Orthopedics

Feature Article Supplemental Data

Patient Risk Profile for Unplanned 90-Day Emergency Department Visits Differs Between Total Hip and Total Knee Arthroplasty

Johannes F. Plate, MD, PhD; Sean P. Ryan, MD; Michael A. Bergen, MD; Cierra S. Hong, MD; Michael A. Mont, MD; Michael P. Bolognesi, MD; Thorsten M. Seyler, MD, PhD

Abstract

Numerous studies have explored 90-day readmissions following total joint arthroplasty; however, there is a paucity of literature concerning 90-day emergency department (ED) visits. The authors aimed to characterize the risk factors for ED presentations and to determine the primary reasons for return, hypothesizing that certain medical comorbidities would account for resource utilization. The institutional database was queried for primary total hip arthroplasty (THA) and total knee arthroplasty (TKA). Patients were stratified based on return visits to the ED within 90 days postoperatively. Univariable and multivariable analyses were performed to determine the factors most predictive of ED return for each THA and TKA. A total of 10,479 procedures resulted in 1234 90-day ED visits made by 937 patients. Significant predictors of 90-day ED return after THA included black race, age older than 80 years, congestive heart failure, valvular heart disease, metastatic disease, peripheral vascular disease, alcoholism, drug use, depression, and discharge to a skilled nursing facility. In contrast, only black race, liver insufficiency, cancer, and pulmonary hypertension were predictive of ED return following TKA. The primary risk factors for ED return differ for THA and TKA, and this is not currently reflected in the medical severity diagnosis-related group system. Specifically, black patients with multiple comorbidities are at high risk for unplanned ED visits following THA. This should be considered in patient counseling and outreach programs when attempting to mitigate the postoperative risks and to decrease 90-day resource utilization in this patient population. [Orthopedics. 2020;43(5):295–302.]

Abstract

Numerous studies have explored 90-day readmissions following total joint arthroplasty; however, there is a paucity of literature concerning 90-day emergency department (ED) visits. The authors aimed to characterize the risk factors for ED presentations and to determine the primary reasons for return, hypothesizing that certain medical comorbidities would account for resource utilization. The institutional database was queried for primary total hip arthroplasty (THA) and total knee arthroplasty (TKA). Patients were stratified based on return visits to the ED within 90 days postoperatively. Univariable and multivariable analyses were performed to determine the factors most predictive of ED return for each THA and TKA. A total of 10,479 procedures resulted in 1234 90-day ED visits made by 937 patients. Significant predictors of 90-day ED return after THA included black race, age older than 80 years, congestive heart failure, valvular heart disease, metastatic disease, peripheral vascular disease, alcoholism, drug use, depression, and discharge to a skilled nursing facility. In contrast, only black race, liver insufficiency, cancer, and pulmonary hypertension were predictive of ED return following TKA. The primary risk factors for ED return differ for THA and TKA, and this is not currently reflected in the medical severity diagnosis-related group system. Specifically, black patients with multiple comorbidities are at high risk for unplanned ED visits following THA. This should be considered in patient counseling and outreach programs when attempting to mitigate the postoperative risks and to decrease 90-day resource utilization in this patient population. [Orthopedics. 2020;43(5):295–302.]

The introduction of bundled payments by the Centers for Medicare & Medicaid Services and episode-of-care reimbursement for total hip arthroplasty (THA) and total knee arthroplasty (TKA) has shifted financial risk toward the institutions performing these procedures.1,2 These payment models reimburse the participating institutions for an entire episode of care, including the index procedure and preoperative and postoperative services. In addition, the participating institution assumes financial responsibility for any readmission, emergency department (ED) visit, and subsequent treatment within 90 days from surgery.1

The mean costs for unplanned ED visits have been reported as $4450 for THA and $5950 for TKA.3 In the United States, the current financial impact of joint arthroplasties—$15 billion annually—is predicted to increase by 600% by 2030.4,5 Although several studies have examined the reasons, risk factors, and costs associated with 90-day hospital readmissions,6 information regarding unplanned ED visits that could help prevent these events and associated health care costs is lacking. Studies using large administrative databases have mainly focused on the causes, insurance status, and patient comorbidities associated with unplanned ED visits.7,8 Institutional studies have reported on costs associated with unplanned ED visits9; however, a comprehensive patient risk profile has not been established.

