- Email: [email protected]
Patterns of readmission and reoperation within 90 days after Roux-en-Y gastric bypass
Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
Original article
Patterns of readmission and reoperation within 90 days after Roux-en-Y gastric bypass Todd Andrew Kellogg, M.D., Therese Swan, B.S., Daniel A. Leslie, M.D., Henry Buchwald, M.D., Ph.D., Sayeed Ikramuddin, M.D. Division of Gastrointestinal Surgery, Department of Surgery, University of Minnesota Medical Center, Minneapolis, Minnesota Received June 20, 2008; revised November 29, 2008; accepted January 20, 2009
Abstract
Background: Health insurance payors harbor concerns regarding the cost of bariatric procedures that are chiefly related to early readmissions and reoperations. We have attempted to identify the avoidable causes of readmission. Methods: We retrospectively reviewed the indications for short-term (⬍90-d) emergency department (ED) visits, readmissions, and reoperations from August 2004 through May 2007 for patients undergoing primary Roux-en-Y gastric bypass (RYGB) for morbid obesity at a tertiary care teaching hospital. The electronic medical record of the primary hospital was reviewed, as well as the electronic medical records of 9 local hospitals serving the area, allowing the incorporation of data from 35 locoregional hospitals. Results: A total of 1222 consecutive patients underwent RYGB, 1051 laparoscopically. Of these 1222 patients, 173 had 252 ED visits, readmissions, and/or reoperations; 147 (58%) visits were to the primary institution and 105 (42%) occurred at a local or regional hospital. No age difference was found between the patients who underwent ED visits, hospital readmissions, or reoperations and those who did not (mean age 43 yr for both groups, P ⬎ .05). Patients who were seen in the ED, readmitted to the hospital, or underwent reoperation had had a greater body mass index (50 kg/m2 versus 48 kg/m2, P ⫽ .001). On average, the readmissions occurred 27.3 days (range 2– 88) postoperatively, and the mean hospital length of stay for readmitted patients was 3.3 days (range 1–16). Patients who presented for ED visits, readmission, or reoperations were more likely to have undergone open RYGB than laparoscopic RYGB (P ⫽ .002). The ⬍90-day all-cause ED visit, readmission, and reoperation rate was 21% (n ⫽ 252). Considering all 1222 patients, the incidence of nausea, vomiting, and dehydration, abdominal pain, and wound issues was 5% (n ⫽ 65), 4% (n ⫽ 50), and 2% (n ⫽ 21), respectively. Considering only the 173 patients with ED visits, readmissions, or reoperations (n ⫽ 252), the admitting diagnosis was nausea, vomiting, and dehydration in 26%, abdominal pain in 20%, and wound issues in 8%. The unemployed, disabled, or retired were more likely to have been seen in the ED or readmitted compared with the employed, nondisabled, or not retired (P ⫽ .01). Conclusion: A considerable number of patients are affected by nausea, vomiting, and dehydration, abdominal pain, and wound issues ⬍90 days postoperatively. Socioeconomic and functional status might have an effect on the rate of ED visits and readmissions. By ensuring that the appropriate outpatient mechanisms for management of these problems are available, early ED visits and readmission rates should significantly decrease. (Surg Obes Relat Dis 2009;5:416 – 424.) © 2009 American Society for Metabolic and Bariatric Surgery. All rights reserved.
Keywords:
Roux-en-Y gastric bypass; Readmission; Reoperation; Rates; Outcomes; Complications
Presented at the 25th Annual Meeting of the American Society for Metabolic and Bariatric Surgery, June 15–20, 2008, Washington, DC. Reprints not available from the authors. 1550-7289/09/$ – see front matter © 2009 American Society for Metabolic and Bariatric Surgery. All rights reserved. doi:10.1016/j.soard.2009.01.008
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
The reported outcomes for Roux-en-Y gastric bypass (RYGB) performed for morbid obesity have been extremely good. Evidence is increasing that bariatric surgery improves the overall health and longevity of morbidly obese patients [1– 4]. Despite these beneficial outcomes, concerns regarding the cost/benefit ratio of bariatric care have been expressed by health insurance payors and others [3,5]. Some aspects of cost that can be easily overlooked are the readmission and reoperation rates associated with bariatric surgery. Little is known about the need for readmission, reoperation, or emergency department (ED) visits for these patients. A major barrier to this type of analysis is that patients with ED issues often do not return to the institution where the primary operation was performed. This can result in gross underreporting of rates of readmission and reoperation, as well as redundant testing and unwarranted admissions. Another common form of readmission and reoperation reporting comes from multi-institutional administrative data, which is subject to significant inaccuracies based on its collection, not for clinical or research purposes, but for insurance analyses [6]. Moreover, studies examining readmission and reoperation are often limited to the first 30 days after operation, which is considered the interval for procedure-related complications. A review of our data (Fig. 1) and data from others [7] has suggested that this limitation might not capture all relevant procedure-related ED visits, readmissions, and reoperations. The objective of this study was to determine, with as much accuracy as possible, the frequency of ED visits, readmissions, and reoperations for patients who had undergone RYGB and to identify potentially avoidable causes.
417
Methods Study design This study is a retrospective analysis of the data from 1222 consecutive patients who had undergone laparoscopic or open RYGB as the index procedure. The records from patients undergoing primary bariatric surgery from August 1, 2004 to May 31, 2007 were reviewed. Of these patients, 977 (80%) were women and 245 (20%) were men. The cohort of patients who underwent ED visits, readmission, or reoperation was identified and analyzed further. All index bariatric operations were performed at the University of Minnesota Medical Center in an American Society for Metabolic and Bariatric Surgery Centers of Excellence program [8]. After obtaining permission from the institutional review board, the data were collected by review of the electronic medical records (EMRs) from the University of Minnesota Medical Center, the tertiary care teaching hospital at which patients underwent their index bariatric operation (primary medical center), as well as a comprehensive review of the EMRs of the 9 local hospitals serving patients from the 3 dominant health insurance payors that serve the region. This approach allowed the incorporation of information from 35 regional hospitals and included ED and ambulatory same-day surgery (SDS) visits, in addition to hospital readmissions and reoperations. Scheduled bariatric follow-up appointments or patient clinic visits to a primary care provider were not included. The data from these patients were then entered and maintained in a bariatric database.
Fig. 1. Frequency of ED visits, readmissions, and reoperations after primary RYGB at University of Minnesota (unpublished data).
418
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
Definitions of readmission and reoperation
Results
Readmissions were defined as any admission to the primary medical center or a local or regional hospital for any reason, including ED visits, if this resulted in hospitalization. Reoperations were defined as operations performed after discharge from the hospital for any reason, either related or unrelated to the index bariatric procedure, and included operations associated with readmission and SDS. ED visits not resulting in hospital admission were defined as such and not as an admission.
Patient demographics
Definitions of related and unrelated to RYGB ED visits, readmissions, and reoperations were considered related to RYGB if the medical condition was directly or indirectly related to RYGB for weight loss or its known complications and adverse effects, both short and long term. This included all abdominal pain, except when confirmed to be of a known unrelated cause, all nausea and vomiting, cholecystitis and cholecystectomy, nephrolithiasis and associated procedures, and early (⬍30 d postoperatively) pancreatitis. The cause was considered unrelated if associated with an exacerbation of a previous medical condition or when no known association of RYGB with the medical condition requiring the ED visit, readmission, or reoperation was found. Gastrointestinal hemorrhage associated with a marginal ulcer was considered related at any point postoperatively. Gastrointestinal hemorrhage associated with iatrogenic coagulopathy not occurring in the immediate postoperative period was considered unrelated.
Operations Three surgeons performed all RYGBs using either the open (n ⫽ 171) or laparoscopic (n ⫽ 1051) technique. A partially hand-sewn, partially linearly stapled, 2-layer gastrojejunal anastomosis, with a Roux limb length that varied from 75 to 150 cm and biliopancreatic limb length that ranged from 30 to 100 cm, was used. Revision operations were excluded from the analysis.
