[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Readmission Rates by Demographic and Clinicopathological Characteristic Among Total Hip Replacement Patients of Derivation Cohort
Readmission Rates by Demographic and Clinicopathological Characteristic Among Total Hip Replacement Patients of Derivation Cohort
Figure 2.
Readmission Percentage vs Readmission After Total Hip Replacement Risk Scale Scores Among Validation Cohort Patients
Readmission Percentage vs Readmission After Total Hip Replacement Risk Scale Scores Among Validation Cohort Patients
Table 1.  
Components of the RATHRR Score
Components of the RATHRR Score
Table 2.  
Demographic and Comorbidities Data Involving 268 518 Total Hip Replacement Patients of Derivation Cohort
Demographic and Comorbidities Data Involving 268 518 Total Hip Replacement Patients of Derivation Cohort
Table 3.  
Univariate and Multivariate Analysis of Factors Associated With Readmission After Total Hip Replacement of the Derivation Cohort
Univariate and Multivariate Analysis of Factors Associated With Readmission After Total Hip Replacement of the Derivation Cohort
1.
Soohoo  NF, Farng  E, Lieberman  JR, Chambers  L, Zingmond  DS.  Factors that predict short-term complication rates after total hip arthroplasty.  Clin Orthop Relat Res. 2010;468(9):2363-2371.PubMedGoogle ScholarCrossref
2.
Mahomed  NN, Barrett  JA, Katz  JN,  et al.  Rates and outcomes of primary and revision total hip replacement in the United States medicare population.  J Bone Joint Surg Am. 2003;85-A(1):27-32.PubMedGoogle Scholar
3.
Canale  ST, Beaty  JH.  Canale & Beaty: Campbell’s Operative Orthopaedics. 12th ed. Philadelphia, PA: Mosby; 2012.
4.
HCUPnet. Healthcare Cost and Utilization Project. http://hcupnet.ahrq.gov/HCUPnet.jsp?Id=3C546A58544D3669&Form=DispTab&JS=Y&Action=Accept. Published 2012. Accessed July 24, 2013.
5.
Kim  S.  Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997-2004.  Arthritis Rheum. 2008;59(4):481-488.PubMedGoogle ScholarCrossref
6.
Harms  S, Larson  R, Sahmoun  AE, Beal  JR.  Obesity increases the likelihood of total joint replacement surgery among younger adults.  Int Orthop. 2007;31(1):23-26.PubMedGoogle ScholarCrossref
7.
Vasarhelyi  EM, MacDonald  SJ.  The influence of obesity on total joint arthroplasty.  J Bone Joint Surg Br. 2012;94(11)(suppl A):100-102.PubMedGoogle ScholarCrossref
8.
Ellimoottil  C, Miller  S, Davis  M, Miller  DC.  Insurance expansion and the utilization of inpatient surgery: evidence for a “woodwork” effect?  Surg Innov. 2015;22(6):588-592.PubMedGoogle ScholarCrossref
9.
Kurtz  S, Ong  K, Lau  E, Mowat  F, Halpern  M.  Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030.  J Bone Joint Surg Am. 2007;89(4):780-785.PubMedGoogle ScholarCrossref
10.
Mont  MA, Issa  K.  Updated projections of total joint arthroplasty demands in America: commentary on an article by Steven M. Kurtz, PhD, et al: “impact of the economic downturn on total joint replacement demand in the united states. updated projections to 2021.”  J Bone Joint Surg Am. 2014;96(8):e68.PubMedGoogle ScholarCrossref
11.
Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group. National health care expenditures data. https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpenddata/downloads/proj2012.pdf. Published 2012. Accessed July 22, 2013.
12.
Fuchs  VR.  The gross domestic product and health care spending.  N Engl J Med. 2013;369(2):107-109.PubMedGoogle ScholarCrossref
13.
Murray  CJL, Frenk  J.  Ranking 37th–measuring the performance of the US health care system.  N Engl J Med. 2010;362(2):98-99.PubMedGoogle ScholarCrossref
14.
The Patient Protection and Affordable Care Act, HR 3590, 111th Cong (2010).
15.
Williams  MV.  A requirement to reduce readmissions: take care of the patient, not just the disease.  JAMA. 2013;309(4):394-396.PubMedGoogle ScholarCrossref
16.
Henry J. Kaiser Family Foundation. Readmissions cost Medicare billions. Kaiser Health News. May 29, 2009.
17.
Jiang  HJ, Wier  LM; Healthcare Cost and Utilization Project.  All-Cause Hospital Readmission among Non-Elderly Medicaid Patients, 2007. Rockville, MD: Agency for Healthcare Research and Quality; 2010.
18.
Weiss  AJ, Elizhauser  A, Steiner  C; Healthcare Cost and Utilization Project.  Readmissions to US Hospitals by Procedure, 2010. Rockville, MD: Agency for Healthcare Research and Quality; 2013.
19.
Jencks  SF, Williams  MV, Coleman  EA.  Rehospitalizations among patients in the Medicare fee-for-service program.  N Engl J Med. 2009;360(14):1418-1428.PubMedGoogle ScholarCrossref
20.
Cram  P, Lu  X, Kaboli  PJ,  et al.  Clinical characteristics and outcomes of Medicare patients undergoing total hip arthroplasty, 1991-2008.  JAMA. 2011;305(15):1560-1567.PubMedGoogle ScholarCrossref
21.
Jordan  CJ, Goldstein  RY, Michels  RF, Hutzler  L, Slover  JD, Bosco  JA  III.  Comprehensive program reduces hospital readmission rates after total joint arthroplasty.  Am J Orthop (Belle Mead NJ). 2012;41(11):E147-E151.PubMedGoogle Scholar
22.
Cullen  C, Johnson  DS, Cook  G.  Re-admission rates within 28 days of total hip replacement.  Ann R Coll Surg Engl. 2006;88(5):475-478.PubMedGoogle ScholarCrossref
23.
Zhan  C, Kaczmarek  R, Loyo-Berrios  N, Sangl  J, Bright  RA.  Incidence and short-term outcomes of primary and revision hip replacement in the United States.  J Bone Joint Surg Am. 2007;89(3):526-533.PubMedGoogle ScholarCrossref
24.
Agabiti  N, Picciotto  S, Cesaroni  G,  et al; Italian Study Group on Inequalities in Health Care.  The influence of socioeconomic status on utilization and outcomes of elective total hip replacement: a multicity population-based longitudinal study.  Int J Qual Health Care. 2007;19(1):37-44.PubMedGoogle ScholarCrossref
25.
Rajgopal  R, Martin  R, Howard  JL, Somerville  L, MacDonald  SJ, Bourne  R.  Outcomes and complications of total hip replacement in super-obese patients.  Bone Joint J. 2013;95-B(6):758-763.PubMedGoogle ScholarCrossref
26.
Bozic  KJ, Lau  E, Ong  K,  et al.  Risk factors for early revision after primary total hip arthroplasty in Medicare patients.  Clin Orthop Relat Res. 2014;472(2):449-454.PubMedGoogle ScholarCrossref
27.
White  RH, Gettner  S, Newman  JM, Trauner  KB, Romano  PS.  Predictors of rehospitalization for symptomatic venous thromboembolism after total hip arthroplasty.  N Engl J Med. 2000;343(24):1758-1764.PubMedGoogle ScholarCrossref
28.
Namba  RS, Paxton  L, Fithian  DC, Stone  ML.  Obesity and perioperative morbidity in total hip and total knee arthroplasty patients.  J Arthroplasty. 2005;20(7)(suppl 3):46-50.PubMedGoogle ScholarCrossref
29.
Willis-Owen  CA, Konyves  A, Martin  DK.  Factors affecting the incidence of infection in hip and knee replacement: an analysis of 5277 cases.  J Bone Joint Surg Br. 2010;92(8):1128-1133.PubMedGoogle ScholarCrossref
30.
Wallace  G, Judge  A, Prieto-Alhambra  D, de Vries  F, Arden  NK, Cooper  C.  The effect of body mass index on the risk of post-operative complications during the 6 months following total hip replacement or total knee replacement surgery.  Osteoarthritis Cartilage. 2014;22(7):918-927.PubMedGoogle ScholarCrossref
31.
Guide  U.  HCUP Supplemental Values for Revisit Analyses. Healthcare Cost and Utilization Project (HCUP). Rockville, MD: Agency for Healthcare Research and Quality; 2015.
32.
Dharmarajan  K, Hsieh  AF, Lin  Z,  et al.  Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia.  JAMA. 2013;309(4):355-363.PubMedGoogle ScholarCrossref
33.
Scott  KW, Jha  AK.  Putting quality on the global health agenda.  N Engl J Med. 2014;371(1):3-5.PubMedGoogle ScholarCrossref
34.
Ng  VY, Lustenberger  D, Hoang  K,  et al.  Preoperative risk stratification and risk reduction for total joint reconstruction: AAOS exhibit selection.  J Bone Joint Surg Am. 2013;95(4):e191-e15.PubMedGoogle ScholarCrossref
Original Investigation
August 2016

