aimed00212_v2.dvi

Size: px

Start display at page:

Download "aimed00212_v2.dvi"

ありおきしげい
5 years ago
Views:

1 AIM SIG-AIMED-2-2 DPC Classiicati Cases with the sae DPC cde assiged based Cluster Aalysis 2 3 Shusaku Tsut, Shi Hira Haruk Iwata 2 Thir Kiura 3 Departet Medical Iratics, Faculty Medicie Shiae Uiversity 2 2 Ceter r Bed-ctrl, Shiae Uiversity Hspital 3 3 Tshiba Medical Systes Abstract: Abstract. This paper shws a prcess hw cliical cases with the sae DPC cde ca be characterized by ixture del clusterig, ad aalysis discharge suaries. First, cases are classiied it a give uber clusters by ixture dels, ad the keywrds iprtat r classiicati are extracted by crrespdece aalysis ad SVM. The experieetal results shw that all the prcess gives gd perrace r autatic classiicati exaples. DPC[] ICD[8] WHO ICD, DPC DPC SVM % 2 : DPC Wrdclud, SVM, () Wrdclud, SVM E-ail: tsut@ed.shiae-u.ac.p : 2-

2 2. [9, 4] 25 DPC [7] EM R 3.3. ixtls[2], lexix[6] DPC ixtls Desity Curves Data Desity : ixtls 3 : 3. Wrdclud (grup) (grup2) lexix (RMeCab)[] ter atrix Wrdclud[3] 3 4 MRI CT D R DWI MRA Na CTR FLAIR Barre K NIHSS L i HR dier CRP (-) ALT AST Bil Crea APTT INR CK PT LDH kg (-) BNP bp + M Cl SpO Hg ADL Ca egfr BP LDA HbA V c S M y y { day s GTP ug { { BSA A BT Migazzii c list y ADC LDL MCA Ž Alp Babiski O k I y { Fib Œ ˆ h Ž HDL y Œ { Aylase Xp h i DAPT DM E ICA JCS NGSP FT { d Chaddck ECG EF MMT ST gd * i { y prble A PICA Trer llw high %( - y Hb dœ Ž l Ha ull RS s sharp % (+) )( ) (-) d ChE VA g plaque p Ž p y MCV WI )= -/- { MCHC Neutr e i u Migazzii ER LV MR P RR ild sius % (+), +/+ y y ATR Ly ph OD PTR lw icrbleeds t e ˆ t t CBC NSR Œ BAD GCS HT PVH (-,-) ),, { CP MCH TSH ax g i Prble y e CCA CPA HCV LA OK RA TP asy ergy lx (+), (+,+) ) ) ± -/-, +,+ ±/± Ã ),, Œ AR BIF Bas Es M NH PCA Sips agle drug pulse rhy th scre y Œ l h g h h STT lair d ttal d F HDS TIA TR bruit dilatati l p A MA (++,++) (-) ): ) -/- -) -), -) +/+ ±/± l Žd i BUN Biceps Brachiradialis JDS LAD LAVI MB N a reg sep r i s l e { ABI HBs Rberg Vital v v Œ Œ g ARB B CAVI DTR Ddier FAB FNF GPA IVC LVEF LVH MMSE SOCRATES clear r tpa Ž % ) (-) +/+ + ] / +,+ -,- -/-, y Œ i i Ž Ž y Alb Alzheier BRTR CA CTA HK Mg Plt Q SPECT TTR Tibial Triceps WBC air ad eusi hal lat lesi e y i d {i e e i ˆ Ž u Ž z g eck t prbre N Ž Œ Œ l Œ k AS Agi BTR HOT Ha Ma PA PH PR Pa SAH SDS Sural VC div e RS l 3: Wrdclud ter atrix R MASS

3 : DPC DPC 2xx97xxx xx97xxx xxxxxxxx xx97xxx xxx99x4x xx99x xx2xxxx xx994xx g 4x99xx xxxxxx xx99xxx xx99x4x x99xxx xx99xx xx99x xx97xxx xx3xx JCS 6x993x xx97xxx xxxxxx

