Functional independence describes the capacity of an individual to carry out activities of daily life. In the aging population, chronic or acute pathological processes can cause impairments leading to functional limitations. People older than 65 yrs with lower functional independence make disproportionate use of hospital services (40% of acute hospital stays), continuing care services (95% of people in residential care and 85% of people in hospital-based continuing care), and home care services (82% of home care clients).1
Understanding the functional independence of individuals can assist with determining the amount of assistance needed later in life,2 but there are a variety of ways to measure it. Many functional assessment scales capture performance with basic activities of daily living (ADLs) (e.g., bathing, grooming, dressing, eating), mobility (e.g., indoor/outdoor locomotion, transferring), and instrumental ADLs (e.g., housework, preparing meals). Some of these assessments, such as the Functional Autonomy Measurement System (SMAF), are administered by health professionals or trained raters who obtain information about individuals' functional performance using a combination of observation and questioning.3 In certain settings, however, the reliance on healthcare professionals or trained raters limits the measurements' practicality, especially when longitudinal data are being collected. A less resource-intense option involves use of self-report measures such as the self-reported Functional Independence Measure (FIM-SR),4 which was derived from the original, clinician-scored FIM to expand its utility in research settings.5 The FIM-SR requires individuals to rate their own level of independence performing basic and instrumental activities, and it has been shown to be valid for assessing functional independence when compared with the original FIM4 and other self-reported measures of functional independence.6,7
A third approach to measuring functional independence involves the use of proxy respondents, for circumstances when individuals are unable or unavailable to self-report. Proxy respondents for older adults typically include family, friends, and caregivers and can also include physicians and health professionals. Previous research suggests that agreement between participant and proxy reports is affected by factors including their relationship8 as well as the domain of the outcome measure9; for example, agreement is generally higher when the participant and proxy have a close relationship and when the dimensions being measured are more observable (i.e., physical dimensions are better than cognitive/emotional dimensions).
Given the nature of their relationship and the amount of time spent providing care, it is reasonable to assume that family caregivers might have the best judgment of their care recipients' performance. Whereas clinicians tend to evaluate the capacity of an individual within the brief time frame of one or more assessments, family caregivers are able to report about individuals' functional abilities in life because they are often the ones providing assistance.
The present study sought to examine the agreement between a self-report version of an FIM that assessed physical functioning and the observations of family caregivers and to determine how the care recipient and caregiver reports correlated with an evaluator evaluation over time. This study adds to the existing literature of addressing whether caregivers are suitable proxy respondents when care recipients are unavailable by looking a heterogeneous population of community-dwelling older adults with disabilities.
Participants and Setting
This study drew on data obtained from a randomized controlled trial examining the effects of a caregiver-inclusive intervention on the outcomes of 90 care recipient and family caregiver dyads in three Canadian regions.10 Ethical approval was obtained from the ethics review boards at each site. All participants provided informed written consent.
Eligible care recipients were 55 yrs or older, lived at home, had one or more limitations with mobility or ADLs, had been referred for homecare services, and were deemed able to provide their own consent by homecare staff. Eligible family caregivers provided 4 or more hours of assistance per week and were able to provide their own consent. Participants needed to be able to speak either French or English.
The outcomes measures, which were available in French or English, were completed by the participants in the language of their choice. The functional independence of care recipients was reported by both the care recipients and their family caregivers using the physical functioning scales of an FIM-SR.6 The independence of care recipients on tasks involving self-care (6 items), sphincter control (2 items), transfers (3 items), and locomotion (2 items) was reported; each item was rated on a seven-point response scale, with 1 representing complete dependence and 7 indicating total independence. The social functioning scales (i.e., communication, social cognition, and cognition scales) were not included in the original study because they were unlikely to be affected by the interventions. The caregiver FIM-SR was identical to the care recipient version, with minor rephrasing of the instructions.