In this study, the authors sought to perform a comprehensive assessment of patient demographics, comorbidities, and specific reasons for ED visits within 90 days from primary THA and TKA to establish a patient risk profile. The authors hypothesized that older patients with an increased number of comorbidities would be more likely to present to the ED.

Materials and Methods

The institutional database at a tertiary academic medical center was retrospectively queried from January 1, 2012, to June 30, 2017, for primary THA and TKA. All surgeries were performed by either fellowship-trained adult reconstruction surgeons or surgeons with 10 or more years of experience in total joint arthroplasty. Patients who underwent bilateral procedures during the same hospitalization, unicompartmental arthroplasties, or revision surgeries were excluded from the study. Patient demographics including age, sex, race, body mass index, procedure type, insurance type, tobacco use, marital status, and discharge disposition were collected, along with both Charlson and Elixhauser comorbidities.10,11 Emergency department visits within the authors' health care system including reasons for the ED visit were noted based on medical coding in the institution's health record. The day and length of time spent in the ED were recorded based on the time stamps in the patient's health record. Time was recorded from the patient's checking in at the front desk until the patient's after-visit summary was printed. Patients were then stratified based on return visits to the ED within 90 days postoperatively.

Univariable analysis was performed to determine the independent demographic and comorbidity risk factors associated with return to the ED. Both Charlson and Elix-hauser comorbidities were used to increase the number of potential risk factors evaluated. Continuous variables, including age and body mass index, were made binary. Several risk factors are collinear between indexes (eg, congestive heart failure, diabetes mellitus, HIV/AIDs, metastatic disease, peripheral vascular disease, peptic ulcer disease, pulmonary insufficiency, and renal insufficiency). Despite their presence in both comorbidity profiles, they were considered only once in the authors' analysis. Variables with significant association (P<.1) on univariable analysis were then included in a multivariable model to determine which were most associated with hospital return. Multiple International Classification of Diseases codes (Table A, available in the online version of the article) were then used to categorize reasons for ED return into mechanical, pain, wound complication, cardiopulmonary, urinary, hematologic, and other. Statistical analysis was performed with Wizard Pro for Mac (E Miller, Chicago, Illinois) and RStudio, version 2.1 (R Foundation, Vienna, Austria) software. Data were presented as number (percent) or as median (lower, upper quartile). P<.05 was considered statistically significant from multivariable logistic regression data.

International Classification of Disease (ICD) codes utilized for reasons for ED returnInternational Classification of Disease (ICD) codes utilized for reasons for ED return

Table A:

International Classification of Disease (ICD) codes utilized for reasons for ED return

Results

A total of 10,479 separate procedures (4649 THAs and 5830 TKAs) were included in the study. A similar number of THA patients (n=421, 9.1%) and TKA patients (n=516, 8.9%; P=.715) presented to the ED at least once within 90 days from surgery, with a total of 581 ED visits for THA patients and 653 ED visits for TKA patients. The median time to ED visit from surgery among THA patients was 23 (11, 47) days, with 46.8% occurring within 3 weeks. For TKA patients, the median time to ED visit was 18 (8, 40) days, with 55.9% occurring within 3 weeks. Total knee arthroplasty patients were significantly more likely to present earlier in the postoperative period relative to THA patients (Figure 1; P<.001). Most ED visits (198, 16.0%) were on Saturdays, followed by Fridays (194, 15.7%) and Sundays (184, 14.9%).

Time to return to the emergency department (ED) in days, after the index procedure. Abbreviations: THA, total hip arthroplasty; TKA, total knee arthroplasty.

Figure 1:

Time to return to the emergency department (ED) in days, after the index procedure. Abbreviations: THA, total hip arthroplasty; TKA, total knee arthroplasty.

For THA patients, discharge to a skilled nursing facility, unmarried status, tobacco use, black race, and Medicare and Medicaid status were associated with ED return (Table 1). Similarly, there was an association between ED return among TKA patients and skilled nursing facility disposition, unmarried status, black race, and Medicaid insurance status (Table 2).

