Reducing hospital readmissions, which are frequent, costly, and inconvenient, has become a major focus of the national health care quality improvement effort. In 2010, The Patient Protection and Affordable Care Act mandated a Hospital Readmissions Reduction Program. The Centers for Medicare & Medicaid Services has recently expanded this program to include select surgical procedures (joint replacement and coronary artery bypass grafting). This penalty program is expected to include additional operations in the future.1,2 For the fiscal year 2015–2016, 54% of hospitals will be subjected to payment reductions that average 0.61% as a result of excess readmission ratios.3 Among surgical patients, readmission rates are also now widely used as a metric for measuring quality of care during the index hospitalization.4
Hysterectomy is the most common nonobstetric operation performed on American women.5 Although predictors of readmissions after hysterectomy have been studied,6–12 there is limited understanding of the underlying reasons for unplanned readmissions. Merkow et al13 recently examined the readmission diagnoses for several operations, including hysterectomy; however, surgical indication, approach, and variations of diagnoses by time were not analyzed.
Clear delineation of the timing of and reasons for readmission and understanding differences based on surgical approach could help answer questions that are pertinent when devising targeted interventions to reduce readmissions. The development of effective strategies aimed at decreasing these rates can, in turn, improve quality of care4 and reduce costs. The objective of this study is to characterize readmission timing and diagnoses after hysterectomy performed for benign indications, taking into account the surgical approach.
MATERIALS AND METHODS
A retrospective analysis was performed using the American College of Surgeons National Surgical Quality Improvement Program database participant user files for 2012 and 2013 (user files before these years did not have detailed readmission diagnoses). The National Surgical Quality Improvement Program database collects detailed preoperative, operative, and 30-day outcome data on surgical patients from participating academic and community hospitals across the United States. Participation in the National Surgical Quality Improvement Program database is voluntary, and hospitals can review their own risk-adjusted outcomes for quality improvement. Details of the sampling strategy, data abstraction procedures, and outcomes have been documented extensively.14,15 Trained data abstractors at participating institutions extract the data from hospital records and follow the patient postoperatively for 30 days after the index surgery. The data abstractors obtain details of postoperative complications and readmissions, even if these events occur in the outpatient setting or in a different hospital. Clinical reviewers examine the medical records, query the involved clinicians, and contact the patient if necessary.
The National Surgical Quality Improvement Program database was queried to examine the reasons for readmission after hysterectomy performed for benign indications. We utilized Current Procedural Terminology (CPT) codes to identify patients aged 18 years or older who underwent nonlaparoscopic abdominal hysterectomy (CPT codes: 58150, 58152), nonlaparoscopic vaginal hysterectomy (CPT codes: 58260, 58262, 58263, 58267, 58270, 58275, 58280, 58290, 58291, 58292, 58293, 58294), and laparoscopic abdominal or vaginal hysterectomy (CPT codes: 58550, 58552, 58553, 58554, 58570, 58571, 58572, 58573).16 Patients with a malignancy diagnosis based on International Classification of Diseases, 9th Revision codes 179–185 were excluded from this analysis. We also excluded patients who died during the index hospitalization and those with length of stay greater than 30 days after the index operation, because the National Surgical Quality Improvement Program database does not collect readmission data for these cases.
The following demographic data, medical comorbidities, and perioperative data were abstracted: age (defined as younger than 65 or 65 years or older), body mass index (calculated as weight (kg)/[height (m)]2, defined as nonobese if less than 30 or obese if 30 or greater), race (defined as white or nonwhite), functional status (defined as a patient's ability to perform basic daily activities to maintain her health, reported as dependent or independent), American Society of Anesthesiologists physical status score (defined as less than 3 or 3 or greater), Charlson Comorbidity Index score (defined as 0–1 or 2 or greater), and discharge destination (defined as home or elsewhere). Surgical complexity was calculated by adding the relative value units for each surgical procedure recorded for the patient. Surgeries with total relative value units above 2 standard deviations above the mean were categorized as having “complex” procedures; all other procedures were categorized as having “routine” complexity procedures. Operative time was measured from skin incision to skin closure (categorized as less than 3 hours or 3 hours or greater). Postoperative complications captured in the National Surgical Quality Improvement Program database included the following: superficial surgical site infection, deep surgical site infection, organ or space surgical site infection, wound dehiscence, pneumonia, unplanned intubation, ventilator dependency for 48 hours, pulmonary embolism, progressive renal insufficiency, acute renal failure, stroke, coma lasting for 24 hours, peripheral nerve injury, cardiac arrest requiring cardiopulmonary resuscitation, urinary tract infection, bleeding requiring transfusion of packed red blood cells, deep venous thrombosis, sepsis, and septic shock. The National Surgical Quality Improvement Program database provides the day the postoperative complication occurred and the total length of stay after the index procedure. Using these two variables, we captured whether a complication was diagnosed before discharge from the index hospitalization.