A Preoperative Scale for Determining Surgical Readmission Risk After Total Hip Replacement

Author Affiliations
  • 1Department of Surgery, Saint Barnabas Medical Center, Livingston, New Jersey
  • 2Duke University, Durham, North Carolina
  • 3Department of Surgery, Rutgers University–New Jersey Medical School, Newark
  • 4Saint George’s University School of Medicine, Grenada, West Indies
 

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

JAMA Surg. 2016;151(8):701-709. doi:10.1001/jamasurg.2016.0020
Abstract

Importance  Total hip replacement is a commonly performed orthopedic procedure for the treatment of painful arthritis, osteonecrosis, or fracture.

Objective  To develop and verify a scale for predicting readmission rates for total hip replacement patients and allow for the development and implementation of readmission risk-reduction strategies.

Design, Setting, and Participants  Discharge data on 268 518 patients from New York and California (derivation cohort) and 153 560 patients from Florida and Washington (validation cohort) were collected from the State Inpatient Database, a part of the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality (2006-2011). Analysis of the derivation cohort was performed in July 2013 and analysis of the validation cohort was performed in August 2014. Demographic and clinical characteristics of patients undergoing total hip replacement were abstracted. The Readmission After Total Hip Replacement Risk Scale was developed to predict readmission risk.

