-
uations or readmissions. Also, because of the study’s retrospective nature, the specific reasons why older age and history of myocardial infarction may have led to increased inpatient stay are not directly available; however, we hypothesize that these variables may be surrogates for patients requiring more intensive postoperative monitoring or more complicated discharge planning. Finally, there is a suggestion that patients requiring hospitalization spanning at least 2 midnights were more likely to be readmitted within 90 days of discharge. The median interval to readmission for the entire study population was short (9.5 days). Thus, patients who are discharged as inpatients may benefit from a different postdischarge plan of care, which could include a closer follow-up appointment, the addition of a home health visit, or a next-day phone call. Our cohort of almost 400 patients adds additional information to the literature regarding longer than usual hospital stays after robotic surgery and highlights multiple contributing factors. A recent study evaluated prolonged hospitalization in patients undergoing laparoscopic gynecologic surgery for benign and malignant indications, without including robotic surgery patients [10]. Interestingly, the results of that study resemble our results in that age .54 years, estimated blood loss .120 mL, perioperative blood transfusion, and postoperative complication were independent predictors of length of stay .2 days. In another study evaluating a series of 325 robotic surgery patients, older age, increasing uterine weight, and higher Charlson Comorbidity Index score were associated with length of stay .1 day, which is equivalent to our defined inpatient stay group [11]. An analysis of approximately 800 consecutive total laparoscopic hysterectomies revealed that longer duration of surgery and increased operative blood loss were associated with longer hospitalization, and appeared to improve with increasing surgeon experience [12]. The importance of the present study is that despite the rapid incorporation of robotic surgery into gynecologic oncology, there is limited literature on length of stay after these procedures. Our single-institution study provides a reliable dataset with a relatively consistent surgical approach
588
Journal of Minimally Invasive Gynecology, Vol 22, No 4, May/June 2015
Table 2 aRR of inpatient stay spanning at least 2 midnights including preoperative and operative risk factors associated with robotic surgery for endometrial cancer
Risk factor Age, yr ,55 55–59 60–64 65 and over Year of surgery 2006–2009 2010 History of myocardial infarction No Yes Surgery start at or after noon No Yes Perioperative transfusion No Yes
Multivariable model including preoperative characteristics only, aRR (95% CI)
Multivariable model including preoperative and operative characteristics, aRR (95% CI)
1.00 1.8 (0.9–3.5) 2.7 (1.5–4.8) 2.1 (1.2–3.7)
1.00 1.8 (0.9–3.4) 2.6 (1.5–4.5) 2.0 (1.1–3.6)
1.00 0.4 (0.3–0.7)
1.00 0.5 (0.3–0.7)
1.00 1.9 (1.0–3.5)
1.00 2.0 (1.0–3.7)
1.00 1.6 (1.2–2.3)
1.00 1.7 (1.2–2.4)
–
1.00 3.2 (2.3–4.5)
among the gynecologic oncologists. The study has several limitations, however. Although the ‘‘two-midnight rule’’ only applies directly to Medicare beneficiaries, our acute care hospital also serves uninsured patients and those insured by third-party payers. Also, among robotic surgical patients who remain hospitalized for 2 midnights or longer, a minority of patients occasionally may be discharged as outpatients if there is no documented medical indication for continued hospitalization. Furthermore, a learning curve during earlier surgical years and the integration of resident and fellow surgical training to our practice are potential contributing factors to longer patient stays. It is noted that compared with patients who underwent surgery between 2006 and 2009, patients who underwent surgery in 2010 were significantly less likely to have an inpatient stay. This finding may in part reflect the expected trends in discharging physician comfort level and surgeon experience over time with the incorporation of any new surgical technology. Future studies will need to better define the specific indications for inpatient stay, given that information on inadequate pain control, desire to confirm a stable postoperative hemoglobin, and social aspects, such as availability of transportation home, were often difficult to glean from the available data. Moreover, it is important to consider that physicians’ bias toward longer observation of patients subjectively considered to be at increased risk for complications, such
as older patients and patients with larger estimated blood loss, may play a role in prolonging hospitalization independent of the patient or surgical factor itself. Robotic surgery has become standard practice in the management of endometrial cancer. It has helped expand the use of minimally invasive surgery for this disease, resulting in overall decreased length of stay. In the present study, we have further defined the factors in patients with endometrial cancer undergoing robotic surgery that contribute to discharge as an inpatient, which ultimately can be used to enhance preoperative counseling and postoperative resource utilization and discharge planning. Avoidance of unnecessary inpatient stays after robotic surgery will help take advantage of the cost savings obtained from decreased length of stay compared with laparotomy. Since 2010, there has been a trend toward the institution of rapid recovery protocols in gynecologic surgery, and even reports of the safe and feasible same-day discharge of patients undergoing minimally invasive surgical staging for gynecologic malignancy [13,14]. As the surgical landscape continues to evolve, our reported data also provide further impetus for continued collection of data on factors predicting inpatient stay after robotic surgery and prospective analysis of strategies to reduce the length of stay for these patients, particularly those with risk factors.
