Reasons for 30-day Postoperative Readmissions for Medicare Patients at a Community-based Teaching Hospital

Healthcare reform initiatives have proposed reducing reimbursement for certain 30-day readmissions among Medicare patients. Our objective was to evaluate the incidence and reasons for 30-day postoperative readmissions at our institution. The medical records of Medicare patients who underwent surgery from January 1, 2010, through May 16, 2011, were reviewed. Statistical analysis included ?2, Wilcoxon rank sum, and t tests. Two thousand eight hundred sixty-five patients were included; 199 (7%) had a 30-day readmission. The readmission group included a higher proportion of men (53.8 vs 43.6%, P = 0.005), and patients with an American Society of Anesthesiologists (ASA) Class 3 or greater (84 vs 66%, P < 0.001) versus the nonreadmission group. Mean index length of stay and operative time were longer in the readmitted versus nonreadmitted group (4.8 vs 2.8 days, P < 0.001; 122.8 vs 98.2 minutes, P < 0.001). Readmission reasons were surgically related (53%), surgically unrelated (35%), planned (7%), and patient-related (5%). Higher 30-day postoperative readmission rates were associated with male sex, higher ASA class, and longer index length of stay and operative time. Reasons for readmission included surgical- and patient-related factors. Decreased reimbursement should be discouraged for readmissions directly related to patient noncompliance.

IN 2010, THEPATIENTProtection and Affordable Care Act (PPACA) was signed into law, establishing some of the most significant healthcare reform initiatives in recent decades. Increased transparency with regard to patient outcomes and, specifically, hospital readmissions has been targeted for quality metrics and an area for cost-savings by identifying preventable readmissions. In 2009, Jencks and colleagues1 published reported rehospitalization rates among Medicare patients from 2003 to 2004. In that study, approximately 12 million Medicare beneficiaries had a hospital stay, and one in five had a readmission within 30 days. Readmission within 30 days of a surgical procedure occurred in 15.6 per cent of patients. Readmission rates have increased when comparing recent data with decades past.1 The reasons for this are multifactorial and can include the shorter hospital length of stay (LOS) that patients now experience as well as the increase in outpatient surgical procedures. The Medicare Payment Advisory Commission submitted a report to Congress that showed a 17.6 per cent hospital readmission rate at 30 days; 76 per cent of these were deemed potentially preventable and accounted for over $17 billion in Medicare spending.2, 3 This report had a major influence in forming the policies behind PPACA.4

With the recent increase in transparency for readmission rates as well as the potential economic uncertainty with reimbursement for surgeons and hospitals, we sought to evaluate the overall incidence of 30-day postoperative readmissions among inpatient and outpatient Medicare patients at our community hospital with the goal of identifying the reasons for surgical readmissions and any associated risk factors.

Methods

Our institution is a 485-physician multispecialty group medical center serving 19 counties over a tristate region with an accredited general surgery residency and minimally invasive bariatric surgery and advanced laparoscopy fellowship. Additionally, it is an American College of Surgeons (ACS) Committee on Trauma-verified Level 2 Trauma Center and a fully accredited comprehensive bariatric surgery center by the American College of Surgeons and American Society for Metabolic and Bariatric Surgery. After receiving Institutional Review Board approval, the electronic medical records of all Medicare patients who underwent a surgical procedure from January 1, 2010, through May 16, 2011, were retrospectively reviewed. A readmission was defined as any new admission within 30 days of the discharge date from the index surgery. Patients with less than 30 days of follow-up data available or those who died within 30 days of surgery were excluded. Each readmission was independently reviewed by a physician (J.L.K.) outside of a surgical specialty and classified into one of four categories: 1) surgical-related reason; 2) patient-related reason; 3) unrelated to surgery; or 4) planned patient care (staged surgery, or treatment resulting from findings on a previous diagnostic surgical procedure). Admissions were considered related to surgery unless there was sufficient evidence to suggest it was unrelated to surgery or patient-related. Indications for readmissions were broadly based on categoriesusedbyHealthGrades^ (Table 1).5 Patients who had a readmission were compared with those who were not readmitted to identify any potential risk factors for readmission. Statistical analysis included x2, Wilcoxon rank sum, and t test. A P value < 0.05 was considered significant.

