A comparison of physical activity (PA) assessment tools across levels of frailty

Abstract

Purpose

Determine which PA assessment tools are most closely related to frailty and whether PA is different across levels of frailty.

Methods

Fifty community-dwelling Greek older women (63–90 years) participated in this study. PA was measured objectively over 10 h using an accelerometer, a heart rate (HR) monitor, a portable electromyography (EMG) unit, and a global positioning system (GPS) and subjectively using the short version of the Minnesota Leisure Time Activity Questionnaire (MLTAQ). Participants were divided into three tertiles based on level of frailty as calculated from a Frailty Index (FI): low FI group (<0.17 FI); intermediate FI group (0.17–0.38 FI); and high FI group (>0.38 FI).

Results

Accelerometer step counts had the strongest correlation with frailty and were different across levels of frailty. The percentage of time engaged in PA was 31 ± 15% when PA was determined using an accelerometer. Forty-five percent of the variability in the FI was explained by a combination of PA assessment tools including; accelerometer, EMG, GPS, and MLTAQ. The individual contribution of EMG determined activity from the biceps brachii (BB) to the FI prediction was 16%. Accelerometer contributed an additional 10% and time engaged in PA, as assessed with the MLTAQ, added an additional 6% to the prediction of FI score.

Conclusions

PA assessment tools, when used in combination, provide important information about the PA accumulation of older women across levels of frailty.

Introduction

Epidemiological studies demonstrate a strong relationship between low levels of self-reported PA and functional decline, comorbidity, and mortality in healthy older adults (Gregg et al., 2003, Paterson et al., 2007). Although a large proportion of older adults consider themselves to be healthy a significant proportion may be considered frail (Fried et al., 2001). Frailty is an age-related state of vulnerability to adverse outcomes, caused by cumulative declines across multiple physiological systems and ranges from mild to severe (Fried et al., 2001, Klein et al., 2005, Theou and Kloseck, 2007, Gallucci et al., 2009). Definitions of frailty that take into account multiple domains can more accurately predict adverse outcomes compared with unidimensional definitions (Forti et al., 2012).

Low level self-reported PA is considered a key indicator of frailty (Rothman et al., 2008) and increased PA may prevent or reverse frailty (Peterson et al., 2009). Greece has one of the oldest populations in Europe (19.2% of the population is over the age of 65) (Central Intelligence Agency, 2009) and within the Greek older adult population 45% are at risk for frailty and 15% are already considered frail (Santos-Eggimann et al., 2009). Self-reported PA levels are lower in Greek older adults as compared to other European countries (Tzorbatzakis and Sleap, 2007). Therefore, it is imperative that suitable assessment tools be developed to support PA interventions for Greek older women.

The identification of frailty using performance-based measures of mobility has previously been examined (Kim et al., 2010, Davis et al., 2011); yet the association of frailty syndrome with PA remains to be elucidated. A range of objective and subjective PA assessment tools have been proposed to measure duration and intensity of PA and these methods are validated in older adults but not for those considered to be frail (Kochersberger et al., 1996, Strath et al., 2000, de Bruin et al., 2008, Pruitt et al., 2008, Copeland and Esliger, 2009, Harris et al., 2009a, Miller et al., 2010). Self-report questionnaires are the most common method to evaluate PA but these are limited due to memory/cognition problems of frail adults and their inability to recall what tasks they did during the day (Jørstad-Stein et al., 2005). Objective measures of PA include; pedometers, accelerometers, HR monitors and GPS (Strath et al., 2000, Pruitt et al., 2008, Harris et al., 2009a, Webber and Porter, 2009, Spierer et al., 2011). Recent evidence has suggested that recordings of daily muscle activity using portable EMG, either singularly (Harwood et al., 2008, Howe and Rafferty, 2009) or in combination with accelerometers (Kishimoto et al., 2009, Theou et al., 2010) may provide additional information regarding the intensity of daily PA in older adults and potential differences between upper and lower limb muscles. Ultimately, it offers a means to determine how hard specific muscles are working while performing PA. Each of these methods has strengths and limitations for the evaluation of PA, but the unique measures each affords, when used in combination, might permit a more comprehensive evaluation during daily life, especially in slow moving frail older adults.

The purpose of this study was to first examine the association of frailty with five common PA assessment tools: (1) accelerometer, (2) HR monitor, (3) portable EMG, (4) GPS, and (5) short version of the MLTAQ. A second purpose was to determine if PA is different across levels of frailty as established by the FI. We hypothesized that a combination of PA measures would explain the greatest proportion of variance in the FI and that PA would be different across all levels of frailty.