Comparison of Back-Propagation Neural Network, LACE Index and HOSPITAL Score in Predicting All-Cause Risk of 30-Day Readmission.
30-day readmission
BPNN
HOSPITAL
LACE
back-propagation neural network
healthcare quality
Journal
Risk management and healthcare policy
ISSN: 1179-1594
Titre abrégé: Risk Manag Healthc Policy
Pays: England
ID NLM: 101566264
Informations de publication
Date de publication:
2021
2021
Historique:
received:
05
05
2021
accepted:
27
08
2021
entrez:
22
9
2021
pubmed:
23
9
2021
medline:
23
9
2021
Statut:
epublish
Résumé
The main purpose of this study is to predict the all-cause risk of 30-day readmission by employing the back-propagation neural network (BPNN) in comparison with traditional risk assessment tools of LACE index and HOSPITAL scores. This was a retrospective cohort study from January 1st, 2018 to December 31st, 2019. A total of 55,688 hospitalizations from a medical center in Taiwan were examined. The LACE index (length of stay, acute admission, Charlson comorbidity index score, emergency department visits in previous 6 months) and HOSPITAL score (hemoglobin level at discharge, discharge from an Oncology service, sodium level at discharge, procedure during hospital stay, Index admission type, number of hospital admissions during the previous year, length of stay) are calculated. We employed variables from LACE index and HOSPITAL score as the input vector of BPNN for comparison purposes. The BPNN constructed in the current study has a considerably better ability with a C statistics achieved 0.74 (95% CI 0.73 to 0.75), which is statistically significant larger than that of the other two models using DeLong's test. Also, it was possible to achieve higher sensitivity (70.32%) without penalizing the specificity (71.76%) and accuracy (71.68%) at its optimal threshold, which is at the 20% of patients with the highest predicted risk. Moreover, it is much more informative than the other two methods because of a considerably higher LR+ and a lower LR-. Our findings suggest that more attention should be paid to methods based on non-linear classification systems, as they lead to substantial differences in risk-scores.
Sections du résumé
BACKGROUND
BACKGROUND
The main purpose of this study is to predict the all-cause risk of 30-day readmission by employing the back-propagation neural network (BPNN) in comparison with traditional risk assessment tools of LACE index and HOSPITAL scores.
METHODS
METHODS
This was a retrospective cohort study from January 1st, 2018 to December 31st, 2019. A total of 55,688 hospitalizations from a medical center in Taiwan were examined. The LACE index (length of stay, acute admission, Charlson comorbidity index score, emergency department visits in previous 6 months) and HOSPITAL score (hemoglobin level at discharge, discharge from an Oncology service, sodium level at discharge, procedure during hospital stay, Index admission type, number of hospital admissions during the previous year, length of stay) are calculated. We employed variables from LACE index and HOSPITAL score as the input vector of BPNN for comparison purposes.
RESULTS
RESULTS
The BPNN constructed in the current study has a considerably better ability with a C statistics achieved 0.74 (95% CI 0.73 to 0.75), which is statistically significant larger than that of the other two models using DeLong's test. Also, it was possible to achieve higher sensitivity (70.32%) without penalizing the specificity (71.76%) and accuracy (71.68%) at its optimal threshold, which is at the 20% of patients with the highest predicted risk. Moreover, it is much more informative than the other two methods because of a considerably higher LR+ and a lower LR-.
CONCLUSION
CONCLUSIONS
Our findings suggest that more attention should be paid to methods based on non-linear classification systems, as they lead to substantial differences in risk-scores.
Identifiants
pubmed: 34548831
doi: 10.2147/RMHP.S318806
pii: 318806
pmc: PMC8449689
doi:
Types de publication
Journal Article
Langues
eng
Pagination
3853-3864Informations de copyright
© 2021 Lin et al.
Déclaration de conflit d'intérêts
The authors declare that they have no competing interests.
Références
Sci Rep. 2020 Jan 24;10(1):1111
pubmed: 31980704
BMJ Open. 2017 Jul 13;7(7):e016921
pubmed: 28710226
Health Aff (Millwood). 2006 Mar-Apr;25(2):510-20
pubmed: 16522605
Med Care Res Rev. 2019 Apr;76(2):229-239
pubmed: 29148348
Health Aff (Millwood). 2019 Apr;38(4):585-593
pubmed: 30933582
BMJ Open. 2016 Jun 27;6(6):e011060
pubmed: 27354072
IEEE/ACM Trans Comput Biol Bioinform. 2018 Nov-Dec;15(6):1968-1978
pubmed: 29993930
PLoS One. 2020 Apr 15;15(4):e0221606
pubmed: 32294087
Med Care. 2006 Nov;44(11):972-81
pubmed: 17063128
Cancer. 1950 Jan;3(1):32-5
pubmed: 15405679
JAMA Intern Med. 2016 Apr;176(4):496-502
pubmed: 26954698
Circ Cardiovasc Qual Outcomes. 2011 Mar;4(2):243-52
pubmed: 21406673
Int J Environ Res Public Health. 2021 May 12;18(10):
pubmed: 34065894
CMAJ. 2010 Apr 6;182(6):551-7
pubmed: 20194559
ESC Heart Fail. 2019 Apr;6(2):428-435
pubmed: 30810291
PLoS One. 2020 Jun 25;15(6):e0235117
pubmed: 32584872
J Hosp Med. 2014 Nov;9(11):743-4
pubmed: 25303367
JAMA. 2017 Oct 10;318(14):1377-1384
pubmed: 29049590
BMC Med Inform Decis Mak. 2020 Jan 6;20(1):4
pubmed: 31906931
Health Syst (Basingstoke). 2017 Nov 07;7(2):120-134
pubmed: 31214343
BMJ. 2014 Mar 31;348:g2329
pubmed: 24687184
J Biomed Inform. 2015 Aug;56:229-38
pubmed: 26044081
J Eval Clin Pract. 2017 Jun;23(3):524-529
pubmed: 27696638
Med Care. 2005 Nov;43(11):1130-9
pubmed: 16224307
Health Care Manag Sci. 2015 Mar;18(1):19-34
pubmed: 24792081
Open Med. 2012 Jul 19;6(3):e80-90
pubmed: 23696773
N Engl J Med. 2016 Apr 21;374(16):1543-51
pubmed: 26910198
PLoS One. 2017 Jul 14;12(7):e0181173
pubmed: 28708848
JAMA Intern Med. 2013 Apr 22;173(8):632-8
pubmed: 23529115
J Hosp Med. 2014 Nov;9(11):681-7
pubmed: 24945696
Sci Rep. 2019 Feb 20;9(1):2362
pubmed: 30787351
Circ Cardiovasc Qual Outcomes. 2016 Nov;9(6):629-640
pubmed: 28263938
PLoS One. 2015 Mar 04;10(3):e0118432
pubmed: 25738806
N Engl J Med. 2009 Apr 2;360(14):1418-28
pubmed: 19339721
Am J Med. 2017 Oct;130(10):1220.e1-1220.e16
pubmed: 28606799
JAMA. 2011 Oct 19;306(15):1688-98
pubmed: 22009101
BMC Health Serv Res. 2011 Jun 17;11:149
pubmed: 21679471
Biometrics. 1988 Sep;44(3):837-45
pubmed: 3203132
Clin Pediatr (Phila). 2014 Aug;53(9):845-53
pubmed: 24899633
PLoS One. 2019 Mar 12;14(3):e0210714
pubmed: 30861004
Artif Intell Med. 2019 Apr;95:27-37
pubmed: 30213670