I... ' ~. .. '( 'r":'":,:''' ::'~"'IIH

.'.. ,;,-,.,(l/,I!,..I:."":''''''','h'''''''IJ:''~fI.1

/ :,I1.,_..~'.."I.':,..!:hl\'..! ;l\ """'. I IIII-II~

Within the field oi aging, the conceptualization and measurement of iunctioning has beendominated by the disability model. In this paper, one limitation of that model is described bycalling attention to a distinClion oetween three "tenses" of functioning. Inadequate attention

has heen paid to the distinction between the capacity 10 function in Ihe abstract (hypothetical(ense) and actual periormance in daily life (enacted tense). Failure to attend to this distinction

has obscured considerable discordance between what people say they arc able to do instandard functional disabilily .1ssessmenls, and what they actually do al home. To illustrate

this point, dala from the MacArthur Studies of Successful Aging comparing the hypotheticalto (he en.lcted tenses are presented. These data show a consistent pattern of discordance

between these two tenses.Key Words: Functional status, Theory, Successful aging

Conjugating the "Tenses" of Function:Discordance Among Hypothetical,Experimental, and Enacted Functionin Older Adults

Substantial discordance can be observed be-tween the level of functioning older persons are ca-pable of (based on self-reports), and how activelyand independently .they function in the "real"world. Recognition of this discordance has givenrise to the distinction between pathology, impair-ment, disability, and handicap featured in the WorldHealth Organization's taxonomy of functioning, theInternational Classification of Impairments, Disabil-ity, and Handicap (ICIDH) (World Health Organ-ization, 1980).This article argues thatthp WJ-!()nie:_tinction has "not gone far enough in sensitizing~rontologists to the complp1Citynf late life func-tioning. This has resulted in a gap in conceptualiza-tion and research due to the failure to differentiateamong three "tenses" of functioning (hypothetical,experimental, and enacted). This article presents amodel to elucidate the distinctions among thesetenses in an effort to fill this gap. Empiricalevidenceof the existence of this phenomenon will also beprovided through a report on a data set which asksquestions of high-functioning elderly persons fromthe perspective of more than one tense.

The Dominance of the Disability Model

For the past several decades, the predominant

'The author would like [0 [hank Drs. Rosalie A. Kane. Lisa F. Berkman.. and Carlos Mendes de Leon for invaluable comments on an earlier drait.

This work was ~upported by [he Brookdale Foundation through the Brook.dale Fellowship Program. Send all correspondence to Dr. Thomas Glass,Harvard School of Public Health. Department of Health and SQciala+!havior. ;;"7 Hl.;nilngton .~ve.. BIJi~on. .\1A 1)2115. E-mail: tglass@hsph.narvard.edu

Vol. 38, No.1, 1998

Thomas A. Glass, PhO'

conceptual scheme for thinking about functioninhas been the disability model (Branch, Katz, Kniep-mann, & Papsidero, 1984; Clark, Stump, & Wolinsky,1997; Hubert, Bloch, & Fries, 1993; jette, 1987; jette,1994; jette & Branch, 1985; Katz, Downs, Cash, &Grotz, 1970; Katz, Ford, Moskowitz, jackson, & jaffe,1963; Lawrence & Jette, 1996; Nagi, 1965; Pinsky,Branch. jette, Haynes, Feinleib, Cornoni-Huntley, &Bailey, 1985; Verbrugge & jette, 1994; Verbrugge,Lepkowski, & Konkol, 1991). An indication of the im-portance of this model is the fact that the term.ills.:!.abili has become nearly s non mous with th n-c~pt ot functioning or unctional status (McDowell& Newell, 1987). The disability mOdel grew out of thework of Nagi in sociology (Nagi, 1965; Nagi, 1976;Nagi, 1991) and Katz and others in medicine (Katz etaI., 1970; Katz et aI., 1963) as an attempt to movebeyond the narrow focus on the presence or absenceof disease. Katz developed the activities of daily liv-ing (ADl) index to measure both the effectiveness oftreatments and the underlying mortality risk "in frail"elderly populations. That the ADL index would be-come synonymous with functioning was foretold inthe subtitle of Katz's first article, "The index of ADL:A standardized measure of biological and psvchoso-cial function" (Katz et aI., 1963). Today more thanfifty ADL scales have been developed and most, in- £C)c1udlng the ~ and the Barthel Indices, werespecifically intended to measure functioning amonginstitutionalized patients (Mahoney & Barthel, 1965).

The disability model, as exemplified by the IClDH,is pictured as a linear pathway ieading from diseaseto handicap (in Nagi's framework) {see Appendix,

101

DISEASE: IMPAIRMENT:

InstrinsicDart/otonv ordisorder.

Any ~ULabnormalirv otpsychological.physiological oranatomicalstructure ortunction.

DISABILITY: HANDICAP:

Any resrricrion orlack of abiliN toperform anactivity in themanner or in tnerange considerednormal tor ahuman being.

A disadvanraoeresulting trom animpairment or adisability that limitsor prevents thetullullment at anomal role(depending on age.sex. and social. andcultural tactors) forindividuals.

FigurE:'1. \.\'orld Health Organiz.1tion Typolog\' oi He.lllh .1nd FunctIoning: Int~rnational Classiiication oi Impairment. Disability. andH.1nclic.1p',ICIDH).

Reuben, Valle, Hays, & Siu, 1995;Tinetti, 1986). Thesemeasures have sought to minimize the difficultiesarising from the use of self-reports, and are anattempt to move beyond the assessment of the hy-pothetical tense to which most standard self-reported ADl scales are limited. Yet, while the shiftto performance-based assessments has beert an im-portant extension into the experimental fense, thishas not lead to a generalized interest in perfor-mance outside the laboratory.

Still others have called for a broadening of our

concepts of function toward social functioning Q(Bortz, 1989; Jette & Branch, 1981), Verbrugge and ..I .Jette (1994) have recently called for the inclusion of ~~lilg er- eve activi les In u In so la I re 4,,J'ac IVlty,tnps. c I care, and recreation. Other inves- ~;f:tlgators have begun to examine productive activity~ ~as one dimension of higher-level functioning (tSutler& Gleason, 1985; Glass, Seeman, Herzog,-.Kahn, &Berkman, 1995; Herzog & House, 1991; Kahn, 1986; \.\0"Myers, Manton, & Bacellar, 1986). 'v~'

A second area of criticism of the disability model !has been that the emphasis on capabilities and diffi-tt\ .cultieshas the unintendedeffectof removingtFiein- .

dlvldual from the social and physicalcontext in whichfunctioning actually takes place. In this way, the mea-surement of functional status has resulted in a de-contextualized view of functioning in older adults. Asa result. there has been a tendency to de-emphasizecompensatory strategies, work-arounds, and the vari-ous coping strategies employed by persons in theirhome settings_Feinstein, Josephy, and Well (1996)re-viewed forty three indices of functional disabilityconcluding that ". . . the most strikin omission in in-dexes of disabilitv is tea sence of attention eco a orative role of the patient. . . . With uncommonexception, the roles of personal effort and externalsupport are constantly overloo~d . . ." (p.414).

Note 1). The emphasis is on the probabilistic ratherthan deterministic nature of those transitions (Fig-ure 1). The model begins with disease (pathology),which predisposes to impairment (loss or structureor function), which predisposes to disability (re-striction or lack of ability), and finally to handicap(disadvantage). Research which is informed by thedisability model shares the idea that functioningshould be understood in terms of functional capa-bilities in relation to the constraints placed on theindividual by the outside world, In both its conceptsand its measures, this model reflects an emphasison functional ability as a hypothetical construct,rather than actual performance in daily life. As anexample, in a recent test of the disability modelusing data from the Longitudinal Study of Aging,Lawrence and Jette (1996) excluded subjP,t<: WOreported that the did not en a e in ADL activities

ecause "It was not ossi Ie to determme It t vwere a e or (sable." This tendency to emphasizefunctional capability and to ignore functional per-formance reflects the history of the disability modelas grounded in studies of functional status amongolder patients in hospitals and nursing homes, farremoved from the real-world constraints which bothenable and dampen functional performance out-side of the institutional setting.

While the disability model continues to predomi-nate the research and policy arenas, several impor-tant criticisms of this model have emergeQ. Amongffiese concerns is that it remains focused on a.aar-

*"' row band of functioning, too closely linked to phys-ical function and the presence of disease. Numerousresearchers have sought to broaden the definitionof functioning beyond the limited scope of self-care(ADLs). First among the revisions was instrumentalactivities of daily living (IADU, developed in the1970s, which were designed to better capture do-mains of functioning among persons living in thecommunity through the addition of domestic main- .. .tenance tasks (Lawton & Brody 1969. Lawton Moss Conjugating the Tenses of Function:Fulcomer, & Kleban, 1982). " , , Enacted Function Explained .