The revised version of the SMAF was used as an additional measure to assess care recipients' autonomy performing activities from an evaluator's perspective.3 Two subscales concerning mobility (6 items) and ADLs (7 items) were applied; each subscale was rated on a five-point scale, ranging from −3 (total dependence) to 0 (complete independence). Both the SMAF and the FIM-SR evaluate subjects according to what they do and not what they could do in other environments or circumstances.
Outcome measures were administered by bilingual research assistants to participants at baseline and 6, 22, and 58 wks after the interventions. The FIM-SRs were administered separately to care recipients and family caregivers, and scores were never shared. The SMAF was subsequently administered by trained data collectors who scored the measures based on information obtained from the care recipient and family caregiver, as well as their personal observations. Sociodemographic data were collected for care recipients and caregivers and included items such as age, sex, level of education, relationship of the dyad, health status, duration of functional problems (care recipient), duration of caregiving (years), and hours of caregiving provided per week. The Montreal Cognitive Assessment (MOCA) was used to measure cognitive impairment; a score of 26 or higher was considered as normal.11
To describe the samples, continuous variables were expressed as means and nominal variables were expressed as frequencies and percentages. Care recipient FIM-SR, caregiver FIM-SR, and SMAF scores were assessed for normal distribution using the one-sample Kolmogorov-Smirnov test and visual inspection of histograms. Statistical analyses were performed using IBM SPSS Statistics, Version 23 (IBM Corp, Armonk, NY). Values of P < 0.05 were considered to be statistically significant.
Given departures from normality, Spearman rank correlations were conducted between care recipient FIM-SR, caregiver FIM-SR, and SMAF outcomes across all time points, to determine the strength and direction of the monotonic relationship between the variables. The study examined correlations of four physical functioning categories as well as total scores: self-care, sphincter control, transfers, and locomotion. This was accomplished by correlating FIM-SR self-care, sphincter control, transfers, and locomotion subscales with the associated SMAF items; for example, FIM-SR “self-care” items were correlated with SMAF ADL items “eating,” “washing,” “dressing,” “grooming,” and “toileting.”
Correlation explores the association between outcome measures, but it does not assess the differences between measures, and high correlation does not guarantee good agreement.12 Therefore, Bland-Altman plots were constructed to examine the agreement between care recipient FIM-SR and caregiver FIM-SR scores by plotting the difference of care recipient FIM-SR and caregiver FIM-SR total scores against the mean of care recipient and caregiver FIM-SR total scores.13 The value of the mean difference provides an estimate of bias and the 95% limits of agreement (mean + 1.96 SD, mean – 1.96 SD) provide an estimate of the influence of random variation.
The baseline characteristics of the care recipients and their family caregivers are presented in Table 1. The mean age of the recipients and their caregivers was 75 and 65 yrs, respectively. The care recipients had functional limitations from various conditions, many of them relating to osteoarthritis and neurological diseases. Most of the family caregivers were spouses and the majority lived with the care recipients.
Table 2 displays the correlation coefficients between the associated categories of the care recipient FIM-SR, caregiver FIM-SR, and the SMAF. The coefficients indicate moderate to very strong correlations between corresponding categories of the measures (self-care, sphincter control, transfers, locomotion; rS = 0.45–0.91) as well as the total scores (rS = 0.79–0.91) across all time points. All correlations were significant (P < 0.01), but in general, the self-care category produced the strongest correlations (rS = 0.78–0.91), whereas the transfers category produced marginally weaker correlations (rS = 0.45–0.82).
Figure 1 depicts the Bland-Altman plots showing the agreement between care recipient FIM-SR and caregiver FIM-SR total scores at all time points. The mean difference between care recipient FIM-SR and caregiver FIM-SR total scores ranged narrowly across the four time points (mean difference = 2.00–2.97). In addition, the mean difference scores at each time point differed significantly from exact agreement, or y = 0.00 units (one-sample t test, P < 0.001). At baseline, the two measures exhibited better agreement (i.e., difference scores approaching y = 0.00 units) when functional independence was rated as poor or very good, compared with midrange scores, which were markedly more variable. The smallest variation in agreement between the care recipient and caregiver FIM-SRs occurred at weeks 22 and 58 (95% limits of agreement = −7.1 to 11.7) and the largest variation in agreement occurred at baseline (95% limits of agreement = −11.5 to 17.5). Three participants were outliers at more than one time point. Among all outliers, the average scores from the MOCA was 20.8 (SD ±5.6).