Statistical analysis The data are presented as the median and range. Groups were compared using Pearson’s chi-square analysis. A chisquare probability of ⱕ .05 was interpreted as justification for rejecting the null hypothesis. The alternate hypothesis was not rejected when the variables had an associated relationship. All statistical operations were performed using MS Excel.
Of the 173 patients seen in the ED, readmitted, or who underwent reoperation, 137 were women and 36 were men. No difference was found in gender between the patients requiring ED visits, readmission, or reoperation and those who did not (P ⫽ .9). The mean patient age was 42 years (range 15–71). The mean preoperative body mass index was 50 kg/m2 (range 35–96). Of the 173 patients, 125 were white, 27 were black, 4 were Hispanic, and 17 were of unknown ethnicity. No difference was found in patient age between those who underwent ED visits, hospital readmissions, or reoperations and those who did not (43 versus 43 yr, respectively; P ⬎ .05). The patients who required ED visits, hospital readmissions, or reoperations had a significantly greater body mass index (50 versus 48 kg/m2, respectively; P ⫽ .001). Compared with whites, African Americans were more likely to require ED visits, readmissions or reoperations (P ⫽ .03). Co-morbidities The most frequent preoperative co-morbidities among the 173 patients are listed in Table 1. No statistically significant difference was found in the incidence of co-morbid disease, either in combination or independently, between those who required ED visits, readmissions, or reoperations and those who did not (P ⫽ .8). Primary operations: laparoscopic versus open The index bariatric procedures were performed at the University of Minnesota Medical Center. All patients underwent RYGB; 1051 (86%) and 171 (14%) were performed using the laparoscopic and open technique, respectively. The patients who presented for ED visits, readmissions, or reoperations ⬍90 days after RYGB were more likely to have undergone open RYGB (n ⫽ 37) than laparoscopic RYGB (n ⫽ 136) compared with those not presenting to the ED or requiring readmission or reoperation
Table 1 Frequency of co-morbid illness in 173 patients Co-morbid disease
n (%)
Musculoskeletal Hypertension Obstructive sleep apnea Dyslipidemia Gastroesophageal reflux disease Lower back pain Depression Type 2 diabetes mellitus Asthma Skin fold rashes
99 (57) 97 (56) 90 (52) 81 (47) 74 (43) 59 (34) 57 (33) 53 (31) 42 (24) 24 (14)
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
(open technique, n ⫽ 134; laparoscopic technique, n ⫽ 915; P ⫽ .002). Overall rate of ED visits, readmissions, and reoperations occurring ⬍90 days Of the 1222 consecutive patients who underwent RYGB, 173 required 252 ED visits, hospital readmissions, or reoperations and 1049 did not. Thus, the all-cause rate of ED visit, readmission, and reoperation occurring within 90 days of the primary operation was 21% (252 of 1222). Of these 252 ED visits, admissions, or reoperations, 127 (50%) occurred within the first 30 days postoperatively, 75 (30%) occurred 31– 60 days after RYGB, and 50 (20%) occurred 61–90 days after RYGB. Visits per patient A total of 252 ED visits, readmissions, or reoperations were required by 173 patients for an average of 1.5 visits per patient. Of these 173 patients, 120 were seen once, 34 twice, 12 were seen 3 times, and 7 were seen 4 times. Location of patient ED visits, readmissions, and reoperations Of the 252 patient ED visits, readmissions, or reoperations, 147 (58%) were to the primary institution and 105 (42%) were to a local or regional hospital. A tendency was seen for patients to stop returning to the primary medical center with time, with 63% returning to the primary medical
419
center during the ⬍30-day interval, 56% during the 31– 60day interval, and 50% during the 61–90-day interval. However, this trend was not statistically significant. Type of visit occurring within 90 days of operation Of the 252 ED visits, readmissions, reoperations, and SDS visits occurring within 90 days of the primary surgery, 79 (31%) were ED visits not resulting in readmission, 164 (65%) were readmissions to the hospital, 35 (14%) were reoperations, and 9 (4%) were SDS visits. ED visits, readmissions, and reoperations: related versus unrelated to RYGB Of the 79 ED visits, 48 (61%) were related to the primary operation. Of the 164 readmissions to the hospital, 126 (77%) were considered related to RYGB. Of the 35 reoperations, 25 (71%) were related to RYGB. Finally, 5 (56%) of the 9 SDS visits were considered related to RYGB. The rate of ED visits, readmission, and reoperation for indications related to RYGB for those occurring at ⬍90 days was 4%, 10%, and 2%, respectively. For those occurring at ⬍30 days, the rate was 3%, 5%, and .7% for ED visits, readmissions, and reoperations, respectively. Fig. 2 depicts the type of visit by month postoperatively and the proportion of those visits that were related or unrelated to RYGB.
Fig. 2. ED visits, readmissions, reoperations, and ambulatory SDSs (A) ⬍30 days (n ⫽ 127), (B) 31– 60 days (n ⫽ 75), and (C) 61–90 days (n ⫽ 50) after RYGB. Black bars indicate portion related to RYGB; gray bars, proportion unrelated to RYGB. (Reoperations extracted from readmissions so total number reduced by number of reoperations.)
420
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
Table 2 Indications for 252 ED visits, readmissions, or reoperations Indication
n (%)
Nausea, vomiting, dehydration Abdominal pain Wound issues Gastrointestinal bleeding Electrolyte imbalance or nutritional deficiency Marginal ulcer Pulmonary embolus Cholelithiasis Urinary tract infection Stoma stricture Anemia Deep venous thrombosis Bowel obstruction Psychiatric illness or substance abuse Pancreatitis Pneumonia Death
65 (26) 50 (20) 21 (8) 13 (5) 12 (5) 11 (4) 8 (3) 8 (3) 8 (3) 6 (2) 6 (2) 5 (2) 5 (2) 4 (2) 3 (1) 3 (1) 2 (⬍1)
ED ⫽ emergency department.
Indications for ED visits, readmissions, and reoperations The diagnoses associated with the 252 ED visits, hospital admissions, reoperations, and SDS visits within 90 days after RYGB are listed in Table 2. Nausea, vomiting, and dehydration, benign abdominal pain, and wound issues accounted for 58% of all ED visits or readmissions within the first month, 56% in the second month, and 49% in the third month after RYGB (Fig. 3). Readmission hospital length of stay On average, readmissions occurred 27.3 days (range 2– 88) postoperatively, and the mean hospital length of stay for readmitted patients was 3.3 days (range 1–16). Reoperations Of the 1222 patients undergoing RYGB, 35 required reoperation for an overall rate of reoperation within 90 days after RYGB of 3%. Of these 35 reoperations, 25 (71%) were related to RYGB and 10 (29%) were not. Of the 25 related reoperations, 9 (36%) occurred within the first 30 days, 7 (28%) within the second 30 days, and 8 (32%) within the final 30 days after RYGB. All 9 reoperations within the first 30 days after RYGB were related to the primary operation; the most frequent reoperation during this period was abdominal exploration (laparoscopic or open, n ⫽ 6). Of the 12 reoperations performed 31– 60 days after RYGB, 7 were related. The most frequent reoperations related to RYGB during this interval were wound related (n ⫽ 3), followed by cholecystectomy (n ⫽ 2). During the 61–90-day interval after RYGB, 14 reoperations were performed, 9 of which were related to RYGB. The most frequent related reoperation during this period was diagnostic laparoscopy/laparotomy (n ⫽ 4). Reoperations not associated with RYGB were
mainly gynecologic and orthopedic related. Table 3 summarizes the reoperations ⬍90 days after RYGB that were related and unrelated to the primary procedure. Same-day surgery A total of 9 SDSs were performed within 90 days of RYGB, 5 of which were considered related and 4 unrelated to RYGB. None occurred ⬍30 days after the primary bariatric operation, 5 occurred 31– 60 days after RYGB, and 4 occurred 61–90 days after RYGB. The most frequent reoperation in the ambulatory setting related to RYGB was wound related (n ⫽ 4), and the most frequent unrelated procedure was gynecologic related (n ⫽ 2; Table 3). Effect of employment and disability status More unemployed patients were seen in the ED or readmitted compared with employed patients (P ⫽ .036). Those who were unemployed or disabled were more likely to have been seen in the ED or readmitted compared with employed or nondisabled patients (P ⫽ .036). When considering all patients who were retired, disabled, or unemployed, the level of statistical significance increased (P ⫽ .010). Discussion The increase in bariatric surgery for the management of the obesity epidemic in the United States has resulted in increased scrutiny of the safety and cost of such procedures, led principally by insurance payors and individual states. However, it is not possible to get a clear sense of the outcomes using claims data alone. Consequently, it is imperative to provide accurate data to assess the origin of the increased cost associated with bariatric surgery and to help provide insight into potential methods for lowering this cost. Of principle concern is the readmission and reoperation rates associated with the primary operation. Because RYGB is the most commonly performed weight loss procedure for morbid obesity in the United States, it was logical
Fig. 3. Nausea, vomiting, and dehydration, benign abdominal pain, and wound issues accounted for 58% of all ED visits or readmissions within first month, 56% in second month, and 49% in third month after RYGB.
