[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.
Total Discharges From the Surgery Service
Total Discharges From the Surgery Service
Table 1.  
Patient Characteristics of All Discharges
Patient Characteristics of All Discharges
Table 2.  
Risk Factors for Readmission on Univariate Analysis
Risk Factors for Readmission on Univariate Analysis
Table 3.  
Risk Factors for Readmission on Multivariate Logistic Regression Analysis
Risk Factors for Readmission on Multivariate Logistic Regression Analysis
1.
Centers for Medicare and Medicaid Services. Readmissions reduction program. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Readmissions-Reduction-Program.html. Accessed March 34, 2016.
2.
Echeverry  LM, Lamb  KV, Miller  J.  Impact of APN home visits in reducing healthcare costs and improving function in homebound heart failure.  Home Healthc Now. 2015;33(10):532-537.PubMedGoogle Scholar
3.
Ziaeian  B, Fonarow  GC.  The prevention of hospital readmissions in heart failure.  Prog Cardiovasc Dis. 2016;58(4):379-385.PubMedGoogle ScholarCrossref
4.
Prescott  HC, Sjoding  MW, Iwashyna  TJ.  Diagnoses of early and late readmissions after hospitalization for pneumonia: a systematic review.  Ann Am Thorac Soc. 2014;11(7):1091-1100.PubMedGoogle ScholarCrossref
5.
Arnold  SV, Smolderen  KG, Kennedy  KF,  et al.  Risk factors for rehospitalization for acute coronary syndromes and unplanned revascularization following acute myocardial infarction.  J Am Heart Assoc. 2015;4(2):e001352.PubMedGoogle ScholarCrossref
6.
Hansen  LO, Young  RS, Hinami  K, Leung  A, Williams  MV.  Interventions to reduce 30-day rehospitalization: a systematic review.  Ann Intern Med. 2011;155(8):520-528.PubMedGoogle ScholarCrossref
8.
van Walraven  C, Dhalla  IA, Bell  C,  et al.  Derivation and validation of an index to predict early death or unplanned readmission after discharge from hospital to the community.  CMAJ. 2010;182(6):551-557.PubMedGoogle ScholarCrossref
9.
Billings  J, Dixon  J, Mijanovich  T, Wennberg  D.  Case finding for patients at risk of readmission to hospital: development of algorithm to identify high risk patients.  BMJ. 2006;333(7563):327.PubMedGoogle ScholarCrossref
11.
Edwards  ST, Abrams  MK, Baron  RJ,  et al.  Structuring payment to medical homes after the affordable care act.  J Gen Intern Med. 2014;29(10):1410-1413.PubMedGoogle ScholarCrossref
12.
Fawcett  VJ, Flynn-O’Brien  KT, Shorter  Z,  et al.  Risk factors for unplanned readmissions in older adult trauma patients in Washington State: a competing risk analysis.  J Am Coll Surg. 2015;220(3):330-338.PubMedGoogle ScholarCrossref
13.
Sacks  GD, Dawes  AJ, Russell  MM,  et al.  Evaluation of hospital readmissions in surgical patients: do administrative data tell the real story?  JAMA Surg. 2014;149(8):759-764.PubMedGoogle ScholarCrossref
14.
van Walraven  C, Bennett  C, Jennings  A, Austin  PC, Forster  AJ.  Proportion of hospital readmissions deemed avoidable: a systematic review.  CMAJ. 2011;183(7):E391-E402.PubMedGoogle ScholarCrossref
15.
Merkow  RP, Ju  MH, Chung  JW,  et al.  Underlying reasons associated with hospital readmission following surgery in the United States.  JAMA. 2015;313(5):483-495.PubMedGoogle ScholarCrossref
16.
Schmeida  M, Savrin  R.  Surgical rehospitalization of the medicare fee-for-service patient: a state-level analysis exploring 30-day readmission factors.  Prof Case Manag. 2015;20(3):130-137.PubMedGoogle ScholarCrossref
17.
Calvillo-King  L, Arnold  D, Eubank  KJ,  et al.  Impact of social factors on risk of readmission or mortality in pneumonia and heart failure: systematic review.  J Gen Intern Med. 2013;28(2):269-282.PubMedGoogle ScholarCrossref
18.
Barnett  ML, Hsu  J, McWilliams  JM.  Patient characteristics and differences in hospital readmission rates.  JAMA Intern Med. 2015;175(11):1803-1812.PubMedGoogle ScholarCrossref
19.
Gentilello  LM, Rivara  FP, Donovan  DM,  et al.  Alcohol interventions in a trauma center as a means of reducing the risk of injury recurrence.  Ann Surg. 1999;230(4):473-480.PubMedGoogle ScholarCrossref
20.
Zatzick  D, Roy-Byrne  P, Russo  J,  et al.  A randomized effectiveness trial of stepped collaborative care for acutely injured trauma survivors.  Arch Gen Psychiatry. 2004;61(5):498-506.PubMedGoogle ScholarCrossref
21.
Schermer  CR, Moyers  TB, Miller  WR, Bloomfield  LA.  Trauma center brief interventions for alcohol disorders decrease subsequent driving under the influence arrests.  J Trauma. 2006;60(1):29-34.PubMedGoogle ScholarCrossref
22.
Hunt  GE, Siegfried  N, Morley  K, Sitharthan  T, Cleary  M.  Psychosocial interventions for people with both severe mental illness and substance misuse.  Schizophr Bull. 2014;40(1):18-20.PubMedGoogle ScholarCrossref
23.
Thew  J. Hospital cuts readmissions in half with help from college students. HealthLeaders Media. September 1, 2015. http://www.healthleadersmedia.com/nurse-leaders/hospital-cuts-readmissions-half-help-college-students. Accessed October 29, 2015.
Original Investigation
Pacific Coast Surgical Association
September 2016

