Age and Ageng I997;26-S2: 19-25 Development of the nursng home Resdent Assessment Instrument n the USA CATHERINE HAWES, JOHN N. MORRIS 1, CHARLES D. PHILLIPS 2, BRANT E. FRIES 3, KATHERINE MURPHY 1, VINCENT MOR 4 Research Trangle Insttute, Research Trangle Park, NC, USA 1 Hebrew Rehabltaton Center for the Aged, Boston, MA, USA 2 Myers Research Insttute, Menorah Park Center for Agng, Beachwood, OH, USA Insttute of Gerontology, Unversty of Mchgan and Veterans Admnstraton Medcal Center, Ann Arbor, Ml, USA 4 Brown Unversty, Center for Gerontology and Health Care Research, Provdence, Rl, USA Address correspondence to: C. Hawes, Research Trangle Insttute, 8450 Whsperng Pnes, velty, OH 44072, USA. Fax: (+1) 440 338 1624. E-mal: MCH@rt.org Abstract Background: the nursng home Resdent Assessment Instrument (RAI) ncludes a set of core assessment tems, known as the Mnmum Data Set (MDS), for assessment and care screenng and more detaled Resdent Assessment Protocols n 18 areas that represent common problem areas or rsk factors for nursng home resdents. Its prmary use s clncal, to assess resdents on admsson to the nursng home, at least annually thereafter and on any sgnfcant change n status and to develop ndvdualzed, restoratve plans of care. Am: to descrbe the content and development of the RAI, ncludng US testng for MDS tem relablty and valdty of the RAI, and the results of a 4-year evaluaton of the effects of ts clncal use. Conclusons: the evaluaton found that mplementaton of the RAI was assocated wth sgnfcant mprovements n a varety of measures of process qualty, resdent functonal outcomes and reduced hosptalzaton. Other uses of the RAI data n the USA ncludng payment usng resdent classfcaton systems and, wth RAI-based outcomeorented qualty ndcators, qualty assurance actvtes and the status of RAI use n other countres are also summarzed. Keywords: assessment nursng homes, Mnmum Data Set, Resdent Assessment Instrument Introducton Ths paper descrbes the content and development of the nursng home Resdent Assessment Instrument (RAI), ncludng the testng for relablty and valdty n the USA. It also summarzes the effects of ts clncal use to assess nursng home resdents and develop ndvdualzed care plans. In addton, t mentons other uses of the RAI data n the USA, such as for payment usng resdent classfcaton systems and to generate outcome-orented qualty ndcators. Fnally, t brefly summarzes the process by whch ths nnovaton has been dffused across other natons and the status of RAI use n other countres. Geness of the RAI The RAI was part of a set of reforms enacted by the Unted States Congress n the Omnbus Budget Reconclaton Act of 1987 (OBRA-87). The RAI and other OBRA-87 provsons were the most sweepng reforms to how nursng homes were regulated snce the onset of federal payment for nursng home care wth the passage of Medcare and Medcad programmes n the md-1960s. Ther enactment was generated by recognton of changng standards of clncal care n the ndustry, concerns about contnung problems n nursng home qualty and wdespread recognton that exstng federal and state regulatory systems were neffectve. In an effort to address these problems, n 1983 Congress asked the Natonal Academy of Scences and ts Insttute of Medcne to examne nursng home qualty and report on how to mprove nursng home regulaton. The Insttute of Medcne formed a commttee of experts and after a 2.5-year study and a seres of hearngs, the commttee ssued ts report [1, 2]. 19
C. Hawes et al. One of ts central recommendatons was the development of a unform, comprehensve resdent assessment system. The Insttute of Medcne commttee argued that a unform, comprehensve assessment of each resdent was essental to mprovng the qualty of care n the naton's nursng homes [1]. The commttee vewed comprehensve functonal assessment as the cornerstone of ndvdualzed care plannng that would focus on helpng each resdent attan and mantan ther maxmum practcable functonng and well-beng. In addton, the commttee argued that the resdent-level data from such assessments were essental to the development of outcome-orented measures of qualty and the mplementaton of resdent-focused qualty assurance systems. Congress enacted most of the commttee's recommendatons, ncludng resdent assessment, as part of OBRA-87. It gave authorty to develop the resdent assessment nstrument