Motivated primarily by concerns over measure- According to Webster's dictionary, to conjugate ament problems, another group of researchers have verb is to "recite or display all or some subsets ofargued for performance-based measures of func- the inflected forms of a verb in a fixed order." The

(~i~ning (Guralnik, Branch, Cummings, & Curb, 1989; conjugation of a verb describes the different forms

~r1r-~k~~, 102 The Gerontologist

.~

of ,Klion as they relate to the temporal context ofU'<.'. In lingLJistics, distinctions are drawn bet\'I.'ccnP,l';t. present. <1nd future iorms of verbs. For thepurpose of this artic:le. " diiferent type of conjuga-lion is proposed in order to elucidate the meaning{If functioning for the field of aging. For the pur-poses of this disCLIssion, the idea of conjugation isnot used in ils temporal meaning (past. present. andfuture). Rather, the idea of conjugation refers todifferent aspects of what might be termed "actionpotenti;tI." The three tenses described below mayeach exist in past, present, and future forms.

Figure 2 provides a typology of how the conceptof functioning might be conjugated in this manner.Traditional measures of functiona~ status, as a con-sequence 111part at the disability model. are primar-

ily examples of the hypothetical te~. Subjects areasked a range of questions about whether they areable to perform tasks, how difficult those tasksmight be, and whether they require help from peo-ple or from assistive technologies. In each case, ther c;' nswer is the product of the respondent'sabstract caleu u based, presumably, on that sub-ject s a i (tv to average across situations and con-texts. This calculation is most often unmeasured~nobserved.

The second tense might be called the experimen-tal tense in that instead of asking hypothetical ques-tions about presumed ability, respondents areasked to demonstrate functional capability in a clin-ical (or experimental) setting. In the experimentaltense, respondents are asked to simulate activitiesthat are thought to be representative ot true tunc-

,}1r .

Hypothetical

~\o\:;).'. /JCan do..." 1

Measures:Self-report (orproxy}"Are you able to...""Any difficulty ""Do you requireassistance... ."How difficult is itto .Q.O.l. items

(inning. I\S h<1sbeen observed, this approach hasthe adv<1ntage of offering direct observation of thesubject's ability and avoids some of the problems of:'l~lf-reports. However. while this tense may ade-quately describe what a person is able to do underIhe inspection of the clinical examiner, it tells us lit-tle about the realitv of that res ondent's fun tion-ing at home. r e expenmental tense refers, then, tothe demonstrated perform<1nceof what is hypothet-ically possible, albeit subject to the conditions andlimitations of the experimental setting.

A third tense. referred to here as the enactedtense, refers neither to hypothetical ability nortowhatthe individual is able to perform under specialand circumscribed conditions. Instead, enacted func-tion refers to the functional performance reali]:ed - )...rb~'\C>bv an individual wltnin the home context. Enactedfunction, whether physical, cognitive, emotional orsocial, refers to the end result of the confluence ofsituational and ~rs which shape themoment to mOT'fTeoA.tperform.mce of functional tasksin the real world. As such, enacted function is a lower \level of abstraction. It requires no abstract calculus C::--s.~'\ ~-t:of what is possible, but requires instead the accurate ~recollection of what has been or is now.

The central argument of this article ISthat withinthe field of aging, the study of functioning has fo-cused almost exclusively on just one tense. Wehave conceptualized functioning in terms of the in-dividual's capacity in the abstract, rather than whatthat individual is doing now or h~s actually don~in -r,aLLthe past. Researchers have constructed tests which ~~C:assess the capabiiity or difficulty of certain actions, ~

EnactedExpedrnental \ r-......

.~_\~.~) .I NCoulddo...", l NDo do..." ,

Measures:Performance-basedtesting (orobservation)Timed performanceQualitativeassessmentTask completionStandardized testscores

Measures:Self-report, proxy, orobservationFrequency/durationof activity/role/taskProductive outputCalories expendedPhysiologicalmarkersOccupational status

Vol. 38, No.1, 1998

Figure 2. Conceptual scheme for conjugating the tenses of functioning.

103

"

but exclude the quesliun of whether thosc ,,(tions,11L',Ktu"lIy pL.'rformcd, In short, a conccptual Ir"me-wurk h"s emerged within gerontology b<lsed un thehypothetical rather than the enactec!tense oj func-tion, rurther, the pn~domin"nt nWdsurement 10 Isin common US!' in t le field n,'ili'c! and are driven bvthis frJme\\'or\.;, '

The usefulness ollhis "prroJch to conccptualiz-ing iunction is not \0 argue that <1I1Yune perspec-tive is inherently superior to the others. In iact. thevery quality that makes one tense better suited toone application may be its weakness in another. ForE'xample, Guralnik and colleagues (1989) correctly

~s5ert th;!t one advantage ot performance-b~es (experlmenral tense) over ADL scales (hy-pothetical tense) is that the former are less influ-enced by "cognition, culture. language, and educa-tion," However, in another research context. thesesame factors may be of primary interest. If.an inves-tigator wishes to research how factors such as lan-guage, culture. education. cognition. social support.or other host characteristics impact levels of func-tional performance. measures based on either thehypothetical or the experimental tense may be insuf-ficient. All three tenses may represent different andindependent ways in which individuals think aboutfunctioning. If that is true, gerontolo~ists npprl todevelop parallel and complementary measurf'mf'nttools to accommodate those perspectives.

To a large extent. these alternati.ves already exist.Any standard textbook on questionnaire design willinclude a discussion of how question constructionimpacts the resulting data that is generated (thiswas pointed out by an anonymous reviewer of anearlier draft of this paper!. Research on functioning.as well as public policy, however, continues to relyalmost exclusively on the hypothetical tense. Onlyrecently have attempts to further elucidate the ex-perimental tense become widespread. Rarely hasthis led to extensive investigation of the enactedtense. The purpose of this article. moreover. is todraw out the importance of the distinction amongthe three, and to make a case for the importance ofthe.enacted tense.

Emphasis on the hypothetical tense has producedconsiderable progress in the field. The ADL disabil-ity scales were designed to capture gross depar-tu res from optimal functioning for purposes of pre-dicting those at increased risk for nursing homeadmission or death (Applegate. Blass, & Williams,1990). Most of these scales. including Katz, theBarthel index. and the origina :.pULSESs a es weredeveloped for use among h . . ed patients(Wagoer, 1987). In the hospital setting. the hypo-

~etical nature of these instruments is clearly more\ J,appropriate; although as several commentators have

.tV! noted, the emphasis ought to shift to actual perfor-

Y'~ mance of functional tasks as the patient moves fromv~ the hospital to home (Fuhrer. 1987). As a predictor,

rather than as an outcome. the disability indices arewell suited to this task.

Traditional disabilitv scale~ also have advantagesfor puhlic porlcy because they allow for the identifi-

calion oi environmental barriers and impedimentsmore c<lsily.As <Il)<lsislor deciding who is eligiblelor he<llth or social services, it may be best to ex-clude the contingencjp<; oi the home environment,or at leJst to minimize their influence as a way ofensuring equity in resource <llIocation. In this sense.asking subjects ii they are able to perform a task inthe hypothetical allows us to separate. or to hold

\

constant. the complexity which arises from the in-troduction of the social. physical. and coltural con-texts that shape and modify actual performance.

r\ Graphical Representation of Two Tenses

To further illustrate the importance of this distinc-tion; Figure 3 depicts a simple 2-by-2 table whichcontrasts the hypothetical tense (functional capac-ity). as measured by most commonly used ADLscales. with the enacted tense. Enacted functionrefers to the extent and quality of functional perfor-mance realized by an individual within a particularsocio-physical context. Individuals who fall on thediagonal (Types I and IV) may be said. to be concor-dant between the hypothetical and enacted tensesof functioning. For these individuals, ADL scaleswould be valid 2nc ~e!i~b!e me3sures of both func-tional capability and actual performance. The term

-H;

~tll;!1 rprformance here refers to functional perfor- jQc+~ ;mance within a naturalistic social setting, as con- -JJ~trasted to performance achieved"in a clinical setting "tt~or under experimental conditions. !U

Types II and III. on the other hand, represent a isubstantial measurement challenge that has seldom I

been explicitly addressed in aging res~~rch. Type (( iindividuals enact high levels of fun.ctional perfor- !

mance despite significant difficulty. disability. or di-'minished capacity. The reasons for this discordanceare many, and are not well understood. These indi-viduals might be thought of as "over-achievers" inthe sense that within the context of their home envi-ronment. they are functioning ata higher level thanthat which they are presumed to be capable. Type IIIindividuals, on the other hand. actually function at

Functional performance at home(Enacted -tense-)

I

t.

I

I

I

II

I

III

I

Low High

Low

Functional capacity(Hypothetical "tense-)

High

Figure 3. Cross.classification 01 "hypothetical" and "enacted"tenses oi iunctioning.