Analysis of Risk Factors for Patient Readmission 30 Days Following Discharge From General Surgery

Author Affiliations
  • 1Department of Surgery, University of Washington Medical Center, Seattle
  • 2Division of Trauma and Burn Surgery, Harborview Medical Center, Seattle, Washington
  • 3Department of Quality Improvement, Harborview Medical Center, Seattle, Washington
 

Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.

JAMA Surg. 2016;151(9):855-861. doi:10.1001/jamasurg.2016.1258
Abstract

Importance  Previous studies investigating patients at risk for hospital readmissions focus on medical services and have found chronic conditions as contributors. Little is known, however, of the characteristics of patients readmitted from surgical services.

Objective  Surgical patients readmitted within 30 days following discharge were analyzed to identify opportunities for intervention in a cohort that may differ from the medical population.

Design, Setting, and Participants  Medical record review of patients readmitted to any service within 30 days of discharge from the general surgery service to characterize index and readmission data between July 1, 2014, and June 30, 2015, at a Level I trauma center and safety-net hospital.

Main Outcomes and Measures  Reasons for readmission identified by manual medical record review and risk factors identified via statistical analysis of all discharges during this period.

Results  One hundred seventy-three patients were identified as being unplanned readmissions within 30 days among 2100 discharges (8.2%). Of these 173 patients, 91 were men. Common reasons for readmission included 29 patients with injection drug use who were readmitted with soft tissue infections at new sites (16.8% of readmissions), 25 with disposition support issues (14.5%), 23 with infections not detectable during index admission (13.3%), and 16 with sequelae of their injury or condition (9.2%). Sixteen patients were identified as having a likely preventable complication of care (9.2%), and 2 were readmitted owing to deterioration of medical conditions (1.2%). On univariate and multivariate analyses, female sex (men to women risk of readmission odds ratio [OR], 0.5; 95% CI, 0.37-0.71; P < .001), presence of diabetes (OR, 1.7; 95% CI, 1.1-2.6; P = .009), sepsis on admission (OR, 1.7; 95% CI, 1.05-2.6; P = .03), or intensive care unit stay during index admission (OR, 1.7; 95% CI, 1.2-2.4; P = .002), as well as discharge to respite care (OR, 2.3; 95% CI, 1.2-4.5; P = .01) and payer status (Medicaid/Medicare compared with commercial OR, 2.0; 95% CI, 1.3-3.0; P = .002) , were identified as risk factors for readmission.