and regulatons governng ts use to the federal agency responsble for settng nursng home standards, the Health Care Fnancng Admnstraton. The RAI was developed by a research consortum under contract wth that agency [3]. (The contract was between the Health Standards Qualty Bureau, Health Care Fnancng Admnstraton and Research Trangle Insttute n rth Carolna. The Insttute's collaborators ncluded the Hebrew Rehabltaton Center for Aged n Boston, MA, the Center for Gerontology and Health Care Research at Brown Unversty, Provdence, RI and the Insttute of Gerontology at the Unversty of Mchgan at Ann Arbor.) The new assessment regulatons, orgnally slated for mplementaton n October 1990, were fully mplemented n Sprng 1991 and now apply to more than 90% of all nursng homes n the USA. They requre use of the RAI when the resdent s frst admtted to a nursng home and at least annually thereafter to assess the resdent and develop the resdent's plan of care. Process of RAI development The RAI development team began work n October 1988. To ad t, the project team establshed 18 clncal work groups. These work groups ncluded geratrcans, gero-psychatrsts, nurses, socal workers, detcans, physcal, occupatonal and speech therapsts, recreatonal therapsts, dentsts, nursng home operators, resdent advocates and researchers. Our frst task was to artculate the goals that would gude development of the RAI and be responsve to the congressonal mandate for a unform, comprehensve functonal assessment. Thus, we wanted an nstrument whose man use was clncal to focus attenton on a vew of the 'whole' person, to encourage restoratve and rehabltatve care and to gude care plans. Further, the RAI was ntended to facltate communcaton and problem-solvng among a mult-dscplnary team of caregvers (e.g. nurse, physcan, socal worker, therapst, detcan) by creatng a common 'language' and understandng of the resdent. We also recognzed that the RAI should be feasble for use n the average nursng home, whch often lacked access to geratrcans, mental health professonals and sometmes even lcensed therapsts. Fnally, the RAI tems had to be relable across users (nter-rater relablty). Achevng these goals nvolved several major actvtes, ncludng: () revewng exstng nstruments; () determnng the domans to be ncluded and the tems, defntons and response categores to be ncluded n each doman; () establshng the relablty and valdty of the nstrument; and (v) developng tranng materals to accompany the assessment nstrument. As shown n Table 1, ths was an teratve and nteractve process. Revewng exstng nstruments The project team started by revewng more than 80 Table I. Content of the Resdent Assessment Instrument: components and major domans or areas Mnmum Data Set for resdent assessment and care screenng Background and customary Cogntve patterns routnes Vson patterns Communcaton/hearng Contnence patterns Actvty pursut patterns Physcal functonng and Health condtons structural problems Oral/dental status Mood and behavour Medcaton use patterns Dsease dagnoses Oral/nutrtonal status Skn condton Specal treatments and procedures Resdent Assessment Protocols Delrum Cogntve loss/dementa Vsual functon Communcaton ADL functonal/ Urnary ncontnence rehabltatve potental and ndwellng Psychosocal well-beng catheter Behavour problem Mood state Falls Actvtes Feedng tubes Nutrtonal status Dental care Dehydraton/flud Psychotropc drug use mantenance Pressure ulcers Physcal restrants ADL, actvtes of daly lvng. 20
Development of the RAI n the USA exstng geratrc assessment nstruments. Through ths process, we dentfed the domans or areas whch others felt were essental for assessment and care plannng n nursng homes. In addton, we found that for the most part these nstruments recognzed the realty of good geratrc care for a populaton wth multple chronc llnesses and dsabltes. Thus, the domans and tems that were ncluded led them to focus less on dseases and dagnoses and more on the functonal consequences. Whle these assessment nstruments were llumnatng, none completely addressed the range of functonal areas or provded the comprehensve and 'holstc' vew of the resdent that the project team felt was essental. In partcular, many neglected to focus on the resdent's strengths and preferences, although understandng these s essental to developng an ndvdualzed care plan that focuses