Type I Type II

Low Over-

functioning achievers

Type III Type!V

Under- Highachievers functioning

. . .

lowcr levels than would be predicted by AO\. scales.Despite the absence of limiting factors such as clis-(~.1se.impairments, or disability, these individuals donut Jc(u.ll1y periorm the tasks <1bout\.vhich they arebeing asked. These "under-achievers" may lead sed-L'nt<1ry,immobile lives. and potentially are quite de-pendent on others. Again, the reasons for this dis-cordance are numerous.

E\..idence from Prel/ious Research

While a general awareness of this issue can beseen throughout the literature on functioning, veryfew investigators have studied the problem directly.This author's own interest in this phenomenon wascatalyzed during qualitative interviews with survivingmembers of the New Haven EPESEcohort. Amongthem was a gentleman in his late 80s who reported ahigh level of ADL impairment, yet spent most of hisdays sailing his boat up and down the Connecticutriver. When asked about how he maintained theboat. he demonstrated how he would use duct-tapeto affix a paint brush to his cane to apply a new coatof paint to his boat every Spring. Equally commonwas the story of a woman in her 70s who reportedthat she was able to do everything for herself. whilemaintaining that all she really did was watch televi-sion all day while her 60 year old son with conges-tive heart failure "waited on her hand-and-foot."During these interviews. it became increasinglyclear that it is an error to assume that functional ca-pacity and performance are the same.

Several commentators have noted the lack ofstudies which investi e the reliabilit and validityof functlona Isability measures an enera c-

owell & Newell. 198 . eyon t ese general criti-cisms. a number of researchers have noted a differ-ence between functional ca acity and Qerformance(Hickey. i980; eere. 19 4; KeHy-Hayes. fette. WOlf.O'Agostino, & Odell. 1992; Patrick, Darby. Green,Horton, Locker, & Wiggins, 1981; Verbrugge & Jette.1994; Ziebland, Fitzpatrick, & Jenkinson. 1993). Forexample, some evidence suggests that those who re-port impaired capacity in one or more domains aresimultaneously quite active and high-functioning inother domains which are assumed to require highdegrees of capacity. In a study of religious participa-tion among members of the New Haven EPESEco-hort, Idler (1987) concludes that ". . . functional dis-ability does not entirely determine behavior: 98 ofthe sample's 358 severely disabled women (who re-ported that they needed assistance in crossing a

I~07r:'~ fe.eding ~h:~;:~v.,es~~a;~i~~: ~~c;)a~~o_r~~~rt,_ .r -"I". _: ;..; ,._-o.~__ _~ _,_: ''-"___ _: ~", ._ at '!\IOoo k ('r ..,...0 '00 O ..::> ~t' nQ,' . -~" n ~ ~,.- .~'1 93).. '__,- __. 'I 1-_ 1"'-.1. ".1 _;, '-"-',:5.-., ~ ,It-,L..:.. ,'J

also found that functional disability is unrelated tofrequency of attending senior centers. Findings fromthe Health Insurance Study (HIS) have shown thatsome subjects who report limitations in physicalactivity report no corresponding limitations in roleactivities (Stewart, Ware, & Brook. 1981). In eachcase, these findings were seen as anomalies. andwere not interpreted as signs of the weakness of amore general model of functioning.

Vol. 38. No.1, 1998

In their influential study oi the (orrcl"lion ;1ll1ong Jf7self.reported. interviewer-administered. and perior. -1'-f;://mance-base? testing of fUl~ction?1abilitie~~ReUb(~n

F-'dh/

et .11.1"1995)round the relatlonshl)$ I mconSIStent an f <;u esting that these in tsare not mea uring the same construct." It is pos- .sible that differences etween self-reports, inter-views. and performance testing reflect more thanmeasurement error. They may also reflect differ.enccs in the meanings oi different questions askedin various ways. Several investigators have shownthat subjects will tend to interpret questions aboutfunctioning in ways that are often different thanwhat researchers intend (Keller. Kovar, lobe, &Branch, 1993), and that the same subjects within thesame disease group, who are given different disabil-ity questionnaires, will evidence substantially differ-ent levels of disability (Ziebland et aI., 1993). Thesediscrepancies may also reflect the inherent unrelia-bility of asking questions in the hypothetical tense(Anderson, Bush, & Berry, 19 also reflect, inpart. differences in the tacit models' (Ziebland et ~aI., 1993) which underlie the construction of disabil-.ity instruments.

The few studies that have examined multiple I

tenses suggest that reduced function in the enacted ~)ur.c:ttr;tense Olav be more prevalent than is observed in i",ICoh~'tne hypothetical tense (Type III discordance). In a 1C"..,.c.{..recent study, Kelly-Hayes et al. (1992) used Nagi's \ _

distinction between functional limitation (capacity) 1:' *~~and disability (closest to enacted) to look at data cvt~:"'~from the Framingham study. Their findings suggest ~substantial discrepancies between the two and, in \most cases in which differences were observed, re- ~~spondents reported a disability (enacted tense) and t.d' ~not a functional limitation (hypothet~al tense). Inaddition. they found that psvchosocial factors weremore strongly associated with disabilities than theywere with limitations. This last finding is importantbecause it suggests that the impact of psychosocialfactors may be greater in determining the enactedrather than hypothetical tense of function. This ob-servation is supported by Mendes de leon et a

D

'

(1996) who studied both the experimental tense(physical performance) and the hypothetical tense(ADL questions) over time and found the relation-ship to be mediated by self-efficacy.

In another study, Patrick et al. (1981) looked atdata from two community surveys of a London bor-ough and found considerable differences across thetwo waves based on what they termed "capacitywording" versus "performance wording" of theirdisability questions. In the study that used the per-formance wording (enacted tense), investigators .

found as much as a two-fold increase in the preva- J 'lence of disability. leading the authors to suggest V-that studies based on capacity wording may sub- ~stantially under-estimate the true prevalence of dis- ~/1

ability in populations. A similar conclusion was ~ .reached by Anderson et al. (1989) who found that

"capacity .wording~' of ~uestions produced an under-reporting of dvsfunct' n ("If::It Ip;:!c;;.J.J)%. ~seu UIIthe studies by Patrick. Kelly-Hayes, Anderson, and

105

.''j

OIIH'r~. it is likely thaI tliscrl'p,1f)cic$ between thetwo It.'nses (J" function Ill.1\' 1)(. 'lIh$lnlial. ancllh,llrL'li.H1Ce I~XclllSi\'t'lv on instrunH~I!ls \\ hid! ,h~(;':-:>only Ihe hvpolheli("'al lense m,lY 1t',1c11(JbiasL'd csti-males or disability rates .1Swcll ,1:-,111undl~n~jlim,l'tion 01' the importance 01 psychosoci,ll fac.tors inlate-life functioning.

Three Examples of Discordance: Data Fromthe lvlacArthur Studies at' Successful Aging

We now turn to an empirical example that illus-trates the importance or what has been argued so far.In order to estimate the magnitude of discordancebetween reported capacity and actual runctional per-formance. we would ideallv choose data that weredrawn from a random Iv c'hosen and community-based study of file elderly. We would hope to ha~e

indicators ot fUnCtiOnal capacity and enacted func-tionin that would be senSitive to higher-level activ-ities as we a ~~ I palrments an Isa I ity inthe physical domains. n .we would hope for mea-sures ot both capacltv and actual performance thatwere directly observed or performance based (onboth countsl rather than sel(-.reports alone. Datathat meet all of these criteria are currently unavail-able. However. for the purpose of illustration. datafrom the MacArthur Studies of Successful Aginghave several important advantages. First. althoughnot intentional. the MacArthur battery asked twosets of questions about several domains of functionin both the hvpotheticaLand enacted tenses. First,traditional functional disability questions wereasked about the extent to which the respondentfeels that they are able to accomplish a task. Then,much later in the interview. under items which arepart of a battery of pro.;:luctive activity questions.subjects were asked whether and how often thosesame tasks were actually performed. These datathen afford the opportunity to explore the extent ofdiscordance between the hypothetical and the en-acted tenses of functioning. It should be stressedthat while these data offer a rich opportunity to esti-mate the magnitude or concordance or discor-dance, the MacArthur study was not designed toinvestigate this question and there are several no-table limitations associated with these data. For ex-ample. both sets of questions are self-reports andare theretore Subject to the problems associatedtherewith.

Aging researchers have begun to study produc-tive activities - or those activities which generategoods and services of economic value - as a markerof "successful" aging (Cutler & Danigelis. 1993;Glass et al.. 1995; Herzog, Antonucci, Jackson, Kahn.& '''\organ. 1987). For the purpose of this study,questions about productive activity are good indica-tors of enacted functioning because they ask thefrequency and duration of routine activities as theywere actually performed. Productivp ;1rtivitv is alsoan ideal indicator of global enrtrtpri functioning be-cause it represents high-level functioning which re-quires performance in the social. physical. and cog-nitive dom;Jins.