Conclusions and Relevance  Many readmissions may be unavoidable in our current paradigms of care. While medical comorbidities are contributory, a large number of readmissions were not caused by suboptimal medical care or deterioration of medical conditions but by confounding issues of substance abuse or homelessness. Identification of the highest risk cohort for readmission can allow more targeted intervention for similar populations with socially challenged patients.

Introduction

Thirty-day readmissions to hospitals are coming under increasing scrutiny by the Centers for Medicare and Medicaid Services (CMS), policymakers, and hospitals. Since 2012, the Medicare Hospital Readmissions Reduction Program (MHRRP), established by the Affordable Care Act, financially penalizes hospitals with higher than expected readmission rates, using this metric as a marker of quality of care.1 While decreasing readmissions is consistent with improved quality within the medical home model of the Affordable Care Act, the validity of using a blanket metric of all readmissions to determine quality of care assumes that all are avoidable. The challenge of identifying potentially preventable readmissions and formulating protocols to reduce their rate has been the subject of many previous publications and reviews, but most have focused on medical conditions such as heart failure, pneumonia, and acute myocardial infarction, the 3 conditions specifically targeted by the MHRRP.2-6 Starting in fiscal year 2015, the MHRRP expanded the conditions that would incur financial penalties for hospitals to include 2 surgical procedures, total hip arthroplasty and knee arthroplasty, with future inclusion of additional specific operations anticipated.

Surgical conditions differ significantly from medical conditions, and the reasons for readmission of patients with surgical conditions can differ, with further distinction between those who did or did not undergo an operation. Process improvements that may decrease the rate of readmissions in patients with chronic medical conditions may not apply to the surgical population and many readmissions may not be preventable. To better inform policymakers as well as hospitals, a better understanding of reasons why patients are readmitted following discharge for surgical conditions is needed. We examined a consecutive convenience cohort of patients who were readmitted to any service within 30 days of discharge to better identify opportunities for intervention, improve care, and reduce the need for readmission.

Box Section Ref ID

Key Points

  • Question What are the reasons for readmission within 30 days of discharge from a general surgery service?

  • Findings In this study of 2100 patients who were discharged from the general surgery service during a 1-year period, 8.2% were readmitted within 30 days, with 31% owing to problems related to substance abuse, mental health comorbidities, or low socioeconomic status and only 9.2% owing to a complication of medical care.

  • Meaning Confounding issues of mental health, substance abuse, or homelessness need to be addressed to reduce readmissions.

Methods

We performed a retrospective medical record review of all patients discharged from the general surgery service at Harborview Medical Center in Seattle, Washington, from July 1, 2014, to June 30, 2015, and readmitted to any hospital service within 30 days of discharge. Harborview is the only Level I trauma center serving Washington, Alaska, Montana, and Idaho and also serves as a safety-net hospital for King County, Washington. Readmitted patients were identified by an administrative database that used criteria for readmission based on the CMS Hospital-Wide All-Cause Unplanned Readmission Measure.7 A subsequent manual medical record review process was completed to ensure that patients were accurately identified as discharges from the general surgery service (correct attribution) and that readmissions excluded those patients with planned readmission (eg, planned second-stage surgery). All valid readmissions underwent further medical record review to ascertain the primary clinical reason for the index admission and the reason for readmission. The reasons for readmission were categorized and assessed for preventability based on being related to a complication of previous medical care or a gap in care transitions. This study was approved as exempt by the institutional review board of the University of Washington because all data was deidentified for analysis. There was therefore not a need for consent to be obtained.