Demographics of Patients Undergoing Total Hip Arthroplasty

Table 1:

Demographics of Patients Undergoing Total Hip Arthroplasty

Demographics of Patients Undergoing Total Knee Arthroplasty

Table 2:

Demographics of Patients Undergoing Total Knee Arthroplasty

Univariable analysis for Charlson and Elixhauser comorbidities was used to identify significant independent predictors of ED return for THA and TKA patients (Table 3; P<.1), which were then incorporated into a multivariable logistic regression model. Multivariable analysis of THA patients revealed several comorbidities and demographic variables as predictors of 90-day ED return (Table 4). In contrast, there were a limited number of predictors for ED return among TKA patients (Table 4). Black race was the only predictor in both THA and TKA.

Comorbidity Univariable Analysis (From Charlson and Elixhauser Comorbidities) for Emergency Department ReturnComorbidity Univariable Analysis (From Charlson and Elixhauser Comorbidities) for Emergency Department Return

Table 3:

Comorbidity Univariable Analysis (From Charlson and Elixhauser Comorbidities) for Emergency Department Return

Significant Variables From the Multivariable Model

Table 4:

Significant Variables From the Multivariable Model

The most common reasons for ED return among THA and TKA patients were wound complications, cardiopulmonary problems, and pain (Table 5). The overall median time spent in the ED was 5.72 (3.82, 8.42) hours. Shortest time in the ED was recorded for patients with deep venous thrombosis. Longest time was recorded for patients who underwent a cardiac workup that was ultimately negative for acute myocardial infarction. Overall, approximately 59% of patients were readmitted to the hospital following the ED visit.

Reasons for Emergency Department Visits

Table 5:

Reasons for Emergency Department Visits

Discussion

This study revealed that, overall, approximately 8.9% of patients returned for an unplanned ED visit within 90 days from surgery performed at a tertiary medical center. This is comparable to previous reports of incidences ranging from 5.3% to 15.9% for THA and TKA.6–8,12 The current authors found that for THA, unmarried black patients older than 80 years with multiple comorbidities, including congestive heart failure, valvular heart disease, metastatic disease, peripheral vascular disease, depression, and alcohol or drug use, receiving Medicare or Medicaid were most likely to return for an unplanned ED visit within 90 days from surgery. This was in contrast to the findings for patients following TKA, which only revealed black race, liver insufficiency, cancer, and pulmonary hypertension. Although some of these demographics and comorbidities have been identified previously,7–9,13 the current study found that each variable was an independent predictor of an unplanned ED visit. Therefore, patients with these characteristics may require increased resources and preoperative optimization to avoid an unplanned ED visit.

The findings of this study further suggest that the risk factors for ED return differ after THA and TKA; however, the Centers for Medicare & Medicaid Services groups arthroplasty patients in one medical severity diagnosis-related group (469, or 470 if major complication/comorbidity present).14 It appears that separate grouping for THA and TKA patients as part of the Hospital Readmission Reduction Program by the Centers for Medicare & Medicaid Services would be more appropriate because of the differences in patient risk profiles. Interestingly, only black race was an independent predictor for 90-day ED visit among both cohorts (ie, THA and TKA). Strategies to improve treatment plan adherence among patients at risk for unplanned ED visits and readmissions have been described using access to personal health records with the ability to contact providers directly through the electronic medical record.15 Specifically, however, older, unmarried black patients discharged to skilled nursing facilities on Medicare or Medicaid were found to underutilize the electronic patient portal and were at highest risk for unplanned ED visits and readmissions within 90 days from joint arthroplasty.15 Although the underlying reasons for this apparent disparity in health care delivery to joint arthroplasty patients who are black remain under investigation, novel strategies to improve access to postoperative care are warranted.

This study revealed that ED visits were more frequent on weekends (Friday to Sunday) compared with during the week. Similar to the current study, a previous study found that unplanned ED visits occurred mostly within the first 2 weeks from discharge.7 Therefore, it appears that efforts invested in patient call backs should be directed toward the first weekend following discharge to anticipate possible patient questions and avoid unplanned readmissions. In addition, approximately 59% of ED returns in the current study resulted in readmission to the hospital under inpatient status. A previous study highlighted that readmissions are more frequent during weekends because of restricted resources.16 Therefore, alternative access to providers specifically during weekends appears crucial to avoid unplanned ED visits and potential readmissions due to lack of resources. Based on the current study, a reduction of 25% of ED visits may lead to saving up to $1 million. However, a recent study evaluating the use of an electronic patient portal as an alternative access pathway did not find that secure provider messaging to providers led to a decrease in unplanned ED visits and readmissions.15 Therefore, further studies are necessary to elucidate the most beneficial form of patient communication and access to providers during weekends to avoid unplanned ED visits and subsequent readmissions.