Results

There were 3041 patients who underwent a surgical procedure over the study period. The case volumes included 37 per cent general surgery, 37 per cent orthopedic surgery, 11 per cent neurosurgery, 10 per cent cardiothoracic surgery, and 5 per cent obstetric/gynecology surgery. The admission priority was elective in 77 per cent of the cases, urgent in 19 per cent, and emergent in 4 per cent. A total of 2865 patients met inclusion criteria.

One hundred ninety-nine (7%) patients were readmitted within 30 days. Mean age and body mass index were similar in the readmitted and the nonreadmitted groups (Table 2). However, the readmitted group included a higher proportion of men compared with the nonreadmitted group (Table 2). Mean LOS during the index hospitalization and mean operative time were increased in the readmitted group versus the nonreadmitted group (Table 2). Comorbidities were common, because 168 (84%) patients in the readmitted group had at least one chronic medical condition (Table 3). There were more patients classified as American Society of Anesthesiologist (ASA) Class 3 or 4 in the readmitted versus nonreadmitted group (84 vs 66%, P < 0.001).

Readmission categories included cardiac, digestive, psychological, urinary/renal, infection, hematology, respiratory, and other indications (Fig. 1). Fourteen patients (7%) were readmitted as part of a care plan and are not included in Figure 1. Of the readmitted patients, 105 (53%) were classified as surgical-related. The most common indications for surgical-related readmissions were infections (32%), medication side effects (12%), and respiratory reasons (9%). Sixty-nine (35%) patients were readmitted for reasons unrelated to surgery. These readmissions were related to chronic medical conditions (62%) such as renal failure, heart failure, or cancer; cholecystitis (12%), community- or nursing home-acquired pneumonia (7%); and other indications (falls, fractures, chest pain, syncope; 19%). Eleven (5%) patients were readmitted for a patient-related reason. These included noncompliance with medications or discharge instructions (36%); psychiatric-related issues (36%) such as suicidal intentions or intentional drug overdoses; and other indications (27%) including leaving the hospital against medical advice.

In a subanalysis of the data, among patients who were readmitted, 64.3 per cent were discharged to home, 4.5 per cent to home with home health care, and 31.2 per cent to a care facility compared with 76.1, 0.9, and 22.9 per cent, respectively, for those not readmitted (P < 0.001).

Imaging was performed on 60 per cent of patients readmitted within 30 days of discharge. Reoperation was required for 19 per cent of patients, an interventional radiology procedure was performed in 10 per cent of patients, and 9 per cent of the readmitted patients required a new operation during their stay. Among patients in the readmitted group, the index surgery was elective in 69 per cent, urgent in 25 per cent, and emergent in 6 per cent.

Discussion

With the healthcare reform initiatives in recent years, hospital readmissions have been a measure of quality and a target for cost-savings. There are three Medicare payment reform proposals detailed in PPACA, including pay-for-performance (P4P) initiatives, bundled payments, and warranties. Under the P4P model, either a fraction or the entire amount of the Medicare reimbursement could be eliminated for certain potentially preventable conditions. In contrast to the P4P model, the bundled payment plan would establish an average cost for the medical treatment and stay for Medicare patients. The bundled payment would cover the hospital stay plus a limited time after discharge. In this plan, the hospital would receive one payment and it would then be responsible for distributing payment to the inpatient providers, skilled nursing facilities, and home healthcare providers. Warranties occur when a hospital provides a guarantee on the quality of services it will provide. Medicare would pay a flat rate to cover both the initial treatment and any needed follow-up treatments. If the patient suffers from any complications within a 30- to 90-day timeframe, no additional payments would be made to the hospital for services. In all, these three changes have been estimated to raise costs for hospitals by $19 billion over 10 years.3

The surgical readmission rate in our series was found to be 7 per cent as compared with the 15.6 per cent reported by Jencks and colleagues.1 Thirty-day postoperative readmissions were associated with male sex, higher ASA class, longer index LOS, and longer operative time in our population. In addition, patients with pre-existing chronic cardiac disease, renal disease, or diabetes mellitus were more likely to be readmitted within 30 days. Reasons for readmission included surgical-related, unrelated to surgery, and patient-related factors.