References 1. Ramirez PT, Adams S, Boggess JF, et al. Robotic-assisted surgery in gynecologic oncology: a Society of Gynecologic Oncology consensus statement. Developed by the Society of Gynecologic Oncology’s Clinical Practice Robotics Task Force. Gynecol Oncol. 2012;124:180–184. 2. Krill LS, Bristow RE. Robotic surgery: gynecologic oncology. Cancer J. 2013;19:167–176. 3. Wright JD, Burke WM, Wilde ET, et al. Comparative effectiveness of robotic versus laparoscopic hysterectomy for endometrial cancer. J Clin Oncol. 2012;30:783–791. 4. Bell MC, Torgerson J, Seshadri-Kreaden U, Suttle AW, Hunt S. Comparison of outcomes and cost for endometrial cancer staging via traditional laparotomy, standard laparoscopy and robotic techniques. Gynecol Oncol. 2008;111:407–411. 5. Barnett JC, Judd JP, Wu JM, Scales CD Jr, Myers ER, Havrilesky LJ. Cost comparison among robotic, laparoscopic, and open hysterectomy for endometrial cancer. Obstet Gynecol. 2010;116:685–693. 6. Wright KN, Jonsdottir GM, Jorgensen S, Shah N, Einarsson JI. Costs and outcomes of abdominal, vaginal, laparoscopic and robotic hysterectomies. JSLS. 2012;16:519–524. 7. Centers for Medicare and Medicaid Services. Medicare program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective hospital prospective payment system and Fiscal Year 2014 rates; quality reporting requirements for specific providers; hospital conditions of participation; payment polices related to patient status. Final rules. Fed Regist. 2013;78:50495–51040. 8. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159:702–706. 9. World Health Organization. Implementation Manual: WHO Surgical Safety Checklist. Geneva, Switzerland: World Health Organization; 2008. 10. Zand B, Frumovitz M, Jofre MF, et al. Risk factors for prolonged hospitalization after gynecologic laparoscopic surgery. Gynecol Oncol. 2012;126:428–431.
Liang et al.
Predicting Inpatient Stay After Robotic Surgery for Endometrial Cancer
11. Woelk JL, Casiano ER, Weaver AL, Gostout BS, Trabuco EC, Gebhart JB. The learning curve of robotic hysterectomy. Obstet Gynecol. 2013;121:87–95. 12. O’Hanlan KA, Dibble SL, Garnier AC, Reuland ML. Total laparoscopic hysterectomy: technique and complications of 830 cases. JSLS. 2007; 11:45–53.
589
13. Nelson G, Kalogera E, Dowdy SC. Enhanced recovery pathways in gynecologic oncology. Gynecol Oncol. 2014;135:586–594. 14. Penner KR, Fleming ND, Barlavi L, Axtell AE, Lentz SE. Same-day discharge is feasible and safe in patients undergoing minimally invasive staging for gynecologic malignancies. Am J Obstet Gynecol. 2015;212: 186.e1–186.e8.
Predicting Inpatient Stay Lasting 2 Midnights or Longer After Robotic Surgery for Endometrial Cancer Margaret I. Liang, MD, Maggie A. Rosen, MD, Kellie S. Rath, MD, Erinn M. Hade, PhD, Aine E. Clements, MD, Floor J. Backes, MD, Eric L. Eisenhauer, MD, Ritu Salani, MD, David M. O’Malley, MD, Jeffrey M. Fowler, MD, and David E. Cohn, MD* From the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology (Liang, Rosen, Rath, Clements, Backes, Eisenhauer, Salani, O’Malley, Fowler, Cohn), and Center for Biostatistics (Hade), The Ohio State University College of Medicine, Columbus, Ohio.