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
421
Table 3 Related and unrelated reoperations in patients after Roux-en-Y gastric bypass Follow-up (d)
⬍30 (n ⫽ 9)
31–60 (n ⫽ 12)
61–90 (n ⫽ 14)
Reoperations
SDS
Related (n ⫽ 20)
Unrelated (n ⫽ 6)
Related (n ⫽ 5)
Unrelated (n ⫽ 4)
Cholecystectomy (1) Diagnostic laparotomy (2) Duodenal ulcer perforation repair (1) Gastrojejunal anastomosis perforation repair (1) Small bowel resection (2) Stenting for obstructing ureteropelvic junction stone (1) Dilation of gastrojejunal stoma (1) Dilation of gastrojejunal stoma (1) Cholecystectomy (1) Diagnostic laparoscopy with cholecystectomy (1) I&D of abdominal wound (1) Diagnostic laparoscopy with cholecystectomy (3) Diagnostic laparotomy with adhesiolysis (1) Dilation of gastrojejunal stoma (1) Cholecystectomy (2)
None (0)
None (0)
None (0)
Breast lumpectomy (1) Rotator cuff repair (1) TAH-BSO with bilateral pelvic and para-aortic lymphadenectomy (1) Total hip arthroplasty (1) Total knee arthroplasty (1) D&C (1)
Gastropexy ring removal (1) I&D of abdominal wound (2)
Cautery of nasal telangiectasis (1) D&C, cervical polyp removal (1)
I&D of abdominal wound (2)
Intraocular lens implant (1) Vaginal wall mass excision (1)
SDS ⫽ same-day surgery; I&D ⫽ irrigation and debridement; D&C ⫽ dilation and curettage; TAH-BSO ⫽ total abdominal hysterectomy with bilateral salpingo-oophorectomy. Data in parentheses are numbers.
to focus on the readmission and reoperation rates associated with this procedure. The present study is unique in several ways. First, we have reported on the readmission and reoperation rates within the first 90 days after RYGB. In contrast, most studies have reported on the 30-day readmission and reoperation rates. Second, our study has taken into account ED visits without readmission, hospital readmission, reoperation, and elective reoperations in an ambulatory surgery setting, all of which can affect the cost associated with bariatric surgery. Third, the present study is the largest analysis of its kind examining RYGB readmission and reoperation rates using clinical data from the EMR. Finally, the present study is the first, to our knowledge, to incorporate data from the EMRs of not only the primary medical center at which patients underwent the index bariatric procedure, but also the EMRs of the local and regional hospitals, with subsequent access to the medical records encompassing a moderate-size metropolitan area. We found that ED visits, readmissions, and reoperations occurred an average of 27.3 days postoperatively. Consequently, studies examining only readmissions and reoperations occurring ⬍30 days postoperatively could be missing a significant number of patients. However, this finding has not been universal. In their study examining readmission rates within 30 days after surgery that included 3 different bariatric procedures, Saunders et al. [9] noted that most patients were readmitted within 2 weeks after discharge from the hospital.
Using regional, multi-institutional EMR data to postoperatively follow-up 1222 consecutive patients who had undergone RYGB, we found an overall all-cause rate of ED visits, readmissions, and reoperations of 21% within 90 days of the index RYGB. When separated, the ⬍90-day ED visit rate was 7%, the readmission rate was 11%, and the reoperation rate was 4%. A direct comparison was not possible because no previous studies have reported the ⬍90-day data or have not reported the rates that included ED visits. However, these rates are within the realm of what would be expected. Published studies have reported ⬍30-day readmission rates of .6 –11.3% [9 –16]. None of these studies used EMR data from multiple local and regional hospitals. We noted a ⬍30-day overall all-cause rate of ED visits, readmissions, and reoperations after RYGB of 11%. When analyzed further, the ⬍30-day ED visit rate was 4%, the readmission rate was 7%, and the reoperation rate was .7%. It was expected that the ⬍30-day readmission rate would be greater than what has been reported in published studies for other EMR-based studies of readmission and reoperation because of the inclusion of multiple regional databases. The 7% rate strictly for readmissions supports the accuracy of the previous studies. However, patients might tend to return to the primary medical center early after surgery, with later ED visits and readmission more likely to occur at other hospitals. Our data have supported this hypothesis, because 63% returned to the primary medical center during the ⬍30-
422
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
day interval, 56% during the 31– 60-day interval, and 50% during the 61–90-day interval. Although this trend was not statistically significant, such a tendency was not unexpected; thus, the accuracy of studies examining ⬎90-day readmission and reoperation will likely be compromised if a multi-institutional approach has not been considered. The addition of the rates of ED visits to the rates of readmissions and reoperations has provided a more accurate picture of the potential costs associated with bariatric procedures. Because many of the issues leading to ED visits also lead to readmission, it follows that examining the rates and causes of ED visits not leading to readmission would provide important information that could be used to avoid future ED visits and readmissions. In an attempt to identify avoidable causes of ED visits and readmissions, we analyzed the indications associated with the ED visits and readmissions within 90 days of RYGB. Nausea, vomiting, dehydration, benign abdominal pain (as defined by no etiology identified, normal radiographic study findings, and no systemic illness), and wound issues accounted for 58% of all ED visits or readmissions within the first month, 56% in the second month, and 49% in the third month after RYGB. These finding were similar to those of other studies reporting on readmission ⬍30 days after bariatric surgery [9,11,12]. Saunders et al. [9] used EMR data to report on various bariatric procedures, including 1185 RYGB patients. Gastrointestinal complaints, including vomiting, dehydration, and abdominal pain was the cause for readmission in nearly one third of all readmitted RYGB patients, and wound issues was the cause in 5.4% of readmitted patients. In that study, the results were stratified by the bariatric procedure. RYGB had a readmission rate that was greater than both laparoscopic adjustable gastric banding and open conversion of vertical banded gastroplasty to RYGB. The complex of nausea, vomiting, and consequent dehydration is common after RYGB [17]. The etiology and true incidence of nonmechanical persistent nausea and vomiting in postoperative RYGB patients is unknown. Potential causes of nonmechanical nausea and vomiting include dysfunctional eating habits such as those that result in dumping syndrome, medication intolerance, and thiamine deficiency. In the absence of these diagnoses, it could be that nausea and vomiting that persists after RYGB develops from a central cause resulting from interruption of the neurohormonal circuitry between the gut and the brain. Although the beneficial aspect of this change can be satiety or lack of hunger, persistent nausea and vomiting could represent an undesirable consequence in some patients. This theory is speculative, and, clearly, more research is needed to better define the etiology of nonmechanical persistent nausea and vomiting after RYGB. Notable differences between our findings and those of other published studies include the incidence of gastrojeju-
nal stoma stricture and bowel obstruction ⬍30 days after RYGB. Gastrojejunal stoma stricture and bowel obstruction comprised 15.1% and 10.5% of all readmissions, respectively, in one study [9]. In the present study, 2.4% of the ED visits, readmissions, or reoperations were for gastrojejunal stoma stricture and 2.0% for bowel obstruction. It has been argued that the most accurate data come from review of the EMRs [18,19]. Although dependable in diagnosis and procedure description, a major weakness of studies that have used data from the EMRs is that only the EMRs of the institution where the index bariatric procedure was performed have been included in the analysis. In the present study, 42% of patients received their healthcare at a medical center other than the location of their primary bariatric operation. The degree to which this occurs probably depends on the geographic location and the density of the healthcare resources in the region. However, this large percentage of patients not returning to the primary medical center suggests that previous reports using data from only the primary medical center’s EMRs might have been underreporting. Nevertheless, as previously noted, the ⬍30day readmission rate of the present study agreed with what has been previously reported. Alternatively, multi-institutional administrative data using diagnosis and procedure codes encompass a large number of sites at which patients receive medical care. The main benefit of these studies is access to large numbers of patients. A critical analysis of administrative data has arisen in response to the use of these types of data in the introduction of hospital and physician report cards, and its accuracy has been questioned, particularly with respect to the correct interpretation of the clinical situation [18 –21]. We found that patients who were unemployed, disabled, or retired were more likely to be seen in the ED or readmitted within 90 days of RYGB, suggesting a socioeconomic component to the probability of ED visits or readmission. Thus, early identification of these patients with appropriate social services interventions could decrease unnecessary ED visits and readmissions for this group of patients. A limitation of this study is that the geographic residence of the patients relative to the primary institution is unknown. It can be assumed that the further away a patient resides, the less likely it is that the patient will return to the primary institution. Although the present study was more inclusive than most in this regard, it remains probable that at least some patients residing more distantly from the primary institution visited a local ED. It is less likely that a significant number of readmissions and reoperations were missed, because it is the general practice in our region to return patients with significant medical issues back to the primary institution. However, given this potential shortcoming, combining EMR data with administrative data may well provide optimal accuracy.
