Archives of Physical Medicine and Rehabilitation
Volume 88, Issue 6 , Pages 732-739, June 2007

Reliability, Construct Validity, and Clinical Feasibility of the Activities-Specific Fall Caution Scale for Residential Living Seniors

Presented in part to the Gerontological Society of America, November 2005, Orlando, FL.

Department of Health Studies and Gerontology, University of Waterloo, Waterloo, ON, Canada.

Article Outline

Abstract 

Blanchard RA, Myers AM, Pearce NJ. Reliability, construct validity, and clinical feasibility of the Activities-specific Fall Caution Scale for residential living seniors.

Objective

To examine the reliability, validity, and feasibility of the Activities-specific Fall Caution (AFC) Scale.

Design

Cross-sectional studies with test-retest and interrater reliability.

Setting

Residential care facilities in Ontario, Canada: 10 in study 1 and 6 in study 2.

Participants

Convenience samples of 101 and 31 residents.

Interventions

Not applicable.

Main Outcome Measures

In study 1, the AFC Scale was readministered to 44 residents, 64 were assessed using the Berg Balance Scale, Timed Up & Go, and Self-Paced Walk Test, and the Nursing Home Life-Space Diameter was completed for 80 residents. In study 2, staff administered the AFC Scale to 31 residents on 2 occasions.

Results

In study 1, test-retest reliability (intraclass correlation coefficient [ICC]) was .87 (95% confidence interval, .78−.93). AFC scores were associated with physical performance and mobility patterns (P<.001) and able to discriminate on the basis of gait aid use (P<.001), balance disorders (P<.05), and transfer assistance and walk speed (P<.01). Comparatively, general fear of falling showed weaker associations and a sex bias. In study 2, staff administration was fairly consistent (ICC=.71) and similar associations emerged for AFC scores.

Conclusions

The AFC Scale shows good reproducibility, convergent and discriminative validity, and is feasible for clinical as well as research use.

Key Words: Fear, Rehabilitation, Reliability and validity, Risk assessment

 

PSYCHOLOGIC FALL CONCERNS are an important component of risk assessment and a modifiable, clinical indicator in prevention and rehabilitation interventions with community seniors.1, 2, 3, 4 Such concerns are quite common even in nonfallers, associated with balance and mobility impairments and activity curtailment, and predictive of functional decline, future fall risk, and nursing home admission.5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 A plethora of tools have been developed to measure this phenomenon, ranging from single-item, general fear-of-falling measures to multi-item scales encompassing a range of balance challenging situations.3 In general, this field of measurement has been criticized for: not involving older adults in tool development; not reporting the practicality or feasibility of administration; under-representing men and failing to examine sex differences; equating general fear of falling and falls self-efficacy; and not directly comparing measures.3 Self-efficacy and fear of falling are related, but distinct constructs.4 Compared with general fear of falling, the efficacy measures—notably the Falls Efficacy Scale (FES)5 and the higher difficulty Activities-specific Balance Confidence (ABC) Scale7—are theoretically based17 and, in head-to-head comparisons, show less evidence of sex bias and stronger relationships with gait aid use, physical balance and mobility tests, and activity curtailment in community seniors.5, 6, 8, 14

Although older adults living in residential care are at higher risk for falls18 and physical decline,19, 20 currently little is known about their related psychologic concerns. Challenges measuring subjective states in this population, together with the lack of appropriate tools, have hampered such investigations. The few studies that we have found that have assessed subjective fall concerns in nursing home seniors have used either single fear-of-falling measures19, 21 or modified versions of the FES,22, 23 without providing psychometric support for their use with this population or comparisons on the relative usefulness of these 2 distinct approaches.

A new tool—the Activities-specific Fall Caution (AFC) Scale—is now available for this population. As described elsewhere,24 the AFC Scale was developed inductively with residents and staff from nursing home and assisted living facilities, based on Bandura’s social cognitive theory.17 The content is specific to the residential care context and representative of progressively challenging situations pertaining to sitting and standing balance, postural change, and possible perturbations (eg, in crowded hallways or rooms). Resident terminology (eg, “just go” vs being more “cautious” or “careful” in particular situations) and qualifications (eg, whether alone and proximity of gait aid) were used to develop instructions and a color response card (similar to a traffic light). Those unable to respond appropriately to the 2 practice questions, invariably, could not complete the tool itself. Generally, residents with Mini-Mental State Examination (MMSE)25 scores of 12 or higher were able to respond. The final 13-item scale was shown to be internally consistent, unidimensional, and hierarchic in structure, with good person and item reliabilities.24

