Functinal-fitness

Development and Validation of a Functional Fitness Test for Community-Residing Older Adults\nRoberta E. Rikli and C. Jessie Jones\nPreventing or delaying the onset of physical frailty is an increasingly important goal because more individuals are living well into their 8th and 9th decades. We describe the development and validation of a functional fitness test battery that can assess the physiologic parameters that support physical mobility in older adults. The procedures involved in the test development were (a) developing a theoretical framework for the test items, (b) establishing an advisory panel of experts, (c) determining test selection criteria, (d) selecting the test items, and (e) establishing test reliability and validity. The complete battery consists of 6 items (and one alternative) designed to assess the physiologic parameters associated with independent functioning—lower and upper body strength, aerobic endurance, lower and upper body flexibility, and agility/dynamic balance. We also assessed body mass index as an estimate of body composition. We concluded that the tests met the established criteria for scientific rigor and feasibility for use in common community settings.\nKey Words: aging, physical performance, functional capacity, assessment, reliability, validity Finding ways to prevent or delay the onset of physical frailty in later years has become a major goal for gerontology researchers and practitioners throughout the world. According to previously published terminology, physical frailty is defined as the loss of physiologic reserve that increases the risk of disability (Buchner & Wagner, 1992). The proposed battery of tests has been developed to provide a means of assessing the key physiologic parameters that support functional mobility in older adults. Although physical decline during aging is due to multiple causes—a combination of biological aging, disease, and certain lifestyle patterns such as low levels of physical activity (American College of Sports Medicine [ACSM], 1998; Buchner & Wagner, 1992)—it is believed that much of this loss is preventable and even reversible through early detection of physical weakness and appropriate activity intervention (Gill, Williams, Richardson, & Tinetti, 1996; Gururaj, Ferrucci, Simonsick, Salive, & Wallace, 1995; Jackson et al., 1995; Lawrence & Jette, 1996; Morey, Pettee, & Cornoni-Huntley, 1998). Many independent older adults, often because of their sedentary lifestyles, are functioning dangerously close to their maximum capacity when performing normal everyday activities (e.g., climbing stairs, getting out of a chair, lifting objects; Chandler & Hadley, 1996; Evans, 1995b; Shephard, 1993; Shephard, 1997). Any further decline or small physical setback could easily cause them to move from independent to disabled status, in which assistance is needed for daily activities and the risk of falling and costly fall-related injuries is greatly increased.\nUnfortunately, a limiting factor in evaluating and managing physical decline during aging has been the lack of suitable measurement tools to assess the underlying physical parameters associated with functional mobility—strength, endurance, flexibility, balance, and agility (Chandler & Hadley, 1996; Chodzko-Zajko, 1994; Fried et al., 1996; Rikli & Jones, 1997; Spiriduso, 1995; Verbrugge & Jette, 1994). Most traditional protocols for assessing fitness (treadmill and cycle ergometer tests, bench step tests, 1RM strength tests, etc.) were developed and validated for younger people and generally are inappropriate or unsafe for the majority of older adults, particularly without medical release and close monitoring of medical conditions. Even the least demanding treadmill and aerobic step test protocols, for example, are too difficult for the majority of the community-residing older population, many of whom are quite sedentary and already have experienced substantial declines in physical capacity. In addition, many of these protocols require expensive equipment or extensive training for test technicians and are not, therefore, feasible for use in the clinical or community setting where fitness assessment of older adults generally takes place. 132 • Functional Fitness Test\n\nmuscular, cardiovascular, neurologic, etc.) to functional limitations (restrictions in physical behavior such as lifting, stooping, walking, or climbing stairs) to disability (the inability to perform normal daily activities such as bathing oneself, housework, or shopping).\n\nAlthough traditional models (Nagi, 1991) indicate that all disability originates directly from disease or pathology (with disease leading to impairment, impairment to functional limitation, and functional limitation to disability—see Figure 1(a)), recent evidence suggests that disuse or physical inactivity can be just as responsible for the physical decline leading to disability and that perhaps the model should be revised as in Figure 1(b) (Chandler & Hadley, 1996; DiPietro, 1996; Morey et al., 1998). Such a modification has important implications both for prevention and for intervention strategies, as well as for developing relevant assessment tools. The growing body of evidence supporting the role of physical inactivity in the loss of function independent of the disease process comes from several large epidemiologic and longitudinal studies (ACSM, 1997; Chandler & Hadley, 1996; Gill et al., 1996; Kaplan, Strawbridge, Camacho, & Cohen, 1993; Lacroix, Guralnik, Berkman, Wallace, & Satterfield, 1993; Lawrence & Jette, 1996; More et al., 1989; Morey et al., 1998; Seeman et al., 1995; Stewart et al., 1994).