[PDF] Prediction of fall events during admission using eXtreme gradient

View - Download Prediction of fall events during admission using eXtreme gradient

Previous PDF

Next PDF

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports p rediction of fall events during admission using eXtreme gradient boosting: a comparative validation study Yin- c hen Hsu & Yuan-Hsiung tsai

in identifying patients at risk of falling. this study proposes an automatic fall risk prediction model

models. this machine learning method provided a higher sensitivity than the standard method

Department of Neuro-

Rehabilitation, Department of Surgery, cardiovascular medication use, admission from the Emergency

Department, and bed rest

practice1 .InTaiwan,theDepart which51.9%resultedin injuries 2 mortality 3 fallrisk 4 validityand reliability 5,6 involved7 offallriskintheUnited

States

8 Department of Diagnostic Radiology, Chang Gung Memorial Hospital Chiayi Branch, Chiayi, Taiwan.

College

of Medicine, Chang Gung University, Taoyuan, Taiwan. Department of Nursing, Chang Gung Memorial Hospital

Chiayi Branch, Chiayi, Taiwan.

email: russell.tsai@gmail.com

Vol:.(1234567890)

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports/

Results

Demographics and clinical characteristics.

1 Table 1. mean a ?egeneralizablefactors b ?e P-valuesindicateasigni?cant di?erencebetweenthetwocohorts.?e P squaretest,respectively.

CharacteristicsDerivation cohortValidation cohort

P b (N

507)(N = 132)

Sex a

Age(y)

a 69.30

13.7469.32±15.180.988

Admissionday

Source

Department

Surgery,n(%)138(27.22%)37(28.03%)0.853

Other,n(%)6(1.18%)2(1.52%)0.754

Wardtype

Vitalsigns

a a 14.41

1.9514.36±2.010.794

Musclepower

a

Activity

a

Normal,n(%)249(49.11%)61(46.21%)0.553

Weak,n(%)207(40.83%)60(45.46%)0.337

Bedrest,n(%)51(10.06%)11(8.33%)0.550

Educationlevel

a

Maritalstatus

a

Single,n(%)42(8.28%)8(6.06%)0.398

Married,n(%)421(83.04%)111(84.09%)0.774

Caregiver

a

None,n(%)5(0.99%)1(0.76%)0.808

Family,n(%)464(91.52%)124(93.94%)0.361

Nursingworker,n(%)38(7.49%)7(5.30%)0.381

Ambulatoryaid

a

Free,n(%)476(93.89%)121(91.67%)0.360

Caneorwalker,n(%)20(3.94%)4(3.03%)0.624

Wheelchair,n(%)11(2.17%)7(5.30%)0.053

FRID a

CNSdrugs,n(%)19(3.75%)11(8.33%)0.027

Vol.:(0123456789)

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports/ showninTable 2.?ereweresigni?cantdi?erencesbetweenthefallandnonfallgroupsinthefallriskassessment performance of the prediction models. Table 2presentsthecontingencytableforthepredictionmod-

65.0%,respectively).Figure 1presentstheROCcurvesandareasunderthecurves(AUCs)toassesstheoverall

and0.700,respectively;P=0.09).Table 3presentsthesensitivity,speci?city,PPV,NPV,LR+,andLR-ofthetwo calculatedattheoptimalcut-o?value.?e featureset. f eature importance according to XGB. Figure 2showstherankingoffeatureimportanceaccordingto onpredictinginpatientfallevents. Table 2.

ModelMFSXGBXGB-GF

Nofall33.21 ± 16.380.4121 ± 0.21520.4099 ± 0.2273 Fall41.63 ± 17.770.5899 ± 0.27910.5852 ± 0.2802 p< 0.01< 0.001< 0.001

NoYesSumNoYesSumNoYesSum

Nofall642892692392642892

Fall202040142640152540

Sum844813283491327953132

Figure 1.

ROCcurvesfortheMFSandXGBmodels.

Vol:.(1234567890)

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports/

Discussion

risk 9-11 MFS 12 actualfallevent. 13 .Tobetterassessthe Table 3. interval;LR

Scale/modelMFSXGBXGB-GF

Cuto?>45>0.53>0.58

Accuracy(%)63.6471.9767.42

LR

Figure 2.

FeatureimportanceoftheXGBmodel.

Vol.:(0123456789)

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports/ ?ndingsoncardiovasculardrugsandfallriskareinconformitywiththerecentmeta-analysisbydeVrieset al. 14 15 .Despiteourinherent c onclusions

Methods

Study design and participants.

theregistrationsystemofthepatientsafetycommittee.Patientswhowereunder20 yearsoldandthosewhohad graphicdatawereobtained.Figure 3 showsthe?owchartofthestudy.

Morse fall scale (M

f S). group 9

Development of XGB.

Chenet al.demonstratedtherobustpowerofXGBsystemtocontrolover-?tting inavarietyofdataminingchallenges 16

XGBinthiswork.