To be considered scientifically credible, the tool must also show evidence of test-retest reliability and amass support for construct validity, including replication with different samples.26, 27 To become widely used in fall risk assessments and clinical interventions, the tool must also be practical for different staff members to administer, score and interpret, and show evidence of interrater reliability.26 The 2 studies reported here were conducted to further examine the AFC Scale with respect to credibility and clinical feasibility, respectively. Specifically, the purposes of study 1 were to examine: (1) the reproducibility (test-retest reliability) of resident scores on 2 occasions; (2) the convergence of AFC scores with physical measures of balance and mobility, as well as the frequency and extent of actual mobility behavior; (3) the ability of AFC scores to discriminate between factors considered clinically important (use of gait aids, presence or absence of balance disorders, walking and transfer assistance, and other known fall risk indicators); and (4) to compare the convergent and discriminative abilities of the multi-item AFC Scale relative to a single-item general fear-of-falling measure. The purposes of study 2 were to examine: (1) the consistency (interrater reliability) and feasibility (ie, training and time requirements, ease of use) of staff administration of the AFC Scale; as well as (2) the similarity of findings (ie, associations with AFC scores) when the tool was administered by different assessors to a new sample.

Back to Article Outline

Methods 

Study 1 

Participants and procedures 

To examine the tool’s reliability and construct validity, we collected additional data from one of the field samples used in the structural examination of the AFC Scale by Pearce et al.24 As described by Pearce, this convenience sample of 101 residents was recruited from 5 nursing home and 5 assisted living facilities, following ethics approval. First, staff approached residents considered able to respond to a face-to-face interview (in English), weight bear, and walk at least 3m (10ft) (allowing for use of gait aids and personal assistance). The primary researcher then met, individually, with interested residents to obtain signed or verbal (tape recorded) study consent and permission to access their records.24 Additionally, 67 of these residents were asked to do 3 simple physical tests of walking and balance, taking about 15 minutes in total (3 people refused), and 44 were asked if we could return in a week or 2 to do the AFC Scale again (none refused). Thirty residents were asked to do both the physical tests and the AFC Scale again.

The primary researcher administered the AFC Scale to all 101 residents at least once, via interview, using practice questions to verify understanding, standardized prompts, and the laminated, color response card.24 Subsequently, the MMSE25 was administered and background information collected (from residents, their charts, or staff). During the interview, residents were asked the specific question: “Are you generally afraid or worried about falling (not at all, a little, very)?” Charts were examined for documentation of health problems (including dementia, depression, balance, and vision disorders), medications, and falls. Because falls were not systematically recorded by all the facilities,24 these data were not examined further. Resident use of gait aids, as well as personal assistance with walking or transfers (including supervision, cuing, or weight-bearing support), was observed and verified with staff.

Immediately after the interview, 64 residents underwent physical testing, usually in the facility’s activity room or, in a few instances, in hallways. Within 2 weeks (average, 9.9d), the same researcher readministered the AFC Scale to 44 residents. At the second administration, these participants were queried about any fall incidents or health changes over the interim. Thirteen staff members (6 nurses, 3 kinesiologists, 1 recreation therapist, 1 personal care worker, 2 others) were asked to complete the Nursing Home Life-Space Diameter (NHLSD)28 for the participating residents in their facilities.

Instruments 

We scored each of the 13 AFC items from 0 to 3 (0, go or green; 1, a little careful or yellow; 2, very careful or orange; 3, wait or stop or red). Total scores could range from 0 to 39 (higher scores denote greater caution due to worry or concern about falling).24 Of the 101 residents, 20 (19.8%) had “not applicable” values for at least 1 item (but no more than 2), that is, the resident responded that they never encountered the situation, and despite prompting, could not imagine themselves doing the activity. In such cases, the person’s median score on the items immediately above and below the one in question (in hierarchical order) was substituted for the not applicable response.24

The 3 physical assessments, conducted using normal gait aids, consisted of the Berg Balance Scale (BBS),29 the Timed Up & Go (TUG) test,30 and the Self-Paced Walk Test (SPWT).20 The BBS is a widely used, 14-task balance assessment tool with good psychometric properties.29 Total scores on the BBS can range from 0 (unable to do) to 56 (considered safe and independent)29; scores below 45 are highly specific (96%) in identifying community seniors prone to falling.31 The TUG assesses functional mobility by measuring how long it takes for a person to stand from a seated position, walk 3m, turn, walk back to the chair, and sit down. Similar to the BBS, the TUG has good psychometrics30 and is suitable for frail seniors.20 Scores above 14 seconds have been shown to be both sensitive (87%) and specific (87%) in predicting future fall risk in community seniors.32 Normal gait or walk speed in meters per second was assessed using the SPWT protocol adapted for residential care, that is, the middle 7m of a 13-m distance was timed to eliminate the effects of acceleration and deceleration.20 Prior research with frail community seniors has shown that walk speeds below 0.9m/s are associated with mobility impairments, lower self-efficacy, and more restricted mobility patterns.12, 33, 34

Each test was demonstrated and residents were given a practice trial, if needed, on the TUG. If a resident stopped or deviated from the path on either the TUG or the SPWT, the protocol was repeated.20 Only 2 subjects were unwilling or unable to do all 3 tests. One person was unable to complete the TUG despite repeated attempts, and another did not want to do any of the BBS tasks. Both, however, completed the other 2 tests.