\n\nThe functional ability framework (Figure 2), based on the models shown in Figure 1, provided a useful guide during the process of defining functional fitness and in identifying relevant physical parameters for measurement. The framework depicts the progressive relationship among physical parameters, functional performance, and activity goals. The common activities in the far right column of figure 2 (e.g., caring for personal needs, shopping, traveling) require the ability to perform the functions listed in column two (e.g., walking, stair climbing, lifting, reaching). These functions, in turn, require adequate reserve in the physical parameters\n\n(a)\nDisability Model (Nagi, 1991)\nDisease/ Pathology ➞ Impairment ➞ Functional Limitation ➞ Disability\n\n(b)\nDisability Model (amended)\nDisease/ Pathology ↕ ↕ Impairment ➞ Functional Limitation ➞ Disability\nLifestyle/ Inactivity\n\nFigure 1. (a) Nagi's (1991) model of the progression leading to disability and (b) an amended version suggesting that an inactive lifestyle can have comparable effects on the disabling process. Adapted from \"Assessing Physical Performance in Independent Older Adults: Issues and Guidelines\" by R.E. Rikli and C.J. Jones, 1997, Journal of Aging and Physical Activity, 5, p. 247. © 1997 by Human Kinetics. Adapted with permission. Rikli and Jones • 133\n\nidentified in column one—muscular strength, endurance, flexibility, power, speed, agility, and balance—as well as an optimal (or at least manageable) body mass index. Supporting rationale for the physical fitness categories (column 1) relative to their importance in functional mobility has been well documented in major reports and publications (ACSM, 1997; ACSM, 1998; Bouchard, Shepard, & Stephens, 1994; Buchner, 1995; Chandler & Hadley, 1996; Hurley & Hagberg, 1998; Morey et al., 1998; U.S. Dept. of Health and Human Services, 1966).\n\nBased on the framework outlined in Figure 2, functional fitness has been defined as having the physiologic capacity to perform normal everyday activities safely and independently without undue fatigue. In support of this definition, appropriate items on a functional fitness test battery would be those that reflect the physiologic attributes that support the behavioral functions necessary to perform activities of daily living. The phrase without undue fatigue is included in the definition to emphasize the importance of maintaining an adequate physiologic reserve. In fact, evidence suggests that a direct relationship between physiologic impairment (in strength, endurance, etc.) and functional limitation exists not only in the lower ranges of the performance spectrum, at least with respect to basic activities such as walking and stair climbing (Buchner & deLateur, 1991; Buchner, Larson, Wagner, Koepsell, & deLateur, 1996; Ferrucci et al., 1997; Jette, Assmann, Rooks, Harris, & Crawford, in press; Judge, Underwood, & Gennosa, 1993). Increasingly, however, gerontologists are recognizing the need for screening tools that can assess physical declines early enough not only to prevent disability—that is, to detect and treat declines before reserves have been depleted and the threshold\n\nPHYSICAL PARAMETERS\nMuscle strength/ endurance\nAerobic endurance\nFlexibility\nMotor ability power speed/agility balance\nBody composition\n\nFUNCTIONS\nWalking\nStair climbing\nStanding up from chair\nLifting/reaching\nBending/kneeling\nJogging/Running\n\nACTIVITY GOALS\nPersonal care\nShopping/ errands\nHousework\nGardening\nSports\nTraveling\n\nPhysical impairment ➞ Functional limitation ➞ Reduced ability/ Disability\n\nFigure 2. A functional ability framework indicating the physiologic parameters associated with functions required for basic and advanced everyday activities. Adapted from \"Assessing Physical Performance in Independent Older Adults: Issues and Guidelines\" by R.E. Rikli and C.J. Jones, 1997, Journal of Aging and Physical Activity, 5, p. 246. © 1997 by Human Kinetics. Adapted with permission. 134 • Functional Fitness Test.\n\nis reached at which ADL behavior is affected—but also to facilitate continued involvement in advanced exercise activities (Buchner & Wagner, 1992; Fried et al., 1996; Morey et al., 1998).\n\nADVISORY PANEL OF EXPERTS\n\nBoth local and national advisory panels were assembled to serve as consultants throughout the project. Local panel members, all exercise specialists in southern California, assisted with the development and pilot testing of the test items. The national advisory panel, comprising noted researchers and program leaders in the fields of gerontology or exercise science, reviewed the test development materials and the reliability and validity analysis procedures. All pilot testing for the project and other background research was conducted at the Ruby Gerontology Center at California State University, Fullerton, and in the surrounding communities.