This study contributes to a growing body of research examining the associations among different perspectives of functional independence. The findings demonstrate that the care recipient FIM-SR, caregiver FIM-SR, and SMAF are significantly correlated. The care recipient and caregiver perspectives on care recipients' functional independence displayed good agreement with only a small bias across all time points. Although participants needed to be able to provide their own consent, the majority of them (66%) did not have normal MOCA scores. The average score for the overall group is closer to that reported for people with mild cognitive impairment (22.1),11 suggesting the results do generalize to that population.
The correlations in the present study support previous research, which determined that SMAF and FIM assessments, had strongly correlated total scores (r = 0.93–0.95) and corresponding category scores (r = 0.87–0.96) when applied to participants undergoing poststroke rehabilitation.14 Compared with Desrosiers et al.,14 this study found slightly more variation among the correlation coefficients. Desrosiers et al.14 had clinicians deliver both the FIM and SMAF, which might partially explain the narrower range of coefficients. Desrosiers et al.14 attributed the strong correlations to the physical components of functional independence (mobility, locomotion, sphincter control, and self-care), as well as to the similarity of items in the corresponding categories between outcome measures.
The findings of this study provide support for the compatibility of the FIM-SR in a heterogeneous population of community-dwelling adults with physical disabilities. Previous studies examined the validity of FIM-SR in populations of individuals with specific diagnoses or a narrower scope of functional limitations, such as spinal cord injury and amputation7 or neuromuscular disease and chronic pain.6 The present study encompasses a diverse group of older adults with functional limitations resulting from one more disabling conditions of various etiologies and thus provides validation of the FIM-SR in a population that is more representative of the general community.
The Bland-Altman analyses indicate that care recipients systematically rate their functional independence to be higher than do their caregivers and that this inclination persists over time. The mean difference of 2.00–2.97 can be considered as being significant because the line of agreement (y = 0.00 units) does not fall within the confidence interval of the mean difference. However, the mean difference between care recipient and caregiver scores can also be viewed in the context of a minimal clinically important difference (i.e., the smallest change in a person's outcome that is identified as being clinically important), a change of two or three points in that context might not be so relevant. For example, the minimal clinically important difference for the FIM motor scales has been examined with stroke survivors undergoing rehabilitation, and one study estimated that change scores on the motor FIM would need to be 11 points,15 whereas another study suggested 17 or more points.16 It should be noted, however, that at each time point 3–5 scores fell outside the 95% level of agreement, indicating a significant disagreement for a small number dyads. The present finding that most outlier scores were between 50 and 80 suggests that some respondents may have had difficulty distinguishing among moderate assistance, supervision, and device use as response options for the measure. Cognitive impairment did not seem to be a primary explanation for these outliers, because the average MOCA scores for these outliers were similar to the sample as a whole.
The tendency for proxy respondents to report lower functional independence is supported by other research related to this topic. Andresen et al8 found that family, friend, and healthcare proxies tended to overestimate impairment and dependency for ADLs and instrumental ADLs. Another study evaluated the agreement between the FIM scores of patients with hip fractures and those of family caregiver proxy respondents. Although agreement was high, the proxy respondents total FIM scores were lower (indicating worse functional independence) across the four study time points.9 Researchers and clinicians may benefit from awareness of this tendency for proxy respondents to report lower functional independence when interpreting their ratings.