ABSTRACT Objective: To estimate the rate of inpatient stay and the factors predicting inpatient status after robotic surgery for endometrial cancer following the change in the Medicare definition of ‘‘inpatient’’ to include hospitalization spanning 2 midnights. Design: Retrospective chart review (Canadian Task Force classification II-1). Setting: Academic hospital. Patients: All patients (n 5 395) with endometrial cancer who underwent robotic surgical management between 2006 and 2010. Intervention: The outpatient stay group with hospitalization spanning 1 midnight was compared with the inpatient stay group with hospitalization spanning 2 midnights or longer through estimation of the adjusted relative risk (aRR) for various characteristics of interest. Results: Ninety-six of 395 patients (24.3%) stayed at least 2 midnights and thus were deemed inpatients. Clinical factors associated with inpatient stay were increasing age, history of myocardial infarction (aRR, 2.0; 95% confidence interval [CI], 1.0–3.7), surgery start time at or after 12 noon (aRR, 1.7; 95% CI, 1.2–2.4), perioperative blood transfusion (aRR, 3.2; 95% CI, 2.3–4.5), and surgery performed in the year 2010 (aRR, 0.5; 95% CI, 0.3–0.7). Age R60 years was associated with at least a 2-fold adjusted risk of prolonged hospitalization. Body mass index, other medical comorbidities, operative duration, estimated blood loss, and performance of lymphadenectomy or additional surgical procedures were not identified as significant risk factors. Conclusion: Approximately 75% of the patients undergoing robotic surgery for endometrial cancer were discharged as outpatients. Recognition of factors predicting inpatient stay can improve hospital resource allocation and throughput in women undergoing robotic surgery for endometrial cancer. Journal of Minimally Invasive Gynecology (2015) 22, 583–589 Ó 2015 AAGL. All rights reserved. Keywords:
DISCUSS
Endometrial cancer; Inpatient stay; Prolonged hospitalization; Robotic surgery
You can discuss this article with its authors and with other AAGL members at http://www.AAGL.org/jmig-22-3-JMIG-D-14-00586
Use your Smartphone to scan this QR code and connect to the discussion forum for this article now* * Download a free QR Code scanner by searching for ‘‘QR scanner’’ in your smartphone’s app store or app marketplace.
The authors declare no conflicts of interest. This paper was presented as an oral presentation at the Western Association of Gynecologic Oncologists Annual Meeting, Seattle, Washington, June 2013. Corresponding author: David E. Cohn, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, The Ohio State University College of Medicine, M210 Starling Loving, 320 West 10th Ave, Columbus, OH 43210. E-mail: [email protected] Submitted November 25, 2014. Accepted for publication December 30, 2014. Available at www.sciencedirect.com and www.jmig.org 1553-4650/$ - see front matter Ó 2015 AAGL. All rights reserved. http://dx.doi.org/10.1016/j.jmig.2014.12.167
Increasing attention is being given to improving the delivery of cost-effective health care. In the field of gynecologic oncology, robotic surgery has become an accepted and popular approach for the surgical management of endometrial cancer [1,2]. It is estimated that more than 60% of minimally invasive hysterectomies performed for endometrial cancer were completed robotically in 2010 [3]. Shorter hospital stay is among the many advantages of robotic surgery compared with laparotomy [4–6]. The high technology costs associated with the robotic platform remain a barrier to its widespread use, however [1].
584
As with any new scientific advancement, capital costs likely will decrease over time. Minimizing the length of hospital stay after robotic surgery has the potential to help offset the overall higher supply costs. Several studies have compared the costs associated with various surgical approaches for the management of endometrial cancer, including laparotomy, laparoscopy, and robotic surgery [3–6]. The majority of the cost of laparotomy has been attributed to longer average hospital stay, whereas that of robotic surgery is related to disposable equipment. For all minimally invasive surgical techniques, shorter hospital stays offered significant cost savings compared with laparotomy [4]. Recent statutory changes implemented to satisfy provisions contained in the Patient Protection and Affordable Care Act and the Health Care and Education Reconciliation Act of 2010 (collectively known as the Affordable Care Act) will have an effect on the care of robotic surgery patients. On August 19, 2013, the Centers for Medicare & Medicaid Services (CMS) revised the Medicare hospital inpatient prospective payment systems (IPPS) applying to the designation of appropriate level of care and subsequent costs in acute care hospitals [7]. As part of the IPPS revision, hospital discharges occurring on or after October 1, 2013, are classified as inpatient if the hospital stay crosses 2 midnights (known