on maxmzng ndependence. In addton, many exstng nstruments were 'professon-specfc', developed for use by physcans, nurses or rehabltaton specalsts. Also, several key areas that affect functonng, such as mood and psychosocal functonng, were often omtted or nadequately addressed n exstng nstruments used n nursng homes. Further, many used response categores that were too general to provde the nformaton needed to develop ndvdualzed rehabltatve or restoratve care plans. Fnally, some otherwse excellent nstruments or scales were nadequate for use wth a nursng home populaton because they were not structured to assess the large number of resdents who have sgnfcant levels of cogntve mparment and are unable to report on ther own condton. Thus, the project team used the revew to help nform the selecton of key domans, but t was clear that a new nstrument was needed to meet the goals of comprehensveness, unformty and relablty. Content of the RAI To acheve the goals set by Congress, the Health Care Fnancng Admnstraton and the project team, we concluded that three thngs were needed [3, 4]. The frst was a core set of assessment tems that would provde a comprehensve pcture of each resdent's functonal status, ncludng the resdent's strengths, preferences and needs. Ths s known as the Mnmum Data Set for resdent assessment and care screenng (MDS). The second element was a set of specalzed assessment protocols that are ntended to more drectly lnk the MDS nformaton to care plan decsons. These are 18 condton-focused Resdent Assessment Protocols (RAPs) whch specfy an addtonal, hghly focused assessment f the resdent's status, as revealed by the MDS, suggests a problem, rsk for development of a problem, or potental for mproved functon. Such condtons are dentfed by applyng a set of algorthms specfed n each RAP area to a resdent's MDS data. Faclty staff then use the more specalzed assessment gudelnes found n the RAPs to dentfy potentally treatable causes and focus decsons about the resdent's plan of care and servces. (Table 1 summarzes the MDS domans and the 18 RAP areas.) The thrd element of the RAI s a user's manual wth detaled specfcatons about how to complete the MDS and RAP assessment process (e.g. ntervewng staff, resdents and famly members, revewng records), tem defntons, examples of codng optons and clncal gudelnes for usng the RAPs to develop care plans [5]. In addton, the RAI ncludes a quarterly revew that specfes a subset of MDS assessment tems and s ntended to montor the resdent's response to the care plan and determne whether suffcent change has occurred to trgger a more comprehensve assessment. Development and testng of the RAI We determned the domans to be ncluded, usng the areas specfed by Congress, those reflected n other comprehensve nstruments and those dentfed by expert clncans as key to the functonal wellbeng of nursng home resdents. The clncal/research work groups then specfed the tems, defntons and response categores that were essental n each doman. After creatng the frst drafts, whch were nternally revewed, the project team dentfed addtonal staff from nursng homes around the country and other well-known clncans and researchers who were asked to revew and comment on varous drafts. Ths became an teratve process, wth revews leadng to further revsons by the clncal workgroups and revews of the new drafts. Before the frst feld test of the MDS, 27 drafts and revsons were completed, based on revews by hundreds of clncal experts and nursng home provders across the country. Durng the process of testng and retestng, elements of the RAJ went through an addtonal 15 revsons wth clncal revews. In addton to clncal revews, we tested two versons of the MDS and RAPs n a total of 28 nursng homes n sx states, usng faclty and research nurses, wth dual assessments of more than 600 resdents. These tests, ncludng debrefng of faclty staff who tested the RAI, were used to establsh the face valdty and nter-rater relablty of the RAI tems and assessment protocols, to lmt tems to those consdered essental to care plannng by faclty provders and expert clncans, and to mprove the tranng manual. The results of these tests and relablty of thefnalrai/ MDS tems are reported elsewhere [3, 6-8]. Fgure 1 summarzes ths development process. Subsequent testng has establshed the valdty of the MDS through creaton of scales of MDS tems (.e., the cogntve performance scale and psychosocal wellbeng scale) and comparson of these to exstng 21