To assess actual mobility patterns, we asked staff members familiar with residents to complete the NHLSD following the developers’ instructions.28 The NHLSD assesses the extent and frequency of movement over the previous 2 weeks in 4 zones: the resident’s room; outside the room or within the unit; outside the unit; and outside the facility. Frequency of movement for each zone is rated on a 6-point scale ranging from 0 (never) to 5 (>3 times/day).28 Initial evidence for test-retest and interrater reliability of the NHLSD28 has since been replicated in both nursing home and assisted living facilities.34 Primary raters were instructed to consult with staff in other parts of the facility to verify resident movement patterns outside the unit.34 Completed forms were returned for 80 residents (79% of the total sample).

Analysis 

We performed data analyses using SPSS,a with significance level set at .05. Descriptive statistics included frequencies, means, standard deviations (SDs), and ranges, with related 95% confidence intervals (CIs). To examine test-retest reliability of the AFC Scale, the intraclass correlation coefficient (ICC2,1) was calculated using a 2-way analysis of variance (ANOVA), considering administrators as a random effect.35 ICC values above 0.9 were considered high, .80 to .89 good, and .70 to .79 fair.36

Because AFC scores were not normally distributed for this sample, we examined the associations with physical test scores and mobility patterns (NHLSD scores), using Spearman ρ coefficients. With respect to resident characteristics and clinically important factors (use or nonuse of gait aids; presence or absence of balance disorders; assistance or no assistance with transfers and walking; high or low fall risk and fast or slow gait speed), differences in mean AFC score and presence or absence of general fear of falling (scored as some vs none) were examined using Mann-Whitney U or chi-square tests, respectively. The sample was classified as high versus low fall risk according to established BBS (−45 vs ≥45)31 and TUG (>14s vs ≤14s)32 cutoffs. Residents were also categorized as “fast” (>0.9m/s) versus “slow” (≤0.9m/s) walkers based on their scores on the SPWT.

Similar to community studies, we expected to find good convergence of the AFC Scale for residential living seniors (from both nursing home and assisted living facilities) with physical measures and mobility patterns. Specifically, we hypothesized that higher AFC scores (indicating greater caution due to worry or concern about falling) would be associated with poorer physical performance and more restricted mobility patterns. We also postulated that scores on the AFC Scale would distinguish between groups classified by known indicators of mobility and balance impairment; AFC scores were expected to be higher in persons who used gait aids, had balance disorders, and received personal assistance with walking or transfers, as well as those considered to be “slow” walkers, or at high fall risk (based on BBS and TUG scores). Comparatively, we presupposed weaker associations and poorer discrimination for the single-item, general fear-of-falling measure. We also expected to find a significant gender difference for general fear of falling (with women more likely to acknowledge such fear), but not for AFC scores.

Study 2 

Participants and procedures 

Prior to undertaking this study, we asked 5 staff members to administer the AFC Scale to 1 or 2 residents to obtain feedback on each step of the protocol (explaining the tool to residents using the color response card, doing the practice questions to determine whether to proceed, and then administering the tool, using the standardized prompts as needed). This pilot group verified that the written protocol was clear and suitable for clinical use.

Study 2 took place several months after study 1. Although some of the same facilities were approached again, we recruited new convenience samples of 31 residents and 17 staff members. A brief training session (≈10min) to review the administration and scoring of the AFC Scale was held for staff members from each of the 6 facilities (1 or 2 staff attended each session). These staff, in turn, recruited further volunteers among their colleagues, explained the tool and distributed copies of the accompanying protocol and laminated color response card. The 17 staff members who agreed to participate in study 2 consisted of 9 kinesiologists, 2 recreation therapists, 4 physiotherapy and kinesiology assistants, and 2 personal care workers.

Staff used the same criteria and process used in study 1 to recruit resident volunteers. Two different staff members administered the AFC Scale to each volunteer on 2 occasions at their convenience (but no longer than 48h apart). Assessors were asked not to discuss resident performance when coordinating readministration of the AFC Scale. They were instructed, however, to ask the resident if he/she was having a good day, and to record this information (as well as recent falls or health incidents) on the form itself, in addition to time for administration. Subsequently, the primary researcher met with participating staff members to discuss ease of administration and scoring of the AFC Scale, and to obtain information on residents (eg, use of gait aids). The researcher also met with participating residents to administer the MMSE25 and the general fear-of-falling question, and obtain background and chart information, similar to study 1.

Instruments 

The 13-item AFC Scale was successfully administered to all 31 residents on 2 occasions. For the 2 residents (6.5%) with “not applicable” responses (each to a single item), the person’s median score was substituted for scoring, as described previously. Scoring of the MMSE, the general question on fear of falling, resident background and chart information (eg, use or nonuse of gait aids) was the same as for study 1.

Analysis 

Staff members were not aware of residents’ scores on other measures such as the MMSE or the general fear-of-falling question and results were not analyzed until all data collection was completed.