Patient characteristics that were collected from the administrative database were based on previously described risk assessment tools, LACE (length of stay, acuity of admission, comorbidities, and emergency department visits in last 6 months) and PARR (patients at risk for rehospitalization), and included sex, age, the presence of certain chronic medical conditions including heart disease and diabetes, and the presence of sepsis on admission and whether a patient had an intensive care unit (ICU) stay during the index admission.8,9 Additionally, we collected secondary diagnoses of mental health or substance abuse (as defined by Agency for Healthcare Research and Quality Comorbidity categories) and socioeconomic status, measured by presence or absence of housing and insurance status.10

The administrative database was then evaluated by means of univariate and multivariate logistic regression. First, we identified variables that were associated (P < .20) with readmission, the dependent variable. These potential confounders were then entered in multivariate stepwise (backward elimination) logistic regression, with readmission as the dependent variable. A logistic regression model was constructed to identify patient factors associated with readmission. All pairwise interaction terms of patient social conditions, such as homelessness, drug abuse history, and discharge disposition, were also considered to assess for potential effect modification. All statistical analyses were performed using Stata, version 14.0 (StataCorp). Statistical significance was defined as a P value less than .05.

Results

A total of 2328 patients were identified as being discharged from the general surgery service between July 1, 2014, and June 30, 2015, based on an administrative database (Figure). On medical record review, 198 patients were excluded owing to being identified as being transferred to other care facilities or leaving against medical advice. Of those remaining, another 30 were identified as being discharged by a service other than general surgery (attribution correction), leaving 2100 discharges from the general surgery service during this period. Of this group, 200 were identified as readmissions based on the criteria used by the MHRRP. Twenty-seven of these patients were further identified as being planned readmissions (eg, readmitted for planned delayed surgical fixation of a fracture after edema resolution or the second stage of a staged procedure), leaving 173 patients (8.2% of discharges) as the readmission group of interest.

Based on medical record review of those readmitted, the most common cause included 29 patients who were initially admitted with soft tissue infections from injection drug use requiring operative drainage and who were then readmitted with new soft tissue infections at other sites (16.8% of readmitted patients). Twenty-five readmitted patients (14.5%) were found to have lack of adequate social support leading to issues surrounding the discharge and follow-up process (eg, lack of home for postdischarge telephone calls, follow-up appointments not scheduled or not attended, postdischarge care needs underestimated). Together, these 2 groups made up almost a third of the readmissions (n = 54, 31.2%). Only 2 patients (1.2%) were readmitted owing to deterioration of a nonpsychiatric chronic condition (chronic obstructive pulmonary disease exacerbation and acute kidney injury owing to gout), but 23 readmissions (13.3%) were owing to some type of new infection not detectable during index admission, primarily surgical site infection (n = 11). Sixteen patients (9.2%) were readmitted owing to known sequelae of their initial injury or condition (eg, biloma formation after a grade 4 liver laceration, delayed hemothorax, and dehydration after ileostomy creation). Another 16 patients (9.2%) were identified as having a likely preventable complication of care leading to readmission (eg, development of a pulmonary embolism in a patient who was not receiving adequate prophylaxis and surgical site infection in a patient who had a leukocytosis on discharge), and represented less than 1% of all discharges. More than a third of the readmitted patients (n = 62, 35.8%) had no common theme and included reasons as varied as gastrointestinal bleeding after trauma (unrelated to injuries), pancreatitis following partial colectomy, and delayed development of partial small-bowel obstruction.

The characteristics of all discharged patients as well as the subset who were readmitted are shown in Table 1. When the administrative database was queried, on univariate analysis, female sex was noted to be significantly associated with risk of readmission (men to women risk of readmission odds ratio [OR], 0.5; 95% CI, 0.37-0.71; P < .001), as were physiologic factors such as presence of sepsis on the index admission (OR, 2.38; 95% CI, 1.55-3.68; P < .001), diabetes as a comorbidity (OR, 2.12; 95% CI, 1.44-3.13; P < .001), heart disease as a comorbidity (OR, 1.93; 95% CI, 1.35-2.77; P < .001), and whether the index admission included an ICU stay (defined as at least 1 night in the ICU) (OR, 1.65; 95% CI, 1.18-2.30; P = .002) (Table 2). Additionally, payer mix was significant, with those having Medicaid or Medicare having more than twice the likelihood of being readmitted compared with those with commercial insurance (OR, 2.06; 95% CI, 1.37-3.09; P < .001). Univariate analyses also demonstrated lack of housing (OR, 1.66; 95% CI, 1.07-2.57; P = .02) and presence of substance abuse (OR, 1.41; 95% CI, 1.02-1.96; P = .04) or mental health diagnoses (OR, 1.80; 95% CI, 1.25-2.58; P < .001) as risk factors. Patients discharged to respite care, who in our system are those patients who are both homeless and require nonacute care nursing needs, such as dressing changes of wounds, were also found to be at significantly higher risk of readmission (OR, 2.0; 95% CI, 1.3-4.4; P = .006). However, neither age nor whether the index admission was elective or nonelective was associated with any increased risk for readmission.