Readmission events are defined as any hospitalization within 30 days of the surgery date (not date of discharge). These events were also classified as planned or unplanned readmission. Based on the National Surgical Quality Improvement Program database definition, a readmission episode was categorized as an “unplanned readmission” if the readmission was not planned as part of the index procedure.17 Our analysis focuses on unplanned readmissions only. Because few patients experienced two or more readmissions within 30 days of surgery, we analyzed only data related to the first readmission.
The National Surgical Quality Improvement Program database provides information on the day the readmission occurred. By subtracting the length of stay from the readmission postoperative day provided (because the days to readmission provided in the database are calculated from the date of surgery), we calculated the number of days between discharge and first readmission. This variable was labeled as “time to readmission.” We then examined the frequency of readmissions as a function of time to readmission. We also examined the median time to readmission based on readmission diagnosis.
Data abstractors used the National Surgical Quality Improvement Program database postoperative complication schema to categorize the primary reason for readmission.17 When the reason for readmission was not contained within the set of standard outcomes collected, data abstractors assigned an International Classification of Diseases, 9th Revision code. These reasons for readmission were abstracted from chart review and not from administrative billing data. We then categorized the reasons for readmission into 10 distinct categories: surgical site infections, including superficial, deep, and organ space; infectious, nonsurgical site infection (urinary tract infection, pyelonephritis, sepsis, septic shock, pneumonia, and cellulitis); gastrointestinal (ileus, intestinal obstruction, nausea, vomiting, dehydration, and electrolyte disturbance); surgical (visceral injury including ureteral, bladder, bowel, and vascular injury); wound, noninfectious (dehiscence, herniation, and seroma formation); venous thromboembolism (deep vein thrombosis and pulmonary embolism); genitourinary (acute kidney injury and urinary retention); medical, other (myocardial infarction, pulmonary edema, and shortness of breath); and other or unknown (the reason did not fit into one of the categories previously mentioned or the reason was not specified).
Outcomes for the study included patient demographic characteristics associated with readmission; readmission diagnoses; readmission timing; and readmission diagnoses by time after discharge. A χ2 test was used to evaluate variables (patient demographics, functional status scores, operative characteristics, discharge destination, and inpatient complications) associated with readmission. Factors predicting readmission with a P value of <.05 or with clinical significance were included in a multivariable logistic regression. The results of the logistic regression are reported as adjusted odds ratios (ORs) with 95% confidence interval (CI). Summary statistics were calculated for readmission diagnoses by categorical variables. We also calculated summary statistics for readmission timing by day after discharge (days 0–30). Median time to readmission was determined for each readmission diagnosis category. STATA 14.0 SE for Macintosh was used for all analyses.
The National Surgical Quality Improvement Program database does not contain any patient identifiers and therefore the study was considered exempt by the University of Michigan institutional review board (HUM00107660).
There were a total of 1,118 readmissions after 40,580 hysterectomies performed for benign indications, resulting in an overall unplanned readmission rate of 2.8%. Table 1 outlines factors associated with readmission. In a bivariable analysis, nonwhite race, obesity (body mass index 30 or higher), higher Charlson Comorbidity Index score, higher American Society of Anesthesiologists physical status score, operative time 3 hours or longer, discharge destination other than home, and having a surgical complication diagnosed before discharge from the hospital were associated with increased rates of unplanned readmission (Table 1). In a multivariable model (adjusting for factors predicting readmission with a P value of <.05 as outlined in Table 1), these factors were associated with readmission except for obesity and Charlson Comorbidity Index score (Table 2).