J. J. Gleysteen / Surgery for Obesity and Related Diseases 5 (2009) 423– 424
Conclusion The results of our study have shown that nausea, vomiting, and dehydration are frequent reasons for ED visits and readmission within 3 months of RYGB. The unemployed and disabled were more likely to present to the ED or be readmitted. Outpatient support systems specifically designed to provide intravenous fluids and antiemetic therapy might reduce the number of ED visits and readmissions. In addition, social service interventions for the unemployed or disabled patients could reduce unnecessary ED visits and readmissions.
Disclosures The authors claim no commercial associations that might be a conflict of interest in relation to this article.
References [1] Salem L, Devlin A, Sullivan SD, Flum DR. Cost-effectiveness analysis of laparoscopic gastric bypass, adjustable gastric banding, and non-operative weight loss interventions. Surg Obes Relat Dis 2008; 4:26 –32. [2] Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292:1724 –37. [3] Christou NV, Sampalis JS, Liberman M, et al. Surgery decreases long-term mortality, morbidity, and health care use in morbidly obese patients. Ann Surg 2004;240:416 –23. [4] Flum DR, Dellinger EP. Impact of gastric bypass operation on survival: a population-based analysis. J Am Coll Surg 2004;199:543–51. [5] Craig BM, Tseng DS. Cost-effectiveness of gastric bypass for severe obesity. Am J Med 2002;113:491– 8. [6] Naessens JM, Ruud KL, Tulledge-Scheitel SM, Stroebel RJ, Cabanela RL. Comparison of provider claims data versus medical records review for assessing provision of adult preventive services. J Ambul Care Manage 2008;31:178 – 86. [7] Saunders JK, Ballantyne GH, Belsey S, et al. One-year readmission rates at a high volume bariatric surgery center: laparoscopic adjustable gastric banding, laparoscopic gastric bypass, and vertical banded gastroplasty–Roux-en-Y gastric bypass. Obes Surg 2008;18:1233–40.
423
[8] Pratt GM, McLees B, Pories WJ. The ASBS Bariatric Surgery Centers of Excellence program: a blueprint for quality improvement. Surg Obes Relat Dis 2006;2:497–503. [9] Saunders JK, Ballantyne GH, Belsley S, et al. 30-Day readmission rates at a high volume bariatric surgery center: laparoscopic adjustable gastric banding, laparoscopic gastric bypass, and vertical banded gastroplasty–Roux-en-Y gastric bypass. Obes Surg 2007;17:1171–7. [10] Nguyen NT, Paya M, Stevens CM, Mavandadi S, Zainabadi K, Wilson SE. The relationship between hospital volume and outcome in bariatric surgery at academic medical centers. Ann Surg 2004;240: 586 –94. [11] Weller WE, Rosati C, Hannan EL. Relationship between surgeon and hospital volume and readmission after bariatric operation. J Am Coll Surg 2007;204:383–91. [12] Encinosa WE, Bernard DM, Chen CC, Steiner CA. Healthcare utilization and outcomes after bariatric surgery. Med Care 2006;44: 706 –12. [13] McCarty TM, Arnold DT, Lamont JP, Fisher TL, Kuhn JA. Optimizing outcomes in bariatric surgery: outpatient laparoscopic gastric bypass. Ann Surg 2005;242:494 – 8. [14] Mehrotra C, Serdula M, Naimi TS, Khan LK, Miller J, Dietz W. Population-based study of trends, costs, and complications of weight loss surgeries from 1990 to 2002. Obes Res 2005;13:2029 –34. [15] Bradley DW, Sharma BK. Centers of excellence in bariatric surgery: design, implementation, and one-year outcomes. Surg Obes Relat Dis 2006;2:513–7. [16] Baker MT, Lara MD, Larson CJ, Lambert PJ, Mathiason MA, Kothari SN. Length of stay and impact on readmission rates after laparoscopic gastric bypass. Surg Obes Relat Dis 2006;2:435–9. [17] Shikora SA, Kim JJ, Tarnoff ME. Nutritional and gastrointestinal complications of bariatric surgery. Nutr Clin Pract 2007;22:29 – 40. [18] Shahian DM, Silverstein T, Lovett AF, Wolf RE, Normand SL. Comparison of clinical and administrative data sources for hospital coronary artery bypass graft surgery report cards. Circulation 2007; 115:1518 –27. [19] Maclean JR, Fick DM, Hoffman WK, King CT, Lough ER, Waller JL. Comparison of 2 systems for clinical practice profiling in diabetic care: medical records versus claims and administrative data. Am J Manag Care 2002;8:175–9. [20] Romano PS, Schembri ME, Rainwater JA. Can administrative data be used to ascertain clinically significant postoperative complications? Am J Med Qual 2002;17:145–54. [21] Best WR, Khuri SF, Phelan M, Hur K. Identifying patient preoperative risk factors and postoperative adverse events in administrative databases: results from the Department of Veterans Affairs National Surgical Quality Improvement Program. J Am Coll Surg 2002;194:257– 66.
Editorial comment
Comment on: Patterns of readmission and reoperation within 90 days after Roux-en-Y gastric bypass The authors from the University of Minnesota [1] have compiled and analyzed data from their medical center that shows the incidence and continuing occurrence of patient return to medical facilities during the first 3 months after successful bariatric surgery (primary Rouxen-Y gastric bypass, in particular). The national relevance of this report lies in revealing a postoperative need to seek medical help that has been a frequent financial concern and drain on health insurance carriers.