,\\cthods

SuIJ;ects

Till;' data used in thesc analyses are from the,\\acArthur Research Network on Successful AginaField Study, a longitudinal "tlldy of hi~h func~tioning men and women aged 70-79. Details of thestudy design are available elsewhere (Berkman,Seeman, Blazer, Kahn, Mohs, Finch, Schneider, Cot-man, McClearn. Nesselroade, Featherman, Gar-mezy, McKhann, Brim. Pragar, & Rowe, 1993). Sub-----jects were drawn from three sites of the EstablishedPopulations lor the Epidemiologic Studies of the El-derly (EPESE):Durham, NC; East Boston, MA; andNew Haven, CT. Men and women aged 70-79 wereselected on the basis of physical and cognitive func-tion at the time of their 1988 EPESEinterview. Age-eligible men and women (N =4,030) were screenedon the basis of six criteria to identify a relatively"high functioning" cohort, representingapproxi-mately the top third of their age peers in terms ofphysical and cognitive function. Of the 1,313 sub-jects who met all criteria, 1,192 (90.8%) agreed toparticipate and provided intormed conse t. Base-me ata co ectlon was comp eted between May

1988 and December 1989 and included a ninetvITiiiiUte, face-to-face interView covering detailed as-sessments of physical and cognitive performance aswell as information regarding health status. A_threeyear follow-up interview was conducted in .1~91.

owever, this study uses only data from the base-line interview.

For comparison purposes; "asmaller subsample ofsubjects was interviewed at baseline (n = 163) whowere chosen to represent the middle and low t~r-tiles of functioning using the same criteria. Thisgroup was chosen to permit comparisons with thehigh functioning subjects in order to determinethose physical and psychosocial factors that dis-criminate "successful" from "usual" patterns ofaging. The medium and low functioning groupswere not re-interviewed at follow-up.

Measures

Features of the data collection and sampling stra-tegies for the EPESEstudies have been well de-scribed regarding sociodemographic characteristics(Cornoni-Huntley, Ostfeld, Taylor, Wallace, Blazer,Berkman, Evans, Kohout, lemke, Scherr, & Korper,1993).Variables used in these analyses include ~~sex, race, and marital status. Age was recorded as of[he date of {he sUbJect's 1988 MacArthur interview,and is dichotomized in these analyses (0 = 70-74,, =75-79). Marital status is also dichotomized (1 ifmarried at baseline, ° if not). .

Our objective in this study was to cross-classifyindicators of functioning across two distinct tenses.To accomplish this goal within the limitations of theMacArthu r data, th ree comparisons were madeusing data from questions both about hypothetical.and enacted tenses. Functional disability scales 10-.

duded in the MacArthur battery include a ~

..

riie.1$lJrCof limitJtiOlls ill activities of daily living asdevelopt~d by 1<,1t~el .11.(1l)70),a measure of grossmobilily based on the work of Nagi (1976), and,\ <;cll-n!fJortcd measure of phvsical periormance(Rosow & Breslc1U, 1%6), .111oi which use weJrding1/1,11relleets the hYPolhetical tcnsr-~nong high-lunctioning respondents. there \'IIa~;ariance on

:::;'"'thl~ KJI4!:.A.DLindex. and only limited vanance onthe .'-Jagi ,md Rosow-Brcsl.llJ sCc11es(See Appendix,:\lole 21.

First. we focused un two item-5from the Rosow-Greslau index which includp.: .1)"Are you able to doheaV\' housework like washin wmdows walls orI oors 'Nit out eIQ?" [coded yes, no, or refused];and 2i "Are you able to walk 1/2 mile without help?"[coded yes, no, and relusedl. For the third compari-son, summary indices of potential .versus enactedtenses were used, including self-reported limita-tions in ability to function in either the Nagi and theRosO\v-Breslau scales and a summary index of pro-ductive activities (enacted tense).

To measure enacted function, or the level offunctional performance as realized by each subjectin their home, 18 items from the productive activi-ties battery developed by the MacArthur SuccessfulAging network in conjunction with colleagues fromthe Americans Changing lives study (Herzog. Kahn,i"~organ, Jackson, & Antonucci. 1989). For our firsttwo comparisons, individual items from the produc-

A. tive activity battery.

which aralleled items from the

~\ 1I~ Rosow- res au ca e were use , mc u m~questions~iP which asked "uo you.currentlv do any heavy house-

w..Q!.k?"and "Do you walk at least 1/2 citv block per

~?" [both it~~ed yes, no, refuses, don'tknow). For the ird c mparison. a summary indexof productive ac I ies developed in an earlierstudy was used (Glass et al.. 1995). This scale com-bines information on five dimensions of productiveactivity chosen to represent the activities commonly

-? L...engaged in by older people which produce eco-( \ nomic goods or services including: 1) housework, 2)

yard work, 3) child care, 4) paid work, and 5) volun-teer work. In the present study, the summary indexwas divided into tertiles for the high-functioninggroup, as well as the medium- to low-fuctioningcomoa.rlson group. j / ~

-~~T("2V'~ ~,." ~1> C<..?

LotOResults

Table 1 shows the socioeconomic and selectedfunctional characteristics of the two in total as wellas by gender. In comparing the high-functioningsample (labeled successful agers in the MacArthurstudy) to the medium- to low-functioning group(usual agers), the high-functioning group, as ex-pected, is younger, less likely to be of minority de-scent and less cognitively impaired. With the ex-ception of a difference in education among women,the two groups are not notably different in educa-tion or income. Among the high-functioning group,women were more likely to be married. In bothgroups. just over half (55%) of the sample is female.

The results of the three comparisons between hy-

Yol. 38, No.1, 1998

Index oi Productive

Activity 6.7. 4.8 (y..'Means presented ior continuous variables, freque:1cies, and ~,

percents for categorical variables. LI' v'T-tests are shown for continuous variables and chi-square ~ II~

~tests ior categorical variables. ~ 4'~ .

\ ~r:tr

~~LTable 2. Discordance Between Hypotheti ~ .and Enacted Functioning:HeavyHousew (1988)' .1i, '?. I

No

Do you ~ny heavy housework? ~No Yes Totals ~~>o>-p

~_..c>r tJ..~73@./84(52%) \..-

M~l~.. .......101.00 .

107

Type I

highs 40%lows 56%Are you able to do

heavy housework?

Yes 110 (94%)78 (48%)

'MacArthur Studies of Succe,\-tedium-Lows =162)..

pathetical and enacted iunction are presented inTables 2, 3, and 4. In Table 2, the abilily to performheavy housework is cross-tabulated with the actualperformance oi heavy housework. Note that thedala Me presented for the hig-fi.1flctiQouP(!iF) and the medium- to low-functionin roupIMLr) eparate y. n y .... ot t e Hf group say theyare not able 10 do heavy housework_ Of that group,

T.1hle I. Comparison (Ii "Successful" nd "Usul" Agerson Demographic Features: Means, Frequencies and Percentages

with Statistical Tests of Difference"'

High Mid-lowfunctioning functioning

\'aridble (N =1191\ (n =1621 Test" Sig.

SexMen 531 (45%) 73 (45%) n.a.Women 661 (55%) 89 (55%) n.a.

Race (non-white! 228 (19%) 55 (34%) 19.0 .000Men 92 (17.3%) 21 (29%1 5.52 .019Women 136 (20.6%1 34 (38%) 14.0 .001

Marital Status(.. married) 564 (47.4%) 62 (38.3%1 . 4.77 .029

Men 367 (69.2%) 46 (63.0"10) 1.16 .283Women 197 (29.9%1 16 (18.0%) 5.43 .020

Age 74.3 75.0 3.29 .001.\olen 74.1 75.0 2.46 .014\"tomen 74.4 75.1 2.21 .028

Education (highest. completed) 10.6 9.8 ...{).99 .321

.\olen 10.7 10.8 0.10 .921'Nomen 10.6 8.9 -4.93 .000

Income (categorical) 13.7 15.7 0.78 .439."Ien 13.4 15_1 0.48 .633Women 14.0 16.2 0.68 .498

Cognitive Stdtus(Pfeiifer) 8.2 7.3 -6.67 .000

ten 8.2 7.4 -4.23 .000Women 8.1 73 -5.17 .000

..;).