C. Hawes et al. Determnaton of Goals and Crtera 4 Revew of Exstng Instruments and Determnaton of Assessment Domans Development of the MDS and User's Manual I Revew by Clncans/Provders Revsons to the MDS (27 Revsons) and User's Manual Feld Testng n Facltes (n rth Carolna and Massachusetts) Revson to the MDS and Development of the RAPs Revew by Clncans/Provders Revsons to the MDS (> 20), RAPs and User's Manual I Feld Testng n Facltes (n Oho, Mnnesota) Revsons to some RAPs Feld Testng of Revsed RAPs (n Tennessee and Connectcut) I Fnal Specfcaton of the RAI Fgure I. Process of Resdent Assessment Instrument development. 'gold-standard' nstruments, such as the Folsten Mn- Mental State Examnaton, that address varous domans n the RAI [9, 10]. Fnally, as summarzed below, a 4-year evaluaton of the effects of the RAI further establshed ts clncal valdty and utlty n mprovng qualty of care for elderly people [11]. Use n the USA Facltes must use the RAI to assess resdents upon admsson n order to develop ther plan of care. The RAI s also used to assess resdents annually after admsson and upon any sgnfcant change n ther health status. Quarterly assessments are performed to montor the effects of care and the need for modfcatons to the care plan. In addton, n about one-quarter of the states, data from the RAI are used to dentfy resdents wth complex medcal or rehabltatve care needs, or hgh dsablty n the actvtes of daly lvng (ADLs). These data are then used to adjust facltes' rembursement by the Medcad programme. Most of these states are usng what s known as a Resource Utlzaton Group (the RUGS-III system) based resdent classfcaton system to determne the nursng home's payment [12] and the current verson of the RAI contans all the tems needed for the RUGs-III system. RAI data are also beng used to determne elgblty of nursng home resdents for State and federal payments for nursng home care. Fnally, the RAI data are beng used to develop qualty ndcators, based on resdent outcomes and key process qualty varables on the MDS. These qualty ndcators are beng used to mprove the survey process whch s part of the federal qualty assurance system [13]. Effects of the RAI on qualty The Health Care Fnancng Admnstraton commssoned a comprehensve evaluaton of the mpact of the RAI on facltes and the qualty of resdent care. The RAI evaluaton desgn was a pre/post-test desgn that examned changes n the process of care and longtudnal resdent outcomes as measures of qualty. The major fndngs, summarzed below, can be found n a fnal report to the Health Care Fnancng Admnstraton [11] and n a seres of papers [14-16]. The evaluaton was conducted n 10 states n 269 randomly selected nursng homes. It nvolved assessment of the qualty of care and resdent status for more than 4000 resdents. Comparsons of process qualty and resdent outcomes were performed between a pre-rai perod (1990 and early 1991) and a post-rai-mplementaton perod n the Sprng and Autumn of 1993- The major fndngs nclude such statstcally sgnfcant mprovements as: 1. Increase n the comprehensveness and accuracy of the nformaton avalable n resdents' medcal records; 2. Increase n the comprehensveness of care plannng, wth care plans n the post-rai perod addressng a greater percentage of resdents' health problems, rsk factors and ther potental for mproved functon; 3. Improvements n a wde array of other care processes that affect resdents' qualty of care and qualty of lfe, ncludng ncreased nvolvement of famles and resdents n care plannng, ncreased use of advance drectves, ncreased use of behavour management programmes, ncreased nvolvement n actvtes and decreased use of problematc nterventons, such as ndwellng urnary catheters and physcal restrants; 4. Sgnfcant reductons n declne among resdents n such areas as physcal functonng n ADLs, cogntve status and urnary contnence; and 5. A sgnfcant reducton n the number of nursng home resdents who were hosptalzed, wth no ncrease n mortalty. Dscusson Strengths of the RAI Clncans have long recognzed that comprehensve functonal assessment of elderly subjects s central to 22