On multivariate analysis, female sex (men to women risk of readmission OR, 0.5; 95% CI, 0.36-0.69; P < .001), presence of diabetes (OR, 1.7; 95% CI, 1.1-2.6; P = .009), sepsis on admission (OR, 1.7; 95% CI, 1.05-2.6; P = .03), ICU stay during index admission (OR, 1.7; 95% CI, 1.2-2.4; P = .002), discharge to respite care (OR, 2.3; 95% CI, 1.2-4.5; P = .01), and payer mix were found to be risks for readmission (Medicaid/Medicare compared with commercial OR, 2.0; 95% CI, 1.3-3.0; P = .002) (Table 3).

Discussion

Although readmissions are an important target for care improvement, cost containment, and quality, there are significant limitations with using the Hospital-Wide All-Cause Unplanned Readmission Measure developed by CMS as an indicator of quality and the basis by which financial incentives or penalties are determined. This study used an administrative database based on the MHRRP criteria to identify and characterize patients who were readmitted within 30 days of discharge from a surgical service but added individual medical record review to more accurately define reasons for admission and readmission. The addition of medical record review also revealed many cases of planned readmissions and cases that were incorrectly attributed to the surgical service as the discharging team, identifying a truer cohort of unplanned, potentially avoidable readmissions to our team and confirming the well-known inaccuracy of administrative data sets.

Although we found varying reasons for readmission based on medical record review, a sizable group had conditions related to ongoing substance abuse. Another distinct group was those who appeared to have issues surrounding the discharge process, primarily owing to some aspect of insufficient social support after discharge such as lower socioeconomic status, lack of stable housing, or cognitive challenges leading to inability to keep follow-up appointments (eg, substance abuse or mental health diagnoses). Close follow-up is critical to identify and treat problems early to prevent decompensation that could require readmission to treat (eg, superficial surgical site infection progressing to fascial necrosis without early detection and treatment). These 2 groups were confirmed as having significantly increased risk for readmission on univariate analyses of all discharged patients, with substance abuse, payer mix, presence of mental health conditions, lack of housing, and those discharged to respite care identified as risk factors. The population going to respite care is additionally socially disadvantaged because not only are these patients without stable homes, they also lack family or friends that can aid in postdischarge care. This also argues in support of the Affordable Care Act concept of shared risk and need to optimize ongoing medical care postdischarge: the concept of Patient-Centered Medical Homes.11 Directed interventions at these groups could have prevented up to a third of our readmissions.

Physiological risk factors for readmission were also found on statistical analysis, including presence of sepsis or ICU stay on index admission and diabetes as a comorbidity, consistent with medical record review findings of new infections as a significant reason for readmission. Because these factors were still present as risks on multivariate analysis, they are important contributors leading to readmission, and continued efforts to critically analyze in-hospital and postdischarge medical care of these groups is a necessity. However, the fact that these risk factors were not evident as contributing risks for readmission based on medical record review also supports the premise that medical record review is a more accurate assessment of a patient’s admission and readmission condition than administrative databases that use International Classification of Diseases, Ninth Revision codes.