To examine interrater reliability of the AFC Scale, the ICC1,1 was calculated using a 1-way ANOVA. This ICC model is appropriate for single ratings, in which subjects are rated by multiple administrators, who do not, in turn, rate all subjects.35

Because AFC scores were normally distributed for this sample, we examined associations with resident characteristics with Pearson correlation coefficients. The correlation between AFC and general fear-of-falling scores was examined via the Spearman ρ. We used t tests and chi-square tests to compare AFC and fear-of-falling scores, respectively, for men versus women, gait aid users versus nonusers, persons with and without balance disorders, and so forth.

Back to Article Outline

Results 

Study 1 

Resident characteristics 

Characteristics of the total sample, as well as each subgroup, are shown in table 1. All participants were ambulatory, although the majority used a walking aid (most commonly a walker), with some using wheelchairs for long distances. Compared with assisted living residents, the nursing home group had lower cognitive functioning (as evident by poorer MMSE scores and greater incidence of diagnosed dementia in their charts, P<.01), and were more likely to have balance disorders and require assistance with transfers (P<.001).

Table 1. Characteristics of Resident Samples
CharacteristicStudy 1 (N=101)Study 2 (N=31)
TotalTest-Retest Assessment (n=44)Physical Assessment (n=64)Mobility Pattern Assessment (n=80)Total
Facility type
Nursing home52(51.5)29(65.9)33(51.6)43(53.8)26(83.9)
Assisted living49(48.5)15(34.1)31(48.4)37(46.3)5(16.1)
Women62(61.4)25(56.8)37(57.8)48(60.0)20(64.5)
Depression17(17.0)9(20.5)10(15.6)13(16.5)5(17.2)
Dementia26(26.0)10(22.7)16(25.0)24(30.4)7(24.1)
Vision problems32(31.7)15(34.1)19(29.7)28(35.0)12(41.4)
Vertigo14(13.9)4(9.1)9(14.1)11(13.8)5(17.2)
Balance disorders23(22.8)9(20.5)16(25.0)20(25.0)11(37.9)
Use gait aid71(70.3)30(68.2)42(65.6)55(68.8)23(74.2)
Transfer assistance⁎14(14.1)5(11.4)5(7.8)11(14.1)10(34.5)
Walking assistance⁎3(3.0)0(0.0)0(0.0)3(3.8)6(20.7)
Fear of falling
A little/very50(50.5)19(43.2)31(49.2)40(51.3)12(41.4)
None49(49.5)25(56.8)32(50.8)38(48.7)17(58.6)
Age (y)82.6±7.384.6±6.482.5±7.983.2±6.681.3±8.4
MMSE score24.2±5.2 (5−30)25.7±3.9 (15−30)24.2±5.6 (5−30)24.4±4.9 (10−30)22.7±5.9 (10−30)
AFC score13.1±7.8 (0−29)12.5±7.9 (0−29)13.9±7.6 (0−29)13.0±7.6 (0−29)17.5±8.0 (2−34)
95%CI11.6−14.010.1−14.912.0−15.811.4−14.714.6−20.5

NOTE. Values are n (%) or mean ± SD (range).

⁎Any personal assistance including supervision, cuing, or weight-bearing support.

The subgroups were comparable with the rest of the sample, with a few exceptions. The test-retest group had higher MMSE scores (P<.01) and were older (P<.05). Those who underwent physical testing were less likely to require transfer (P<.05) and/or walking assistance (P<.05).

Reliability of the AFC Scale 

Average AFC scores were similar at each time point (time 1, 12.8±8.2; time 2, 12.3±7.7) and showed good consistency over 7 to 14 days (ICC2,1=.87; 95% CI, .78−.93). Three residents had scores greater than ±10 scale points apart between time 1 and time 2. The resident with the largest discrepancy (−15, less cautious at time 2) fell the day before the first administration of the tool. This person’s fall may also have affected his best friend, who scored 11 points higher (more cautious) at time 1 compared with time 2.

Convergent validity 

Scores for each of the physical performance measures, as well as mobility patterns (NHLSD scores), are shown in table 2. Residents from nursing home and assisted living facilities differed significantly on only one of these measures; NHLSD scores were higher for the assisted living group (P<.01). Only 13 (20%) residents in total had normal walk speeds above 0.9m/s. Most of the sample (95%) was classified as being at “high” fall risk according to either the TUG or the BBS cutoff scores, whereas over half (57%) were considered at risk according to both indicators.