Main outcome and measure  Readmission rate.

Results  Of the 268 518 patients from New York and California (derivation cohort), 151 009 (56.2%) were women and 216 477 (80.6%) were white. Of the 153 560 patients from Florida and Washington (validation cohort), 86 534 (56.3%) were women and 120 591 (78.5%) were white. Overall 30-day readmission rate was 5.89% for the derivation cohort and 5.82% for the validation cohort. Readmission rates for men and women were 5.79% and 6.08% for the derivation cohort (odds ratio [OR], 1.05; 95% CI, 1.02-1.09) and 5.80% and 5.84% for the validation cohort (OR, 0.99; 95% CI, 0.95-1.04), respectively. The following were all determined to be associated with increased risk of readmission after total hip replacement: being older than 71 years (OR, 1.83; 95% CI, 1.77-1.89), African American (OR, 1.23; 95% CI, 1.15-1.31), and in the lowest income quartile (OR, 1.18; 95% CI, 1.12-1.24); revision replacement (OR, 1.82, 95% CI, 1.75-1.90); liver disease (OR, 1.57; 95% CI, 1.39-1.77); congestive heart failure (OR, 1.49; 95% CI, 1.38-1.61); chronic pulmonary disease (OR, 1.33; 95% CI, 1.27-1.39); renal failure (OR, 1.26; 95% CI, 1.18-1.36); diabetes (OR, 1.21; 95% CI, 1.16-1.27); fluid and electrolyte disorder (OR, 1.21; 95% CI, 1.14-1.27); anemia (OR, 1.19; 95% CI, 1.15-1.25); rheumatoid arthritis (OR, 1.19; 95% CI, 1.10-1.29); coagulopathy (OR, 1.19; 95% CI, 1.08-1.32); hypertension (OR, 1.17; 95% CI, 1.12-1.21); and obesity (OR, 1.15; 95% CI 1.09-1.21). They were used to create the Readmission After Total Hip Replacement Risk Scale, which was applied to the validation cohort and explained 89.1% of readmission variability in that cohort.

Conclusions and Relevance  Data derived from patients in the New York and California State Inpatient Database were reliably able to explain readmission variability for patients in the Florida and Washington State Inpatient Database at a rate of 89.1% based on known preoperative risk factors. Risk-stratification models, such as the Readmission After Total Hip Replacement Risk Scale, can identify high-risk patients for readmission and permit implementation of patient-specific readmission-reduction strategies to reduce readmissions and health care expenditures.

Introduction

Total hip replacement (THR) is the most common orthopedic procedure performed in the United States to treat patients experiencing hip arthritis, osteonecrosis, or fracture to reduce pain and improve function.1-3 Approximately 345 860 total or revision hip replacements were performed in the United States in 2013.4 The number of total or revision hip replacements performed annually has increased exponentially, from 159 580 in 1993 to 343 095 in 2012.4 The rapidly growing arthritic aging and obese population, along with the expansion of health insurance access, has further increased the need for THR.5-7 Ellimoottil et al8 demonstrated that insurance reform in Massachusetts in 2007 led to a 9.3% increase in discretionary surgery, including THR, and similar increases are to be expected with the implementation of the Affordable Care Act. The number of primary and revision hip replacements is estimated to exceed 668 700 by 2030 if these trends continue.9,10

In 2012, US health care costs approximated $2.8 trillion, accounting for 17.2% of the gross domestic product, compared with $260 billion and 9.18% of the gross domestic product in 1980.11 If these trends continue, health care spending could reach 30% of the gross domestic product by 2040.12 Charges for primary and revision hip replacements were $9.9 billion in 2009 and were projected to exceed $25 billion by 2015.5 Current US health care spending exceeds all other countries in the world, but ranks 37th in performance, demonstrating no benefits in quality of care.13 Other countries provide better health care at a fraction of the cost, leaving the United States with the difficult task of simultaneously reducing spending while improving quality.

In 2010, President Barack Obama signed the Affordable Care Act,14 which contained the Hospital Readmissions Reduction Program. Beginning in 2012, this program imposes financial penalties for excess readmissions for specific procedures and diagnoses.15 The initial program aimed to reduce readmissions for heart failure, pneumonia, and acute myocardial infarction. Strict penalties were imposed for failing to reduce readmission rates in an attempt to reduce the 25% of annual Centers for Medicare and Medicaid Services budget spent on inpatient care costs. In 2015, the program expanded to include THR.

In 2004, nearly 1 in 5 Medicare beneficiaries was readmitted within 30 days of discharge, costing approximately $17.5 billion.16-19Quiz Ref ID Previous studies reported 30-day readmission rates for THR patients as 3.5% to 8.5%.18,20-23 Studies have identified advanced age, comorbid diseases, male sex, African American race, obesity, lower income status, and revision procedures as specific risk factors for readmission.1,2,17,23-30 Given the wide range of published THR readmission rates, conflicting reports on factors affecting readmission risk, and pending reduced reimbursements for excessive readmissions, it is imperative to understand readmission risk and develop readmission-reduction strategies.