as the ‘‘two-midnight rule’’) and thus are paid under Medicare Part A. In contrast, if the stay is unanticipated to or does not span 2 midnights, then the hospitalization is classified as outpatient, with appropriate payment under Medicare Part B [7]. The rationale for this IPPS revision includes the fact that outpatient and/or observation hospital encounters were not being appropriately converted to inpatient status, leading to higher out-of-pocket expenses for patients. With the more clearly defined classification for inpatient stay, hospitals also will have a further incentive to expedite safe medical discharges owing to differences in reimbursement. The CMS estimated that under the new guidelines for determining inpatient status, there would be a net shift of 40,000 additional inpatient admissions in fiscal year 2014, at a cost of $220 million [7]. With these changes, identifying patients at increased risk for hospitalization lasting 2 midnights or longer after robotic surgery for endometrial cancer is imperative for hospitals to achieve optimal resource allocation. Although the relevance of the ‘‘two-midnight rule’’ to other health systems is uncertain, it is likely that utilization of healthcare resources will continue to be an important subject internationally as well. Therefore, in response to the new CMS revision of the IPPS using the ‘‘two-midnight rule’’ to define inpatient status, we set out to evaluate the factors associated with a hospital stay spanning at least 2 midnights following robotic surgery for endometrial cancer. Materials and Methods The Ohio State University Institutional Review Board approved this study. All patients who underwent robotic sur-
Journal of Minimally Invasive Gynecology, Vol 22, No 4, May/June 2015
gical management of endometrial cancer between January 1, 2006, and December 31, 2010, at The Ohio State University Wexner Medical Center in Columbus, Ohio were identified. All patients underwent hysterectomy and bilateral salpingooophorectomy, with lymph node staging performed in 86% of cases based on the surgeon’s clinical judgment. Five robotically trained gynecologic oncologists at Ohio State were included in the analysis. Exclusion criteria included conversion to laparotomy or mini-laparotomy for delivery of the specimen, so that only patients who underwent robotically completed surgery were included in the analysis. Hospital and office charts were reviewed for patient characteristics, including age, height, weight, body mass index (BMI), and presence of prespecified medical comorbidities. The major comorbidities examined were hypertension, diabetes, hyperlipidemia, congestive heart failure, history of myocardial infarction, history of stroke, chronic renal disease, asthma, and obstructive sleep apnea. Operative data, including surgical start time, operative duration, estimated blood loss, perioperative blood transfusion (including transfusions administered during and after surgery), performance of pelvic lymphadenectomy with or without para-aortic lymphadenectomy, performance of other surgical procedures, and intraoperative complications, were collected. Procedure start times were stratified as before 12 noon or at 12 noon and after. Operative duration was defined as the time of skin incision to incision closure. Procedures performed concurrently included ureterolysis, enterolysis, omentectomy, hernia repair, cholecystectomy, and cystoscopy. Intraoperative complications were categorized as vessel injury, genitourinary injury, gastrointestinal injury, or vaginal laceration. Information on postoperative complications was gathered, including postoperative morbidity and readmission within 90 days. The postoperative morbidities were classified as gastrointestinal, genitourinary, infectious, cardiopulmonary, thromboembolism, and wound complications. Standard postoperative discharge criteria were used. Before discharge, patients were expected to be tolerating a regular diet, achieving adequate pain control with a combination of oral narcotics and nonsteroidal anti-inflammatory drugs, ambulating without difficulty, and voiding spontaneously with adequate urine output. Although the ‘‘twomidnight rule’’ was not in effect during study years 2006 to 2010, we applied the definition to our cohort of robotic surgical patients based on length of stay. As a result, patients who were discharged following a hospitalization spanning a single midnight (i.e., discharged on postoperative day 1) were designated as the outpatient stay group, and those discharged after a hospitalization spanning at least 2 midnights (i.e., discharged on postoperative day 2 or later) were classified as the inpatient stay group. Demographic and clinical preoperative and operative characteristics were summarized through descriptive statistics. Categorization of these variables was determined by clinical relevance, historically used cutpoints, and natural breaks in the data owing to collection methodology. For
Liang et al.
