Download PDF
J Knee Surg 2022; 35(01): 007-014
DOI: 10.1055/s-0040-1710573
Original Article

Construction and Comparison of Predictive Models for Length of Stay after Total Knee Arthroplasty: Regression Model and Machine Learning Analysis Based on 1,826 Cases in a Single Singapore Center

Hui Li‡   *
1   Department of Bone and Joint Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China
,
Juyang Jiao‡   *
1   Department of Bone and Joint Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China
,
Shutao Zhang
1   Department of Bone and Joint Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China
,
Haozheng Tang
1   Department of Bone and Joint Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China
,
Xinhua Qu
1   Department of Bone and Joint Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China
,
Bing Yue
1   Department of Bone and Joint Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People's Republic of China
› Author Affiliations Funding This research was supported by the National Natural Science Foundation of China (grant nos. 81972086, 81672196, and 51971222); Shanghai “Rising Stars of Medical Talent” Youth Development Program (Youth Medical Talents—Specialist Program); National Key Research and Development Project; “Technology Innovation Action Plan” Key Project of Shanghai Science and Technology Commission (grant no. 19411962800); Shanghai Municipal Education Commission—Gaofeng Clinical Medicine Grant Support (grant no. 20161423); Clinical Scientific Innovation and Cultivation Fund of Renji Hospital Affiliated School of Medicine, Shanghai Jiaotong University(grant no. PY2018-I-02).
› Further Information
    Permissions and Reprints

Abstract

The purpose of this study was to develop a predictive model for length of stay (LOS) after total knee arthroplasty (TKA). Between 2013 and 2014, 1,826 patients who underwent TKA from a single Singapore center were enrolled in the study after qualification. Demographics of patients with normal and prolonged LOS were analyzed. The risk variables that could affect LOS were identified by univariate analysis. Predictive models for LOS after TKA by logistic regression or machine learning were constructed and compared. The univariate analysis showed that age, American Society of Anesthesiologist level, diabetes, ischemic heart disease, congestive heart failure, general anesthesia, and operation duration were risk factors that could affect LOS (p < 0.05). Comparing with logistic regression models, the machine learning model with all variables was the best model to predict LOS after TKA, of whose area of operator characteristic curve was 0.738. Machine learning algorithms improved the predictive performance of LOS prediction models for TKA patients.

Keywords

length of stay - total knee arthroplasty - predictive model - machine learning

Note

The approval from Institutional Review Board (Sing-health CIRB 2014/651/D) was obtained prior to the start of the study.


* These authors contributed equally to this work.




Publication History

Received: 06 September 2019

Accepted: 31 March 2020

Article published online:
08 June 2020

© 2020. Thieme. All rights reserved.

Thieme Medical Publishers, Inc.
333 Seventh Avenue, 18th Floor, New York, NY 10001, USA

 