Development of the RAI n the USA maxmzng ther physcal and cogntve functonng and ther qualty of lfe. Indeed, studes have establshed ts utlty n mprovng qualty, reducng unnecessary nursng home placement and mprovng the functonal well-beng of elders n hosptal, ambulatory and clnc settngs [17-21]. Wth the RAI, we fnd that ths process acheves smlar results n long-term care facltes. The evaluaton of the RAI demonstrated ts utlty n mprovng the qualty of care n US nursng homes. Further, we learned more about why t s effectve from the perspectve of long-term care admnstrators and drectors of nursng. In two studes, the vast majorty of nursng home admnstrators and drectors of nursng or unt nurses who were surveyed reported that t mproved assessment and care plannng n ther facltes, mproved ther ablty to dentfy problems and mproved ther ablty to determne whether care plans were achevng desred effects [22-25]. Further, t enhanced clncans' knowledge and ablty to ntervene effectvely n such areas as dentfcaton of delrum, dehydraton and mood problems areas that were typcally under-dagnosed and under-treated pror to the mplementaton of the RAI. Smlarly, another survey of nursng home staff n sx states found that staff vewed the MDS as a useful tool for clncans [26]. At the same tme that admnstrators and nursng home staff were generally postve about the mpact of the RAI on qualty, about one-thrd of staff felt t dd not mprove qualty. Moreover, between half and twothrds of the staff felt that completng the RAI added to ther paper-work burden, ncludng even some of those who felt t mproved qualty of care. In addton, a recent study ndcates that some facltes are experencng sgnfcant dffcultes usng the RAI approprately, resultng n naccurate MDS data and poor use of the RAPs [27]. Thus, ensurng effectve use of the RAI to mprove qualty of care and lfe for resdents unformly across facltes and mplementng practces that mnmze the burden on facltes requres addtonal work. At the same tme, a summary of nursng home staff's vews on why the RAI has proved clncally useful (Table 2) s nstructve. Fnally, we fnd that because of ts relablty, the RAI/MDS data can be used for a varety of purposes n addton to ts orgnal and prmary goal of mprovng care. Facltes n the USA use t to determne staffng needs and for contnuous qualty mprovement ntatves. Staff use t to support ther clncal decsonmakng. Famles and long-term care ombudsmen use t to help them understand and evaluate the care beng provded to elders. Polcy-makers use t to set nursng home payment rates that encourage access for heavy care resdents and acheve greater equty across provders. State and federal nspecton agences use t to evaluate facltes' performance, usng resdent outcomes and key process qualty ndcators whch allow them to more effectvely target ther qualty assurance Table 2. Nursng home staff's vews on why the Resdent Assessment Instrument (RAI) has proved useful The RAI provdes a holstc vew of the resdent as a person, not merely a lst of nursng care needs, by ncludng a focus on strengths, preferences and customary routnes The Mnmum Data Set provdes a comprehensve summary across the major functonal domans, so that, for example, the resdent's cogntve status can be consdered n relaton to hs or her actvtes of daly lvng status, when decdng on the approprate rehabltaton or restoratve care plan nterventon The use of standardzed defntons and response categores provdes a common language across dscplnes whch facltates mult-dscplnary team assessment and care plannng The Resdent Assessment Protocols provde a systematc way to lnk assessment nformaton to care plan decsons The focus of the RAI assessment s on restorng and mantanng functon, so t not only dentfes current problems but also the rsk for the development of new problems and dentfes the potental for mproved functon. Ths facltates more aggressve and affrmatve care plan nterventons actvtes. Increasngly, researchers are also usng t to examne the effects of varous clncal nterventons and to study polcy-relevant ssues, such as the relatonshp between cost and qualty. The nternatonal communty and the RAI The nternatonal long-term care communty has also found multple uses for the RAI. Startng n 1990, geratrcans and researchers n other countres began expressng nterest n the RAI and ts development n the USA. Geratrc assessment has a long hstory n European countres; thus, a systematc way of brngng ths process to the care of elders n nursng