The finding that older age was not a risk factor was surprising, although this could be because the elderly population is at highest risk of dying before having a chance for readmission, as described by Fawcett et al12 in the trauma population. However, the finding of female sex as a significant risk factor for readmission is difficult to explain and possibly caused by as yet unidentified biases involving discharge criteria for female patients vs male patients. Other possibilities include men dying at a higher rate before readmission than women or loss of spousal support being more prevalent and influential for women. Importantly, on medical record review, our data found that only a small percentage of readmitted patients was determined to have a potentially preventable complication of medical care at their index admission, further questioning the metric of using readmissions as a surrogate for quality, although arguably the groups that returned with new infections and progression of surgical disease could also be categorized as being potentially preventable and is worth further examination.

The inaccuracies of using administrative databases that use International Classification of Diseases, Ninth Revision codes for reasons for readmission have been described previously, with the CMS readmission tool having difficulty accurately accounting for planned readmissions and failing to correctly identify the reason for hospital readmission in nearly one-third of cases.13 Moreover, the validity of hospital readmissions as an indicator of quality of care depends on the assumption that readmissions are avoidable if medical care is optimized. However, there is little agreement regarding the true proportion of readmissions that are potentially avoidable; it is certainly not zero. A systematic review of the literature by van Walraven et al14 examining readmission rates demonstrated ranges from 5% to 79%, further calling into question the method (“higher than expected”) by which financial penalties are calculated by the MHRRP.

Previous studies have tried to identify patient groups at risk for readmission but have used administrative databases that lack the qualitative data necessary to be applicable to individual institutions. Using data from the American College of Surgeons National Surgical Quality Improvement Program, which is a large administrative database that includes more clinically abstracted information, Merkow et al15 in 2015 were able to ascertain with more granularity the reasons for readmissions in this group, finding surgical site infections, postoperative ileus or obstruction, and bleeding as the most common reasons for readmission within 30 days of surgery (not discharge). While these diagnoses could arguably be preventable conditions, owing to the targeted patient sample that included only those undergoing a limited number of operations as defined by American College of Surgeons National Surgical Quality Improvement Program, these conclusions may not be generalizable to most patients admitted to our surgical service, many of whom may include trauma patients who may not have undergone an operation. In our series, there were 11 readmissions for surgical site infections (6.4% of all readmissions), 7 for postoperative ileus or obstruction (4.0%), and 2 for bleeding (1.1%). A 2015 state-level analysis of surgical readmissions by Schmeida and Savrin16 revealed many different categories of variables were associated with readmission, including demographic (age and language), clinical process, hospital capacity, and patient need, most of which are not included in calculations for risk adjustment by the readmissions tool.

The importance of social factors surrounding certain patient groups cannot be underappreciated. Calvillo-King et al17 performed a systematic review assessing the effect of social factors on readmission and mortality on heart failure and pneumonia and found a broad spectrum of social factors, such as housing stability and lack of social support, that were associated with worse outcomes in these 2 common diseases, again suggesting that strategies aside from improving index admission medical care were important to consider. Barnett et al18 demonstrated that when several patient characteristics, such as education, income, and functional status (which again are not included in usual risk adjustments for readmission), are accounted for, the differences between hospitals with the lowest readmissions and those with the highest readmissions are nearly cut in half. Those authors conclude that “differences in patient characteristics between hospitals may contribute substantially to the penalties levied by Medicare on hospitals with high readmissions rates.”18 The burden of correcting societal inequities cannot be borne by those few safety-net hospitals least able financially to provide this support.

Regardless of the validity of using readmissions as a surrogate for quality, there is little disagreement that readmissions are costly to hospitals from both financial and resource use perspectives; any efforts to identify at-risk populations and implement systems to try to reduce the revolving door of care for these patients are critical for an institution’s fiscal health. While processes to improve in-hospital and postdischarge care have proven to be helpful for the care of medical conditions, few data exist regarding what interventions can be helpful to decrease readmissions in socially vulnerable populations. Evidence from several randomized studies suggests that interventions derived from motivational interviewing principles may reduce alcohol consumption and related alcohol use problems among patients presenting to trauma centers.19-21 Such a strategy may be beneficial for nontrauma patients with substance abuse problems as confounders of their admission diagnoses, although to our knowledge, none have been described in the literature. Additionally, a Cochrane review in 2014 assessing whether psychosocial interventions for substance use reduction in people with the “dual diagnosis” of substance abuse and mental illness did not find convincing evidence that these interventions were advantageous over treatment as usual, suggesting that behavioral therapy alone is not sufficient to effect sustained changes when mental illness is also present.22 For those patients otherwise socially disadvantaged, novel ideas, such as using trained college students as community health workers to help vulnerable patients after discharge with home visits, goal setting, and ensuring they have proper resources to maintain their health, demonstrated decreases in both readmissions and emergency department visits.23 Indeed, a combination of behavioral therapies as well as programs and processes to include significant postdischarge social support may be necessary for populations at the highest risk.