Previous studies have used administrative databases to identify pain as the major cause of an unplanned ED visit 90 days from surgery.7,8 The current study retrieved the cause of ED visits from the individual patient record International Classification of Diseases codes used during the ED return. Using this methodology, wound complications followed by pain and cardiopulmonary complications were the major causes of ED returns. This finding may be related to the methodology of the current study compared with large state and national databases relying on correct coding for ED visits. Furthermore, prior studies have been limited in their ability to characterize risk factors for ED returns as part of administrative databases.

This study is the first to examine the time spent in the ED prior to discharge or admission. Patients with deep venous thrombosis, myocardial infarction, or bleeding from the postoperative wound spend the shortest times in the ED. In the case of deep venous thrombosis and myocardial infarction, this may be due to established care protocols to expedite admission of these patients for further care. The longest ED times were associated with medications and non-myocardial infarction cardiac workup, likely related to exhaustive testing and the need for further subspecialty consultation. However, the numbers of tests or consultations were not evaluated in this study and may be of interest for future studies related to the implementation of standardized care pathways for postoperative arthroplasty patients.

The patients included in this study underwent surgery and were followed at a single tertiary medical center. Therefore, previously noted limitations of large database studies may not apply. Additionally, the reasons for ED return were determined through the primary diagnosis code at the time of the ED visit rather than manual chart review. This potentially explains why “pain,” if not the primary reason for ED return and observed secondarily, was not included for as many patients during analysis. However, the study was limited to ED visits within the authors' health care system and did not include ED visits at outside institutions. Therefore, the number of actual ED visits may have been underestimated because patients may have presented to outside hospitals for care. This study analyzed the number of readmissions to the hospital following ED visits; however, the status (inpatient vs 24-hour observation) was not assessed and requires further inquiry regarding differences in reimbursement in future studies. The length of time spent in the ED was based on health record time stamps and included the time from checking in until the after-visit summary printed. This may have underestimated the time the patient was physically in the ED because the time spent in the waiting room until checking in and the time from printing the after-visit summary until physically leaving the ED were not accessible. The effects of the individual surgeon or socioeconomic status on ED visits were not assessed.

Conclusion

Black race, age older than 80 years, and multiple comorbidities, including heart disease, metastatic disease, depression, and alcohol or drug use, were independent risk factors for unplanned ED visits within 90 days from THA. In contrast, only liver insufficiency, cancer, pulmonary hypertension, and black race were risk factors for ED return following TKA. Wound complications, pain, and cardiopulmonary complications were the most common reasons for return. Studies focusing on decreasing health care disparities for the identified patients at risk for ED returns are needed to decrease the burden on the health care system. Future strategies could include scheduling follow-up calls or text messages, using telemedicine follow-up appointments, assessing the influence of orthopedic urgent care locations, and implementing 24-hour orthopedic hotlines for postoperative questions and issues staffed by an orthopedic physician assistant or nurse practitioner to decrease unplanned postoperative ED visits, especially for patients at risk.