4 d l clear Œ Ha h HDS { Ž VPC d abral BS ral TTR TP sharp MR ˆ sth Sips sius OD q RR h p Ch Plt -/- { d Ha derate DSEP Aylase MCV Œ LA i sepe p s { g VA LAP drug LDA t Hg c + Ca h y Cl Babiski bp egfr u v l BAD i CK ext { l ractiv e HR INR HbA Œ BNP RS wall { Neutr APTT h ADL t llw ALT FT h CPA ST Fib PH Barre d NGSP % RBC SpO SVPC ATR K CT p scre e h lx PT z CTA Œ BP i e SDS (-,-) M D ˆ M DWI F ECG E Ab h Rt e + s { Eq (+,+) ph HK (-) Ž AST * CEA s % ):??/ y -# EF y N MMT { { (+) FLAIR -) O GCS l z GTP V I -, i CTR ADC s { Hb )= BSA BRTR c L Bil Xp i AR Lt ttal) JCSCrea Ã HCV DAPT y MRA q Es MRI CRP WBCHPF y Ž MCA i LDL S R M LV BT ŒŒ -/- i A Chl Œ ug high list rhy th lw HDL A LDH RA P d i t k pause k ) -/- CCA d i PTR y A CA (++/+) bruit day { d Bruit h ADCap ˆ " N y AM HBs Trer Ly ph NH + Œ aximt +/+ Chaddck dœ TSH r ech Ž y (+/+) k p { s plaque OK Œ %( JDSChE -/- p Bas PSV r asy ergy +/+ Suary Œ lz { q Œ Migazzii ICA Prble FNF Š hr DM EOM BUN i FAB (++,++) IVC l d Œ CAS PSA FDP (+,+) LAVI PSL MCHC HT PPI Œ (-,+) Ž prble (-) air PCA HL TR dier Na ull Alp B g TG reg MMSE a IU NIHSS %), r ABI ti BMI kg Migazzii (+/+), BTR Pa y g (-/-) Alb ild s CBC MCH y SPO CAVI h Œ + ACAap DTR Grup- gazzii rble ADL day BT Xp A + FLAIR L (-) Barre CTR HR Hg LDA MRA egfr list PT RBSA i APTT dier SpO CKCl Na INR bp Fib DWI kg CRPCT BP Crea CaChaddck AST ALT M MRI BNP LDH D GTPHbA Aylase K NGSP * NIHSS O MAlp Bil S Vug (-) c MCA c FT LDL Babisk Tre Grup- GCS MMT HDL (+,+) JCS (-,-) EI : Wrdclud 2: t-id ADL day BT Migazzii : 3: t-id Babiski HDL MMT JCS I GCS : 2 SVM 2 2-4

5 4: : SVM R3.3. rpart ld crss validati 4.2 SVM R3.3. kerlab [5] 25 5-ld crss validati 78.8% ld crss validati Grup-2 Grup- Nde 3 ( = 38) 2 <.5.5 <.5.5 Grup-2 Grup- Nde 4 ( = 2) Grup-2 Grup- Nde 6 ( = 7) 6: 5.5 <.5 Grup-2 Grup- Nde 7 ( = 23) ,22 25 SVM verittig 2-5

6 5 DPC DPC [6] Friedrich Leisch. Flexix: A geeral raewrk r iite ixture dels ad latet class regressi i r. Jural Statistical Stware, (8): 8, 24. [7] G. J. McLachla ad D. Peel. Fiite Mixture Mdels. Wiley, New Yrk, 2. [8] Wrld Health Orgaizati. ICD-. ICD- / Wrld Health Orgaizati. Wrld Health Orgaizati, 993. [9] Shusaku Tsut, Haruk Iwata, Shi Hira, ad Yuk Tsut. Siilarity-based behavir ad prcess iig edical practices. Future Geerati Cp. Syst., 33:2 3, 24. []. Recab. idex.php?rmecabfuctis, 26. 5lk3h [] DPC., 29. [2] Tatiaa Beaglia, Didier Chauveau, David R. Huter, ad Derek Yug. ixtls: A r package r aalyzig iite ixture dels. Jural Statistical Stware, 32(6): 29, 29. [3] Ia Felllws. Wrd cluds. https: //cra.r-prect.rg/web/packages/ wrdclud/idex.htl, 24. [4] Haruk Iwata, Shi Hira, ad Shusaku Tsut. Maiteace ad discvery dai kwledge r ursig care usig data i hspital irati syste. Fuda. Ir., 37(2): , 25. [5] Alexadrs Karatzglu, Alex Sla, Kurt Hrik, ad Achi Zeileis. kerlab a S4 package r kerel ethds i R. Jural Statistical Stware, (9): 2,

387 ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )

386 B B () () ( ) ( ) ( ) 387 ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) 388 ( ) ( ) ( ) 389 ( ) 2 ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) 390 ( ) ( ) ( ) () () ( ) Hot flush 391 ( ) ( ) ( ) 392 ( ( )