~"'+ .'"J ~1 o/'v~

-:!. t\:Jn .(j ~::.>'

Table:1.DiSCI~rd~nceneh~een'.-iYPolhc:liul -f i 1 Approximately one-third of the sample of both.tnd En.te\edFunctIon",!;:W.tlkln!:CItyI!lucks(1938)' I fir-~ high- and low-iunctioning groups report hypotheti-

T C<11and enacted tenses which are consistent withone another. Again, however. our focus is on theoli-diagonal cells of the Table. Among those HFwho report 1-2 Junctional limitations, nearly onethird c~r;\) performed in the highest tertile of pro-ductive activity. Similarly, among those with no limi-tations, :>0':';,are in the lowest tertile of productivity.Among medium- to low-functioning respondents,31'Yoare in the highest terti Ie of productivity despiterepo~f three or more functional limitations. Fi-

nail ...28°;~C!(th~se wit~ no limitations are amongthe east productive subjects.

The purpose of presenting these data has been tocall attention to the possibility of sl.1J1c;tantialrii<:-- " {cordance betwppn qnswers to questions about fllnr- ~~

tional capacity and actual performanf.~e h~me r-~context. A thorough and comprehensive. examma- ...-tion of this discordance is beyond the scope of these :

1-2 highs41~0 highs32°/~~hsJ7"~.v104(26:~) data ~n? this study. These data are clearly not :-vith- ilows 41,0 lows 27"/0 I~ ~ (27Yo) out limits. However, although others have pOintedhighs 30% highs 33% highs 36% 888(75%) out the importance of question design in studies oflows 28% lows 30% lows 43% 40(25%) functional status (Anderson, Bush, & Berry, 1989;

, . . _~, ": Hickey, 1980; Kovar, 1991; Robine. Michel, & Branch,Note. Percent~ges do ~ot add to~Ec: -/.;0;Iu;:tu 4 1992),few investigators have taken seriously the idea~1~''() Or"~~--.;:>.' () that functional performance may be as Important as

, (: C,,\ 0 .I~ .' .. functional capacity. In the absence of bette~ data,'

(

however, 60 Yoreport later In the questionnaire that 't. these analyses suggest a relatively consistent PI~ur

.

e.1 they do.perform heavy housework. Among the 1\.\LF~Vftcross multiple comparisons, the degree of dlscor-~ group, Just over half .say t~ey are not able to dO"!itl}ance between hypothetical and enact:~ tenses ~f~ heavy housework, while 38 Yoot that group reports function appears to be between the range of 25Yoi that they do heavy housework. Conversely, among and 35%. That is, many older persons .appeJ those who said that they were able to ~o heavy 'I'. ooth overachieving and underachieving in their

housework, 7% of the HFgr~:)Upand ~% ot the MLF~ functional performance relative to self-reports ofgroup reported that they, In fact, did not do any functional capacity as measured by standard in-heavy housework. . . . dic.es. This discordance appears to be relatively

Ta?le 3 pre~ents results of ~ similar analY:;lsfor symmetriCal in that the magnitude of overachieving :~alkrng. In this case, the wordrng of the two ques- roughly matches the magnitude of underachieving.

Itlons was not exactly the same. Here we contrast This suggests that failure to include separate assess- I, those who said they were or were not able to walk ments of performance may lead to considerable'\ one-half mile without help, with those who walk at misclassification and measurement error~ ~ !-L

least one-half city block per day. Among the HF . ~ -group, only 3% said they were not able to walk one- .? ~ ~ - . "'~"Jhalf mile without help. Among that group 0.£37 re- Why Does ThIs Matter.spondents, 65% said they walk one-hplf city block Is the distinction between hypothetical, experi-per day. Among the MLF group, 40% of those who mental, and enacted tenses simply a linguistic trick,

"-. said they are not abl~ to walk one-half mile, re- or are the implications of this argument important?ported that they .did 'N.alkat least one-half city block Is there anything here beyond simple measurementeach day. Among those who said they are able to error? In the remainder of this discussion, eight rea-

", walk one-half ~ile, nearly one-~uarter walk less sons for the significance and importance of this~ than one-half dty hlnc~ per day. phenomenon are presented.

Finally, Table 4 presents a similar analysis whichlooks at any self-reported impairment in the hypo-thetical tense with a summary index of enacted

. function. We contrast those who report 0 versus 1-2versus 3 or more functional limitations (Nagi orRosow-Breslau), with the index of productive activi-ties broken into tertiles. In first looking at the diago-nal cells of the Table. a consistent pattern is visible.

" . ,

Type III Typ~ IV

Yes highs 15':;,lows :!4%

highs 85'~~.lo\vs 511~'~t

1104 (97%)

104 (67";,1

Table 4. Discordance Between Hypotheticaland Enacted Function: Disability and Productive Activity

Tertile of Productive Activity?

Low Medium High Totals

3+ highs n.a. highs n.a.lows 32% lows 37"!n

n.a.78 (48%)

Anyiunctionallimitations

o

-- ---

'f

Discussion

MissingInformation.- Socialand behavioral sci-entists have increasingly moved toward a broaderand richer conception of aging. The new themeshave been heterogeneity, fluidity, adaptation, modi-fiability, and "successful aging" (Berkman, 1988;Maddox, 1987; Rowe.& Kahn, 198?).From large-scaleepidemiological data sets, much has been learned

108

'.

about the pl<lsticity of human aging. Old notions ofaging <IS<In inevitable process of decline <lnd lossof cap<lcity Me being repl<lced. ""\<lnyof these ideasconiirm and reiterate the convcntiOl1<l1wisdom thai.1:>wc age, Ihe mind matters. Regrettably. the devel-opment oi new multidimensional Instruments ior11~'essment 01 global functioning in all its iacetshas lagged bell Ind. Instead. aging researchers haverelied on measures designed in the '1960s largely jorclinical purposes. which are based "almost exclu-sively on the hypothetical tense. Designed to cap-ture the physiological substrata underlying func-tional status. these instruments have been lesssuccessful at capturin howelderlv eo Ie are a u-ai,' unc Ion 111 In t ir i v ives. As a result, a rig-orous assessment and subsequent investigation ofthe contribution of psychosocial factors to function-ing has been lacking A new sensitivity to the impor-tance of multiple tenses of functioning is needed aswell as new and innovative tools for disaggregatingthose tenses,

Will the Real Disability Please Stand Up? - It iswidely recognized that the definition and measure-ment of disability has a profound impact on therates of disability observed in a given population.For exam Ie, Jette (1994) shows that wordin bas~don " IttlCUty' as oppose to the need for "humanassistance" resulted In disability estImates that were1.2 to 5 times greater. Colven the Importance ofwording, it may be especially important to measureactual performance of self-care tasks independentlyot difficulties and help. bQ.th human and non-huma~.,:.'-;0 'Nhiie s.).-:1~5twa:e,,; .:;;:;c:;;yfc,cus .:opiopriateiyc:: ,he ;C:en,lfication or Impairments at the hypo-thetical tense, it may be equally important to begint rize that what reaffy matters most, both inreia-tion to ualit of lite. an as a true marker of un-deriving risk of poor outcome, is not w at people~. but what thev do do. .

Evaluation of Interventions: A Better Mousetrap.- The main goal of health interventions is the opti-mization or restoration of the actual. functional per-formance of life activities. As Kane and Kane (1987)observe, "If long-term care is directed at improvingor maintaining functioning, then its success can bemeasured by functional outcomes" (p. 107).Yet,ourevaluation tools often pertain only to whether help

~

is needed. not whether it I received: whe.1ber diffi-culties are encountered, not whether'they are over-come; and whether performance is ossibl otwhether I IS ac leve , y e inition. outcome mea-sures which tap into performance will be more ap-propriate and def.initive than those that ask onlyabout the hypothetical ability to function, The re-verse is also true. As observed by Anderson et al.(1989). "Treatment programs should not take credit(and resources) for restoring function that does notexist." (p. 15). Interventions which do not translate

\reductions in capacity into corresponding incre-ments in performance are clearly less effective thanthose that do.

Vol. 38, No.1, 1998

In addition, interventions targeted at the enactedtense may require different strategies and ap-proaches. As the work of Mendes de Leon dnd othershave shown (Mendes de leon. hllenbaum. Williams.Brock, Beckell, & Berkman, 1995). attention to issuessuch as self-efficacv may be necessary 111order totranslate the benefits oj increased C<1p<1rityintn theultimate goal nf inrreased function as realized in theresidential context.

The Underestimation of Life-Context. - In addi-tion to the survey data shown here, this research isbased on in-depth interviews conducted by the au-thor in an effort to determine the meanings of vari-ous types of questions about functioning and theirrelation to quality of life in late life. As a conse-quence of this qualitative work, it became increas-ingly clear that functional performance is substan-tially shaped by the social and cultural context ofthe home-world and that tunctioning cannot be un-derstood 111the absence of that context (Glass, 1991;Keith, 1985.1988). However, because aging researchhas relied so heavily on large-scale survey studies.that.include little information about that social con- k::-text data. it is perhaps not surprising that the fieldhas tended to underestimate its importance (Kast,1995). By integrating qualitative data collected innaturalistic settings with large scale survey datil, alevel of richness and insight can be achieved whichis lackmg When either type ot data is examinedalone. In many mtervlews, subjects answered ques-tions about the hypotheticaJ tense, then gave muchdifferent answers when asked to describe what theya~v dId Ihere eXists a need for more qualitativestudies of functioning in the real world which ex-plore explicitly the different tenses of .function andhow discrepancies are understood as meaningful byolder persons.

