Home > Journals > European Journal of Physical and Rehabilitation Medicine > Past Issues > Europa Medicophysica 2001 September;37(3) > Europa Medicophysica 2001 September;37(3):129-34

CURRENT ISSUE
 

JOURNAL TOOLS

eTOC
To subscribe
Submit an article
Recommend to your librarian
 

ARTICLE TOOLS

Reprints
Permissions

 

EDITORIAL   Freefree

Europa Medicophysica 2001 September;37(3):129-34

Copyright © 2001 EDIZIONI MINERVA MEDICA

language: English

The Italian research project for the prospective payment of hospital rehabilitation care. Notes on a work in progress

Saitto C., De Bernardini L. 1, Fletzer D. A. 2, Giustini A. 3, Littera A. M. 4, Mastrilli F. 5, Mastrobuono I. 3, Traballesi M. 5, Di Benedetto P. 6

From the Dipartimento di Epidemiologia RME - Roma 1 Villa delle Quercie-Poligest - Roma 2 ASL RMD - Roma Centro Paraplegici - Ostia 3 Tosinvest Sanità - Roma 4 Agenzia di Sanità Pubblica - Roma 5 Fondazione IRCCS S. Lucia - Roma 6 Azienda «Ospedali Riuniti» - Trieste


PDF


Prospective pay­ment of inpa­tient reha­bil­i­ta­tion ser­vic­es ­has ­been sug­gest­ed or imple­ment­ed in sev­er­al coun­tries. The cur­rent Italian ­system clas­si­fies ­patients ­into 5 ­broad ­groups ­based on MDCs. MDCs ­are ­the par­ent par­ti­tion of ­the ­DRG clas­sifi­ca­tion ­and ­are mean­ing­less in ­the reha­bil­i­ta­tion set­ting, a spe­cif­ic clas­sifi­ca­tion of inpa­tient reha­bil­i­ta­tion ­care is ­thus obvi­ous­ly need­ed. The fun­da­men­tal ­goal of an ­iso-­resource clas­sifi­ca­tion is build­ing clin­i­cal­ly homo­ge­ne­ous dis­charge ­groups ­with ­the small­est pos­sible with­in-­the-­group ­resource con­sump­tion var­i­abil­ity. Existing mod­els of ­patient assess­ment ­and clas­sifi­ca­tion ­are pre­sent­ed ­and ­their ­strengths ­and short­com­ings ­are dis­cussed. The ­basic char­ac­ter­is­tics of a sat­is­fac­to­ry mod­el ­are ­defined ­and meth­ods of mod­el build­ing ­are brief­ly ­described. No assess­ment ­tool ­looks com­plete­ly sat­is­fac­to­ry, ­but to ­accept a lim­it­ed ­set of infor­ma­tion ­means accept­ing ­the var­i­able selec­tion ­and ­the asso­ci­a­tion meth­ods of exist­ing clas­sifi­ca­tions. The selec­tion of var­i­ables ­and ­the attri­bu­tion of var­i­able ­weights in ­the clas­sifi­ca­tion ­system ­must be a ver­i­fi­able ­end ­point ­and ­not ­the start­ing ­point of anal­y­sis. A rela­tion­ship ­between explan­a­to­ry var­i­ables ­and out­come ­must be ­sought, ­not ­only as a ­tool to orga­nize ­and ­group ­the ­costs of ­the ­observed ­events, ­like ­CART, ­but ­also as an instru­ment to fore­see ­the rea­son­able ­costs of ­the ­events to ­come. Several group­ing meth­ods ­must ­then be test­ed ­and com­pared ­against ­CART (Correlation And Regression Trees), ­like Search Partition Analysis (­SPAN), ­main com­po­nents anal­y­sis, clus­ter anal­y­sis, mul­ti­var­i­ate anal­y­sis. The objec­tives of ­our ­research ­are a man­age­able assess­ment ­tool ­where ­only pos­i­tive­ly test­ed infor­ma­tion ­will be ­retained ­and a clas­sifi­ca­tion ­system ­which ­could be con­sid­ered an intel­li­gible ­and equi­ta­ble “­iso-­resource” clas­sifi­ca­tion ­system.

top of page