Our study has limitations. The analysis included only patients readmitted to our own institution, and therefore, we did not account for patients who were discharged from our hospital and readmitted to another within 30 days, which may account for the overall low readmission rate, although it is unlikely in this population. We also did not include data on patients who died within 30 days of discharge, and thus, this group may have influenced the conclusions. As a single-institution study, the degree to which our findings can be generalizable to other hospitals is unknown, especially those that are not designated as safety-net or trauma hospitals. However, while safety-net hospitals treat a larger population of patients with socioeconomic issues, all hospitals admit patients with similar issues, and therefore, the results can be applicable to these populations. Additionally, having a single reviewer to ascertain reason for readmission may be less accurate than using multiple reviewers to confirm.

Conclusions

Many cases of readmissions may truly be unavoidable in our current paradigms of care because we found socially fragile populations to be at as high risk as those that are medically fragile. A large number of readmissions were not owing to suboptimal medical care delivered at the index admission to our surgery service, but many were owing to confounding issues of mental health, substance abuse, or homelessness, issues that likely require more intense in-hospital and postdischarge social support than most hospitals, including our own, can currently provide. Because interventions to reduce the risk of readmission for any group of patients can be costly and labor intensive, identification of the highest risk cohort for readmission can allow more targeted intervention for this population of socially vulnerable patients.

Back to top
Article Information

Corresponding Author: Lisa K. McIntyre, MD, Department of Surgery, University of Washington Medical Center, 325 Ninth Ave, PO Box 359796, Seattle, WA 98104 (lmcintyr@uw.edu).

Accepted for Publication: March 25, 2016.

Published Online: June 15, 2016. doi:10.1001/jamasurg.2016.1258

Author Contributions: Dr McIntyre had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: McIntyre, Maier.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: McIntyre, Arbabi, Robinson.

Critical revision of the manuscript for important intellectual content: McIntyre, Arbabi, Maier.

Statistical analysis: Arbabi.

Administrative, technical, or material support: McIntyre, Robinson, Maier.

Study supervision: McIntyre, Maier.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This paper was presented at the 86th Annual Meeting of the 2016 Pacific Coast Surgical Association; February 14, 2016; Kona, Hawaii.