Predicting Inpatient Stay After Robotic Surgery for Endometrial Cancer
instance, the majority of measurements for estimated blood loss were recorded in 50-mL intervals. Crude relative risk (RR) and adjusted RR (aRR), along with the associated 95% confidence intervals (CIs), of inpatient stay for these factors were estimated via a modified Poisson regression approach [8]. The development of a multivariable model to determine predictive factors of inpatient stay proceeded in 2 stages. First, a model that included clinically important preoperative factors was established. Then operative factors of interest were added to the model, and the model was refined based on the relative information provided by the addition of each variable. Potential interactions between history of myocardial infarction and either age or transfusion status were explored and found to be noninfluential. All reported CIs are 2-sided. Analyses were performed with Stata version 13 (StataCorp, College Station, TX). Results A total of 395 patients met the inclusion criteria for this study (Table 1). The outpatient stay group comprised 299 patients (75.7%), and the inpatient stay group comprised 96 patients (24.3%). The average length of stay was 1.5 6 2.4 days, with a median length of stay of 1 day (range, 1– 46 days) across all patients. One patient with a prolonged hospitalization of 46 days underwent reoperation on postoperative day 2 for ileal perforation, requiring ileal resection with gastrostomy tube placement. This was further complicated by prolonged respiratory failure, bacteremia, and postoperative ileus. This patient was ultimately discharged to a rehabilitation facility and experienced complete recovery. In the inpatient stay group, the average length of stay was 2.9 6 4.6 days, with median length of stay of 2 days (range, 2–46 days). Patient demographic data and clinical preoperative and operative characteristics are summarized in Table 1. In univariable analyses, an increased risk of inpatient stay was associated with older age, history of renal disease, history of myocardial infarction, surgery starting at or after noon, need for perioperative transfusion, and surgery performed during earlier study years (2006 to 2009 vs 2010) (Table 1). Although the estimated unadjusted risk of inpatient stay was not associated with operative duration or estimated blood loss, patients with very long surgeries (R5 hours) or greater blood loss (R150 mL) were found to be at increased risk of longer hospitalization. The overall rate of intraoperative complications was low (2%), and the occurrence of complications was not associated with inpatient stay (Table 1). Adjusted multivariable models identified the demographic and preoperative characteristics of age, history of myocardial infarction, timing of surgery, and year of surgery as important predictors of inpatient stay when included together (Table 2). Patients aged R60 years had at least a twofold increased risk of inpatient stay on average compared
585
with patients aged ,55 years (age 60–64: aRR, 2.7; 95% CI, 1.5–4.8; age R65: aRR, 2.1; 95% CI, 1.2–3.7). Increased risk was also consistently found in patients with a history of myocardial infarction (aRR, 1.9; 95% CI, 1.0–3.5) and those whose surgery began at or after 12 noon (aRR, 1.6; 95% CI, 1.2–2.3). When further considering operative characteristics in this model, perioperative transfusion was found to be the most influential factor of all operative variables in predicting inpatient stay (aRR, 3.2; 95% CI, 2.3–4.5). Inclusion of this covariate with those previously included preoperative covariates had little effect on the previously estimated effects. Patients who experienced any postoperative complication (overall rate, 18.2%) were more likely to be hospitalized for at least 2 midnights and thus be discharged as an inpatient (RR, 1.9; 95% CI, 1.4–2.8). The 2 most common postoperative complications in the inpatient stay group, ileus and postoperative fever, were also independently associated with inpatient status (ileus: RR, 3.7; 95% CI, 2.6–5.3; postoperative fever: RR, 3.0; 95% CI, 1.9–4.7). Patients who required inpatient hospitalization after robotic surgery for endometrial cancer were more likely