References

  1. Karas V, Kildow BJ, Baumgartner BT, et al. Preoperative patient profile in total hip and knee arthroplasty: predictive of increased Medicare payments in a bundled payment model. J Arthroplasty. 2018;33(9):2728–2733.
  2. Sabeh KG, Rosas S, Buller LT, Roche MW, Hernandez VH. The impact of discharge disposition on episode-of-care reimbursement after primary total hip arthroplasty. J Arthroplasty. 2017;32(10):2969–2973.
  3. Sibia US, Mandelblatt AE, Callanan MA, MacDonald JH, King PJ. Incidence, risk factors, and costs for hospital returns after total joint arthroplasties. J Arthroplasty. 2017;32(2):381–385.
  4. Kurtz SM, Ong KL, Lau E, Bozic KJ. 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):624–630.
  5. Cram P, Lu X, Kates SL, Singh JA, Li Y, Wolf BR. Total knee arthroplasty volume, utilization, and outcomes among Medicare beneficiaries, 1991–2010. JAMA. 2012;308(12):1227–1236.
  6. Ramkumar PN, Chu CT, Harris JD, et al. Causes and rates of unplanned readmissions after elective primary total joint arthroplasty: a systematic review and meta-analysis. Am J Orthop (Belle Mead NJ). 2015;44(9):397–405.
  7. Huerfano E, Gonzalez Della Valle A, Shanaghan K, Girardi F, Memtsoudis S, Liu J. Characterization of readmission and emergency department visits within 90 days following lower-extremity arthroplasty. HSS J. 2018;14(3):271–281.
  8. Finnegan MA, Shaffer R, Remington A, Kwong J, Curtin C, Hernandez-Boussard T. Emergency department visits following elective total hip and knee replacement surgery: identifying gaps in continuity of care. J Bone Joint Surg Am. 2017;99(12):1005–1012.
  9. Luzzi AJ, Fleischman AN, Matthews CN, Crizer MP, Wilsman J, Parvizi J. The “bundle busters”: incidence and costs of postacute complications following total joint arthroplasty. J Arthroplasty. 2018;33(9):2734–2739.
  10. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–383.
  11. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27.
  12. Nedza SM, Fry DE, DesHarnais S, Spencer E, Yep P. Emergency department visits following joint replacement surgery in an era of mandatory bundled payments. Acad Emerg Med. 2017;24(2):236–245.
  13. Plate JF, Ryan SP, Goltz DE, Howell CB, Bolognesi MP, Seyler TM. Medicaid insurance correlates with increased resource utilization following total hip arthroplasty. J Arthroplasty. 2019;34(2):255–259.
  14. Ryan SP, Plate JF, Goltz DE, et al. Should medical severity-diagnosis related group classification be utilized for reimbursement? An analysis of Elixhauser comorbidities and cost of care. J Arthroplasty. 2019;34(7):1312–1316.
  15. Plate JF, Ryan SP, Bergen MA, Hong CS, Attarian DE, Seyler TM. Utilization of an electronic patient portal following total joint arthroplasty does not decrease readmissions. J Arthroplasty. 2019;34(2):211–214.
  16. McCusker J, Ionescu-Ittu R, Ciampi A, et al. Hospital characteristics and emergency department care of older patients are associated with return visits. Acad Emerg Med. 2007;14(5):426–433.

Demographics of Patients Undergoing Total Hip Arthroplasty

CharacteristicNo.No Emergency Department Return (N=4228)90-Day Emergency Department Return (N=421)P
Age, median (lower, upper quartile), y464963.4 (55.4, 70.4)65.0 (55.4, 72.8).013
Female, No. (%)46492240 (53.0)255 (60.6).003
Body mass index, median (lower, upper quartile), kg/m2435828.6 (25.1, 32.8)28.8 (25.2, 33.8).249
Disposition, No. (%)4649
  Home3528 (83.4)300 (71.3)<.001
  Skilled nursing facility626 (14.8)115 (27.3)<.001
  Rehabilitation facility65 (1.5)5 (1.2).574
  Other9 (0.2)0 (0.0).343
Married, No. (%)46482785 (65.9)228 (54.2)<.001
Tobacco use (ever), No. (%)42281855 (43.9)221 (52.5)<.001
Race, No. (%)4600
  White3340 (79.9)303 (72.5)<.001
  Black739 (17.7)107 (25.6)<.001
  Asian19 (0.5)0 (0.0).167
  Multiracial34 (0.8)4 (1.0).757
  Other50 (1.2)4 (1.0).666
Insurance type, No. (%)3980
  Managed care1586 (43.8)111 (30.6)<.001
  Medicaid99 (2.7)26 (7.2)<.001
  Medicare1665 (46.0)211 (58.1)<.001
  Private148 (4.1)10 (2.8).214
  Other119 (3.3)5 (1.4).046