homes was appealng to many of these geratrcans who learned of the RAI at professonal meetngs and from colleagues. Other researchers were drawn to the RUGs system as a means of payng nursng facltes approprately for 'heavy care' resdents and dscovered the RAI as part of the process of nvestgatng use of RUGs. Table 3 presents a summary of the man uses of the RAI n countres that partcpate n nterraj, an organzaton of geratrcans, academc researchers and other professonals commtted to developng and usng standardzed assessment nstruments to mprove care for elders. In each of these countres, an nterral member has taken lead responsblty for translatng the MDS, the tranng or user's manual and, n most countres, the RAPs. (Ths has also ncluded a process of reverse translaton' and comparsons between the 23
C. Hawes et al. Table 3. Internatonal use of the Resdent Assessment Instrument (RAI) Country Sweden Italy UK Denmark The Netherlands Japan Iceland Swtzerland France Span Canada Germany rway Fnland Czech Republc Language Swedsh Italan Englsh Dansh Dutch Japanese Icelandc Swss French Swss German French Spansh Englsh German rwegan Fnnsh Czech Relablty testng Man use Intally, to see whether use of RUGs could reduce hosptal length of stay and pay nursng homes more equtably; now, for research To tran nurses n gerontologcal nursng; also mplemented testng effects of full RAI use on qualty of care n two regons and n nursng homes n regons, wth possblty of natonal mplementaton Testng effect on qualty of care and beng evaluated as potental natonal standard assessment for contnung care homes Implementaton of the full RAI for use n nursng homes n Copenhagen; full country mplementaton under consderaton f approprate software becomes avalable Testng effect of use of the full RAI on qualty Testng effect of RAI to provde qualty-based ncentve payments Full country mplementaton of the RAI for assessment and care plannng Testng mpact on qualty n several cantons Prospectve study n a few facltes to determne acceptablty and effect on qualty; testng expanded to more facltes To measure dfferences n facltes' case mx, wth consderaton of usng RUGs for nursng faclty payment Testng utlty of RUGs to dentfy case mx dfferences among facltes; mplemented for chronc care hosptals n Ontaro and beng consdered n Saskatchewan Beng advocated for use n determnng elgblty for the new long-term care beneft under German socal securty law May replace locally developed assessment system for long-term care RUG-III valdaton study completed and evaluaton of MDS/RAI underway Results of a MDS/RAI study beng used to nform polcy development n long-term care MDS, mnmum data set for resdent assessment and care screenng; RUGs, resource utlzaton groups. orgnal and twce-translated versons to reduce ambgutes or ncompatblty across the versons.) Thus, versons of the RAI exst n 14 languages (most recently Chnese) and t s beng actvely used n several natons. Based on the experence of the European and Japanese nvestgators, the followng observatons can be made: 1. The RAI has been found to be relable n several trals n several countres [28] and appears to be a transportable nstrument, rrespectve of dfferences n local cultural habts and nsttutonal settngs; 2. The RAI s very well-accepted by care professonals n Europe who have been traned n ts use and actually used t; 3. The RAI-assocated resdent classfcaton system (RUGs-III) s also transportable, leadng to satsfactory workload estmates across facltes and countres and ther assocated payment systems; and 24
Development of the RAI n the USA 4. The RAI offers a powerful common language across professonals of dfferent dscplnes, across nsttutons and across regons and countres. These fndngs suggest the RAI can facltate crossnatonal research and help dentfy potental for mprovements n qualty of care and greater system equty and effcency. Wdespread use of the RAI can also contrbute to the development of vald process and outcome qualty measures that wll enhance the move to evdence-based health care. References 1. Insttute of Medcne. Improvng the Qualty of Care n Nursng Homes. Washngton, DC: Natonal Academy of Scences Press, 1986. 2. Hawes. C. The Insttute of Medcne study: mprovng qualty of care n nursng homes. In: Katz P, Kane RL, Mezey M, eds. Advances n Long-term Care. New York: Sprnger, 1990. 