During recent years, most federal and commercial insurance carriers have developed respect for the health advantages and appropriateness of surgical care of morbid obesity. Although willing to pay the primary costs to counter the significant medical costs of long-term comorbidities, they have been alarmed by the seemingly high secondary—postoperative— costs and have searched for data to clarify its magnitude. That alarm has caused a reluctance to approve surgical plans and, in
Original article
Patterns of readmission and reoperation within 90 days after Roux-en-Y gastric bypass Todd Andrew Kellogg, M.D., Therese Swan, B.S., Daniel A. Leslie, M.D., Henry Buchwald, M.D., Ph.D., Sayeed Ikramuddin, M.D. Division of Gastrointestinal Surgery, Department of Surgery, University of Minnesota Medical Center, Minneapolis, Minnesota Received June 20, 2008; revised November 29, 2008; accepted January 20, 2009
Abstract
Background: Health insurance payors harbor concerns regarding the cost of bariatric procedures that are chiefly related to early readmissions and reoperations. We have attempted to identify the avoidable causes of readmission. Methods: We retrospectively reviewed the indications for short-term (⬍90-d) emergency department (ED) visits, readmissions, and reoperations from August 2004 through May 2007 for patients undergoing primary Roux-en-Y gastric bypass (RYGB) for morbid obesity at a tertiary care teaching hospital. The electronic medical record of the primary hospital was reviewed, as well as the electronic medical records of 9 local hospitals serving the area, allowing the incorporation of data from 35 locoregional hospitals. Results: A total of 1222 consecutive patients underwent RYGB, 1051 laparoscopically. Of these 1222 patients, 173 had 252 ED visits, readmissions, and/or reoperations; 147 (58%) visits were to the primary institution and 105 (42%) occurred at a local or regional hospital. No age difference was found between the patients who underwent ED visits, hospital readmissions, or reoperations and those who did not (mean age 43 yr for both groups, P ⬎ .05). Patients who were seen in the ED, readmitted to the hospital, or underwent reoperation had had a greater body mass index (50 kg/m2 versus 48 kg/m2, P ⫽ .001). On average, the readmissions occurred 27.3 days (range 2– 88) postoperatively, and the mean hospital length of stay for readmitted patients was 3.3 days (range 1–16). Patients who presented for ED visits, readmission, or reoperations were more likely to have undergone open RYGB than laparoscopic RYGB (P ⫽ .002). The ⬍90-day all-cause ED visit, readmission, and reoperation rate was 21% (n ⫽ 252). Considering all 1222 patients, the incidence of nausea, vomiting, and dehydration, abdominal pain, and wound issues was 5% (n ⫽ 65), 4% (n ⫽ 50), and 2% (n ⫽ 21), respectively. Considering only the 173 patients with ED visits, readmissions, or reoperations (n ⫽ 252), the admitting diagnosis was nausea, vomiting, and dehydration in 26%, abdominal pain in 20%, and wound issues in 8%. The unemployed, disabled, or retired were more likely to have been seen in the ED or readmitted compared with the employed, nondisabled, or not retired (P ⫽ .01). Conclusion: A considerable number of patients are affected by nausea, vomiting, and dehydration, abdominal pain, and wound issues ⬍90 days postoperatively. Socioeconomic and functional status might have an effect on the rate of ED visits and readmissions. By ensuring that the appropriate outpatient mechanisms for management of these problems are available, early ED visits and readmission rates should significantly decrease. (Surg Obes Relat Dis 2009;5:416 – 424.) © 2009 American Society for Metabolic and Bariatric Surgery. All rights reserved.
Keywords:
Roux-en-Y gastric bypass; Readmission; Reoperation; Rates; Outcomes; Complications
Presented at the 25th Annual Meeting of the American Society for Metabolic and Bariatric Surgery, June 15–20, 2008, Washington, DC. Reprints not available from the authors. 1550-7289/09/$ – see front matter © 2009 American Society for Metabolic and Bariatric Surgery. All rights reserved. doi:10.1016/j.soard.2009.01.008
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
The reported outcomes for Roux-en-Y gastric bypass (RYGB) performed for morbid obesity have been extremely good. Evidence is increasing that bariatric surgery improves the overall health and longevity of morbidly obese patients [1– 4]. Despite these beneficial outcomes, concerns regarding the cost/benefit ratio of bariatric care have been expressed by health insurance payors and others [3,5]. Some aspects of cost that can be easily overlooked are the readmission and reoperation rates associated with bariatric surgery. Little is known about the need for readmission, reoperation, or emergency department (ED) visits for these patients. A major barrier to this type of analysis is that patients with ED issues often do not return to the institution where the primary operation was performed. This can result in gross underreporting of rates of readmission and reoperation, as well as redundant testing and unwarranted admissions. Another common form of readmission and reoperation reporting comes from multi-institutional administrative data, which is subject to significant inaccuracies based on its collection, not for clinical or research purposes, but for insurance analyses [6]. Moreover, studies examining readmission and reoperation are often limited to the first 30 days after operation, which is considered the interval for procedure-related complications. A review of our data (Fig. 1) and data from others [7] has suggested that this limitation might not capture all relevant procedure-related ED visits, readmissions, and reoperations. The objective of this study was to determine, with as much accuracy as possible, the frequency of ED visits, readmissions, and reoperations for patients who had undergone RYGB and to identify potentially avoidable causes.
417
Methods Study design This study is a retrospective analysis of the data from 1222 consecutive patients who had undergone laparoscopic or open RYGB as the index procedure. The records from patients undergoing primary bariatric surgery from August 1, 2004 to May 31, 2007 were reviewed. Of these patients, 977 (80%) were women and 245 (20%) were men. The cohort of patients who underwent ED visits, readmission, or reoperation was identified and analyzed further. All index bariatric operations were performed at the University of Minnesota Medical Center in an American Society for Metabolic and Bariatric Surgery Centers of Excellence program [8]. After obtaining permission from the institutional review board, the data were collected by review of the electronic medical records (EMRs) from the University of Minnesota Medical Center, the tertiary care teaching hospital at which patients underwent their index bariatric operation (primary medical center), as well as a comprehensive review of the EMRs of the 9 local hospitals serving patients from the 3 dominant health insurance payors that serve the region. This approach allowed the incorporation of information from 35 regional hospitals and included ED and ambulatory same-day surgery (SDS) visits, in addition to hospital readmissions and reoperations. Scheduled bariatric follow-up appointments or patient clinic visits to a primary care provider were not included. The data from these patients were then entered and maintained in a bariatric database.
Fig. 1. Frequency of ED visits, readmissions, and reoperations after primary RYGB at University of Minnesota (unpublished data).
418
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
Definitions of readmission and reoperation
Results
Readmissions were defined as any admission to the primary medical center or a local or regional hospital for any reason, including ED visits, if this resulted in hospitalization. Reoperations were defined as operations performed after discharge from the hospital for any reason, either related or unrelated to the index bariatric procedure, and included operations associated with readmission and SDS. ED visits not resulting in hospital admission were defined as such and not as an admission.
Patient demographics
Definitions of related and unrelated to RYGB ED visits, readmissions, and reoperations were considered related to RYGB if the medical condition was directly or indirectly related to RYGB for weight loss or its known complications and adverse effects, both short and long term. This included all abdominal pain, except when confirmed to be of a known unrelated cause, all nausea and vomiting, cholecystitis and cholecystectomy, nephrolithiasis and associated procedures, and early (⬍30 d postoperatively) pancreatitis. The cause was considered unrelated if associated with an exacerbation of a previous medical condition or when no known association of RYGB with the medical condition requiring the ED visit, readmission, or reoperation was found. Gastrointestinal hemorrhage associated with a marginal ulcer was considered related at any point postoperatively. Gastrointestinal hemorrhage associated with iatrogenic coagulopathy not occurring in the immediate postoperative period was considered unrelated.
Operations Three surgeons performed all RYGBs using either the open (n ⫽ 171) or laparoscopic (n ⫽ 1051) technique. A partially hand-sewn, partially linearly stapled, 2-layer gastrojejunal anastomosis, with a Roux limb length that varied from 75 to 150 cm and biliopancreatic limb length that ranged from 30 to 100 cm, was used. Revision operations were excluded from the analysis.