Readmission rates varied by surgical approach with 3.7% (361/9,869) of abdominal, 2.6% (576/22,266) of laparoscopic, and 2.1% (181/8,445) of vaginal hysterectomies resulting in an unplanned readmission. After adjusting for potentially confounding factors in the multivariable model, compared with the vaginal approach, readmissions were more likely when hysterectomy was performed through the abdominal route (adjusted OR 1.45, 95% CI 1.20–1.76) but not significantly more likely when performed through the laparoscopic route (adjusted OR 1.10, 95% CI 0.97–1.40) (Table 2).
Figure 1 illustrates the unplanned 30-day readmissions by day from initial hospital discharge (0–30). Approximately one fourth of readmissions occurred within the first 3 days of discharge, no matter the surgical approach: 25% overall (285/1,118), 27.7% (100/361) of abdominal, 24.1% (139/576) of laparoscopic, and 25.4% (46/181) of vaginal hysterectomies. More than three fourths of all readmissions occurred in the first 2 weeks after discharge, independent of approach: 82.3% (920/1,118) overall, 85% (307/361) of abdominal, 82.1% (473/576) of laparoscopic, and 77.3% (140/181) of vaginal hysterectomies.
Median time to readmission (measured in days) by categorized readmission diagnosis is presented in Table 3. Pain-related reasons for readmission had the shortest time to readmission (median 3, interquartile range 2–6.5 days) and noninfectious wound complications had the longest time to readmission (median 10, interquartile range 5–15 days). The percentage of readmissions changed over time in varying patterns depending on the reason for readmission (Table 4). There was no statistically significant difference in time to readmission based on surgical approach (Table 5).
The reasons for unplanned readmission, grouped by category, are presented in Table 6. Surgical complications, specifically, surgical site infection, surgical injuries, and noninfectious wound complications, were the primary reason for admission in 51.5% (186/361) of abdominal, 51.9% (299/576) of laparoscopic, and 50.8% (92/181) of vaginal hysterectomies. Surgical site infection was the most common reason for readmission, regardless of the surgical approach (abdominal, 36.6% [132/361]; laparoscopic, 28.3% [163/576]; vaginal, 32.6% [59/181]). Infectious complications (surgical site infections in addition to infections not related to the surgical site) were responsible for 49% (177/361) of abdominal, 43.6% (251/576) of laparoscopic, and 51.4% (93/181) of vaginal hysterectomy readmissions. Medical complications such as cardiovascular events and venous thromboembolism accounted for 5.8% (21/361) of abdominal, 6.9% (40/576) of laparoscopic, and 8.8% (16/181) of vaginal hysterectomies. There were more readmissions secondary to gastrointestinal complications (eg, ileus, obstruction) after abdominal hysterectomies than observed among laparoscopic (unadjusted OR 2.4, 95% CI 1.67–3.52) or vaginal (unadjusted OR 2.8, 95% CI 1.55–4.87) hysterectomies. Compared with the abdominal approach, there were significantly higher odds of readmission for surgical injury after laparoscopic (unadjusted OR 2.3, 95% CI 1.48–3.65) and vaginal hysterectomies (unadjusted OR 2.3, 95% CI 1.29–3.97).
Using a national surgical outcomes database with more than 40,000 benign hysterectomies from several hundred hospitals, we found that unplanned, 30-day readmissions occur disproportionately soon after discharge and are primarily attributable to surgical issues, regardless of approach, with the most common reason being surgical site infection. The overall readmission rate in this study (2.8%) was comparable with published rates from the National Surgical Quality Improvement Program database for earlier time periods (3.1%, 3.4%).6,7
Our results provide benchmarking data on the timing of readmissions after hysterectomy performed for benign indications. The finding that readmissions tend to occur shortly after discharge validates recent data from a smaller, single-site study, which found that more than 75% of women readmitted after undergoing hysterectomy for benign disease were readmitted within the first 2 weeks.9 This increased vulnerability to readmission soon after discharge raises several key points. First, some patients may have been discharged prematurely or had lapses in care coordination. This is particularly likely in the 25% of patients readmitted within 3 days of discharge. Second, outpatient follow-up could occur sooner after discharge, potentially allowing intervention. Early physician follow-up has been shown to be effective in reducing readmissions in patients with heart failure, especially if a patient experienced a complication in the hospital.18,19 There have been limited data on whether early postdischarge contact is useful in reducing readmissions after hysterectomy (information available is retrospective data on incidental utilization of postacute care services).10 Further investigation into early postdischarge follow-up after hysterectomy is needed.