to be readmitted within 90 days of surgery (RR, 2.1; 95% CI, 1.0–4.2). A total of 30 patients, including 18 (6.0%) from the outpatient group and 12 (12.5%) from the inpatient group, were readmitted within 90 days. Adjustment for age and year of surgery attenuated this effect downward (aRR, 1.7; 95% CI, 0.9–3.4). Discussion Following robotic surgery for endometrial cancer, hospital stays for the majority of patients (approximately 75%) span a single midnight and thus are classified by the revised Medicare hospital IPPS as outpatient hospitalizations. Increasing age, history of myocardial infarction, surgery start time at or after 12 noon, and perioperative blood transfusion were independently associated with stay lasting 2 midnights or longer. Elucidating factors that predict inpatient status following robotic surgery for endometrial cancer is of interest to patients, their caregivers, providers, and hospital systems. Preoperative identification of patients with risk factors for inpatient stay (such as older age or history of myocardial infarction) can help surgeons set appropriate expectations regarding discharge date and course. Alternatively, these data are important to aid in counseling obese patients and patients with multiple comorbidities that the expected postoperative stay following robotic surgery for endometrial cancer is typically similar to that of patients who are of normal weight and with no comorbid conditions. The recognition that operative factors may place a patient at risk for inpatient stay is also important for planning surgery. Surgeries at the extreme of operative duration (R5 hours) or with greater blood loss (R150 mL) were associated with longer hospital stays. Although increasing estimated blood loss was not identified as a very strong predictor, the need for perioperative blood transfusion, a more
586
Journal of Minimally Invasive Gynecology, Vol 22, No 4, May/June 2015
Table 1 Potential clinical factors associated with inpatient stay, defined as length of stay spanning 2 midnights or longer, after robotic surgery for endometrial cancer Clinical factor Preoperative characteristics Age, yr, mean 6 SD Age ,55 yr, n (%) Age 55–59 yr, n (%) Age 60–64 yr, n (%) Age R65 yr, n (%) BMI, kg/m2, mean 6 SD Normal weight (18.5–24.9), n (%) Overweight (25–29.9), n (%) Obese (R30), n (%) Medical comorbidities, n (%) 0 1 R2 History of renal disease, n (%) No Yes History of myocardial infarction, n (%) No Yes Timing of surgery, n (%) Start before 12 noon Start at or after 12 noon Year of surgery, n (%) 2006–2009 2010 Operative characteristics Operative duration, hr, mean 6 SD Duration of ,3 hr, n (%) Duration of 3 hr to 3 hr and 59 min, n (%) Duration of 4 hr to 4 hr and 59 min, n (%) Duration of R5 hr, n (%) Estimated blood loss, mL, median (range) Loss of 0–49 mL, n (%) Loss of 50–99 mL, n (%) Loss of 100–149 mL, n (%) Loss of R150 mL, n (%) Perioperative transfusion, n (%) No Yes Pelvic lymphadenectomy, n (%) No Yes Para-aortic lymphadenectomy, n (%) No Yes Additional surgical procedure, n (%) No Yes Intraoperative complication, n (%) No Yes Postoperative characteristics
Outpatient stay (n 5 299)
Inpatient stay (n 5 96)
59 6 11 88 (88) 58 (76) 55 (66) 98 (72) 35 6 10 39 (76) 55 (67) 205 (78)
64 6 10 12 (12) 18 (24) 28 (34) 38 (28) 35 6 10 12 (24) 27 (33) 57 (22)
1.0 1.4 (0.8–2.5) 0.9 (0.5–1.6)
120 (79) 99 (73) 80 (74)
31 (21) 37 (27) 28 (26)
1.0 1.3 (0.9–2.0) 1.3 (0.8–2.0)
297 (76) 2 (40)
93 (24) 3 (60)
1.0 2.5 (1.2–5.3)
297 (76) 2 (33)
92 (24) 4 (67)
1.0 2.8 (1.6–5.1)
185 (81) 114 (68)
43 (19) 53 (32)
1.0 1.7 (1.2–2.4)
139 (67) 160 (86)
69 (33) 27 (14)
1.0 0.4 (0.3–0.6)
200 6 43 99 (79) 149 (78) 45 (71) 6 (35) 100 (10–1400) 61 (80) 87 (77) 80 (80) 71 (67)
218 6 57 26 (21) 41 (22) 18 (29) 11 (65) 100 (10–500) 15 (20) 26 (23) 20 (20) 35 (33)
297 (77) 2 (20)
88 (23) 8 (80)
1.0 3.5 (2.4–5.0)
43 (80) 256 (75)
11 (20) 85 (25)
1.0 1.2 (0.7–2.1)
98 (76) 201 (76)
31 (24) 65 (24)
1.0 1.0 (0.7–1.5)
271 (77) 28 (67)
82 (23) 14 (33)
1.0 1.4 (0.9–2.3)
292 (75) 7 (88)
95 (25) 1 (13)
1.0 0.5 (0.1–3.2)
RR (95% CI)
1.0 2.0 (1.0–3.9) 2.8 (1.5–5.2) 2.3 (1.3–4.2)
1.0 1.0 (0.7–1.6) 1.4 (0.8–2.5) 3.1 (1.9–5.1) 1.0 1.2 (0.7–2.1) 1.0 (0.6–1.8) 1.7 (1.0–2.8)
(Continued )
Liang et al.