Statistical analysis The data are presented as the median and range. Groups were compared using Pearson’s chi-square analysis. A chisquare probability of ⱕ .05 was interpreted as justification for rejecting the null hypothesis. The alternate hypothesis was not rejected when the variables had an associated relationship. All statistical operations were performed using MS Excel.
Of the 173 patients seen in the ED, readmitted, or who underwent reoperation, 137 were women and 36 were men. No difference was found in gender between the patients requiring ED visits, readmission, or reoperation and those who did not (P ⫽ .9). The mean patient age was 42 years (range 15–71). The mean preoperative body mass index was 50 kg/m2 (range 35–96). Of the 173 patients, 125 were white, 27 were black, 4 were Hispanic, and 17 were of unknown ethnicity. No difference was found in patient age between those who underwent ED visits, hospital readmissions, or reoperations and those who did not (43 versus 43 yr, respectively; P ⬎ .05). The patients who required ED visits, hospital readmissions, or reoperations had a significantly greater body mass index (50 versus 48 kg/m2, respectively; P ⫽ .001). Compared with whites, African Americans were more likely to require ED visits, readmissions or reoperations (P ⫽ .03). Co-morbidities The most frequent preoperative co-morbidities among the 173 patients are listed in Table 1. No statistically significant difference was found in the incidence of co-morbid disease, either in combination or independently, between those who required ED visits, readmissions, or reoperations and those who did not (P ⫽ .8). Primary operations: laparoscopic versus open The index bariatric procedures were performed at the University of Minnesota Medical Center. All patients underwent RYGB; 1051 (86%) and 171 (14%) were performed using the laparoscopic and open technique, respectively. The patients who presented for ED visits, readmissions, or reoperations ⬍90 days after RYGB were more likely to have undergone open RYGB (n ⫽ 37) than laparoscopic RYGB (n ⫽ 136) compared with those not presenting to the ED or requiring readmission or reoperation
Table 1 Frequency of co-morbid illness in 173 patients Co-morbid disease
n (%)
Musculoskeletal Hypertension Obstructive sleep apnea Dyslipidemia Gastroesophageal reflux disease Lower back pain Depression Type 2 diabetes mellitus Asthma Skin fold rashes
99 (57) 97 (56) 90 (52) 81 (47) 74 (43) 59 (34) 57 (33) 53 (31) 42 (24) 24 (14)
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
(open technique, n ⫽ 134; laparoscopic technique, n ⫽ 915; P ⫽ .002). Overall rate of ED visits, readmissions, and reoperations occurring ⬍90 days Of the 1222 consecutive patients who underwent RYGB, 173 required 252 ED visits, hospital readmissions, or reoperations and 1049 did not. Thus, the all-cause rate of ED visit, readmission, and reoperation occurring within 90 days of the primary operation was 21% (252 of 1222). Of these 252 ED visits, admissions, or reoperations, 127 (50%) occurred within the first 30 days postoperatively, 75 (30%) occurred 31– 60 days after RYGB, and 50 (20%) occurred 61–90 days after RYGB. Visits per patient A total of 252 ED visits, readmissions, or reoperations were required by 173 patients for an average of 1.5 visits per patient. Of these 173 patients, 120 were seen once, 34 twice, 12 were seen 3 times, and 7 were seen 4 times. Location of patient ED visits, readmissions, and reoperations Of the 252 patient ED visits, readmissions, or reoperations, 147 (58%) were to the primary institution and 105 (42%) were to a local or regional hospital. A tendency was seen for patients to stop returning to the primary medical center with time, with 63% returning to the primary medical
419
center during the ⬍30-day interval, 56% during the 31– 60day interval, and 50% during the 61–90-day interval. However, this trend was not statistically significant. Type of visit occurring within 90 days of operation Of the 252 ED visits, readmissions, reoperations, and SDS visits occurring within 90 days of the primary surgery, 79 (31%) were ED visits not resulting in readmission, 164 (65%) were readmissions to the hospital, 35 (14%) were reoperations, and 9 (4%) were SDS visits. ED visits, readmissions, and reoperations: related versus unrelated to RYGB Of the 79 ED visits, 48 (61%) were related to the primary operation. Of the 164 readmissions to the hospital, 126 (77%) were considered related to RYGB. Of the 35 reoperations, 25 (71%) were related to RYGB. Finally, 5 (56%) of the 9 SDS visits were considered related to RYGB. The rate of ED visits, readmission, and reoperation for indications related to RYGB for those occurring at ⬍90 days was 4%, 10%, and 2%, respectively. For those occurring at ⬍30 days, the rate was 3%, 5%, and .7% for ED visits, readmissions, and reoperations, respectively. Fig. 2 depicts the type of visit by month postoperatively and the proportion of those visits that were related or unrelated to RYGB.
Fig. 2. ED visits, readmissions, reoperations, and ambulatory SDSs (A) ⬍30 days (n ⫽ 127), (B) 31– 60 days (n ⫽ 75), and (C) 61–90 days (n ⫽ 50) after RYGB. Black bars indicate portion related to RYGB; gray bars, proportion unrelated to RYGB. (Reoperations extracted from readmissions so total number reduced by number of reoperations.)
420
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
Table 2 Indications for 252 ED visits, readmissions, or reoperations Indication
n (%)
Nausea, vomiting, dehydration Abdominal pain Wound issues Gastrointestinal bleeding Electrolyte imbalance or nutritional deficiency Marginal ulcer Pulmonary embolus Cholelithiasis Urinary tract infection Stoma stricture Anemia Deep venous thrombosis Bowel obstruction Psychiatric illness or substance abuse Pancreatitis Pneumonia Death
65 (26) 50 (20) 21 (8) 13 (5) 12 (5) 11 (4) 8 (3) 8 (3) 8 (3) 6 (2) 6 (2) 5 (2) 5 (2) 4 (2) 3 (1) 3 (1) 2 (⬍1)
ED ⫽ emergency department.
Indications for ED visits, readmissions, and reoperations The diagnoses associated with the 252 ED visits, hospital admissions, reoperations, and SDS visits within 90 days after RYGB are listed in Table 2. Nausea, vomiting, and dehydration, benign abdominal pain, and wound issues accounted for 58% of all ED visits or readmissions within the first month, 56% in the second month, and 49% in the third month after RYGB (Fig. 3). Readmission hospital length of stay On average, readmissions occurred 27.3 days (range 2– 88) postoperatively, and the mean hospital length of stay for readmitted patients was 3.3 days (range 1–16). Reoperations Of the 1222 patients undergoing RYGB, 35 required reoperation for an overall rate of reoperation within 90 days after RYGB of 3%. Of these 35 reoperations, 25 (71%) were related to RYGB and 10 (29%) were not. Of the 25 related reoperations, 9 (36%) occurred within the first 30 days, 7 (28%) within the second 30 days, and 8 (32%) within the final 30 days after RYGB. All 9 reoperations within the first 30 days after RYGB were related to the primary operation; the most frequent reoperation during this period was abdominal exploration (laparoscopic or open, n ⫽ 6). Of the 12 reoperations performed 31– 60 days after RYGB, 7 were related. The most frequent reoperations related to RYGB during this interval were wound related (n ⫽ 3), followed by cholecystectomy (n ⫽ 2). During the 61–90-day interval after RYGB, 14 reoperations were performed, 9 of which were related to RYGB. The most frequent related reoperation during this period was diagnostic laparoscopy/laparotomy (n ⫽ 4). Reoperations not associated with RYGB were
mainly gynecologic and orthopedic related. Table 3 summarizes the reoperations ⬍90 days after RYGB that were related and unrelated to the primary procedure. Same-day surgery A total of 9 SDSs were performed within 90 days of RYGB, 5 of which were considered related and 4 unrelated to RYGB. None occurred ⬍30 days after the primary bariatric operation, 5 occurred 31– 60 days after RYGB, and 4 occurred 61–90 days after RYGB. The most frequent reoperation in the ambulatory setting related to RYGB was wound related (n ⫽ 4), and the most frequent unrelated procedure was gynecologic related (n ⫽ 2; Table 3). Effect of employment and disability status More unemployed patients were seen in the ED or readmitted compared with employed patients (P ⫽ .036). Those who were unemployed or disabled were more likely to have been seen in the ED or readmitted compared with employed or nondisabled patients (P ⫽ .036). When considering all patients who were retired, disabled, or unemployed, the level of statistical significance increased (P ⫽ .010). Discussion The increase in bariatric surgery for the management of the obesity epidemic in the United States has resulted in increased scrutiny of the safety and cost of such procedures, led principally by insurance payors and individual states. However, it is not possible to get a clear sense of the outcomes using claims data alone. Consequently, it is imperative to provide accurate data to assess the origin of the increased cost associated with bariatric surgery and to help provide insight into potential methods for lowering this cost. Of principle concern is the readmission and reoperation rates associated with the primary operation. Because RYGB is the most commonly performed weight loss procedure for morbid obesity in the United States, it was logical
Fig. 3. Nausea, vomiting, and dehydration, benign abdominal pain, and wound issues accounted for 58% of all ED visits or readmissions within first month, 56% in second month, and 49% in third month after RYGB.