Table 2. Physical Measures, Fall Risk Classification, and Volitional Mobility Scores
MeasuresTotalNursing HomeAssisted Living
SPWT (n)643331
Mean ± SD (m/s)0.7±0.20.7±0.3>0.7±0.2
Range (m/s)0.3−1.20.3−1.20.3−1.2
95% CI.6−.7.6−.8.6−.8
TUG (n)633231
Mean ± SD (s)24.6±10.326.5±15.322.6±9.1
Range (s)9.8−74.09.8−74.012.4−42.6
95% CI21.5−27.821.2−31.819.4−25.9
High fall risk, n (%)53 (84.1)26 (81.3)>27 (87.1)
BBS (n)633330
Mean ± SD39.8±24.638.6±10.641.2±10.0
Range6.0−56.06.0−56.07.0−52.0
95% CI37.3−42.434.9−42.237.7−44.8
High fall risk, n (%)39 (61.9)23 (69.6)16 (53.3)
NHLSD (n)804337
Mean ± SD34.9±7.732.4±7.337.9±7.2
Range18−5018−5021−50
95% CI33.1−36.330.6−34.335.6−40.2

As expected, BBS scores were inversely related to TUG scores (ρ=−.67, P<.001) and positively related to SPWT scores (ρ=.68, P<.001); better balance was related to better functional mobility and faster self-paced walking. Not surprisingly, TUG scores and walk speed were highly related (ρ=−.85, P<.001); poorer functional mobility was associated with slower walk speed. Physical scores were moderately correlated with NHLSD scores (ρ range, −.41 to .45; P<.01); better balance, functional mobility and faster walk speed were all associated with greater frequency and extent of mobility behavior.

As shown in table 3, scores on the AFC scale were significantly related to each of the physical measures in the expected directions for the sample as a whole, and for both groups of residents (from nursing home and assisted living facilities, respectively). Greater falls-related caution (higher AFC scores) was also associated with more restricted mobility behavior (lower NHLSD scores).

Table 3. Correlations of AFC and Fear-of-Falling Scores With Physical and Behavioral Measures
TestAFCFear of Falling
TotalNursing HomeAssisted LivingTotalNursing HomeAssisted Living
TUG (s).50⁎.52†.41†.31‡.16.46†
BBS scores−.43⁎−.48†−.38‡−.25‡−.37‡−.12
Walk speed (m/s)−.52⁎−.56⁎−.42†−.31†−.23−.48†
Mobility patterns−.45⁎−.42†−.34‡−.17−.23−.04

NOTE. Walk speed as measured by the SPWT; mobility patterns as measured by the NHLSD.

⁎P<.001;

†P<.01;

‡P<.05.

Discriminant validity 

As shown in table 4, mean AFC scores were significantly higher (indicating greater caution) for gait aid users (vs nonusers), residents with balance disorders, those who required transfer assistance, those who walked slower, and those classified as being at high fall risk according to cutoff scores on the BBS and the TUG, respectively. Walking assistance was not analyzed due to the small number of residents who required such assistance (n=3).

Table 4. Discrimination Between Clinically Important Factors
GroupsN (%)Mean AFC Score ± SDMann-Whitney UFear of Falling§Chi-Square Test
NoneSome
Gait aid use −3.8⁎ 4.2‡
Yes71(70.3)15.0±7.6 3040
No30(29.7)8.6±6.3 1910
Balance disorders −1.9‡ .83
Yes23(22.8)15.8±7.3 913
No78(77.2)12.3±7.8 4037
Transfer assistance −3.0† .00
Yes14(14.1)18.8±6.5 77
No85(85.9)12.1±7.7 4242
Walk speed −3.0† 6.3†
Slow (≤0.9m/s)51(79.7)15.4±7.1 2229
Fast (>0.9m/s)13(20.3)7.9±6.9 102
Fall risk (BBS) −3.4⁎ 5.6‡
High (<45)39(61.9)16.7±6.8 1524
Low (≥45)24(38.1)9.8±6.9 167
Fall risk (TUG) −2.2‡ 1.6
High (>14s)53(84.1)14.8±7.7 2527
Low (≤14s)10(15.9)9.1±5.6 73

⁎P<.001;

†P<.01;

‡P<.05.

§Data missing for 2 participants.

Comparison of AFC and general fear-of-falling scores 

Scores on the AFC Scale were moderately related to presence or absence of fear of falling (ρ=.40, P<.001). Although women were significantly more likely than men (59% vs 37%) to acknowledge some fear of falling (P<.05), no sex difference emerged for AFC scores. Other sample characteristics, including: resident age, MMSE scores, presence or absence of diagnosed dementia and depression, number of health problems and medications, as well as type of residential living (nursing home vs assisted living), were not related to either AFC or general fear-of-falling scores.

As shown in table 3, fear of falling was not as strongly associated with the physical measures (lower correlations) and associations were not consistent across both the nursing home and assisted living groups. Moreover, no significant association emerged between fear of falling and resident mobility patterns, as measured by the NHLSD. The AFC Scale also showed better discriminative abilities with respect to several clinically important indicators of balance and mobility impairments. As shown in table 4, general fear of falling was able to discriminate between groups on only 3 of the 6 factors (gait aid use, walk speed, and fall risk, according to BBS cutoff scores).

Study 2 

Resident characteristics 

As shown in table 1, the majority (84%) of study 2 participants were nursing home residents. A significantly greater proportion of this sample required assistance with transfers (P<.05) and walking (P<.001). Documented balance disorders, vertigo, and vision problems were also higher in this sample, although not significantly. The 2 samples were comparable with respect to age, sex, recorded dementia and depression, as well as MMSE scores.