This study sought to evaluate a large, population-based state database to calculate prevalence and outcomes of THR over a 6-year period to identify specific factors associated with readmissions and to create a scale that could reliably stratify preoperative readmission risk. The scale was then validated on a separate and distinct patient cohort to determine its predictability and accuracy to assess an individual’s risk for readmission and allow for implementation of risk-reduction strategies.

Box Section Ref ID

Key Points

  • Question: What factors are associated with readmission following total hip replacement and how can physicians reliably stratify preoperative risk factors to identify high-risk patients?

  • Findings: A total of 268 518 hip replacement patients from the State Inpatient Database were analyzed, and factors associated with an increased risk of readmission were used to create the Readmission After Total Hip Replacement Risk Scale. The Readmission After Total Hip Replacement Risk Scale was applied to a validation cohort of 153 560 total hip replacement patients and predicted readmission risk with 89.1% accuracy.

  • Meaning: Risk-stratification models, such as the Readmission After Total Hip Replacement Risk Scale, identify high-risk patients, permitting implementation of patient-specific readmission-reduction strategies to reduce readmissions and overall health care expenditures.

Methods

Discharge data for 39 667 498 patients from New York and California and 19 472 544 patients from Florida and Washington were obtained from the 2006 to 2011 State Inpatient Database (SID), a part of the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality. Institutional review board approval (#13-36) was obtained from Saint Barnabas Medical Center’s institutional review board prior to review. Because this was a retrospective study using the SID, patient consent was not obtained.

Data on patients who had undergone total hip replacement (International Classification of Diseases, Ninth Revision code 815.1) or revision hip replacement (International Classification of Diseases, Ninth Revision codes 815.3, 00.70, 00.71, 00.72, and 00.73) as a primary procedure were abstracted. Patients with a diagnosis of femur or acetabulum fracture, those younger than 20 years, and those with a missing identifier variable used to track repeated visits were excluded from this study (eFigure 1 in the Supplement). Demographic and clinicopathologic factors associated with 30-day THR readmission in previous literature were abstracted including age group (in 10-year increments starting at 21 years to facilitate application of the scale), sex, race/ethnicity, income quartile, type of procedure (primary or revision), hospital length of stay (LOS), discharge location, material used for bearings, readmission status and diagnosis, and presence of comorbidities (including rheumatoid arthritis, obesity, hypertension, diabetes mellitus, chronic pulmonary disease, anemia, renal failure, fluid and electrolyte disorder, congestive heart failure, coagulopathy, and liver disease).1,2,17,23-30

Readmission status was determined based on a difference of 30 days or less between the initial procedure date and any subsequent inpatient visits. Visits were then screened to exclude any primary diagnosis indicating a follow-up of unrelated procedures or a follow-up of rehabilitation for joint replacement, diagnoses not indicative of a complication from the initial THR. Complication rate data were extracted using International Classification of Diseases, Ninth Revision codes (eTable 1 in the Supplement).

Statistical analysis was performed using χ2 tests, and univariate analysis was performed using binary logistic regression. A P value of less than .01 was considered statistically significant. Multivariate logistic regression analysis assessed factors determined to be statistically significant in univariate analysis. Odds ratios (ORs) and 95% CIs were calculated to determine strength of association within each subgroup. Data analysis was performed using SAS statistical software, version 9.3 (SAS Institute, Inc).31 As per SID reporting guidelines, values below 10 were not reported.

Variables with an OR of 1.1 or greater and a P value of less than .01 for the study population from New York and California (derivation cohort) were deemed significantly associated with readmission after THR and were used to construct the Readmission After Total Hip Replacement Risk (RATHRR) Scale. These values were chosen a priori in an attempt to avoid giving too much weight to any factors. Scores for each OR were calculated using the following formula:

(OR−1/Σ[OR−1])×100% and then calibrated to a 100-point scale to facilitate use by health care physicians (Table 1). To validate the scale, a RATHRR Scale score was generated for patients in a study population from Florida and Washington (validation cohort) and linear regression was performed to determine the fit of the RATHRR Scale to an independent data set. The RATHRR Scale was only applied to individuals without missing variables. Linear regression was performed on the graph of percentage of patients readmitted vs RATHRR score so a 1-point increase in score could be quantified easily and so physicians would be able to compare their patients’ scores with the national standard set by the Centers for Medicare and Medicaid Services. The goal was to create a scale that would have an R2 value of at least 0.80 in the validation cohort.

Results
Demographics and Clinicopathological Data

A total of 268 518 patients from the derivation cohort and 153 560 patients from the validation cohort were identified in the SID between 2006 and 2011 to have undergone either primary or revision total hip replacement. Clinicopathological characteristics for the derivation and validation cohorts are detailed in Table 2 and eTable 2 in the Supplement, respectively. Demographic and clinicopathological distributions for both populations were similar. Quiz Ref IDIn the validation cohort, 86.8% (n = 133 229) of patients underwent primary THR, while 13.2% (n = 20 331) underwent revision THR. Most THRs were performed in patients aged 61 to 70 years (n = 44 471 [29.0% of the validation cohort]). The mean (SD) age for primary THR patients in the validation cohort was 66.4 (12.1) years. For revision THR patients, the mean (SD) age was 68.3 (13.3) years in the validation cohort. In both populations, women made up 56% of the population (n = 86 534 in the validation cohort). The greatest sex difference was observed among patients older than 91 years (male:female, 1:2.98 [derivation cohort] and 1:2.28 [validation cohort]). Total hip replacement rates among all age groups increased over the 6-year study period. Among the validation cohort, 78.5% (n = 120 591) were white, 4.56% (n = 6998) were African American, and 4.89% (n = 7508) were identified as other race/ethnicity. Two hundred seventy-three (0.18%) patients in the validation cohort died during initial hospitalization. The most common comorbidities were hypertension (59.2% of the validation cohort, n = 90 843), anemia (21.9% of the validation cohort, n = 33 646), and chronic pulmonary disease (14.0% of the validation cohort, n = 21 502). Quiz Ref IDLength of stay increased as patient age increased in both populations, with more than 21% of patients older than 91 years requiring a stay of more than 7 days, whereas less than 15% of patients required a stay of more than 7 days in all other age groups.