Demographics of Patients Undergoing Total Knee Arthroplasty

CharacteristicNo.No Emergency Department Return (N=5314)90-Day Emergency Department Return (N=516)P
Age, median (lower, upper quartile), y583066.9 (60.5, 72.6)66.9 (59.3, 73.5).780
Female, No. (%)58303257 (61.3)315 (61.0).913
Body mass index, median (lower, upper quartile), kg/m2547631.3 (27.5, 35.6)31.7 (28.0, 36.6).055
Disposition, No. (%)5830
  Home4069 (76.6)375 (72.7).047
  Skilled nursing facility1120 (21.1)128 (24.8).049
  Rehabilitation facility107 (2.0)12 (2.3).632
  Other13 (0.2)1 (0.2).882
Married, No. (%)58303553 (66.9)298 (57.8)<.001
Tobacco use (ever), No. (%)58302252 (42.4)252 (48.8).005
Race, No. (%)5771
  White4100 (77.9)330 (64.6)<.001
  Black975 (18.5)160 (31.3)<.001
  Asian57 (1.1)7 (1.4).555
  Multiracial60 (1.1)5 (1.0).740
  Other68 (1.3)9 (1.8).378
Insurance type, No. (%)5089
  Managed care1497 (32.3)119 (26.0).006
  Medicaid63 (1.4)17 (3.7)<.001
  Medicare2773 (59.9)294 (64.3).063
  Private200 (4.3)22 (4.8).620
  Other99 (2.1)5 (1.1).133

Comorbidity Univariable Analysis (From Charlson and Elixhauser Comorbidities) for Emergency Department Return

VariableNo. (%)No. (%)


Total Hip ArthroplastyTotal Knee Arthroplasty


No ED Return90-Day ED ReturnPNo ED Return90-Day ED ReturnP
Age >80 y240 (5.7)52 (12.4)<.001344 (6.5)46 (8.9).035
BMI >40 kg/m2149 (3.8)24 (5.9).038402 (8.1)51 (10.2).099
Discharge disposition SNF626 (14.8)115 (27.3)<.0011120 (21.1)128 (24.8).049
Medicare or Medicaid1764 (48.8)237 (65.3)<.0012836 (61.2)311 (68.1).004
MI143 (3.4)18 (4.3).340177 (3.3)17 (3.3).965
Cerebrovascular disease69 (1.6)12 (2.9).07292 (1.7)9 (1.7).983
Dementia13 (0.3)1 (0.2).80316 (0.3)5 (1.0).022
Connective tissue disease224 (5.3)37 (8.8).003298 (5.6)38 (7.4).103
Tumors/cancer94 (2.2)19 (4.5).00497 (1.8)14 (2.7).162
Liver insufficiency (mild)73 (1.7)14 (3.3).023106 (2.0)21 (4.1).003
Liver insufficiency (severe)4 (0.1)1 (0.2).4108 (0.2)3 (0.6).046
CHF134 (3.2)38 (9.0)<.001163 (3.1)28 (5.4).005
Arrhythmia493 (11.7)66 (15.7).016661 (12.4)81 (15.7).034
Valvular heart disease140 (3.3)31 (7.4)<.001206 (3.9)24 (4.7).389
Pulmonary HTN53 (1.3)16 (3.8)<.00168 (1.3)19 (3.7)<.001
PVD115 (2.7)24 (5.7)<.001143 (2.7)17 (3.3).424
HTN2139 (50.6)252 (59.9)<.0013269 (61.5)352 (68.2).003
HTN with complications237 (5.6)41 (9.7)<.001400 (7.5)55 (10.7).012
Paralysis10 (0.2)1 (0.2).9978 (0.2)1 (0.2).812
Neuro75 (1.8)12 (2.9).124120 (2.3)25 (4.8)<.001
Pulmonary584 (13.8)89 (21.1)<.001873 (16.4)102 (19.8).053
DM490 (11.6)64 (15.2).030901 (17.0)113 (21.9).005
DM complication85 (2.0)14 (3.3).078191 (3.6)29 (5.6).022
Hypothyroidism508 (12.0)57 (13.5).362800 (15.1)80 (15.5).785
Renal insufficiency264 (6.2)43 (10.2).002408 (7.7)59 (11.4).003
PUD16 (0.4)4 (1.0).09935 (0.7)3 (0.6).835
HIV18 (0.4)1 (0.2).5697 (0.1)1 (0.2).718
Lymphoma16 (0.4)1 (0.2).65121 (0.4)4 (0.8).216
Metastatic disease22 (0.5)8 (1.9).0027 (0.1)1 (0.2).718
Tumor66 (1.6)15 (3.6).00464 (1.2)7 (1.4).764
Rheumatologic256 (6.1)41 (9.7).004324 (6.1)41 (7.9).099
Coagulopathy116 (2.7)14 (3.3).49199 (1.9)22 (4.3)<.001
Obesity913 (21.6)106 (25.2).0901719 (32.3)177 (34.3).366
Weight loss18 (0.4)4 (1.0).1467 (0.1)3 (0.6).031
Electrolyte abnormalities379 (9.0)54 (12.8).010490 (9.2)62 (12.0).039
Blood loss10 (0.2)1 (0.2).9977 (0.1)5 (1.0)<.001
Anemia61 (1.4)9 (2.1).26790 (1.7)10 (1.9).683
Alcohol70 (1.7)19 (4.5)<.00142 (0.8)6 (1.2).374
Drugs24 (0.6)14 (3.3)<.00124 (0.5)6 (1.2).038
Psychosis20 (0.5)5 (1.2).06523 (0.4)4 (0.8).281
Depression563 (13.3)90 (21.4)<.001774 (14.6)97 (18.8).010