Reliability and feasibility of the AFC Scale 

On average, staff took 8.4 minutes to administer the AFC Scale to residents and reported few problems administering the tool or scoring the items using the protocol provided. Mean AFC scores were similar at time 1 (17.5±8.0) and time 2 (18.8±6.8) and interrater reliability was acceptable (ICC1,1=.71; 95% CI, .49−.85). The average difference in AFC scores between the 2 administrations was +4 (range, −10 to 15). According to staff, the resident with the greatest discrepancy (+15) was not having a good day at the time of the second administration and expressed some discomfort with the second rater (a kinesiology student).

Associations of AFC and general fear-of-falling scores 

Similar to study 1, AFC and fear-of-falling scores were moderately related (ρ=.37, P<.05). Neither AFC nor fear-of-falling scores were related to age, recorded depression or dementia, or MMSE scores. No sex difference emerged for either the AFC or the fear-of-falling scores for this sample.

The residents in this study had significantly higher mean AFC scores (P<.05) than the first sample (values are shown in table 1), although fewer residents expressed general fear of falling (41% vs 51%). Mean AFC scores were significantly higher (P<.05) for residents with documented balance disorders (21.5±8.6; 95% CI, 15.8–27.3) compared with those without such disorders (15.2±7.0; 95% CI, 11.8–18.7). Residents who used gait aids or needed transfer or walking assistance also had higher AFC scores, although not statistically significant. There were no apparent differences in fear of falling for any of these group comparisons.

Back to Article Outline

Discussion 

The present findings support the credibility, as well as the clinical feasibility, of the newly developed AFC Scale. Resident scores were consistent over 1 to 2 weeks when administered by the researcher (study 1) and acceptably reliable when administered and scored by different staff members (study 2). Staff reported little difficulty administering and scoring the AFC Scale on their own, using the guided protocol. Minimal training and time for administration was required, and costs were negligible for printing the tool and laminating the response card. Although critical for clinical adoption, such practical considerations are often not addressed by scale developers.3, 26, 27 Some of the participating facilities have already incorporated the AFC Scale into their routine fall risk assessment protocol for newly admitted residents and plan to use the tool as an outcome indicator in their rehabilitation interventions.

In addition to reliability and practicality, however, an instrument must also show evidence of construct validity—that the tool assesses what it purports to measure. In the absence of a gold standard (criterion validity), subjective rating scales must amass support through an ongoing process of hypothesis testing regarding the underlying construct.26, 27 Based on Bandura’s social cognitive theory,17 the AFC Scale was inductively developed with residents and staff, and the content is representative of progressively challenging activities pertaining to sitting and standing balance, postural change, and possible perturbations (eg, moving around crowded hallways or rooms).24 The AFC Scale assesses the extent of caution or carefulness expressed by residents due to worry or concern about falling across 13 situations frequently encountered in residential care environments.

As hypothesized, scores on the AFC Scale showed good convergence with all 3 physical measures. Greater fall-related caution was associated with poorer balance (BBS scores), more impaired functional mobility (TUG scores), and slower walking (SPWT). Not surprisingly our sample scored lower on these measures than community seniors8, 12, 14, 15, 16; comparable to the performance by other residential seniors.20 The pattern of associations with subjective fall concerns assessed by the AFC Scale, however, was consistent with previous findings using the ABC Scale.8, 12, 14, 15, 16

As expected, greater fall caution was also associated with more restricted behavior in both nursing home and assisted living residents, even though the latter group had more extensive mobility patterns. This association is consistent with Bandura’s theory17 and has been documented in several community studies5, 6, 8, 10, 13 and in 1 prior nursing home study.19 In community studies, activity curtailment has been assessed via self-reports. In residential care settings, the NHLSD provides an objective and reliable method for quantifying frequency and extent of resident mobility patterns.28, 34

Scores on the AFC Scale were also able to differentiate residents on the basis of several clinically important factors, including: use of gait aids, presence of balance disorders, need for transfer assistance, walk speed, and fall risk. Compared with community samples, a substantially greater proportion of our residents was classified as slow walkers (<0.9m/s),12 and at high fall risk according to cutoff scores on both the BBS and the TUG.14 The ability of AFC scores to discriminate between these groups was similar to the ABC and FES tools,12, 14 suggesting that this tool may be useful for identifying residential living seniors who are prone to falling and at risk for functional decline.

As noted at the outset, we have found only 4 previous studies that have assessed subjective fall concerns in residential care; 2 using single item fear-of-falling measures,19, 21 and 2 using the multi-item FES.22, 23 Franzoni et al19 found that fear of falling was not related to fall history, but was predictive of future falls, functional decline, and reduced indoor activity level at 24 months (after controlling for baseline functioning, psychotropic drug use, and chronic symptoms). At the time of their study, use of the binary fear-of-falling measure was the norm; other tools were not available. Although prospective studies are necessary to examine the ability of AFC Scale scores to predict declines in resident functioning, activity curtailment, and future fall risk, the present findings indicate that the AFC Scale is a better tool than the general fear-of-falling measure for future research with this population.