Readmissions

Figure 1 details readmission rates for each demographic and clinicopathological characteristic in the derivation cohort. The overall readmission rate for this population was 5.89%. The validation cohort’s readmission rates for demographic and clinicopathological characteristics are detailed in eFigure 2 in the Supplement. The overall readmission rate among this cohort was 5.82%, with readmission rates highest among patients older than 91 years (14.8%; OR, 3.18; 95% CI, 2.64-3.82), those undergoing a revision THR (10.7%; OR, 2.24; 95% CI, 2.13-2.36), and patients with congestive heart failure (14.6%; OR, 2.88; 95% CI, 2.63-3.16) or renal failure (11.9%; OR, 2.28; 95% CI, 2.10-2.47) on univariate analysis.

The most common reasons for readmission were prosthetic complications (12.9%, n = 4590) and surgical site infections (12.8%, n = 4560) (eTable 3 in the Supplement). The most common complication among patients younger than 65 years was surgical site infection (17.2%), while those older than 65 years most often presented with prosthetic complications (13.3%).

Univariate Analysis

On univariate analysis of the derivation cohort, all analyzed variables significantly affected readmission rates. Odds ratios and 95% CIs are detailed in Table 3. On univariate analysis of the validation cohort, similarly, all analyzed variables significantly affected readmission rates except sex. Age (P < .01), race/ethnicity (P < .01), income quartile (P < .01), revision procedure (OR, 2.24; 95% CI, 2.13-2.36), rheumatoid arthritis (OR, 1.38; 95% CI, 1.26-1.52), obesity (OR, 1.18; 95% CI, 1.10-1.26), hypertension (OR, 1.35; 95% CI, 1.29-1.41), diabetes (OR, 1.38; 95% CI, 1.30-1.46), chronic pulmonary disease (OR, 1.62; 95% CI, 1.54-1.71), anemia (OR, 1.42; 95% CI, 1.35-1.49), renal failure (OR, 2.28; 95% CI, 2.10-2.47), fluid and electrolyte disorder (OR, 1.73; 95% CI, 1.64-1.84), congestive heart failure (OR, 2.88; 95% CI, 2.63-3.16), coagulopathy (OR, 1.63; 95% CI, 1.47-1.81), and liver disease (OR, 1.64; 95% CI, 1.38-1.94) were associated with increased THR readmission rates (eTable 4 in the Supplement).

Multivariate Analysis

Quiz Ref IDOn multivariate analysis of the derivation cohort, sex was the only factor to not independently affect readmission rates and was not included in the RATHRR Scale. Odds ratios and 95% CIs for the derivation cohort are displayed in Table 3. On multivariate analysis of the validation cohort, coagulopathy did not independently affect readmission rates. The following were associated with increased readmission rates: being older than 71 years (OR, 1.83; 95% CI, 1.77-1.89), African American (OR, 1.30; 95% CI, 1.18-1.44), and in the first income quartile (OR, 1.12; 95% CI, 1.04-1.19); revision THR (OR, 1.96; 95% CI, 1.85-2.07); rheumatoid arthritis (OR, 1.25; 95% CI, 1.13-1.39); obesity (OR, 1.23; 95% CI, 1.14-1.32); hypertension (OR, 1.11; 95% CI, 1.05-1.17); uncomplicated diabetes (OR, 1.20; 95% CI, 1.13-1.28); chronic pulmonary disease (OR, 1.43; 95% CI, 1.35-1.52); anemia (OR, 1.17; 95% CI, 1.12-1.24); renal failure (OR, 1.41; 95% CI, 1.29-1.54); fluid and electrolyte disorder (OR, 1.30; 95% CI, 1.22-1.39); congestive heart failure (OR, 1.70; 95% CI, 1.53-1.88); coagulopathy (OR, 1.13; 95% CI, 1.01-1.27); and liver disease (OR, 1.55; 95% CI, 1.29-1.85) (eTable 4 in the Supplement). Collinearity was assessed between variables, and all factors displayed variance inflation factors associated with independence.