Significant Variables From the Multivariable Model

VariableOdds Ratio95% Confidence IntervalP
Total hip arthroplasty
  Black1.4491.097–1.914.009
  Age >80 y1.6351.096–2.439.016
  Discharge disposition SNF1.3681.020–1.835.037
  Medicare/Medicaid insurance1.3361.030–1.732.029
  CHF2.0501.293–3.251.002
  Valvular heart disease1.6641.025–2.702.040
  Metastatic disease3.6211.142–11.479.029
  PVD1.8341.112–3.023.017
  Alcohol2.4901.378–4.499.003
   Drugs3.6281.715–7.677<.001
  Depression1.6711.255–2.224<.001
Total knee arthroplasty
  Black1.8131.438–2.286<.001
  Liver insufficiency (mild)1.9061.120–3.243.017
  Cancer2.5621.105–5.941.028
  Pulmonary hypertension2.1591.189–3.922.011

Reasons for Emergency Department Visits

Reason for ED VisitTotal Hip ArthroplastyTotal Knee Arthroplasty


No. of Readmissions/ED VisitsMedian (Lower, Upper Quartile) ED LOS, hNo. of Readmissions/ED VisitsMedian (Lower, Upper Quartile) ED LOS, h
Mechanical
  Fall9/155.95 (3.35, 7.52)6/137.28 (4.07, 8.87)
  Fracture22/246.52 (5.85, 9.25)4/46.00 (5.52, 9.62)
  Dislocation20/506.90 (4.63, 9.37)0/0NA
  Pain51/1396.13 (4.15, 8.63)40/1605.29 (3.25, 7.65)
Wound complication
  Dehiscence14/177.08 (4.98, 9.18)13/185.02 (4.07, 7.33)
  Infection76/946.45 (4.88, 9.35)62/1006.80 (4.47, 9.07)
  Hematoma12/158.00 (4.20, 17.48)5/84.78 (2.80, 7.23)
  Bleeding19/264.85 (3.63, 6.93)11/215.43 (4.12, 8.90)
  Other wound complication38/476.77 (3.62, 9.18)16/275.02 (3.37, 7.28)
Cardiopulmonary
  Anemia12/176.00 (4.17, 6.75)8/116.07 (5.12, 7.62)
  Cardiac (non-MI)34/587.39 (4.98, 10.17)42/667.56 (4.95, 10.62)
  Cardiac MI6/64.49 (4.03, 8.38)1/12.52
  Hemodynamic instability20/356.23 (4.58, 10.12)24/407.43 (4.72, 10.22)
  Pulmonary insufficiency18/255.28 (4.18, 8.47)24/327.19 (5.65, 8.72)
Urinary
  Acute renal failure8/86.43 (4.37, 12.03)10/116.31 (5.03, 9.98)
  Other urinary complication4/85.00 (3.92, 8.95)10/204.95 (3.65, 9.97)
Hematologic
  DVT4/154.19 (3.28, 6.75)7/215.78 (2.77, 9.85)
  PE10/116.60 (5.45, 8.02)13/146.29 (5.8, 9.3)
  Sepsis7/74.88 (4.03, 7.50)8/86.84 (5.58, 11.72)
Other
  GI complication18/306.00 (4.68, 8.62)17/306.94 (5.22, 9.13)
  Poor glycemic control2/28.93 (8.60, 9.25)5/74.63 (3.98, 7.28)
  Medication related4/78.43 (6.32, 8.93)5/117.03 (5.07, 30.08)