Predicting Inpatient Stay After Robotic Surgery for Endometrial Cancer
587
Table 1 Continued Clinical factor Postoperative complication, n (%) No Yes Ileus, n (%) No Yes Postoperative fever, n (%) No Yes
Outpatient stay (n 5 299)
Inpatient stay (n 5 96)
RR (95% CI)
256 (79) 43 (60)
67 (21) 29 (40)
1.0 1.9 (1.4–2.8)
298 (77) 1 (14)
90 (23) 6 (86)
1.0 3.7 (2.6–5.3)
296 (77) 3 (30)
89 (23) 7 (70)
1.0 3.0 (1.9–4.7)
objective measure of operative blood loss, was independently associated with inpatient stay in multivariate analysis. Perhaps not surprisingly, surgical start time at or after 12 noon was associated with a significantly greater risk of hospital stay of 2 midnights or longer. This information may lead surgeons to consider scheduling older patients or anticipated longer and more difficult cases as a ‘‘first start’’ to reduce the number of robotic endometrial cancer patients discharged as inpatients, thereby decreasing health care costs and improving health care utilization. At the time of surgery, a surgeon may consider the proposed risk factors for inpatient stay to enhance appropriate communication to the healthcare team. To promote surgical patient safety, the World Health Organization recommends surgeon involvement in a ‘‘time out’’ after induction of anesthesia and before the surgical incision, as well as a ‘‘sign out’’ after wound closure and before the patient is transferred from the operating room. The surgeon’s role during the surgical ‘‘time out’’ is to inform all team members of the expected operative duration and blood loss, as well as ‘‘any steps of the anticipated surgery that may put the patient at risk for rapid blood loss, injury, or other major morbidity’’ [8]. During the surgical ‘‘sign out,’’ the surgeon is urged to focus on particular intraoperative issues that might affect the patient, including ‘‘events that present a specific risk to the patient during recovery and that may not be evident to all involved’’ [9]. The aim of this step is the efficient and appropriate transfer of critical information to the entire team; in many institutions (including Ohio State), this step includes information regarding resource utilization (such as the anticipated need for postoperative telemetry), expected postoperative course, and early information regarding discharge planning. As expected, patients who experienced a postoperative complication related to surgery were more likely to be admitted as an inpatient regardless of the timing of the complication. However, this analysis was limited by the inability to reliably differentiate between complications that represented the reason for prolonged hospitalization and those that might have accounted for later medical eval-
uations or readmissions. Also, because of the study’s retrospective nature, the specific reasons why older age and history of myocardial infarction may have led to increased inpatient stay are not directly available; however, we hypothesize that these variables may be surrogates for patients requiring more intensive postoperative monitoring or more complicated discharge planning. Finally, there is a suggestion that patients requiring hospitalization spanning at least 2 midnights were more likely to be readmitted within 90 days of discharge. The median interval to readmission for the entire study population was short (9.5 days). Thus, patients who are discharged as inpatients may benefit from a different postdischarge plan of care, which could include a closer follow-up appointment, the addition of a home health visit, or a next-day phone call. Our cohort of almost 400 patients adds additional information to the literature regarding longer than usual hospital stays after robotic surgery and highlights multiple contributing factors. A recent study evaluated prolonged hospitalization in patients undergoing laparoscopic gynecologic surgery for benign and malignant indications, without including robotic surgery patients [10]. Interestingly, the results of that study resemble our results in that age .54 years, estimated blood loss .120 mL, perioperative blood transfusion, and postoperative complication were independent predictors of length of stay .2 days. In another study evaluating a series of 325 robotic surgery patients, older age, increasing uterine weight, and higher Charlson Comorbidity Index score were associated with length of stay .1 day, which is equivalent to our defined inpatient stay group [11]. An analysis of approximately 800 consecutive total laparoscopic hysterectomies revealed that longer duration of surgery and increased operative blood loss were associated with longer hospitalization, and appeared to improve with increasing surgeon experience [12]. The importance of the present study is that despite the rapid incorporation of robotic surgery into gynecologic oncology, there is limited literature on length of stay after these procedures. Our single-institution study provides a reliable dataset with a relatively consistent surgical approach
588
Journal of Minimally Invasive Gynecology, Vol 22, No 4, May/June 2015
Table 2 aRR of inpatient stay spanning at least 2 midnights including preoperative and operative risk factors associated with robotic surgery for endometrial cancer
Risk factor Age, yr ,55 55–59 60–64 65 and over Year of surgery 2006–2009 2010 History of myocardial infarction No Yes Surgery start at or after noon No Yes Perioperative transfusion No Yes
Multivariable model including preoperative characteristics only, aRR (95% CI)
Multivariable model including preoperative and operative characteristics, aRR (95% CI)
1.00 1.8 (0.9–3.5) 2.7 (1.5–4.8) 2.1 (1.2–3.7)
1.00 1.8 (0.9–3.4) 2.6 (1.5–4.5) 2.0 (1.1–3.6)