Vol:.(1234567890)

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports/ extracteddatasetareshowninTable 1. feature 17

Statistical analysis.

Smirnovtest).Datawerereportedasthemean

index 18 .ROCcurveswerecom paredusingthemethoddescribedbyDeLonget al. 19 bya2 +),andnegativelikelihood ratio(LR-) 20 P

0.05indicatesstatistical

signi?cance. e thics approval and consent to participate. participantsinthemanuscript. c onsent for publication.

Notrequired.

Data availability

Received: 14 March 2020; Accepted: 15 September 2020

References

1.Evans,D.,Hodgkinson,B.,Lambert,L.&Wood,J.Fallsriskfactorsinthehospitalsetting:asystematicreview.Int. J. Nurs. Pract.

Figure 3.

Flowdiagram.

Vol.:(0123456789)

Scientific Repo

R t

S | (2020) 10:16777 |

www.nature.com/scientificreports/ 2.

Taiwan Patient Safety Reporting System Annual Report,https://www.patientsafety.mohw.gov.tw/Content/Downloads/List0

1.aspx?SiteI

D=1&MmmID

=621273303702500244(2017).

3.Coussement,J.et al.Interventionsforpreventingfallsinacute-andchronic-carehospitals:asystematicreviewandmeta-analysis.

J. Am. Geriatr. Soc.

56,29-36(2008).

4.

Sentinel Event Alert: Preventing Falls and Fall-Related Injuries in Health Care Facilities,https://www.jointcommission.org/asset

s/1/18/SEA_55.pdf(2015).

5.5Gates,S.,Smith,L.A.,Fisher,J.D.&Lamb,S.E.Systematicreviewofaccuracyofscreeninginstrumentsforpredictingfallriskamongindependentlylivingolderadults.J. Rehabil. Res. Dev. 45(2008).

6.Park,S.H.Toolsforassessingfallriskintheelderly:asystematicreviewandmeta-analysis.Aging Clin. Exp. Res.30,1-16.https

7.Marschollek,M.et al.Mininggeriatricassessmentdataforin-patientfallpredictionmodelsandhigh-risksubgroups.BMC Med.

Inform. Decis. Mak.12,19.https://doi.org/10.1186/1472-6947-12-19(2012).

8.Kim,K.et al.Developmentofperformancemeasuresbasedonthenursingprocessforpreventionandmanagementofpressure

ulcers,fallsandpain.

J. Korean Clin. Nurs. Res.15,133-147(2009).

9.Morse,J.M.,Morse,R.M.&Tylko,S.J.Developmentofascaletoidentifythefall-pronepatient.Can. J. Aging/La Revue canadienne

du vieillissement8,366-377(1989). hospitalizedpatients.

Appl. Nurs. Res.

BMJ315,1049-1053(1997).

12.O'Connell,B.&Myers,H.Afailedfallpreventionstudyinanacutecaresetting:lessonsfromtheswamp.Int. J. Nurs. Pract.7,

13.Spoelstra,S.L.,Given,B.A.&Given,C.W.Fallpreventioninhospitals:anintegrativereview.Clin Nurs Res21,92-112.https://

doi.org/10.1177/1054773811418106(2012).

14.deVries,M.et al.Fall-risk-increasingdrugs:asystematicreviewandmeta-analysis:Icardiovasculardrugs.J. Am. Med. Dir. Assoc.19(371)e371-e379,https://doi.org/10.1016/j.jamda.2017.12.013(2018).

15.

Aranda-Gallardo,M.et al.Instrumentsforassessingtheriskoffallsinacutehospitalizedpatients:asystematicreviewprotocol.

J Adv Nurs

16.Chen,T.&Guestrin,C.InProceedings of the 22nd ACM Sigkdd International Conference on Knowledge Discovery and Data

Mining.785-794.

17.Guyon,I.&Elissee?,A.Anintroductiontovariableandfeatureselection.J. Mach. Learn. Res.3,1157-1182(2003).

20.Sheskin,D.J.Handbook of Parametric and Nonparametric Statistical Procedures(CRCPress,BocaRaton,2003).

Author contributions

project.Allauthorsreviewedthemanuscript. f undingquotesdbs_dbs14.pdfusesText_20

[PDF] eyfel kulesi basit çizim

[PDF] eyfel kulesi basit çizimi

[PDF] eyfel kulesi çizimi youtube

[PDF] eyfel kulesi çizimleri karakalem

[PDF] eyfel kulesi karakalem çizimi nas?l yap?l?r

[PDF] eyfel kulesinin çizimi

[PDF] e^(a b) math

[PDF] f 35 2019 deliveries

[PDF] f 35 2019 demo

[PDF] f 35 2019 production

[PDF] f 35 2019 sar

[PDF] f 35 air show 2019

[PDF] f 35 block 3f

[PDF] f 35 block 4

[PDF] f 35 cni suite