RATHRR Scale

From multivariate analysis of the derivation cohort, factors associated with an increased risk of readmission following THR were analyzed to create the RATHRR Scale. The scale assigns a value to each factor based on its OR (Figure 2). The scale was then applied to patients in the derivation cohort without missing variables to assess its ability to accurately predict a patient’s risk of readmission. The percentage of patients readmitted for each RATHRR score was plotted, and on linear regression, an R2 of 0.935 was calculated (eFigure 3 in the Supplement). This value implies that the RATHRR Scale score could explain 93.5% of readmission variability and suggests that a higher RATHRR Scale score was associated with increased risk for readmission. The linear regression line calculated demonstrates a 0.36% increase in readmission risk associated with a 1-point increase in RATHRR Scale score. A score greater than 9 was associated with an above average readmission percentage and a score greater than 25 doubles readmission percentage.

The RATHRR Scale was verified by applying it to the validation cohort. Linear regression analysis was performed to assess the scale’s ability to accurately predict readmission risk in an independent population. The linear regression equation from the derivation cohort was also applied to the validation cohort to compare observed and estimated readmission rates for these patients (Figure 2). Linear regression calculated an R2 value of 0.891, suggesting the RATHRR Scale can explain 89.1% of percentage readmission variability. The regression line calculated demonstrates a 0.46% increase in readmission risk associated with a 1-point increase in RATHRR Scale score. In comparing observed and estimated regression lines, the validation cohort follows the regression line for the derivation cohort up to a RATHRR Scale score of 30. At that point, the scatter of the graph increases, skewing the observed regression line. More than 95% of patients in this cohort had a RATHRR Scale score of less than 30 points. Therefore, when those patients with a RATHRR Scale score of more than 30 were removed and the observed and estimated readmission rates were compared, the slopes of the regression lines are almost identical, demonstrating that a 1-point increase in the RATHRR Scale score is associated with a 0.36% increase in readmission risk (eFigure 4 in the Supplement). The new R2 value was 0.943, meaning that the RATHRR Scale can explain 94.3% of the percentage of readmission variability for patients with a RATHRR Scale score of less than 30 points.

Discussion

The RATHRR Scale was shown to be associated with readmission rates more than 80% of the time in both the derivation and validation cohorts based on linear regression analysis of the percentage of patients readmitted at each RATHRR Scale score. Quiz Ref IDThe scale is intended for use during presurgical patient care to identify high-risk patients for readmission prior to surgery. A mobile application facilitating use by physicians should enable better coordination between various health professionals to ensure high-risk patients receive necessary care to prevent readmission. Studies conducted within the past 5 years suggest that the best strategy to reduce readmissions is coordinated care.15,32,33 Williams15 reported that patient-centered approaches coordinating care among physicians were more successful at reducing readmissions than disease-centered approaches involving only 1 physician. Dharmarajan et al32 demonstrated that most readmissions in their cohort of patients with heart failure, pneumonia, and acute myocardial infarction were not attributed to the initial diagnosis. Ng et al34 have suggested specific interventions to reduce complication risks for many comorbidities associated with 30-day readmission risk, including aggressive treatment of chronic obstructive pulmonary disease to optimize the patient’s breathing capacity and strict maintenance of blood glucose levels and use of an insulin bolus rather than a sliding scale for patients with diabetes.

Outcomes data can propel technical and procedural advances, but little data exist regarding controlling 30-day THR readmissions. Weiss et al18 reported 30-day all-cause readmissions for total or partial hip replacements in 2010 as 8.1% using the SID but did not identify risk factors or risk-reduction strategies. Additionally, Cram et al20 reported an increase in 30-day readmission rates in THR patients from 5.9% in 1991 to 8.5% in 2008 owing to a reduction in mean LOS from 9.1 days in 1991 to 3.7 days in 2008. This change may be attributed to new protocols following THR. Our study does not address LOS as a variable affecting readmission, but it is possible that increased LOS offers optimal care for sicker patients and therefore reduces readmission risk.

Jordan et al21 developed a program to reduce readmission rates after total joint replacement in a single institution that enables outpatient workup for deep vein thrombosis, increased efforts to prevent surgical site infection, early follow-up after discharge, and increased physician education of economic ramifications of readmission. Total hip replacement readmission rates were reduced from 3.70% 2 years prior to implementation to 1.78% following implementation.21 The implementation of the RATHRR Scale aims to identify high-risk patients who would benefit from similar patient-specific readmission reduction approaches. Cullen et al22 observed an 8.5% THR readmission rate over a 42-month study at a single institution and identified deep vein thrombosis, dislocation, wound complications, and swollen limbs as the most common reasons for readmission. Our study identifies similar complications as reasons for readmission. Identifying patients at risk for these complications and implementing outpatient management may help avoid readmission. Finally, Zhan et al23 conducted a retrospective study using the National Inpatient Sample Database and SID and reported a 30-day readmission rate of 4.91% for primary THR and 8.48% for revision THR. In concurrence with our study, advanced age and comorbid diseases were associated with worse outcomes.