Consistent with community studies which have directly compared fear-of-falling and the multi-item falls efficacy measures,5, 6, 8, 14 fear-of-falling and AFC scores were found to be moderately related. We also found a sex bias for fear of falling, with women significantly more likely to acknowledge such fear. As expected, general fear of falling showed weaker associations with the physical measures and poorer discrimination of groups based on known indicators of mobility and balance impairment. General fear of falling was not related to resident mobility patterns, consistent with Bandura’s framework that global states such as fear or anxiety are often poor predictors of behavior.17 Global fear-of-falling measures are not theoretically based,3 and limited in their ability to detect varying levels of the construct across different circumstances or change over time.5, 6, 8, 14, 15 In general, a score derived from aggregate responses to multiple items is a better indicator of the “true score” than the response to any single statement.27

Our second study provided the opportunity to re-examine some of the relationships when the AFC Scale was administered by different assessors (clinicians vs researchers) to a new convenience sample of residents. Similar to study 1, AFC and fear-of-falling scores were found to be moderately related. Associations with resident characteristics were also similar in the 2 studies, except that a significant sex bias did not emerge for fear of falling. Mean AFC scores were significantly higher for residents with documented balance disorders. The second sample had significantly higher mean AFC scores than the first sample, consistent with the findings that this group was also significantly more likely to require assistance with transfers or walking.

The second sample was small and comprised predominantly nursing home residents, precluding comparisons of nursing home and assisted living groups. Even though both studies were limited to convenience samples of English-speaking residents, unlike many prior samples,3 men were sufficiently represented. It is possible that staff approached higher functioning residents for study participation, particularly in the nursing homes, although residents presented with a wide range of cognitive and physical abilities (including some with documented dementia or depression). Cognition was assessed using the MMSE25; however, depression was not directly measured. Depression has been previously related to subjective fall concerns in a community sample,12 but not in a nursing home sample.19 Other limitations include the cross-sectional design and the fact that the researchers were not blind to study hypotheses.

Amassing support for a scale, however, is an ongoing process. For instance, several studies were conducted to establish the ABC Scale’s reliability, construct validity, and sensitivity to change.1, 7, 8 Studies by other researchers have further supported the credibility and usefulness of the ABC Scale for community seniors in general,3, 12, 14 and clinical populations.15, 16 Simply put, confidence in a measure increases as the tool is applied in a number of settings with different populations.26, 27

Back to Article Outline

Conclusions 

Previously, we showed that the AFC Scale has good internal consistency and structural properties (unidimensionality, hiearchicality).24 The present findings provide support for test-retest and interrater reliability, as well as convergent and discriminant validity. Further studies, with other samples, are needed to verify and extend the psychometric evidence for the AFC Scale, including the tool’s sensitivity to change and its ability to predict declines in resident functioning, activity curtailment and future falls. Studies have shown that functional decline is not inevitable and that physical abilities of residents can be significantly improved with targeted interventions.20 It is equally important, however, that prevention and rehabilitation studies assess psychologic fall concerns as an important mediator of activity curtailment and functional decline.1, 2, 3, 4 The AFC Scale provides a promising new tool for quantifying subjective fall concerns in residents of both nursing home and assisted living facilities that is practical and reliable for use by clinicians, as well as by researchers.