Similar to previously published data, we found that infectious complications, particularly surgical site infection, are a leading cause of readmission.13 Our study expanded prior literature by showing that infections are the primary cause of readmission for all hysterectomy surgical approaches. Such results highlight the continued need for investigation into methods to reduce infections after hysterectomy. This is particularly important in light of recent implementation of financial penalties for surgical site infection after laparoscopic or abdominal hysterectomy by the Hospital Acquired Condition reduction program.20
Our analysis also demonstrates that readmissions after hysterectomy tend to be related to issues typically considered surgical (eg, visceral injuries and noninfectious wound complications) with less than 10% related to medical issues (eg, venous thromboembolism and pulmonary edema). Unfortunately, unlike venous thromboembolism, there are no clear-cut interventions with proven efficacy to reduce many of these surgical complications leading to readmission. Moreover, not all surgical complications reflect modifiable care. As a starting point, future analyses could focus on assessing preventability of readmissions related to surgical complications and determining quality issues that may be amenable to intervention as has been done in other surgical specialties.21
A significant approach-specific finding was that readmissions related to surgical injury (eg, vascular, ureteral, bowel injury) were higher among the laparoscopic and vaginal cohorts than the abdominal group. Notably, this finding does not mean there were more surgical injuries in the minimally invasive group; rather, readmissions related to injury were higher. Previously published data suggest no significant difference in visceral injuries between surgical approaches with the possible exception of increased urinary tract injuries when comparing laparoscopic with abdominal hysterectomy.22,23 The difference reflected in our data may be the result of a higher rate of occult visceral injuries that are not recognized before discharge among patients undergoing laparoscopic and vaginal hysterectomy. Regardless, anticipating differences in readmission based on approach could be helpful in readmission reduction efforts.
Our study has several strengths. The National Surgical Quality Improvement Program database more reliably captures reasons for readmission than databases that utilize administrative Diagnosis-Related Group codes, because information is collected from medical record extraction. Also, findings based on the National Surgical Quality Improvement Program database data have greater external validity than studies that report data from a single institution. However, there are several limitations. The database captures one diagnosis for readmission when the underlying reason for readmission may be multifactorial. This effect is presumably minor, because readmission data have been previously validated and found to be in agreement with chart reviews.24 Also, the National Surgical Quality Improvement Program database does not track readmissions occurring more than 30 days after surgical procedures. Given that greater than 75% of all readmissions occurred within the first 2 weeks of hospital discharge, the number of readmissions occurring more than 30 days after hysterectomy likely is low.
In conclusion, targeted surgical quality improvement efforts to reduce unplanned 30-day readmissions after hysterectomy should focus on early postdischarge follow-up, preventing infectious complications, and determining preventability of surgical-related reasons for readmission. Readmission reduction initiatives that are specific to these factors will address a large proportion of patients at risk for readmission.
2. Fontanarosa PB, McNutt RA. Revisiting hospital readmissions. JAMA 2013;309:398–400.
3. Rau J. Half of nation's hospitals fail again to escape Medicare’s readmission penalties. Kaiser Health News, August 3, 2015.
4. Tsai TC, Joynt KE, Orav EJ, Gawande AA, Jha AK. Variation in surgical-readmission rates and quality of hospital care. N Engl J Med 2013;369:1134–42.
5. Wilcox LS, Koonin LM, Pokras R, Strauss LT, Xia Z, Peterson HB. Hysterectomy in the United States, 1988–1990. Obstet Gynecol 1994;83:549–55.