  • References

  • 1 Maradit Kremers H, Larson DR, Crowson CS. et al. Prevalence of total hip and knee replacement in the United States. J Bone Joint Surg Am 2015; 97 (17) 1386-1397
    CrossrefPubMedGoogle Scholar
  • 2 den Hertog A, Gliesche K, Timm J, Mühlbauer B, Zebrowski S. Pathway-controlled fast-track rehabilitation after total knee arthroplasty: a randomized prospective clinical study evaluating the recovery pattern, drug consumption, and length of stay. Arch Orthop Trauma Surg 2012; 132 (08) 1153-1163
    CrossrefPubMedGoogle Scholar
  • 3 Husted H, Jensen CM, Solgaard S, Kehlet H. Reduced length of stay following hip and knee arthroplasty in Denmark 2000-2009: from research to implementation. Arch Orthop Trauma Surg 2012; 132 (01) 101-104
    CrossrefPubMedGoogle Scholar
  • 4 Healy WL, Rana AJ, Iorio R. Hospital economics of primary total knee arthroplasty at a teaching hospital. Clin Orthop Relat Res 2011; 469 (01) 87-94
    CrossrefPubMedGoogle Scholar
  • 5 Ayalon O, Liu S, Flics S, Cahill J, Juliano K, Cornell CN. A multimodal clinical pathway can reduce length of stay after total knee arthroplasty. HSS J 2011; 7 (01) 9-15
    CrossrefPubMedGoogle Scholar
  • 6 Kim S, Losina E, Solomon DH, Wright J, Katz JN. Effectiveness of clinical pathways for total knee and total hip arthroplasty: literature review. J Arthroplasty 2003; 18 (01) 69-74
    CrossrefPubMedGoogle Scholar
  • 7 Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270 (24) 2957-2963
    CrossrefPubMedGoogle Scholar
  • 8 Iwashyna TJ, Liu V. What's so different about big data? A primer for clinicians trained to think epidemiologically. Ann Am Thorac Soc 2014; 11 (07) 1130-1135
    CrossrefPubMedGoogle Scholar
  • 9 Churpek MM, Yuen TC, Winslow C, Meltzer DO, Kattan MW, Edelson DP. Multicenter comparison of machine learning methods and conventional regression for predicting clinical deterioration on the wards. Crit Care Med 2016; 44 (02) 368-374
    CrossrefPubMedGoogle Scholar
  • 10 Koyner JL, Carey KA, Edelson DP, Churpek MM. The development of a machine learning inpatient acute kidney injury prediction model. Crit Care Med 2018; 46 (07) 1070-1077
    CrossrefPubMedGoogle Scholar
  • 11 Fernández-Delgado M, Cernadas E, Barro S, Ribeiro J, Neves J. Direct Kernel Perceptron (DKP): ultra-fast kernel ELM-based classification with non-iterative closed-form weight calculation. Neural Netw 2014; 50: 60-71
    CrossrefPubMedGoogle Scholar
  • 12 Pinsky MR, Dubrawski A. Gleaning knowledge from data in the intensive care unit. Am J Respir Crit Care Med 2014; 190 (06) 606-610
    CrossrefPubMedGoogle Scholar
  • 13 Abdullah HR, Sim YE, Hao Y. et al. Association between preoperative anaemia with length of hospital stay among patients undergoing primary total knee arthroplasty in Singapore: a single-centre retrospective study. BMJ Open 2017; 7 (06) e016403
    CrossrefPubMedGoogle Scholar
  • 14 Kaiser SV, Bakel LA, Okumura MJ, Auerbach AD, Rosenthal J, Cabana MD. Risk factors for prolonged length of stay or complications during pediatric respiratory hospitalizations. Hosp Pediatr 2015; 5 (09) 461-473
    CrossrefPubMedGoogle Scholar
  • 15 Modi SY, Dharaiya D, Katramados AM, Mitsias P. Predictors of prolonged hospital stay in status migrainosus. Neurohospitalist 2016; 6 (04) 141-146
    CrossrefPubMedGoogle Scholar
  • 16 Smith ID, Elton R, Ballantyne JA, Brenkel IJ. Pre-operative predictors of the length of hospital stay in total knee replacement. J Bone Joint Surg Br 2008; 90 (11) 1435-1440
    CrossrefPubMedGoogle Scholar
  • 17 Adili A, Bhandari M, Petruccelli D, De Beer J. Sequential bilateral total knee arthroplasty under 1 anesthetic in patients > or = 75 years old: complications and functional outcomes. J Arthroplasty 2001; 16 (03) 271-278
    CrossrefPubMedGoogle Scholar
  • 18 Martínez-Huedo MA, Jiménez-García R, Jiménez-Trujillo I, Hernández-Barrera V, Del Rio Lopez B, López-de-Andrés A. Effect of type 2 diabetes on in-hospital postoperative complications and mortality after primary total hip and knee arthroplasty. J Arthroplasty 2017; 32 (12) 3729-3734.e2
    CrossrefPubMedGoogle Scholar
  • 19 Martinez-Huedo MA, Villanueva M, de Andres AL. et al. Trends 2001 to 2008 in incidence and immediate postoperative outcomes for major joint replacement among Spanish adults suffering diabetes. Eur J Orthop Surg Traumatol 2013; 23 (01) 53-59
    CrossrefPubMedGoogle Scholar
  • 20 Lovecchio F, Beal M, Kwasny M, Manning D. Do patients with insulin-dependent and noninsulin-dependent diabetes have different risks for complications after arthroplasty?. Clin Orthop Relat Res 2014; 472 (11) 3570-3575
    CrossrefPubMedGoogle Scholar
  • 21 Yong CM, Sharma M, Ochoa V. et al. Multivessel coronary artery disease predicts mortality, length of stay, and pressor requirements after liver transplantation. Liver Transpl 2010; 16 (11) 1242-1248
    CrossrefPubMedGoogle Scholar
  • 22 Higuera CA, Elsharkawy K, Klika AK, Brocone M, Barsoum WK. 2010 Mid-America Orthopaedic Association Physician in Training Award: predictors of early adverse outcomes after knee and hip arthroplasty in geriatric patients. Clin Orthop Relat Res 2011; 469 (05) 1391-1400
    CrossrefPubMedGoogle Scholar
  • 23 Khormaee S, Do HT, Mayr Y. et al. Risk of ischemic stroke after perioperative atrial fibrillation in total knee and hip arthroplasty patients. J Arthroplasty 2018; 33 (09) 3016-3019
    CrossrefPubMedGoogle Scholar
  • 24 van den Belt L, van Essen P, Heesterbeek PJ, Defoort KC. Predictive factors of length of hospital stay after primary total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2015; 23 (06) 1856-1862
    CrossrefPubMedGoogle Scholar
  • 25 Otero JE, Gholson JJ, Pugely AJ, Gao Y, Bedard NA, Callaghan JJ. Length of hospitalization after joint arthroplasty: does early discharge affect complications and readmission rates?. J Arthroplasty 2016; 31 (12) 2714-2725
    CrossrefPubMedGoogle Scholar
  • 26 Inneh IA, Lewis CG, Schutzer SF. Focused risk analysis: regression model based on 5,314 total hip and knee arthroplasty patients from a single institution. J Arthroplasty 2014; 29 (10) 2031-2035
    CrossrefPubMedGoogle Scholar
  • 27 Rezaei S, Akbari Sari A, Arab M, Majdzadeh R, Shaahmadi F, Mohammadpoorasl A. The association between smoking status and hospital length of stay: evidence from a hospital-based cohort. Hosp Pract (1995) 2016; 44 (03) 129-132
    CrossrefPubMedGoogle Scholar
  • 28 Chen L, Dubrawski A, Wang D. et al. Using supervised machine learning to classify real alerts and artifact in online multisignal vital sign monitoring data. Crit Care Med 2016; 44 (07) e456-e463
    CrossrefPubMedGoogle Scholar