Recently, several articles have been published exploring various aspects of readmission rates after surgery. Martin et al.6 evaluated 266 prospectively enrolled patients in a university setting to determine the impact of major abdominal operations on 90-day readmission rates. Variables evaluated included patient factors, social factors, perioperative factors, and discharge location. They noted several factors associated with increased rates of readmission, including the total number of complications, discharge to home with home health care, and age older than 69 years. Interestingly, they found that diabetes was not a factor for readmission. Our data also showed a higher proportion of readmitted patients discharged to home with home health care compared with nonreadmitted patients; we hypothesize that the threshold for transfer or readmission is lower among healthcare providers in skilled nursing facilities as their medical insight into ''failure to rescue'' and its concerning morbidity may affect their decision-making. However, in our series, diabetes mellitus was more prevalent among readmitted patients. This has also been reported in other series,1, 7 and the reasons for this variation in comparison to Martin and colleagues6 may have to do with regional differences in the management of diabetes. Clearly, further investigation into these factors is warranted.

Kassin et al.8 analyzed National Surgical Quality Improvement Program (NSQIP) data for all inpatient procedures at a single academic center from 2009 to 2011. They evaluated 1442 patients, of whom 163 (11.3%) were readmitted. The most common reasons were gastrointestinal problems (27.6%) and postoperative surgical complications (22.1%). Interestingly, patients who experienced a complication after discharge were two times more likely to be readmitted when compared with those who experienced a complication during the index admission.

Kazaure et al.7 also used the NSQIP database looking at over 550,000 patients from 2005 to 2010. They categorized 21 groups of general surgical procedures and noted that 41.5 per cent of postoperative complications occurred after hospital discharge. Thirty-one per cent of these were surgical site infections, and almost 14 per cent were organ space infections followed by 11.4 per cent with severe sepsis. Of those with a postdischarge complication, 17.9 per cent required a reoperation. Similarly, the primary indications for surgical-related readmissions in our series were infections (18%). Overall, reoperation was required in 19 per cent of those in the readmitted group. They also reported that a history of diabetes, higher ASA class, increased operative time, and an increased LOS were risk factors for a postdischarge complication; these factors were also associated with postoperative readmissions in our series. However, unlike Kassin et al.8 who reported a decreased readmission rate among those with an inpatient complication, Kazaure et al. found that an inpatient complication increased the likelihood of a postdischarge complication as well as mortality.

Although NSQIP allows for analysis of large sample sizes and several healthcare systems, it only captures 25 per cent of all patients in any single healthcare institution. To our knowledge, this is the largest, most comprehensive, single-institution review of all surgical readmissions. It provides realistic data on both inpatient and outpatient surgical readmissions in a community setting. However, the strength in our study could also be considered a potential weakness. We evaluated both inpatient and outpatient surgical procedures in a variety of surgical disciplines. It is possible that some outpatient surgical procedures in certain specialties with known lower complication rates could explain our lower readmission rate compared with the literature. Additionally, patients may have been readmitted to another healthcare institution for complications within 30 to 90 days postoperatively. These readmissions would not have been captured in this series.

The true reasons for our 7 per cent surgical readmission rate also remains largely unknown. Based on our data as well as others, striving for exemplary surgical outcomes and lower than expected complication rates should, in turn, lower an institution's readmission rate, because surgical complications continue to be the main driver behind readmissions. However, the presence of multiple medical comorbidities in an aging population cannot be ignored. Exacerbation of underlying medical conditions resulted in 35 per cent of the readmissions. In addition, although patient-related factors comprised only 5 per cent of postoperative readmissions, healthcare institutions should not be financially penalized for these readmissions related to patient noncompliance or other direct patient-related factors.