The limitations of our study included those inherent to large administrative databases, such as coding and sampling errors, misclassified variables, and the inability to accurately obtain readmission data on all patients. Readmissions for complications within 30 days of THR are available in the SID, but rates may be underestimated or overestimated owing to exclusions and missing data. Only patients with data for all studied variables were included in the application of this scale, which could result in selection bias. However, this limitation would apply to all age groups and should not negate overall findings. Another potential limitation was the designation of the RATHRR Scale and its overadjustment or underadjustment for age and comorbidity factors. To avoid this, we attempted to limit the amount any 1 factor could contribute to the scale. There is limited information regarding material used in the replacement, technical details on the procedure (type of closure and protection devices), surgeon characteristics (specialty and volume), and discharge and rehabilitation instructions. Multiple testing could have also been an issue in this study, but many of the P values were less than .001, so this is not likely to contribute much error in creating the RATHRR Scale. Failure to assess for interaction terms may have been a limitation of this study, and this assessment may alter the point values assigned to components of the RATHRR Scale. The scale does not take day of discharge into account, and further studies would be needed to determine the effect of day of discharge on readmission rates. An additional possible limitation was the similarity between the derivation and validation cohorts. These 4 populations were chosen a priori based on availability of readmission variables from the SID. These populations were large and from variable geographic locations and should therefore be representative of the entire United States. The next step in validating this scale would be to apply it on smaller, unique populations that may not be representative of the entire United States.

Conclusions

Given scrutiny of health care spending in the United States and the expansion of the Hospital Readmissions Reduction Program to include THR, the RATHRR Scale was developed to identify high-risk THR patients for 30-day readmission to allow precautionary and preventive care prior to or during the index admission. This scale, developed and validated in 2 separate populations, calculates an individual’s readmission risk prior to surgery using demographic and clinicopathological factors. To mitigate readmission risk, the RATHRR scale allows physicians to identify high-risk patients during the index admission and permit individualized care programs before and after surgery. Reducing readmission rates using risk-stratifications models, such as the RATHRR scale, should increase the quality of care patients receive while reducing overall health care expenditures.

Back to top
Article Information

Corresponding Author: Ronald S. Chamberlain, MD, MPA, Department of Surgery, Saint Barnabas Medical Center, Rutgers University–New Jersey Medical School (RU-NJMS), 94 Old Short Hills Rd, Livingston, New Jersey 07039 (rchamberlain@barnabashealth.org).

Accepted for Publication: December 10, 2015.

Published Online: March 9, 2016. doi:10.1001/jamasurg.2016.0020.

Author Contributions: Both authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Both authors.

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

Drafting of the manuscript: Both authors.

Critical revision of the manuscript for important intellectual content: Both authors.

Statistical analysis: Both authors.

Administrative, technical, or material support: Chamberlain.

Study supervision: Chamberlain.

Conflict of Interest Disclosures: None reported.