\n\nTEST SELECTION CRITERIA\n\nIn choosing specific test items to assess the functional fitness components, consideration was given to two major goals: (a) the development of test protocols that meet acceptable scientific standards with respect to test reliability and validity and (b) the development of tests that would be easy to administer and feasible for use in common clinical and community settings in which the majority of assessment is most likely to take place. The following is a list of 12 criteria established for use as guidelines during the test development. It was agreed that all test items should\n\n• Represent major functional fitness components—that is, key physiologic parameters associated with the functions required for independent living.\n• Have acceptable test–retest reliability (≤.80).\n• Have acceptable validity, with documentation to support at least two of the following: content validity, criterion validity, and construct (discriminant) validity. For acceptable criterion validity, correlations between the test item and the criterion measure were to be .70 or greater. For construct, or discriminant, validity, relevant group differences should be significant beyond the .01 level.\n• Reflect normal age-related changes in physical performance.\n• Be able to detect physical changes resulting from training or exercise.\n• Be able to assess on a continuous scale across wide ranges of functional ability—from the low fit/borderline frail to the highly fit. The goal was to avoid “ceiling” and “floor” effects so that all, or most, participants would receive a score.\n• Be easy to administer and score (by paraprofessionals and volunteer technicians, who often assist in administering tests).\n• Require minimal equipment and space (can be administered in typical clinical and community settings).\n• Be capable of being administered by oneself or a partner in the home setting.\n• Be safe to perform without medical release for the majority of community-residing older adults.\n• Be socially acceptable and meaningful.\n• Be reasonably quick to administer, with individual testing time requiring no found for both other test trials.\n\nTable 2 Descriptive Characteristics of Study Participants—Means and Standard Deviations\n\nMen (n = 34)\n M SD\nAge (years) 72.6 6.6\nHeight (cm) 177.0 7.4\nWeight (kg) 83.1 16.6\nNumber of chronic conditions 1.9 1.8\nNumber of medications* 2.0 1.7\nNumber of years of education 18.6 2.1\n\nWomen (n = 48)\n M SD\nAge (years) 69.1 5.1\nHeight (cm) 163.1 5.8\nWeight (kg) 71.2 14.3\nNumber of chronic conditions 1.8 1.6\nNumber of medications* 1.9 2.1\nNumber of years of education 16.0 1.9\n\n*Different prescription medications taken each day. main purpose of the chair-stand test, for example, which was validated against 1RM (1-repetition maximum) leg-press strength, is to assess lower body strength, not just the ability to get out of a chair. Similarly, the 6-min walk is included to assess aerobic endurance, not simply walking ability.\n\nESTABLISHING TEST-RETEST RELIABILITY\n\nParticipants. Eighty-two older adults (48 women and 34 men; mean age = 71.8 years, SD = 6.9) were solicited from a nearby senior housing complex and from enrollees in a university-sponsored exercise program to participate in testing to establish the reliability of the selected items. Each participant signed an informed consent and completed a written questionnaire requesting information about his or her health, functional status, and physical activity level. The criteria for inclusion in the study were that participants be over the age of 60, be community residing, be ambulatory without the use of assistive devices, have no medical conditions that would be contraindicated for submaximal testing (ACSM, 1995), and not have been advised by their physicians to refrain from exercise. Participant characteristics for the reliability studies are presented in Table 2.\n\nReliability Procedures. All of the tests (except for the 6-min walk) were administered on the same day, with retesting occurring 2 to 5 days later. The 6-min walk test was administered separately approximately 3 weeks later, again with 2 to 5 days separating the test trials. Prior to all testing, participants performed 8 min of warm-up and stretching exercises. All tests were administered by trained graduate students and volunteer senior technicians. During the training, technicians practiced on each other until they demonstrated proper procedures to the project coordinators. So that test objectivity would be reflected in the reliability estimates, different technicians were used to conduct the tests on each of the two test days. Day 2 technicians were not aware of the scores obtained on Day 1. Because data indicate that significant increases can occur in mean scores from Day 1 to Day 2 on two of the test items—the 6-min walk (Rikli & Jones, 1998) and the 2-min step-in-place test (Dugas, 1996)—participants received three trials on both of these tests 2 to 5 days apart, with the first trial treated as a practice trial. Although a significant trial, or “practice,” effect (increase in scores from one day to the next) does not always have a large impact on the relative reliability of test–retest scores (i.e., the individual positioning of scores relative to other scores in the group), it does affect absolute scoring consistency and the stability reliability of the test. Scoring consistency (the lack of a practice effect) is of particular concern in intervention studies, in which stable baseline measures are required for accurate interpretation of treatment effects.