Kazaure and colleagues7 noted that the 2-week period after hospital discharge results in a significantly increased risk for complications and readmissions. Jenks et al.1 reported that approximately 50 per cent of patients who developed a complication and were readmitted were not seen by a physician during the interval between hospital discharge and readmission. This transitional period from hospital discharge to follow-up has been referred to as the ''vulnerable period.''9 It is possible that focusing more efforts on more aggressive follow-up during this period may decrease readmission rates. Several proposals have been put forth to explore this idea and include the Transitional Care Model,10 Care Transition Initiatives,9 and Project ReEngineered Hospital Discharge program.11 Patient-related factors comprised 5 per cent of postoperative readmissions in our series. It is possible that poor discharge planning may have contributed to these readmissions. Whether adopting one of these transitional care discharge models in our system would lower our surgical readmission rate even further remains to be determined. In an attempt to further define factors contributing to readmissions after surgery, our institution has established a committee to review the medical record of every surgical patient who is readmitted within 30 days of discharge. Similar to a morbidity and mortality conference, the reason for each readmission is categorized and examined to see if there are any changes that can be made either in the treatment of specific patient subpopulations or within institutional procedures to further decrease the readmission rate. The impact of this newly formed committee on preventing further surgical readmissions also remains to be seen.

In conclusion, an increased risk for 30-day postoperative readmission was associated with male sex, pre-existing chronic cardiac or renal disease, diabetes mellitus, longer index LOS, longer operative time, and higher ASA class. Reasons for readmissions include surgically related, unrelatedtosurgery,and patient-related factors. Further study is needed to address factors for readmission on a prospective, multicenter level. Decreased reimbursement should be discouraged for readmissions directly related to patient noncompliance.

REFERENCES

1. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418-28.

2. Medicare Payment Advisory Commission. Report to the Congress: Promoting Greater Efficiency in Medicare. 2007. Available at: www.medpac.gov. Accessed January 26, 2015.

3. Stone J, Hoffman G. Medicare Hospital Readmissions: Issues, Policy Options and PPACA. CRS Report to Congress. 2010:1-36. Congressional Research Service. Available at: www.crs.gov. Accessed January 26, 2015.

4. Patient Protection and Affordable Care Act. H.R. 3590. Available at: www.gpo.gov/fdsys/pkg/BILLS-111hr3590enr/pdf/ BILLS-111hr3590enr.pdf. Accessed February 4, 2013.

5. HealthGrades^. Available at: www.healthgrades.com. Accessed February 4, 2013.

6. Martin RC, Brown R, Puffer L, et al. Readmission rates after abdominal surgery: the role of surgeon, primary caregiver, home health, and subacute rehab. Ann Surg 2011;254:591-7.

7. Kazaure HS, Roman SA, Sosa JA. Association of postdischarge complications with reoperation and mortality in general surgery. Arch Surg 2012;147:1000-7.

8. Kassin MT, Owen RM, Perez SD, et al. Risk factors for 30-day hospital readmission among general surgery patients. J Am Coll Surg 2012;215:322-30.

9. Coleman EA, Parry C, Chalmers S, et al. The care transitions intervention: results of a randomized controlled trial. Arch Intern Med 2006;166:1822-8.

10. Transitional Care Model. Available at: www.transitionalcare. info. Accessed January 29, 2013.

11. Jack BW, Chetty VK, Anthony D, et al. A reengineered hospital discharge program to decrease rehospitalization: a randomized trial. Ann Intern Med 2009;150:178-87.

JENNIFER L. KIRSCH, M.D., M.B.A.,* SHANU N. KOTHARI, M.D.,[dagger] JANELLE M. AUSLOOS, B.S.,[double dagger] JACOB D. GUNDRUM, M.S.,[double dagger] KARA J. KALLIES, M.S.[double dagger]

From the Departments of *Anesthesiology and [dagger]General and Vascular Surgery, Gundersen Health System, La Crosse, Wisconsin; and the [double dagger]Department of Medical Research, Gundersen Medical Foundation, &lt;location value="LU/us.wi.lacsse" idsrc="xmltag.org">La Crosse, Wisconsin

Portions of these data were presented at the American College of Surgeons 98th Annual Clinical Congress, October 3, 2012, Chicago, Illinois.

Address correspondence and reprint requests to Shanu N. Kothari, M.D., Department of General and Vascular Surgery, Gundersen Health System, 1900 South Avenue, C05-001, La Crosse, WI, 54601. E-mail: [email protected].