References
1.
Soohoo  NF, Farng  E, Lieberman  JR, Chambers  L, Zingmond  DS.  Factors that predict short-term complication rates after total hip arthroplasty.  Clin Orthop Relat Res. 2010;468(9):2363-2371.PubMedGoogle ScholarCrossref
2.
Mahomed  NN, Barrett  JA, Katz  JN,  et al.  Rates and outcomes of primary and revision total hip replacement in the United States medicare population.  J Bone Joint Surg Am. 2003;85-A(1):27-32.PubMedGoogle Scholar
3.
Canale  ST, Beaty  JH.  Canale & Beaty: Campbell’s Operative Orthopaedics. 12th ed. Philadelphia, PA: Mosby; 2012.
4.
HCUPnet. Healthcare Cost and Utilization Project. http://hcupnet.ahrq.gov/HCUPnet.jsp?Id=3C546A58544D3669&Form=DispTab&JS=Y&Action=Accept. Published 2012. Accessed July 24, 2013.
5.
Kim  S.  Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997-2004.  Arthritis Rheum. 2008;59(4):481-488.PubMedGoogle ScholarCrossref
6.
Harms  S, Larson  R, Sahmoun  AE, Beal  JR.  Obesity increases the likelihood of total joint replacement surgery among younger adults.  Int Orthop. 2007;31(1):23-26.PubMedGoogle ScholarCrossref
7.
Vasarhelyi  EM, MacDonald  SJ.  The influence of obesity on total joint arthroplasty.  J Bone Joint Surg Br. 2012;94(11)(suppl A):100-102.PubMedGoogle ScholarCrossref
8.
Ellimoottil  C, Miller  S, Davis  M, Miller  DC.  Insurance expansion and the utilization of inpatient surgery: evidence for a “woodwork” effect?  Surg Innov. 2015;22(6):588-592.PubMedGoogle ScholarCrossref
9.
Kurtz  S, Ong  K, Lau  E, Mowat  F, Halpern  M.  Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030.  J Bone Joint Surg Am. 2007;89(4):780-785.PubMedGoogle ScholarCrossref
10.
Mont  MA, Issa  K.  Updated projections of total joint arthroplasty demands in America: commentary on an article by Steven M. Kurtz, PhD, et al: “impact of the economic downturn on total joint replacement demand in the united states. updated projections to 2021.”  J Bone Joint Surg Am. 2014;96(8):e68.PubMedGoogle ScholarCrossref
11.
Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group. National health care expenditures data. https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpenddata/downloads/proj2012.pdf. Published 2012. Accessed July 22, 2013.
12.
Fuchs  VR.  The gross domestic product and health care spending.  N Engl J Med. 2013;369(2):107-109.PubMedGoogle ScholarCrossref
13.
Murray  CJL, Frenk  J.  Ranking 37th–measuring the performance of the US health care system.  N Engl J Med. 2010;362(2):98-99.PubMedGoogle ScholarCrossref
14.
The Patient Protection and Affordable Care Act, HR 3590, 111th Cong (2010).
15.
Williams  MV.  A requirement to reduce readmissions: take care of the patient, not just the disease.  JAMA. 2013;309(4):394-396.PubMedGoogle ScholarCrossref
16.
Henry J. Kaiser Family Foundation. Readmissions cost Medicare billions. Kaiser Health News. May 29, 2009.
17.
Jiang  HJ, Wier  LM; Healthcare Cost and Utilization Project.  All-Cause Hospital Readmission among Non-Elderly Medicaid Patients, 2007. Rockville, MD: Agency for Healthcare Research and Quality; 2010.
18.
Weiss  AJ, Elizhauser  A, Steiner  C; Healthcare Cost and Utilization Project.  Readmissions to US Hospitals by Procedure, 2010. Rockville, MD: Agency for Healthcare Research and Quality; 2013.
19.
Jencks  SF, Williams  MV, Coleman  EA.  Rehospitalizations among patients in the Medicare fee-for-service program.  N Engl J Med. 2009;360(14):1418-1428.PubMedGoogle ScholarCrossref
20.
Cram  P, Lu  X, Kaboli  PJ,  et al.  Clinical characteristics and outcomes of Medicare patients undergoing total hip arthroplasty, 1991-2008.  JAMA. 2011;305(15):1560-1567.PubMedGoogle ScholarCrossref
21.
Jordan  CJ, Goldstein  RY, Michels  RF, Hutzler  L, Slover  JD, Bosco  JA  III.  Comprehensive program reduces hospital readmission rates after total joint arthroplasty.  Am J Orthop (Belle Mead NJ). 2012;41(11):E147-E151.PubMedGoogle Scholar
22.
Cullen  C, Johnson  DS, Cook  G.  Re-admission rates within 28 days of total hip replacement.  Ann R Coll Surg Engl. 2006;88(5):475-478.PubMedGoogle ScholarCrossref
23.
Zhan  C, Kaczmarek  R, Loyo-Berrios  N, Sangl  J, Bright  RA.  Incidence and short-term outcomes of primary and revision hip replacement in the United States.  J Bone Joint Surg Am. 2007;89(3):526-533.PubMedGoogle ScholarCrossref
24.
Agabiti  N, Picciotto  S, Cesaroni  G,  et al; Italian Study Group on Inequalities in Health Care.  The influence of socioeconomic status on utilization and outcomes of elective total hip replacement: a multicity population-based longitudinal study.  Int J Qual Health Care. 2007;19(1):37-44.PubMedGoogle ScholarCrossref
25.
Rajgopal  R, Martin  R, Howard  JL, Somerville  L, MacDonald  SJ, Bourne  R.  Outcomes and complications of total hip replacement in super-obese patients.  Bone Joint J. 2013;95-B(6):758-763.PubMedGoogle ScholarCrossref
26.
Bozic  KJ, Lau  E, Ong  K,  et al.  Risk factors for early revision after primary total hip arthroplasty in Medicare patients.  Clin Orthop Relat Res. 2014;472(2):449-454.PubMedGoogle ScholarCrossref
27.
White  RH, Gettner  S, Newman  JM, Trauner  KB, Romano  PS.  Predictors of rehospitalization for symptomatic venous thromboembolism after total hip arthroplasty.  N Engl J Med. 2000;343(24):1758-1764.PubMedGoogle ScholarCrossref
28.
Namba  RS, Paxton  L, Fithian  DC, Stone  ML.  Obesity and perioperative morbidity in total hip and total knee arthroplasty patients.  J Arthroplasty. 2005;20(7)(suppl 3):46-50.PubMedGoogle ScholarCrossref
29.
Willis-Owen  CA, Konyves  A, Martin  DK.  Factors affecting the incidence of infection in hip and knee replacement: an analysis of 5277 cases.  J Bone Joint Surg Br. 2010;92(8):1128-1133.PubMedGoogle ScholarCrossref
30.
Wallace  G, Judge  A, Prieto-Alhambra  D, de Vries  F, Arden  NK, Cooper  C.  The effect of body mass index on the risk of post-operative complications during the 6 months following total hip replacement or total knee replacement surgery.  Osteoarthritis Cartilage. 2014;22(7):918-927.PubMedGoogle ScholarCrossref
31.
Guide  U.  HCUP Supplemental Values for Revisit Analyses. Healthcare Cost and Utilization Project (HCUP). Rockville, MD: Agency for Healthcare Research and Quality; 2015.
32.
Dharmarajan  K, Hsieh  AF, Lin  Z,  et al.  Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia.  JAMA. 2013;309(4):355-363.PubMedGoogle ScholarCrossref
33.
Scott  KW, Jha  AK.  Putting quality on the global health agenda.  N Engl J Med. 2014;371(1):3-5.PubMedGoogle ScholarCrossref
34.
Ng  VY, Lustenberger  D, Hoang  K,  et al.  Preoperative risk stratification and risk reduction for total joint reconstruction: AAOS exhibit selection.  J Bone Joint Surg Am. 2013;95(4):e191-e15.PubMedGoogle ScholarCrossref
×