International Classification of Disease (ICD) codes utilized for reasons for ED return

Reason for ED visitICD-9 codesICD-10 codes
Mechanical
  FallE880-8W00-W19
  Fracture820–822; 996.44M96.65-9; M97.0-1; S32; S72; S82; T84.04
  Dislocation718.35; 835; 839.8; 996.42M24.45; S73.0; T84.02
Pain338.11; 338.18; 719.45-6; 719.7; 729.5 780.96G89.18; G89.28-9; G89.4; M25.55-6; M79.1; M79.604-6; M79.65-6; R10.2; R52; T84.84
Wound complication
  Dehiscence890.1; 891; 998.31-3T81.30-2; S71.00
  Infection682.6; 682.9; 686.9; 958.3; 998.59; 730.05; 730.25-6; 996.66-9L02.415-6; L03.115-6; T81.4; T84.5; T84.7
  Hematoma998.11-2L76.32; M25.06; M96.83-4; T84.83
  Bleeding431-2; 459; 531.4; 532.4; 569.3; 578; 599.7; 784.7; 786.3; 852.2; 998.1H11.3; I60-62; I97.62; K25.4; K26.4; K29.31; K57.31; K57.51; K57.91; K92.2; L76.22; M25.06; M96.83-4; R58; T84.83
  Other Wound complication998.12-3; 998.31; 998.83T81.3-4
Cardiopulmonary
  Anemia283; 285D50-64
  Cardiac (non-MI)402; 413-416; 420-429; 441; 785; 786I11; I13; I25.10-1; I34; I36; I44-I51; R00.0-2; R07.8-9
  Cardiac MI410-1I21-3; I46
  Hemodynamic instability401-2; 405; 458E86; E87.0-1; E87.7; I10; I16; I95; I97; R00; R03; R57
  Pulmonary insufficiency278.03; 480-6; 491-6; 512; 514; 518; 786.09; 799.02J12-J18; J20-1; J41-5; J80-81; J90; J96
Urinary
  Acute Renal Failure580; 583; 584N17
  Other Urinary complication590; 599; 604; 788.1-2B37.49; N13.6; N39; R30-3
Hematologic
  DVT453.4-6I82.4-5
  PE415.1I26; I27.82
  Sepsis38; 599; 995.91-2A40-1
Other
  GI complication8.45; 8.8; 535-7; 560; 564; 787A04; G43.A; K56; K59-63; K91-2; R11
  Poor glycemic control250-1; 790.2E10-1; E15-6
  Medication related292.81; 693; 787.01; 960-79; 995.2D52.1; I95.2; K59.03; L27.0; R50.2; T36-50
Authors

The authors are from the Department of Orthopaedic Surgery (JFP), Wake Forest Baptist Health, Winston-Salem, and the Department of Orthopaedic Surgery (SPR, MAB, CSH, MPB, TMS), Duke University, Durham, North Carolina; and the Department of Orthopaedic Surgery (MAM), Lenox Hill Hospital, Northwell Health, New York, New York.

Drs Plate, Ryan, Bergen, and Hong have no relevant financial relationships to disclose. Dr Mont is a paid consultant for Microport, Stryker, Cymedica, DJ Orthopaedics, Flexion Therapeutics, Johnson & Johnson, Orthosensor, Pacira, Peerwell, Performance Dynamics, Pfizer, Skye Biologics, and Tissue Gene and has received grants from Stryker, DJ Orthopaedics, Johnson & Johnson, Orthosensor, and Tissue Gene. Dr Bolognesi is a paid consultant for Zimmer Biomet and Total Joint Orthopedics. Dr Seyler is a paid consultant for Total Joint Orthopedics and Pfizer.

This study was supported by a grant from the Piedmont Orthopedic Society to Dr Plate.

Correspondence should be addressed to: Thorsten M. Seyler, MD, PhD, Department of Orthopaedic Surgery, Duke University, DUMC Box 3000, Durham, NC 27710 ( Thorsten.Seyler@duke.edu).

Received: May 03, 2019
Accepted: July 25, 2019

10.3928/01477447-20200818-02

Sign up to receive

Journal E-contents