Supplier

Back to Article Outline

References 

  1. Myers AM, Fletcher PC, Myers AH, Sherk W. Discriminative and evaluative properties of the Activities-specific Balance Confidence (ABC) Scale. J Gerontol Med Sci. 1998;53:M287–M294
  2. Brouwer BJ, Walker C, Rydahl SJ, Culham EG. Reducing fear of falling in seniors through education and activity programs: a randomized trial. J Am Geriatr Soc. 2003;51:829–834
  3. Jørstad EC, Hauer K, Becker C, Lamb SE. ProFaNE Group (Measuring the psychological outcomes of falling: a systematic review). J Am Geriatr Soc. 2005;53:501–510
  4. Li F, Fisher KJ, Harmer P, McAuley E. Falls self-efficacy as a mediator of fear of falling in an exercise intervention for older adults. J Gerontol B Psychol Sci Soc Sci. 2005;60:34–40
  5. Tinetti ME, Richman D, Powell LE. Falls efficacy as a measure of fear of falling. J Gerontol. 1990;45:P239–P243
  6. Tinetti ME, Mendes de Leon CF, Doucette JT, Baker DI. Fear of falling and fall-related efficacy in relationship to functioning among community-living elders. J Gerontol. 1994;49:M140–M147
  7. Powell LE, Myers AM. The Activities-specific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1995;50:M28–M34
  8. Myers AM, Powell L, Maki B, Holliday PJ, Brawley LR, Sherk W. Psychological indicators of balance confidence: relationship to actual and perceived abilities. J Gerontol A Biol Sci Med Sci. 1996;51:M37–M43
  9. Mendes de Leon CF, Seeman TE, Baker DI, Richardson ED, Tinetti ME. Self-efficacy, physical decline, and change in functioning in community living elders: a prospective study. J Gerontol B Psychol Sci Soc Sci. 1996;51:S183–S190
  10. Howland J, Lachman M, Peterson EW, Cote J, Kasten L, Jette A. Covariates of fear of falling and associated activity curtailment. Gerontologist. 1998;38:549–555
  11. Cumming RG, Salkeld G, Thomas M, Szonyi G. Prospective study of the impact of fear of falling on activities of daily living, SF-36 scores, and nursing home admission. J Gerontol A Biol Sci Med Sci. 2000;55:M299–M305
  12. Kressig RW, Wolf SL, Sattin RW, et al. Associations of demographic, functional, and behavioural characteristics with activity-related fear of falling among older adults transitioning to frailty. J Am Geriatr Soc. 2001;49:1456–1462
  13. Yardley L, Smith H. A prospective study of the relationship between feared consequences of falling and avoidance of activity in community-living older people. Gerontologist. 2002;1:17–23
  14. Hatch J, Gill-Body KM, Portney LG. Determinants of balance confidence in community-dwelling elderly people. Phys Ther. 2003;83:1072–1079
  15. Miller WC, Deathe AB, Speechley M. Psychometric properties of the Activities-specific Balance Confidence scale among individuals with a lower-limb amputation. Arch Phys Med Rehabil. 2003;84:656–661
  16. Botner EM, Miller WC, Eng JJ. Measurement properties of the Activities-specific Balance Confidence Scale among individuals with stroke. Disabil Rehabil. 2005;27:156–163
  17. Bandura A. Social foundations of thought and action: a social cognitive theory. Englewood Cliffs: Prentice Hall; 1986;
  18. Rubenstein LZ, Josephson KR, Osterweil C. Falls and fall prevention in the nursing home. Clin Geriatr Med. 1996;12:881–902
  19. Franzoni S, Rozzini R, Boffelli S, Frisoni GB, Trabucchi M. Fear of falling in nursing home patients. Gerontology. 1994;40:38–44
  20. Lazowski DA, Ecclestone NA, Myers AM, et al. A randomized outcome evaluation of group exercise programs in long-term care institutions. J Gerontol A Biol Sci Med Sci. 1999;54:M621–M628
  21. Schnelle JF, Mac Rae PG, Simmons SF, et al. Safety assessment for the frail elderly: a comparison of restrained and unrestrained nursing home residents. J Am Geriatr Soc. 1994;42:586–592
  22. Schoenfelder DP. A fall prevention program for elderly individuals (Exercise in long-term care settings). J Gerontol Nurs. 2000;26:43–51
  23. Gentleman B, Malozemoff W. Falls and feelings: description of a psychosocial group nursing intervention. J Gerontol Nurs. 2001;27:35–39
  24. Pearce NJ, Myers AM, Blanchard RA. Assessing subjective fall concerns in residential living seniors: development of the Activities-specific Fall Caution Scale. Arch Phys Med Rehabil. 2007;88:724–731
  25. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–198
  26. Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. New York: Oxford Univ Pr; 1989;
  27. Williams JI, Naylor CD. How should health status measures be assessed? (Cautionary notes on procrustean frameworks). J Clin Epidemiol. 1992;45:1347–1351
  28. Tinetti M, Ginter S. The Nursing Home Life Space Diameter: a measure of extent and frequency of mobility among nursing home residents. J Am Geriatr Soc. 1990;45:239–243
  29. Berg KO, Maki BE, Williams JI, Holliday PJ, Wood-Dauphinee SL. Clinical and laboratory measures of postural balance in an elderly population. Arch Phys Med Rehabil. 1992;73:1073–1080
  30. Podsiadlo D, Richardson S. The timed “Up & Go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39:142–148
  31. Bogle Thorbahn LD, Newton RA. Use of the Berg Balance Test to predict falls in elderly persons. Phys Ther. 1996;76:576–584
  32. Shumway-Cook A, Brauer S, Woollacott M. Predicting the probability for falls in community-dwelling older adults using the Timed Up & Go Test. Phys Ther. 2000;80:896–903
  33. May D, Nayak U, Isaacs B. The life-space diary: a measure of mobility in old people at home. Int Rehabil Med. 1985;7:182–186
  34. Myers AM, Cyarto EV, Blanchard R. Challenges in quantifying mobility in frail older adults. Eur Rev Aging Phys Act. 2005;2:13–21
  35. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater ability. Psychol Bull. 1979;86:420–428
  36. Kachingwe AF, Phillips BJ. Inter- and intrarater reliability of a back range of motion instrument. Arch Phys Med Rehabil. 2005;86:2347–2353
  • a SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.

 No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated.Reprints are not available from the author.

PII: S0003-9993(07)00196-7

doi:10.1016/j.apmr.2007.03.003

Archives of Physical Medicine and Rehabilitation
Volume 88, Issue 6 , Pages 732-739, June 2007

Article Tools