Compensatory Strategies, - Aging specialists havelong recognized thar older persons cope differ-ently, and vary in their use of compensatory strate-gies and technologies in dealing'with the challengesof late life. The concept of compensation is particu-larly important for identifying factors which helpdisabled persons overcome those challenges andavoid the handicaps that can result. Forms of com-

pensation might include substitution of ~oals, slowertimetables, hel from e ~ui ment or f om others,envlronmenta mo i ication, avoidance 0 azard,or even greater ettort. The concept of compensa-tion ISappealing if for no other reason than it caststhe disabled person in a heroic light. However, tothe extent that enacted function is not measured in-dependently of the hypothetical tense. it becomesnearly impossible to fully explore the impact of thesecompensatory strategies, It may be argued that thediscordance observed in these data represents little "7more than the effects of these various strat ies(i.e.. elp from people or specialized equipment).Certainly the need to perform productive activitiesdepends on more than the ability to do so, but alsoon the need to produce as dictated by one's socioe-

109

."

.cc.nomic-resources. It is possible that the failure tomeasure the two tenses separately has led to a sys-tematic myopia, whereby the real importance andsignificance or"those Iypes ur"help milYbe Sllbst.:ln-Iiall)' underestirnatecl.

Determinants of Under and Over Achir~vP(71en/.-It is not J novel observation Ihat some elders seemto overcome seemingly extraordinary challenges.while others appear to be devastated by only minordisabilities (Hickey, 1980). This observation should beobvious and compelling to clinicians, as it is to any-one who has known a "super" grandparent. How-ever, the disability model, wb.!ch ends with handicap,leaves off prior to the endpoint QI actual perTor-mance within broad fuctional domains. As such, wecurrently lack the tools, both methodologicallv andconceptuallY1 to study those factors most closely as-sociated with these two discordant types. What are

1::.. the factors that are associated with overcoming im-

~~airments among those living in the community?

10' What are the factors which are associated with low~ levels OTenacted tunctlon, despHe (he relative ab-J sence of dIsease or disabIlity? IhIs may be an espe-

cially important rrsk group as is indicated by studiesshowing the deleterious consequences of inactivity(Hartley, 1985; Kannel, Belanger, O'Agostino, & Israel,1986; Kiely, Wolf, Cupples, Beiser, & Kannel, 1994; Lee& Markides, 1990; Leon, 1985; Paffenbarger, Hyde,Wing, & Hsieh, 1986; Simonsick, Lafferty, Phillips,Mendes, Kasl, Seeman, Fillenbaum, Hebert, & Lemke,1993). By developing new tools for the assessment ofenacted function, the capability to address thesequestions will be substantially enhanced.

Differential Discordance Across Age Categories.- There is evidence that the dIscrepancies betweenhypothetical and enacted tenses may be greater inolder peoQle than in youn~er people Jhat is, mea-sures based on the hypothetical tens~ alone maybe especially poor markers of actual functioning inolder populations, therefore leadin to unwarrantedconclusions about th r v ion. Forexample, in a study of stroke recovery Ferrucci andhis colleagues (Ferrucci, Bandinelli, Guralnik, Lam-poni, Berttini, Falchini, & Baroni, 1993)looked at therelation between motor impairment (an index offunctional capability as it relates to the stroke's sever-ity) and functional performance (measured using theBarthel Index of A~" eir findings, based or: a9 prospective study f 50 fi st-time stroke survivors,

~behow that motor imp Irment predicts most of therecover that is observed amon oun er eople

. «65), while 0 er subjects were "more likely to em-K.I

~IOYcompensatory strategIes to overcome some of

,.J he neural impairment that remains atter stroke ,p."8 00). Older persons kept pace with the recovery of

ounger persons in terms of e..nacted function, de-spite demonstrating significantly less improvementat the motor level. Studies such as this one suggestthat the need to look at enacted and hypotheticalfunc!jon independently may be more acute amongthe oldest old. "

r\ S(ratc~y {or i'vlcilsuringSuccessful Aging. - Re-search in aging over the last decade has emphasizedpopul<1tiol1 helcro~eneitv and a widening of theSCOP(\of (aClors thought to determine functioning(Iv\"dclox, 1979."1987).The distinction between "suc-cessiul,'. .'normal," "nd "pathological" patterns ofaging is nO\\~y recognized (Baltes & Baltes, 1990;Berkman. "198();Roos & Havens, 1981; Rowe & Kahn,"19B7;Ruclinger & Tt-"omae.1990). Yet. the concept ofsuccessl\1I aging has not been fully translated intooperational measures and testable hypotheses, withsome notable exceptions. The distinction betweenh...vpothetical and enacted function may be a key tofurther specifying (and measuring) the mechanismsunderlying successful aging. In this study, discrep-ancies were observed between what respondentssay they are able to do, and how they actually per-form. The observation of these discrepancies isespecially important among a cohort of older sub-jects who were selected to be homogeneous andhigh-functioning. ''''any of the psychosocial factorsthat have been linke Wit u clud-in socia networks and su orts, resilien , er-sona Itv c aracteristics. motivation, an self-e tlcacy.may be more closely associated with enacted tunc-tion than with its more hypothetical forms. A newmethod of measuring enacted function will createadditional avenues for the em1Jirical study of suc-

" cessful aging beyond whc.. '"~_ ;;;;, Deen achieved.

Conclusionsand ImplicationsThis article has proposed a conceptual model to

describe and understand differences between func-tional capacity ai1c actual functional performance inthe home. Although the importance of the-wordingof survey questions is commonly recQgnized, thedevelopment and implementation of new measuresof enacted function have lagged. Gerontological re-search has remained focused on standard measuresof function which are limited to the assessment ofcapability (hypothetical tense). The absence of inde-pendent measu rement of the separate tenses offunction mav lead to misclassification of risk, mea-surement er'ror, and a gap in research on the im-portance of psychosocial factors and compensatory

- strategies. It has been argued that the distinctionamong hypothetical. experimental and enacted func-tion is more than semantic, and more than a techni-cal detail of question design, but addresses theways in which the field of aging has tended to con-ceptualize functioning and how the history of thefield informs our understanding of that conceptual-ization. Several empirical examples were providedfrom the MacArthur Studies of Successful Agingwhich suggest a consistent pattern of substantialdiscordance between the hypothetical and enactedtenses of functioning. The full extent of, and expla-nations for, these discrepancies remain poorly un-derstood and, in large part, masked by the conceptsand measures currently in wide use.

.The implications of this argument for researchers,interventionists, and policy makers are potentially

110 The Gerontologist

.

far-reaching. Research is needed to compare inde-pendent assessments of the three tenses of function.This will require the development of additional mea-surement tools specifically designed to capture en-acted function to complement established methodsfor the assessment of disability (hypothetical tense)and physical performance (experimental tense).Measures of enacted function could be developed inboth self-report and observer-rated formats. In ei-ther case, these new measures must be designed tocapture functional performance in the home setting,and must be sensitive to both the frequency and du-ration of functional performance. These new mea-sures may be quite useful for designers of interven-tions. As an outcome, enacted function can be usedto assess the efficacy of interventions more preciselybecause in tapping actual performance, rather thanjust capability, they capture the complex influencesof the social and physical milieu, all of which are rel-evant to the ultimate efficacy of any intervention. Inthe policy domain, many programs utilize standardADl scales to determine eligibility, despite evidencewhich suggests that "capacity wording" leads to un-derestimation of the true extent of functional diffi-culties (Anderson, Bush, &Berry,1989;Kelly-HayesetaI., 1992;Patrick et aI.,1981).The addition of separatemeasures of enacted function would allow for morevalid and reiiable estimation of true disability in pop-ulations, and would lead to the identification ofunder-served groups. Finally, dual assessments ofhypothetical and enacted function will identify theprevalence of discrepant patterns II and III (over-and under-achieving) which constitute, in one case,a., iiiipcrtant ,l€Wca.tegory of risk and, in the othercase, an important new phenomenon for study. Inshort, much can be learned from those individualsfor whom assessments of how they might functionand how they,do function do not coincide.

References

Anderson, J. P., Bush, J. W., & Berry, C. C. (1989). £"or causes in function as-sessment for quality of life and health outcome measures: Performanceversus capacity and other sources. Technical report of the Health PolicyProject, University of California at San Diego, San Diego, CA.