3. Morrs J, Hawes C, Fres B, Phllps C, Mor V, Katz S. Desgnng the natonal resdent assessment system for nursng homes. Gerontologst 1990; 30: 293-307. 4. Mor V Morrs JR, Hawes C, Fres B, Phllps C. The mnmum data set. In: Maddox G, ed. The Encyclopeda of Agng. New York: Sprnger Publshng Co., 1995; pp. 639-42. 5. Morrs JR, Hawes C, Murphy K et al. Resdent Assessment Instrument Tranng Manual and Resource Gude. Natck, MA: Elot Press, 1991. 6. Hawes C, Morrs JR, Phllps CD, Mor V, Fres B. Relablty estmates for the Mnmum Data Set for nursng home resdent assessment and care screenng (MDS). Gerontologst 1995; 35: 172-8. 7. Hawes C, Morrs J, Phllps C et al. Report on the Small Scale Tral of the Mnmum Data Set for Resdent Assessment and Care Screenng. Research Trangle Park, NC: Research Trangle Insttute, 1989. 8. Hawes C, Phllps CD, Morrs JR, Mor V, Fres B. Relablty and Valdty of the Nursng Home Resdent Assessment Instrument (RAI): report on the feld testng of the RAI. Research Trangle Park, NC: Research Trangle Insttute, 1991. 9. Morrs J, Fres B, Mehr D, et al. MDS cogntve performance scale. J Gerontol Med Sc 1994; 49: 4 M174-82. 10. Mor V Branco K, Fleshman J et al. The structure of socal engagement among nursng home resdents. J Gerontol Psychol Sc 1995; 50: 1, PI-8. 11. Phllps CD, Hawes C, Morrs J, Mor V, Fres B. Effects of the RAI on Qualty of Care: Executve Summary. Research Trangle Park, NC: Research Trangle Insttute, Program on Agng and Long-Term Care, 1996. 12. Fres BE, Schneder D, Foley WJ et al. Refnng a case-mx measure for nursng homes: Resource Utlzaton Groups (RUG-IID- Med Care 1994; 32: 668-85. 13. Zmmerman DR, Karon SI, Arlng G et al. Development and testng of nursng home qualty ndcators. Health Care Fnancng Rev 1995; 16: 107-29. 14. Mor V, Intrator O, Hrs J et al. Impact of the MDS on changes n nursng home dscharge rates and destnatons. J Am Geratr Soc 45: 1002-10. 15. Phllps CD, Morrs JN, Hawes C et al. The mpact of the RAI on ADLs, contnence, communcaton, cognton and psychosocal well-beng. J Am Geratr Soc 45: 986-93. 16. Hawes C, Phllps C, Morrs J et al. The mpact of the RAI on ndcators of process qualty n nursng homes. J Am Geratr Soc 1997; 45: 977-85. 17. Rubensten LZ, Rhee L, Kane RL. The role of geratrc assessment unts n carng for the elderly: an analytc revew. J Gerontol 1982; 37: 513-21. 18. Rubensten LZ, Josephson K, Weland GD, et al. Effectveness of a geratrc evaluaton unt. N Engl J Med 1984; 311: 1664-70. 19. Applegate WB, Akns D, Vanderzwagg R, Thon K, Baker MG. A geratrc rehabltaton and assessment unt n a communty hosptal. J Am Geratr Soc 1983; 31: 206-10. 20. Tulloch AJ, Moore V. A randomzed controlled tral of geratrc screenng and survellance n general practce. J Roy CoU Gen Pract 1979; 29: 733-42. 21. Hendrcksen C, Lund E, Stromgard S. Consequences of assessment and nterventon among elderly people: A threeyear randomzed controlled tral. Br Med J 1984; 289: 1522-4. 22. Hnes M, Mor V, Hawes C, Phllps CD, Morrs J, Fres B. Development of Resdent Assessment System and Data Base for Nursng Home Resdents: Post-Implementaton Telephone Survey Report. Provdence, RI: Center for Gerontology and Health Care Research, Brown Unversty, 1994. 23. Mor V, Phllps CD, Morrs JR, Hawes C, Fres B. Evaluaton of Resdent Assessment System for Nursng Home Resdents: Results of the Baselne Survey of Admnstrators and Drectors of Nursng. Provdence, RI: Center for Gerontology and Health Care Research, Brown Unversty, 1991. 24. Phllps CD, Mor V, Hawes C, Fres B, Morrs J. Development of Resdent Assessment System and Data Base for Nursng Home Resdents: mplementaton report. Research Trangle Park, NC: Research Trangle Insttute, 1993. 25. Lombardo N, Morrs JN, Sherwood S, Levne DA, Morrs SA, Bellevlle-Taylor P. Mental Health Servces for Nursng Home Resdents. A fnal report contract from Amercan Assocaton of Retred Persons, Washngton, DC to Hebrew Rehabltaton Center for Aged, Boston, MA, 1992. 26. Marek KD, Rantz MJ, Fagn CM, Krejc JW. OBRA '87: has t resulted n better qualty of care? J Gerontol Nursng 1996 (October): 28-36. 27. Harrngton C. Faclty Implementaton of the RAI: a workng paper. San Francsco, CA: School of Nursng, Unversty of Calforna at San Francsco, 1996. 28. Sgadar A, Morrs JN, Fres BE et al. Efforts to establsh the relablty of the RAI. Age Ageng 1997: 26 (suppl. 2); 27-30. 25