\n\nData Analysis. Test–retest reliability for all test items was established by calculating the intraclass correlation coefficient (R), using a one-way ANOVA model appropriate for estimating what the reliability would be for a single test (Baumgartner & Jackson, 1995). The one-way ANOVA also provides information on the amount of change in scores, if any, from one day to the next. A lack of significant change indicates scoring stability across trials.\n\nResults. Intraclass reliability values (R) and 95% confidence intervals (CI) for all tests are presented in Table 3. The R values for the test items ranged from .80 to .98—with a majority of the values being .90 or above, indicating that the tests have good relative reliability across trials. The R values in Table 3 reflect intraclass correlations between Trials 1 and 2 on all items except the 6-min walk and 2-min step test. On these tests, for which Trial 1 was treated as a practice trial, the R values reflect intraclass correlations between Trials 2 and 3. In an earlier study on step-test performance in older adults, Dugas (1996) found intraclass correlations of around .80 between Trials 1 and 2 but correlations ranging from .83 to .97 between Trials 2 and 3. The same data also revealed a significant increase in mean scores from Day 1 to Day 2 (p < .001), but not between Trials (Days) 2 and 3, thus suggesting that treating Trial 1 as a practice trial and Trial 2 as the official test trial results in superior relative reliability, as well as stability reliability.\n\nSimilarly, as reported elsewhere (Rikli & Jones, 1998), a significant increase was found in 6-min walk performance from Day 1 to Day 2, but not between Days 2 and 3. However, on the 6-min walk, the intraclass correlations (relative reliability) of .88 to .91 between Trials 1 and 2 were nearly as high as the .91 to .97 correlations 138 • Functional Fitness Test\nbetween Trials 2 and 3. On all other test items, ANOVA results revealed no significant changes in scores from Day 1 testing to Day 2, thus indicating that the tests also have absolute, or stability, reliability across trials.\n\nESTABLISHING TEST VALIDITY\n\nContent validity (or logical validity, as it usually is called when referring to performance measures) is the degree to which a test (or test battery) reflects a defined universe or domain of content (American Psychological Association, 1985). For performance measures, content validity is typically established by examining, through literature review or through expert opinion, the degree to which a test measures the capacity (or capacities) that it is intended to measure. The content relevance (or logical validity) of each item in the battery of tests is discussed in the next section.\n\nCriterion validity represents the degree to which a test correlates with a criterion measure that is already known to be valid. The criterion-related validity of the test items was estimated by calculating an interclass correlation coefficient (Pearson's r) between the test item scores and scores on a criterion measure of performance, whenever a suitable criterion could be identified. Validation procedures are discussed separately for each test item and are summarized in Table 4.\n\nConstruct, or discriminant, validity represents the degree to which a test measures a particular construct an attribute that exists in theory but cannot be directly measured) such as \"functional fitness.\" Assessing construct validity typically involves comparing test results of two or more groups that have known or presumed differences with respect to the construct of interest (Morrow, Jackson, Disch, & Mood, 1995). A test's ability to discriminate among such groups is an indication of its construct, or discriminant, validity. To assess the construct validity of the tests in this battery, comparisons were made between older adults who were regular exercisers and those who were not regularly active, with the assumption that exercisers would possess higher levels of fitness than nonexercisers. Comparisons were also made across different age groups of individuals (those in their 60s, 70s, and 80s) in which performance is generally expected to decline.\nParticipants in the construct validity studies (except for the 6-min walk) consisted of 190 male and female residents of a nearby retirement housing complex (mean age = 76.2, SD = 6.7). Participants in the 6-min walk study were 77 volunteer men and women (mean age = 73.1, SD = 7.2) solicited from another retirement housing complex and from a university-sponsored exercise program. Each participant signed an informed consent and completed a written questionnaire requesting information about his or her age, height, weight, health status, and physical activity level. Participants were ambulatory without the use of assistive devices and not suffering from joint pain, unstable cardiovascular disease, or other medical condition that would be contraindicated for submaximal testing (ACSM, 1995).\n\nTest Items\n\nFollowing is a general description of each test item, along with supporting evidence of the item's validity as an indicator of functional fitness. Full descriptions of each test item and instructions for administering the tests are presented in Appendix A.