References
1.
Centers for Medicare and Medicaid Services. Readmissions reduction program. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Readmissions-Reduction-Program.html. Accessed March 34, 2016.
2.
Echeverry  LM, Lamb  KV, Miller  J.  Impact of APN home visits in reducing healthcare costs and improving function in homebound heart failure.  Home Healthc Now. 2015;33(10):532-537.PubMedGoogle Scholar
3.
Ziaeian  B, Fonarow  GC.  The prevention of hospital readmissions in heart failure.  Prog Cardiovasc Dis. 2016;58(4):379-385.PubMedGoogle ScholarCrossref
4.
Prescott  HC, Sjoding  MW, Iwashyna  TJ.  Diagnoses of early and late readmissions after hospitalization for pneumonia: a systematic review.  Ann Am Thorac Soc. 2014;11(7):1091-1100.PubMedGoogle ScholarCrossref
5.
Arnold  SV, Smolderen  KG, Kennedy  KF,  et al.  Risk factors for rehospitalization for acute coronary syndromes and unplanned revascularization following acute myocardial infarction.  J Am Heart Assoc. 2015;4(2):e001352.PubMedGoogle ScholarCrossref
6.
Hansen  LO, Young  RS, Hinami  K, Leung  A, Williams  MV.  Interventions to reduce 30-day rehospitalization: a systematic review.  Ann Intern Med. 2011;155(8):520-528.PubMedGoogle ScholarCrossref
8.
van Walraven  C, Dhalla  IA, Bell  C,  et al.  Derivation and validation of an index to predict early death or unplanned readmission after discharge from hospital to the community.  CMAJ. 2010;182(6):551-557.PubMedGoogle ScholarCrossref
9.
Billings  J, Dixon  J, Mijanovich  T, Wennberg  D.  Case finding for patients at risk of readmission to hospital: development of algorithm to identify high risk patients.  BMJ. 2006;333(7563):327.PubMedGoogle ScholarCrossref
11.
Edwards  ST, Abrams  MK, Baron  RJ,  et al.  Structuring payment to medical homes after the affordable care act.  J Gen Intern Med. 2014;29(10):1410-1413.PubMedGoogle ScholarCrossref
12.
Fawcett  VJ, Flynn-O’Brien  KT, Shorter  Z,  et al.  Risk factors for unplanned readmissions in older adult trauma patients in Washington State: a competing risk analysis.  J Am Coll Surg. 2015;220(3):330-338.PubMedGoogle ScholarCrossref
13.
Sacks  GD, Dawes  AJ, Russell  MM,  et al.  Evaluation of hospital readmissions in surgical patients: do administrative data tell the real story?  JAMA Surg. 2014;149(8):759-764.PubMedGoogle ScholarCrossref
14.
van Walraven  C, Bennett  C, Jennings  A, Austin  PC, Forster  AJ.  Proportion of hospital readmissions deemed avoidable: a systematic review.  CMAJ. 2011;183(7):E391-E402.PubMedGoogle ScholarCrossref
15.
Merkow  RP, Ju  MH, Chung  JW,  et al.  Underlying reasons associated with hospital readmission following surgery in the United States.  JAMA. 2015;313(5):483-495.PubMedGoogle ScholarCrossref
16.
Schmeida  M, Savrin  R.  Surgical rehospitalization of the medicare fee-for-service patient: a state-level analysis exploring 30-day readmission factors.  Prof Case Manag. 2015;20(3):130-137.PubMedGoogle ScholarCrossref
17.
Calvillo-King  L, Arnold  D, Eubank  KJ,  et al.  Impact of social factors on risk of readmission or mortality in pneumonia and heart failure: systematic review.  J Gen Intern Med. 2013;28(2):269-282.PubMedGoogle ScholarCrossref
18.
Barnett  ML, Hsu  J, McWilliams  JM.  Patient characteristics and differences in hospital readmission rates.  JAMA Intern Med. 2015;175(11):1803-1812.PubMedGoogle ScholarCrossref
19.
Gentilello  LM, Rivara  FP, Donovan  DM,  et al.  Alcohol interventions in a trauma center as a means of reducing the risk of injury recurrence.  Ann Surg. 1999;230(4):473-480.PubMedGoogle ScholarCrossref
20.
Zatzick  D, Roy-Byrne  P, Russo  J,  et al.  A randomized effectiveness trial of stepped collaborative care for acutely injured trauma survivors.  Arch Gen Psychiatry. 2004;61(5):498-506.PubMedGoogle ScholarCrossref
21.
Schermer  CR, Moyers  TB, Miller  WR, Bloomfield  LA.  Trauma center brief interventions for alcohol disorders decrease subsequent driving under the influence arrests.  J Trauma. 2006;60(1):29-34.PubMedGoogle ScholarCrossref
22.
Hunt  GE, Siegfried  N, Morley  K, Sitharthan  T, Cleary  M.  Psychosocial interventions for people with both severe mental illness and substance misuse.  Schizophr Bull. 2014;40(1):18-20.PubMedGoogle ScholarCrossref
23.
Thew  J. Hospital cuts readmissions in half with help from college students. HealthLeaders Media. September 1, 2015. http://www.healthleadersmedia.com/nurse-leaders/hospital-cuts-readmissions-half-help-college-students. Accessed October 29, 2015.
×