1.00 0.4 (0.3–0.7)
1.00 0.5 (0.3–0.7)
1.00 1.9 (1.0–3.5)
1.00 2.0 (1.0–3.7)
1.00 1.6 (1.2–2.3)
1.00 1.7 (1.2–2.4)
–
1.00 3.2 (2.3–4.5)
among the gynecologic oncologists. The study has several limitations, however. Although the ‘‘two-midnight rule’’ only applies directly to Medicare beneficiaries, our acute care hospital also serves uninsured patients and those insured by third-party payers. Also, among robotic surgical patients who remain hospitalized for 2 midnights or longer, a minority of patients occasionally may be discharged as outpatients if there is no documented medical indication for continued hospitalization. Furthermore, a learning curve during earlier surgical years and the integration of resident and fellow surgical training to our practice are potential contributing factors to longer patient stays. It is noted that compared with patients who underwent surgery between 2006 and 2009, patients who underwent surgery in 2010 were significantly less likely to have an inpatient stay. This finding may in part reflect the expected trends in discharging physician comfort level and surgeon experience over time with the incorporation of any new surgical technology. Future studies will need to better define the specific indications for inpatient stay, given that information on inadequate pain control, desire to confirm a stable postoperative hemoglobin, and social aspects, such as availability of transportation home, were often difficult to glean from the available data. Moreover, it is important to consider that physicians’ bias toward longer observation of patients subjectively considered to be at increased risk for complications, such
as older patients and patients with larger estimated blood loss, may play a role in prolonging hospitalization independent of the patient or surgical factor itself. Robotic surgery has become standard practice in the management of endometrial cancer. It has helped expand the use of minimally invasive surgery for this disease, resulting in overall decreased length of stay. In the present study, we have further defined the factors in patients with endometrial cancer undergoing robotic surgery that contribute to discharge as an inpatient, which ultimately can be used to enhance preoperative counseling and postoperative resource utilization and discharge planning. Avoidance of unnecessary inpatient stays after robotic surgery will help take advantage of the cost savings obtained from decreased length of stay compared with laparotomy. Since 2010, there has been a trend toward the institution of rapid recovery protocols in gynecologic surgery, and even reports of the safe and feasible same-day discharge of patients undergoing minimally invasive surgical staging for gynecologic malignancy [13,14]. As the surgical landscape continues to evolve, our reported data also provide further impetus for continued collection of data on factors predicting inpatient stay after robotic surgery and prospective analysis of strategies to reduce the length of stay for these patients, particularly those with risk factors.
References 1. Ramirez PT, Adams S, Boggess JF, et al. Robotic-assisted surgery in gynecologic oncology: a Society of Gynecologic Oncology consensus statement. Developed by the Society of Gynecologic Oncology’s Clinical Practice Robotics Task Force. Gynecol Oncol. 2012;124:180–184. 2. Krill LS, Bristow RE. Robotic surgery: gynecologic oncology. Cancer J. 2013;19:167–176. 3. Wright JD, Burke WM, Wilde ET, et al. Comparative effectiveness of robotic versus laparoscopic hysterectomy for endometrial cancer. J Clin Oncol. 2012;30:783–791. 4. Bell MC, Torgerson J, Seshadri-Kreaden U, Suttle AW, Hunt S. Comparison of outcomes and cost for endometrial cancer staging via traditional laparotomy, standard laparoscopy and robotic techniques. Gynecol Oncol. 2008;111:407–411. 5. Barnett JC, Judd JP, Wu JM, Scales CD Jr, Myers ER, Havrilesky LJ. Cost comparison among robotic, laparoscopic, and open hysterectomy for endometrial cancer. Obstet Gynecol. 2010;116:685–693. 6. Wright KN, Jonsdottir GM, Jorgensen S, Shah N, Einarsson JI. Costs and outcomes of abdominal, vaginal, laparoscopic and robotic hysterectomies. JSLS. 2012;16:519–524. 7. Centers for Medicare and Medicaid Services. Medicare program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective hospital prospective payment system and Fiscal Year 2014 rates; quality reporting requirements for specific providers; hospital conditions of participation; payment polices related to patient status. Final rules. Fed Regist. 2013;78:50495–51040. 8. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159:702–706. 9. World Health Organization. Implementation Manual: WHO Surgical Safety Checklist. Geneva, Switzerland: World Health Organization; 2008. 10. Zand B, Frumovitz M, Jofre MF, et al. Risk factors for prolonged hospitalization after gynecologic laparoscopic surgery. Gynecol Oncol. 2012;126:428–431.
Liang et al.
Predicting Inpatient Stay After Robotic Surgery for Endometrial Cancer
11. Woelk JL, Casiano ER, Weaver AL, Gostout BS, Trabuco EC, Gebhart JB. The learning curve of robotic hysterectomy. Obstet Gynecol. 2013;121:87–95. 12. O’Hanlan KA, Dibble SL, Garnier AC, Reuland ML. Total laparoscopic hysterectomy: technique and complications of 830 cases. JSLS. 2007; 11:45–53.
589
13. Nelson G, Kalogera E, Dowdy SC. Enhanced recovery pathways in gynecologic oncology. Gynecol Oncol. 2014;135:586–594. 14. Penner KR, Fleming ND, Barlavi L, Axtell AE, Lentz SE. Same-day discharge is feasible and safe in patients undergoing minimally invasive staging for gynecologic malignancies. Am J Obstet Gynecol. 2015;212: 186.e1–186.e8.