6. Dessources K, Hou JY, Tergas AI, Burke WM, Ananth CV, Prendergast E, et al. Factors associated with 30-day hospital readmission after hysterectomy. Obstet Gynecol 2015;125:461–70.
7. Catanzarite T, Vieira B, Qin C, Milad MP. Risk factors for unscheduled 30-day readmission after benign hysterectomy. South Med J 2015;108:524–30.
8. Jennings AJ, Spencer RJ, Medlin E, Rice LW, Uppal S. Predictors of 30-day readmission and impact of same-day discharge in laparoscopic hysterectomy. Am J Obstet Gynecol 2015;213:344.e1–7.
9. Lee MS, Venkatesh KK, Growdon WB, Ecker JL, York-Best CM. Predictors of 30-day readmission following hysterectomy for benign and malignant indications at a tertiary care academic medical center. Am J Obstet Gynecol 2016;214:607.e1–607.e12.
10. Clark RM, Growdon WB, Wiechert A, Boruta D, Del Carmen M, Goodman AK, et al. Patient, treatment and discharge factors associated with hospital readmission within 30 days after surgical cytoreduction for epithelial ovarian carcinoma. Gynecol Oncol 2013;130:407–10.
11. Morris MS, Deierhoi RJ, Richman JS, Altom LK, Hawn MT. The relationship between timing of surgical complications and hospital readmission. JAMA Surg 2014;149:348–54.
12. Uppal S, Penn C, Del Carmen MG, Rauh-Hain JA, Reynolds RK, Rice LW. Readmissions after major gynecologic oncology surgery. Gynecol Oncol 2016;141:287–92.
13. Merkow RP, Ju MH, Chung JW, Hall BL, Cohen ME, Williams MV, et al. Underlying reasons associated with hospital readmission following surgery in the United States. JAMA 2015;313:483–95.
14. Khuri SF, Daley J, Henderson W, Hur K, Demakis J, Aust JB, et al. The Department of Veterans Affairs' NSQIP: the first national, validated, outcome-based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. National VA Surgical Quality Improvement Program. Ann Surg 1998;228:491–507.
15. Khuri SF, Henderson WG, Daley J, Jonasson O, Jones RS, Campbell DA Jr, et al. Successful implementation of the Department of Veterans Affairs' National Surgical Quality Improvement Program in the private sector: the Patient Safety in Surgery study. Ann Surg 2008;248:329–36.
16. American Medical Association. CPT 2014 standard edition. Washington, DC: American Medical Association; 2014.
17. American College of Surgeons National Surgical Quality Improvement Program. Chapter 4: variable definitions. Retrieved May 2016.
18. Hernandez AF, Greiner MA, Fonarow GC, Hammill BG, Heidenreich PA, Yancy CW, et al. Relationship between early physician follow-up and 30-day readmission among Medicare beneficiaries hospitalized for heart failure. JAMA 2010;303:1716–22.
19. Riegel B, Carlson B, Glaser D, Romero T. Randomized controlled trial of telephone case management in Hispanics of Mexican origin with heart failure. J Card Fail 2006;12:211–9.
20. Morgan DM, Swenson CW, Streifel KM, Kamdar NS, Uppal S, Burgunder-Zdravkovski L, et al. Surgical site infection following hysterectomy: adjusted rankings in a regional collaborative. Am J Obstet Gynecol 2016;214:259.e1–8.
21. Tejedor-Sojo J, Singleton LM, McCormick K, Wrubel D, Chern JJ. Preventability of Pediatric 30-Day Readmissions following Ventricular Shunt Surgery. J Pediatr 2015;167:1327–33.e1.
22. Johnson N, Barlow D, Lethaby A, Tavender E, Curr L, Garry R. Methods of hysterectomy: systematic review and meta-analysis of randomised controlled trials. BMJ 2005;330:1478.
23. Aarts JW, Nieboer TE, Johnson N, Tavender E, Garry R, Mol BW, et al. Surgical approach to hysterectomy for benign gynaecological disease. The Cochrane Database Systematic Reviews 2015, Issue 8. Art. No.: CD003677.
© 2016 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
24. Sellers MM, Merkow RP, Halverson A, Hinami K, Kelz RR, Bentrem DJ, et al. Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2013;216:420–7.