Applegate, W. B., Blass, J. P., & Williams, T. F. (1990). Instruments for thefunctional assessment of older patients. New England Journal of Medi-cine,322,1207-1214.

Baltes, P. B., & Baltes, M. M. (1990). Psychological perspectives on success-ful aging: The model of selective optimization with compensation. InP. B. Baltes & M. M. Baltes (Eds.), Successful aging: Perspectives fromthe behavioral sciences. Cambridge: Cambridge University Press.

Berkman, L (1988). The changing and heterogeneous nature of aging andlongevity: A social and biomedical perspective. In G. L.Maddox & M. P.Lawton (Eds.), Annual review of gerontology and geriatrics (Vol. 8).New York: Springer Publishing Company.

Berkman, L. F., Seeman, M., Blazer, D., Kahn, R., Mohs, R., Finch, c.,Schneider, E,Cotman, c., McClearn, G., Nesselroade, J., Featherman, D.,Garmezy, N., McKhann, G., Brim, G., Pragar, D., & Rowe, J. (1993). High,usual and impaired functioning in community-dwelling older men andwomen: Findings from the MacArthur Foundation research network onsuccessful aging. Journal of Clinical Epidemiology, 46, 1129-1140.

Bortz, W. M. (1989). Redefining human aging. Journal of the American Geri-atric Society, 37,1092-1096.

t:. Branch, L. G., Katz, S., Kniepmann, K., & Papsidero, J. A. (1984). A prospec-tive study of functional status among community elders. AmericanJournal of Public Health, 74,266-268.

Butler, R. N., & Gleason, H. P. (1985). Productive aging: Enhancing vitality in. later life. New York: Springer.

~Iark, D.O., Stump, T. E;, & Wolinsky, F. D. (1997). A race- and gender-

Vol. 38, No.1, 1998

specific replication of five dimensions of functional limitation and dis-ability. Journal of Aging and Health, 9, 28-12.

Cornoni-Huntley, J., Ostfeld, A, Taylor, J., Wallace. R., BI.1Zer,D., Berkman,L. F., Evans, D., Kohout, F., Lemke, J., Scherr. P., & Korper, S. (1993). Es-tablished populations for the epidemiologic studies of the elderly:Study design and methodology. Aging/Clinical Experimental Research.5, 27-37.

Cutler,S. !., & Danigelis, N. L (1993). Organized contexts of activity. In J. R.Kelly (Ed.), Activity and aging: Staying involved in later life. NewburyPark, CA: Sage Publications.

Feinstein, A. R., Josephy, B. R., & Wells, C. K. (1986). Scientific and clinicalproblems in indexes of functional disability. Annals of Internal ,'"Iedi-cine, 105,413-420.

Ferrucci. L. Bandinelli, 5., Guralnik, J. ,vI., Lamponi. M., Berttini, c., Falchini,M., & Baroni, A. (1993). Recovery of functional status after stroke: Apostrehabilitation follow-up study. Stroke, 24, 200-205.

Fuhrer, M. J. (1987). Overview of outcome analysis in rehabilitation. In M. J.Fuhrer (Ed.), Rehabilitation outcomes: analysis and measurement. Balti-more: Paul H. Brookes Publishing Co.

Glass, T. A. (1991). A healthy state of tension: A review essay on three re-cent works in aging. Journal of Cross-cultural Gerontology, 6, 119-125.

Glass, T. A., Seeman, T. E., Herzog. A. R., Kahn, R., & Berkman, L F. (1995).Change in productive activity in late adulthood: MacArthur Studies ofSuccessful Aging. Journal of Gerontology: Social Sciences, 508, S65-S76.

Guralnik, J. M., Branch, L.G., Cummings, S. R., & Curb. J. D. (1989). Physicalperformance measures in aging research. Journal of Gerontology:Medical Sciences. 44, M141-146.

Hartley, L. (1985). The role of exercise in the primary and secondary pre-vention of artherosclerotic coronary artery disease. CardiovascularClinics, 15,1-8.

Herzog. A., Antonucci. T., Jackson, J., Kahn, R., & Morgan, J. (1987). Produc-tive activities and health over the life course. Paper presented at themeeting of the American Association for the Advancement of Science,Chicago, IL

Herzog. A..,& House, J. (1991). P~oductive activities and aging well. Genera-tions, 15, 49-54.

Herzog. A. R., Kahn, R. L, Morgan, J. N., Jackson, J. 5., & Antonucci, T. C.(1989). Age differences in productive activities. Journal of Gerontology:Social Sciences, 44, 5129-5138.

.Hickey, T. (1980). Health and aging. Monterey, CA: Brooks/Cole Publishing.-rHubert, H. B., Bloch, D. A., & Fries, J. F. (1993). Risk factors for physical dis-

ability in an aging cohort: the NHANES I Epidemiologic FollowupStudy. Journal of Rheumatology, 20, 480-488.

Idler, E. L (1987). Religious involvement and the health of the elderly:Some hypotheses and an initial test. Social Forces, 66, 2:6-238.

';- Jette, A. (1987). The Functional Status Index: reliability and validity of a self-report functional disability measure. Journal of Rheumatology. 14,15-21.

Y Jette, A. (1994). Physical disablement concepts for physical therapy re-search and practice. Physical Therapy, 74, 380-386.

;'Jette, A. M. (1994). How measurement techniques influence estimates of. disability in older populations. Social Science & Medicine, 38, 937-942.

Jette, A. M., & Branch, L G. (1981). The Framingham disabiiity .tudy: II.Physical disability among the elderly. American Journal Public Health,71,1211-1216.

~, Jette, A. M., & Branch, L G. (1985). Impairment and disability in the aged.Journal of Chronic Diseases, 38, S9~5.

Kahn, R. L. (1986). Productive activities and well-being. Paper presented atannual meetings ofThe Gerontological Society of America, Chicago, IL.

Kane. R. A., & Kane, R. L (1987). Long-term care: Principles, programs andpolicies. New York: Springer.

Kannel, W. B.. Belanger, A., D'Agostino, R., & Israel. I. (1986), Physical activ-ity and physical demand on the job and risk of cardiovascular diseaseand death: the Framingham Study. American Heart Journal, 112.820-825.

Kasl, S. V. (1995). Strategies in research on health and aging: Looking be-yond secondary data analysis. Journal of Gerontology: Social Sciences.SOB, 5191-5193.

\. Katz, S., Downs, T. D., Cash, H. R., & Grotz, R. C. (1970). Progress in devel-opment of the index of ADL Gerontologist, 10, 20-30.

~Katz, S., Ford, A. B., Moskowitz, R. W., Jackson, B. A., & Jaffe, M. W. (1963).Studies of illness in the aged. The Index of ADL: A standardized mea-sure of biological and psychosocial function. Journal 0; the AmericanMedical Association, 185,914-919.

Keith, J. (1985). Cross-cultural research on aging. In Binstock & Shanas(Eds.), Handbook on aging and the social sciences. New York: VanNostrand and Reinhold.

Keith, J. (1988). Participant observation. In K. W. Schaie, R. T. Campbell, W.Meredith, & S. C. Rawlings (Eds.), Methodological issues in aging re-search. New York: Springer.

Keller, D. M., Kovar, M. G., Jobe, J. B., & Branch, L.G. (1993). Problems elic-iting elders' reports of functional status. Journal 0; Aging and Health,S,306-318.

Kelly.Hayes, M., Jette, A. M., Wolf, P. A., D'Agostino, R. B., & Odell, P. M.(1992). Functional limitations and disability among elders in the Fram-ingham Study. American Journal of Public Health. 82, 841-84S.

111

..:-t. 0.. \Voll. 1'.. Cllppk'~. 1.00IId~"r. '\00,", ",,,,,,d. W. /1')').11."hY,"C,ll ~t:. seli.adminislered. inlervie\Ver.~dminislered. ~nd perlorm;ance.b~sedli\"l1~.lI1d ~Irul...t'rrsl...:11'It~fr.uuin::;h,lItI Study. ",ueric.,n luurn.lI t)i f/);' mC..lsures. Journal of the Atncr;c:ln Gcr;.1tricsSOC;CI)-..IJ. 17-23.(/(..""oloS)'. 1411,(~lIl-f.".!O. I(ohinl'. I. M.. Michel. I. 1'00.j. IIr.lnch. L.C, (1991). ,"INsuremenl ~nd utiliza.

",)\'.((. .\\. G. ,1'1'111.Fllnt:tin,,;!1 ..hility .",d Ihe "l'ed ior eMe: Is".cs i,.. linn ni h.",lthy lile e~peclancy: Conceplu;al issues. Bulletin of thefUt',hun'ment rt!~"Jr("h. \';1,,1 \~ IIc-,lllh S,.l/;~tit:.. - :\(.,it~" ;. C'U'UJoIf.I' \\'orld H.!o1/tI. Of~.Jn;T..1"on 7rJ 791-800.