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
421
Table 3 Related and unrelated reoperations in patients after Roux-en-Y gastric bypass Follow-up (d)
⬍30 (n ⫽ 9)
31–60 (n ⫽ 12)
61–90 (n ⫽ 14)
Reoperations
SDS
Related (n ⫽ 20)
Unrelated (n ⫽ 6)
Related (n ⫽ 5)
Unrelated (n ⫽ 4)
Cholecystectomy (1) Diagnostic laparotomy (2) Duodenal ulcer perforation repair (1) Gastrojejunal anastomosis perforation repair (1) Small bowel resection (2) Stenting for obstructing ureteropelvic junction stone (1) Dilation of gastrojejunal stoma (1) Dilation of gastrojejunal stoma (1) Cholecystectomy (1) Diagnostic laparoscopy with cholecystectomy (1) I&D of abdominal wound (1) Diagnostic laparoscopy with cholecystectomy (3) Diagnostic laparotomy with adhesiolysis (1) Dilation of gastrojejunal stoma (1) Cholecystectomy (2)
None (0)
None (0)
None (0)
Breast lumpectomy (1) Rotator cuff repair (1) TAH-BSO with bilateral pelvic and para-aortic lymphadenectomy (1) Total hip arthroplasty (1) Total knee arthroplasty (1) D&C (1)
Gastropexy ring removal (1) I&D of abdominal wound (2)
Cautery of nasal telangiectasis (1) D&C, cervical polyp removal (1)
I&D of abdominal wound (2)
Intraocular lens implant (1) Vaginal wall mass excision (1)
SDS ⫽ same-day surgery; I&D ⫽ irrigation and debridement; D&C ⫽ dilation and curettage; TAH-BSO ⫽ total abdominal hysterectomy with bilateral salpingo-oophorectomy. Data in parentheses are numbers.
to focus on the readmission and reoperation rates associated with this procedure. The present study is unique in several ways. First, we have reported on the readmission and reoperation rates within the first 90 days after RYGB. In contrast, most studies have reported on the 30-day readmission and reoperation rates. Second, our study has taken into account ED visits without readmission, hospital readmission, reoperation, and elective reoperations in an ambulatory surgery setting, all of which can affect the cost associated with bariatric surgery. Third, the present study is the largest analysis of its kind examining RYGB readmission and reoperation rates using clinical data from the EMR. Finally, the present study is the first, to our knowledge, to incorporate data from the EMRs of not only the primary medical center at which patients underwent the index bariatric procedure, but also the EMRs of the local and regional hospitals, with subsequent access to the medical records encompassing a moderate-size metropolitan area. We found that ED visits, readmissions, and reoperations occurred an average of 27.3 days postoperatively. Consequently, studies examining only readmissions and reoperations occurring ⬍30 days postoperatively could be missing a significant number of patients. However, this finding has not been universal. In their study examining readmission rates within 30 days after surgery that included 3 different bariatric procedures, Saunders et al. [9] noted that most patients were readmitted within 2 weeks after discharge from the hospital.
Using regional, multi-institutional EMR data to postoperatively follow-up 1222 consecutive patients who had undergone RYGB, we found an overall all-cause rate of ED visits, readmissions, and reoperations of 21% within 90 days of the index RYGB. When separated, the ⬍90-day ED visit rate was 7%, the readmission rate was 11%, and the reoperation rate was 4%. A direct comparison was not possible because no previous studies have reported the ⬍90-day data or have not reported the rates that included ED visits. However, these rates are within the realm of what would be expected. Published studies have reported ⬍30-day readmission rates of .6 –11.3% [9 –16]. None of these studies used EMR data from multiple local and regional hospitals. We noted a ⬍30-day overall all-cause rate of ED visits, readmissions, and reoperations after RYGB of 11%. When analyzed further, the ⬍30-day ED visit rate was 4%, the readmission rate was 7%, and the reoperation rate was .7%. It was expected that the ⬍30-day readmission rate would be greater than what has been reported in published studies for other EMR-based studies of readmission and reoperation because of the inclusion of multiple regional databases. The 7% rate strictly for readmissions supports the accuracy of the previous studies. However, patients might tend to return to the primary medical center early after surgery, with later ED visits and readmission more likely to occur at other hospitals. Our data have supported this hypothesis, because 63% returned to the primary medical center during the ⬍30-
422
T. A. Kellogg et al. / Surgery for Obesity and Related Diseases 5 (2009) 416 – 424
day interval, 56% during the 31– 60-day interval, and 50% during the 61–90-day interval. Although this trend was not statistically significant, such a tendency was not unexpected; thus, the accuracy of studies examining ⬎90-day readmission and reoperation will likely be compromised if a multi-institutional approach has not been considered. The addition of the rates of ED visits to the rates of readmissions and reoperations has provided a more accurate picture of the potential costs associated with bariatric procedures. Because many of the issues leading to ED visits also lead to readmission, it follows that examining the rates and causes of ED visits not leading to readmission would provide important information that could be used to avoid future ED visits and readmissions. In an attempt to identify avoidable causes of ED visits and readmissions, we analyzed the indications associated with the ED visits and readmissions within 90 days of RYGB. Nausea, vomiting, dehydration, benign abdominal pain (as defined by no etiology identified, normal radiographic study findings, and no systemic illness), and wound issues accounted for 58% of all ED visits or readmissions within the first month, 56% in the second month, and 49% in the third month after RYGB. These finding were similar to those of other studies reporting on readmission ⬍30 days after bariatric surgery [9,11,12]. Saunders et al. [9] used EMR data to report on various bariatric procedures, including 1185 RYGB patients. Gastrointestinal complaints, including vomiting, dehydration, and abdominal pain was the cause for readmission in nearly one third of all readmitted RYGB patients, and wound issues was the cause in 5.4% of readmitted patients. In that study, the results were stratified by the bariatric procedure. RYGB had a readmission rate that was greater than both laparoscopic adjustable gastric banding and open conversion of vertical banded gastroplasty to RYGB. The complex of nausea, vomiting, and consequent dehydration is common after RYGB [17]. The etiology and true incidence of nonmechanical persistent nausea and vomiting in postoperative RYGB patients is unknown. Potential causes of nonmechanical nausea and vomiting include dysfunctional eating habits such as those that result in dumping syndrome, medication intolerance, and thiamine deficiency. In the absence of these diagnoses, it could be that nausea and vomiting that persists after RYGB develops from a central cause resulting from interruption of the neurohormonal circuitry between the gut and the brain. Although the beneficial aspect of this change can be satiety or lack of hunger, persistent nausea and vomiting could represent an undesirable consequence in some patients. This theory is speculative, and, clearly, more research is needed to better define the etiology of nonmechanical persistent nausea and vomiting after RYGB. Notable differences between our findings and those of other published studies include the incidence of gastrojeju-
nal stoma stricture and bowel obstruction ⬍30 days after RYGB. Gastrojejunal stoma stricture and bowel obstruction comprised 15.1% and 10.5% of all readmissions, respectively, in one study [9]. In the present study, 2.4% of the ED visits, readmissions, or reoperations were for gastrojejunal stoma stricture and 2.0% for bowel obstruction. It has been argued that the most accurate data come from review of the EMRs [18,19]. Although dependable in diagnosis and procedure description, a major weakness of studies that have used data from the EMRs is that only the EMRs of the institution where the index bariatric procedure was performed have been included in the analysis. In the present study, 42% of patients received their healthcare at a medical center other than the location of their primary bariatric operation. The degree to which this occurs probably depends on the geographic location and the density of the healthcare resources in the region. However, this large percentage of patients not returning to the primary medical center suggests that previous reports using data from only the primary medical center’s EMRs might have been underreporting. Nevertheless, as previously noted, the ⬍30day readmission rate of the present study agreed with what has been previously reported. Alternatively, multi-institutional administrative data using diagnosis and procedure codes encompass a large number of sites at which patients receive medical care. The main benefit of these studies is access to large numbers of patients. A critical analysis of administrative data has arisen in response to the use of these types of data in the introduction of hospital and physician report cards, and its accuracy has been questioned, particularly with respect to the correct interpretation of the clinical situation [18 –21]. We found that patients who were unemployed, disabled, or retired were more likely to be seen in the ED or readmitted within 90 days of RYGB, suggesting a socioeconomic component to the probability of ED visits or readmission. Thus, early identification of these patients with appropriate social services interventions could decrease unnecessary ED visits and readmissions for this group of patients. A limitation of this study is that the geographic residence of the patients relative to the primary institution is unknown. It can be assumed that the further away a patient resides, the less likely it is that the patient will return to the primary institution. Although the present study was more inclusive than most in this regard, it remains probable that at least some patients residing more distantly from the primary institution visited a local ED. It is less likely that a significant number of readmissions and reoperations were missed, because it is the general practice in our region to return patients with significant medical issues back to the primary institution. However, given this potential shortcoming, combining EMR data with administrative data may well provide optimal accuracy.