The present results suggest that family caregivers may function as reliable proxies for care recipients without severe cognitive impairment in clinical and research settings when care recipients are unable or unavailable to respond on their own behalf. That may contribute to reducing clinicians' or researchers' workloads by allowing them to use one FIM instead of several. For example, when evaluating physical functioning, the FIM-SR can substitute for the SMAF, which takes more time and effort to administer, and either the care recipient or caregiver can serve as the respondent. These alternatives may especially benefit future studies dealing with the challenges of conducting dyadic research, including the recruitment and retention of participants.17
There are potential limitations of the present study. Care recipients systematically rate their own functional independence as being marginally better than their family caregivers rated it. However, which of the two assessments most accurately reflects care recipients' actual functional status cannot be determined. It also should be noted that in view of the inclusion criteria for care recipients and the actual sample obtained, the findings may not apply to recipients with more severe cognitive impairment. Another possible limitation of this study is the small risk of contamination, whereby between time points, it is possible that respondents could have compared responses. Finally, given the manner in which the SMAF was administered, it represents a composite rather than independent assessment of care recipient function. Because the SMAF ratings provided by evaluators could be informed from multiple sources including (1) direct observation of the care recipient's functional performance, (2) the care recipient's verbal characterization of the performance, and/or (3) the caregiver's verbal characterization of the performance, this may have inflated the association of the SMAF and SR-FIM scores.
2. Ishizaki T, Watanabe S, Suzuki T, et al: Predictors for functional decline among nondisabled older Japanese living in a community during a 3-year follow-up. J Am Geriatr Soc
3. Desrosiers J, Bravo G, Hébert R, et al: Reliability of the revised functional autonomy measurement system (SMAF) for epidemiological research. Age Ageing
4. Grey N, Kennedy P: The functional independence measure: a comparative study of clinician and self ratings. Paraplegia
5. Hamilton B, Granger C, Sherwin F, et al: A uniform national data system for medical rehabilitation. In: Fuhrer M (ed): Rehabilitation Outcomes: Analysis and Measurement
. Baltimore, MD, Paul H. Brookes, 1987:276
6. Jensen MP, Abresch RT, Carter GT: The reliability and validity of a self-report
version of the FIM instrument in persons with neuromuscular disease and chronic pain. Arch Phys Med Rehabil
7. Masedo AI, Hanley M, Jensen MP, et al: Reliability and validity of a self-report
FIM (FIM-SR) in persons with amputation or spinal cord injury and chronic pain. Am J Phys Med Rehabil
8. Andresen EM, Vahle VJ, Lollar D: Proxy reliability: health-related quality of life (HRQoL) measures for people with disability. Qual Life Res
9. Jones CA, Feeny DH: Agreement between patient and proxy responses during recovery after hip fracture: evidence for the FIM instrument. Arch Phys Med Rehabil
10. Demers L, Mortenson WB, Fuhrer MJ, et al: Effect of a tailored assistive technology intervention on older adults
and their family caregiver: a pragmatic study protocol. BMC Geriatr
11. Nasreddine ZS, Phillips NA, Bédirian V, et al: The Montreal cognitive assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc
12. Altman DG, Bland JM: Commentary on quantifying agreement between two methods of measurement. Clin Chem
13. Bland JM, Altman DG: Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
14. Desrosiers J, Rochette A, Noreau L, et al: Comparison of two functional independence scales with a participation measure in post-stroke rehabilitation. Arch Gerontol Geriatr
15. Wallace D, Duncan PW, Lai SM: Comparison of the responsiveness of the Barthel index and the motor component of the functional independence measure in stroke: the impact of using different methods for measuring responsiveness. J Clin Epidemiol
16. Beninato M, Gill-Body KM, Salles S, et al: Determination of the minimal clinically important difference in the FIM instrument in patients with stroke. Arch Phys Med Rehabil
17. Shatenstein B, Kergoat M-J, Reid I: Issues in recruitment, retention, and data collection in a longitudinal nutrition study of community-dwelling older adults
with early-stage Alzheimer's dementia. J Appl Gerontol