. , ~ III'"'olcm,llioll,lll'il.lI& 11 ,1111SI.II;slic:.~R"pflll".'1:-11I.1. ' 1.:0"'. :0:.1'00S. H~vc~s.II~11911;I. Predictors oi success(ul ~sing:A Iwelve.~r..,\\'n"H:"~. ~. 1-1..Solelte. ..\. .\t ,I'I'}hl. Di,;,('nl.U1~lin!-:111\.'dior..\hl,~rn~nrpru. \'t~Jr ,Iully 01 ""t1nitoIM elderly. 1\lneric.tn /nurnoJlof Public HeoJlth, 81,(,',.../()urntll oiGl'rt)nttJl().~r:)"no1l Sc:;t:t1re..:.;1/:. S!;".t-'\U\l. hi-lJ\.

l.awl,,". .\1. 1'00& limo\'. ~. ,'.1. 1I"'Ihl), '\'<l'Si"'~'" IIi ..Id.., p "Il.: ~..h. I.:'''IJW I..,), B,,:sbu. N. .191"11>, Cu(1l11~nh"Jllh sc;ale ior the ..ged.Jour.1I1"iI1'~1iniI1g~ll1d1I1'lrun1cnl~1I.1(tivlli,.",\,:"cI.1il\'li\'ln~. G~f(mtl,/l}g":. ". :...1Il:"Gc(onto/(J~,...lr. 556-55').

'f 1")_11\(.. Ruwr. I, \\'.. oS."",hn. R. L 0<)1\71.Hum~n ~ging: USUJI"00 suc:cessCul. Sci.'J~L1\\Inn. \,1.I'.. I"\U~"..\1... ulcu",,!r. .\\.. &.KlclMn..\\. H. I Jlm..!J.-\ r~~t'.1rr.:h I..n(-: 1J;. 1~J-'49.

, .lOd service un~nted OIuilil.,\,cI ,""'''111(:111 ""Irume"'. lourn.11 (,; Ruc/ingf:r. t;.. SoThom..e. H. 119901.The 60nn I.ongitudin.al Study oCAging:G,'rw!/o/osr. .I:, '11-1)<). Cupin!;. Iii" Jdjustm"n.. Jnd liCes;atisi.lction. In P. B. &lres & M. M.

Lee. I). I.. $, M~rkidcs. ". 5.11<)90'. Acti\'ity and mor,.llity ~mong dgcd per. B..lies rEds.). Succcss;ul ~ging' Perspecti\'es from the behavioral sci-.>I)I1Sover an eighl \'ear period. 1,>uro.llof {A!r<)nl<tlogv:Social Sciencc. "nces. Clmbridgc: Cambridge University Press.~:;. 539-542. S,monsick. E.M..lalierty, M. E..Phillips, C. L. Mendes. D~ leon, C. F., !<asl,

Leon, S. 1198SI.Ph\"Sical~ctivitv levels and coron,I'" hedrt dise;as" an~I\". S. Seem;an. T. E., Fillenbaum. C.. Hebert. P.. & lemke.. ,. H. (1993).:,is 01 .,pidemiologic and supporting itudies., ,\f.~dic:alQinics 0; ."mi/, Risk due 10 inactivily in physically capable older adults. American lour.AmeriCJ.69.3-10. _:1/ n;,10; Public Health.83.1+J3-1450.

.'.I~ddox, G. L (19791.The sociology oi I~ter liie. Alllllal Re..;e..' Soc,olog)'. ?-~te.vart. L Ware, J. E.. oS<Brook. R. H. (19811.Advances in the Measu.'e-5. 113-135. '..J menl oi functional status: Construction oi aggregate indexes. MedlcaJ

. ; .'.Iaddox. C. L.(19871.Aging diiieren.!!¥ The Gerontologist. 17. 557-564. ure. XIX,.J7J-.oI88.'.Mahone\'. F. I.. & Bar1hel. D. W. /19651f Functional ev;aluation: Th., Barthel Tineni. .\.t. E. (1986). Per(onnance-orienled ;assessment oi mobility prol>-

Inde~.Maryland State Medicar(ournal.14,61-65. . lems in elderly patients. Journ;alof the American GeriatricSociety, 34,McDowell. I.. & Newell. C. (1987). Measuring health: A guide to r;a/'"8 r.::J 119-126. . .

sCJ/esand questionnaires. New York:O~(ord Unive~ty Press.

~. l'erbrugge. L M.. & Jette. A. M. (1994).The disablement process. SOCIal Sc/.

Mendes de Leon. C. F.. Fillenb;aum. C. C.. Williams. C. S.. Brock. D. B.. ence & Med"lCine.38,1-14.Becken. L. A., & Berkman. L F. (1995;; Functional disability ;among el. Verbrugge, L M.. lepkowski, ,. M., & Konkol. L L (1991). levels oCdisabil-derly blacks and whites in two diverse areas: the New Haven and ity among U.s. adults with arth(itis. Journal of Gerontology, 46. S71~.Sorth Carolina EPESE.Established populations ior the epidemiologic Wagner. K.A. (19871.Outcome analysis in comprehensive medical rehabil-studies of the elderlv. American lournal of Public Health. 16. 994-998. itation. In M. J. Fuhrer IEd.l, Rehabilitation outcomes: Analysis and. .' .

Mendesde Leon,C. F.:Seem;an.T.E.,Baker.D. I.. Richardson.E.D..& measurement.Baltimore:PaulH.BrookesPublishingCo. "

Tineni, M. E.11996I.Self-e(ficacy.physical decline. and change in func. World Health Organization. (19801.Internationalclassificationof impair.

tioning in community.living elders: a prospective study. Journals 0; ,:;t.:... ments. disabilities, and handicaps. Geneva: World H~lth Organiza.tion:Gerontology. Social Sciences, 51. 5183-5190.' ,:~pebland. 5.. Firzpatrick, R.. & Jenkinson. C. (1993).Taat models of dlsab.l-

,'.Iiner. 500Logan. I. R.. oS<Spitze. C. (1993). Predicting the irequency oi se. ., ity underlying health status instruments. Social Science & Medicine,nior center attendance. Gerontologist, 33. 650-657. 37,69-i5.

;'Iolvers.C. C.. Manton, K. G.. & Bacell;ar. H. (1'1861.Sododemographic as-pects of future unpaid productive roles. In Institute oi Medicine,'Na-lional Research Council IEds.). America's aging: Producti.-e roles in In

. older sociery.';'\'ashington DC: National Academy Press."<; i':agi. S. Z. 11%5).-Some conceptual issues in disability and rehabilit;ation.. In M. B. Sussman (Ed.l. Sociology and rehabilitarion. Washington. DC:. American Sociological Association.~'Nagi, S. Z. (19761.An epidemiology oi disability among adults in the United. Slates.Milbank MemorialFund Quarterly. 54. 439-168.

:.it'ragi. S. Z.(1991). Disability concepts revisited: Implications (or prevention..' In.A. Pope &A.Tarlov (Eds.),Disability in America: Towarda national

agenda for prevention. Washington, DC: National Ac;ademy Press.Paiienbarger. R. 5.. Ir., Hyde. R. T.. Wing. A. L. & Hsieh. C. C. 11986.. Physi-

cal aClivily, all-cause mortality. and longevity oi collelte ;alumni. NewEng/dnd Journal of Medicine. 6314, 605-013.

Patrick. D. L, Darby. S. CooGreen,S.. Horton, C_ Locker. Doo& Wiggins. R.D. (19B1).Screening for disability in the inner city.Journalof Epidemiol

~, Community Healrh,35.65-70.'? Pinsky. J.. Branch, L. Jelte. A.. Haynes. S.. Feinleib. M.. Comoni.Huntlev. 100

.s. Bailey, K. (19851. Framingham Disability Srudy: Relationship oi dis-ability to cardiovascular risk iactors among persons iree oi diagnosed

~ cardiovascular disease. American Journal of Epidemiology. 122.~56.!;Reuben. D. B.. Valle, L A., Hays. R. D.. & Siu. A. L (19951.Measuring physi-

.¥ cal function in community-dwelling older persons: a comparison 01

.r

Received December 18, 1996Accepted June 20, 1997

AppendixNotes

1. There are minor differences in the language used bydifferent authors in the disability tradition. Where Nagiuses the terms "disability" and "handicap" (Nagi,1965),Verbrugge and Jette use instead the terms "functionallimitations" and "disability" (Verbrugge &Jette, 1994).

2. Subjects were considered to be high-functioning ifthey met several criteria including a) no disability onthe Katz ADL scale, b) not more than one disabilityitem on either the Rosow-Breslau or Nagi scales ofgross mobility and physical performance, and c) aSCDreof 6 or higher on the 9-item test of cDgnitiveper-formance developed by Pfieffer.

112 The Gerontologist