J. J. Gleysteen / Surgery for Obesity and Related Diseases 5 (2009) 423– 424
Conclusion The results of our study have shown that nausea, vomiting, and dehydration are frequent reasons for ED visits and readmission within 3 months of RYGB. The unemployed and disabled were more likely to present to the ED or be readmitted. Outpatient support systems specifically designed to provide intravenous fluids and antiemetic therapy might reduce the number of ED visits and readmissions. In addition, social service interventions for the unemployed or disabled patients could reduce unnecessary ED visits and readmissions.
Disclosures The authors claim no commercial associations that might be a conflict of interest in relation to this article.
References [1] Salem L, Devlin A, Sullivan SD, Flum DR. Cost-effectiveness analysis of laparoscopic gastric bypass, adjustable gastric banding, and non-operative weight loss interventions. Surg Obes Relat Dis 2008; 4:26 –32. [2] Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;292:1724 –37. [3] Christou NV, Sampalis JS, Liberman M, et al. Surgery decreases long-term mortality, morbidity, and health care use in morbidly obese patients. Ann Surg 2004;240:416 –23. [4] Flum DR, Dellinger EP. Impact of gastric bypass operation on survival: a population-based analysis. J Am Coll Surg 2004;199:543–51. [5] Craig BM, Tseng DS. Cost-effectiveness of gastric bypass for severe obesity. Am J Med 2002;113:491– 8. [6] Naessens JM, Ruud KL, Tulledge-Scheitel SM, Stroebel RJ, Cabanela RL. Comparison of provider claims data versus medical records review for assessing provision of adult preventive services. J Ambul Care Manage 2008;31:178 – 86. [7] Saunders JK, Ballantyne GH, Belsey S, et al. One-year readmission rates at a high volume bariatric surgery center: laparoscopic adjustable gastric banding, laparoscopic gastric bypass, and vertical banded gastroplasty–Roux-en-Y gastric bypass. Obes Surg 2008;18:1233–40.
423
[8] Pratt GM, McLees B, Pories WJ. The ASBS Bariatric Surgery Centers of Excellence program: a blueprint for quality improvement. Surg Obes Relat Dis 2006;2:497–503. [9] Saunders JK, Ballantyne GH, Belsley S, et al. 30-Day readmission rates at a high volume bariatric surgery center: laparoscopic adjustable gastric banding, laparoscopic gastric bypass, and vertical banded gastroplasty–Roux-en-Y gastric bypass. Obes Surg 2007;17:1171–7. [10] Nguyen NT, Paya M, Stevens CM, Mavandadi S, Zainabadi K, Wilson SE. The relationship between hospital volume and outcome in bariatric surgery at academic medical centers. Ann Surg 2004;240: 586 –94. [11] Weller WE, Rosati C, Hannan EL. Relationship between surgeon and hospital volume and readmission after bariatric operation. J Am Coll Surg 2007;204:383–91. [12] Encinosa WE, Bernard DM, Chen CC, Steiner CA. Healthcare utilization and outcomes after bariatric surgery. Med Care 2006;44: 706 –12. [13] McCarty TM, Arnold DT, Lamont JP, Fisher TL, Kuhn JA. Optimizing outcomes in bariatric surgery: outpatient laparoscopic gastric bypass. Ann Surg 2005;242:494 – 8. [14] Mehrotra C, Serdula M, Naimi TS, Khan LK, Miller J, Dietz W. Population-based study of trends, costs, and complications of weight loss surgeries from 1990 to 2002. Obes Res 2005;13:2029 –34. [15] Bradley DW, Sharma BK. Centers of excellence in bariatric surgery: design, implementation, and one-year outcomes. Surg Obes Relat Dis 2006;2:513–7. [16] Baker MT, Lara MD, Larson CJ, Lambert PJ, Mathiason MA, Kothari SN. Length of stay and impact on readmission rates after laparoscopic gastric bypass. Surg Obes Relat Dis 2006;2:435–9. [17] Shikora SA, Kim JJ, Tarnoff ME. Nutritional and gastrointestinal complications of bariatric surgery. Nutr Clin Pract 2007;22:29 – 40. [18] Shahian DM, Silverstein T, Lovett AF, Wolf RE, Normand SL. Comparison of clinical and administrative data sources for hospital coronary artery bypass graft surgery report cards. Circulation 2007; 115:1518 –27. [19] Maclean JR, Fick DM, Hoffman WK, King CT, Lough ER, Waller JL. Comparison of 2 systems for clinical practice profiling in diabetic care: medical records versus claims and administrative data. Am J Manag Care 2002;8:175–9. [20] Romano PS, Schembri ME, Rainwater JA. Can administrative data be used to ascertain clinically significant postoperative complications? Am J Med Qual 2002;17:145–54. [21] Best WR, Khuri SF, Phelan M, Hur K. Identifying patient preoperative risk factors and postoperative adverse events in administrative databases: results from the Department of Veterans Affairs National Surgical Quality Improvement Program. J Am Coll Surg 2002;194:257– 66.
Editorial comment
Comment on: Patterns of readmission and reoperation within 90 days after Roux-en-Y gastric bypass The authors from the University of Minnesota [1] have compiled and analyzed data from their medical center that shows the incidence and continuing occurrence of patient return to medical facilities during the first 3 months after successful bariatric surgery (primary Rouxen-Y gastric bypass, in particular). The national relevance of this report lies in revealing a postoperative need to seek medical help that has been a frequent financial concern and drain on health insurance carriers.
During recent years, most federal and commercial insurance carriers have developed respect for the health advantages and appropriateness of surgical care of morbid obesity. Although willing to pay the primary costs to counter the significant medical costs of long-term comorbidities, they have been alarmed by the seemingly high secondary—postoperative— costs and have searched for data to clarify its magnitude. That alarm has caused a reluctance to approve surgical plans and, in