Frail Older Adults

A research assistant was responsible for assessing frailty using: 1) Fried’s phenotype method; and 2) SPPB.

Fried’s phenotype method

Fried’s phenotype method classifies older adults as frail, pre-frail or non-frail based on five criteria [3 (link)]. For each of the criteria, the participant was classified as frail or not frail, using the following cut-offs: 1) Weight loss: more than 10 lbs. lost unintentionally in the last year; 2) Exhaustion: participants stating that they felt that everything they did was an effort or that they could not get going (from the CES-D Depression Scale) a moderate amount of the time or most of the time; 3) Physical activity (Minnesota Leisure Time Activity Questionnaire): energy expenditure <383 kcal per week for men and <270 kcal per week for women; 4) Walk time (15-ft walk): ≥ 7 sec (men height ≤ 173 cm, women height ≤ 159 cm) or ≥ 6 sec (men height > 173 cm, women height > 159 cm); 5) Grip strength (Jamar Dynamometer, Layfayette Instruments, USA) (average of three trials): ≤ 29–32 kg for men (stratified by BMI classifications) and ≤ 17–21 kg for women (stratified by BMI classifications) [3 (link)]. If the participant was unable to answer any questions due to memory problems, the accompanying legally authorized representative provided an answer, which is the approach used in other studies [26 (link)]. Participants were instructed to use an assistive ambulatory aid for the walk test if an aid was used in their normal routine. Frail participants scored below the cut-offs for three or more criteria, pre-frail participants scored below the cut-offs for one or two criteria, and non-frail participants did not score below the cut-offs for any criteria [3 (link)].

Short Performance Physical Battery (SPPB)

The SPPB consists of three assessments: 1) repeated chair stands; 2) balance tests (side-by-side, semi-tandem and tandem balance tests); 3) an eight-foot walk test [23 (link)]. Similar to Fried’s phenotype method, the participant’s scores on each component of the battery were compared to normative data and a score between zero and four was determined for each component. If participants were unable to complete a component of the test, a score of zero was given for that component. A final summary performance score out of 12 is calculated, with higher scores indicating superior lower extremity function [23 (link)]. Regarding the threshold score for frailty, community-dwelling older adults who score ≤ nine on the SPPB are most likely to be classified as frail [16 (link)] and are at risk of losing the ability to walk 400 m [27 (link)] (predictive validity). An SPPB score of ≤9 has the most desirable sensitivity (92%), specificity (80%) and greatest area under the curve (AUC =0.81) for identifying frail adults [15 (link)]. In order to classify participants as frail, pre-frail and non-frail, the following cut-offs were used: SPPB zero–six (frail), SPPB seven–nine (pre-frail), SPPB 10–12 (non-frail) [28 (link)].

We will use an exploratory qualitative design combining focus group interviews, face-to-face individual and telephone interviews with stakeholders (e.g., well, pre-frail, and frail older adults, general practitioners, practice nurses, emergency department physicians, orthopedic surgeons) to understand perceptions of frailty and frailty screening within and between these diverse groups. Arts-based data collection methods incorporating visual elicitation (e.g., drawing the meaning of frailty) will also be used with consumer groups to augment qualitative methods. We will iteratively collect and analyze data over a six-month period across urban and rural regions in South Australia (e.g., Adelaide, Southern Fleurieu Peninsula).

This study is a secondary analysis of an observational cohort study in frail older ED patients that was performed in an ED of a teaching hospital in Finland. In the primary study we included patients who were ≥ 75 years of age, had a score between 4 to 9 on the 9-point Clinical Frailty Scale (CFS) [34 (link)], and were registered residents of the hospital’s service area. ED visit data were collected between December 11^th, 2018 and June 7^th, 2019. The included patients were followed up from electronic health records. Methods for the primary study have been described in detail in our previous article [42 (link)].
The clinical laboratory service of the ED routinely gives RDW values (% value as integer) for all blood counts tested. Besides the clinical laboratory service, the ED has point-of-care testing equipment available, which does not provide RDW values. Point-of care testing is typically preferred, if more extensive laboratory testing is not anticipated based on patient’s chief complaint or condition. For the secondary analysis conducted here, those patient visits from the primary study who had the CFS score 4–8 and had RDW tested 0–48 h after ED admission were included. If more than one blood count was drawn from a patient within 48 h of ED admission, the result of the first laboratory test was used for the analysis. Patients who had a CFS score of 9 were excluded because such patients are defined as having a short life expectancy < 6 months, but otherwise not living with severe frailty.
Nonparametric baseline data were presented with interquartile ranges (IQR). The outcome measure was 30-day mortality. Patients were allocated to six classes based on their RDW value: ≤ 13%, 14%, 15%, 16%, 17%, and ≥ 18%. We used same cut-off values as a recent study to enable comparison of our results in frail ED patients to general older adult ED patient population [42 (link)]. Mortality rate was calculated for each class. The Cochran–Armitage test for trend was used to test the statistical significance of the trend of increasing mortality with higher RDW values.
Crude and adjusted ORs with 95% confidence intervals (CI) of a one-class increase in RDW for 30-day mortality were calculated. Univariate and multivariate models of binary logistic regression analysis were used for crude and adjusted ORs, respectively. Age, sex, and CFS score were considered as potential confounders and were included in the analysis.
As a sensitivity analysis to assess if categorisation of the RDW values has impact on the results, we performed a regression analysis with RDW as continuous variable. We also performed a sensitivity analysis with haemoglobin as a potential confounder, because haemoglobin level is directly related to red blood cells, like RDW is, and may be associated with mortality.
From clinical perspective, we were interested whether RDW is independent of vital parameters. The National Early Warning Score 2 (NEWS2), a widely used prognostic score based on common vital signs, was included in the baseline data for our previous study [42 (link)]. We performed an additional testing by adjusting with the NEWS2 besides other potential confounders used in the regression analysis.
A p value of < 0.05 was considered statistically significant. GraphPad Prism software, version 9.4.1 (Graphpad Software LCC) was used for the Cochran–Armitage test. SPSS software, version 28 (IBM) was used for all other statistical analyses.
The primary study which this secondary analysis was based on, was registered at ClinicalTrials.gov on December 20t^h, 2018, identifier NCT03783234.

Motivations to participate in the evaluation of three different eHealth services were inventoried. The first service, Stranded (see Fig. 1), is a web-based, gamified eHealth service for (pre-)frail older adults. Stranded [27 (link)] consists of two parts: a falls prevention programme based on the OTAGO Programme [28 ], and cognitive minigames. The falls prevention programme consists of physical exercise videos that older adults can perform at home. These exercises focus on improving muscle strength, balance, and flexibility. The minigames are different kinds of puzzle games. The duration of the study evaluating Stranded was four weeks. The second eHealth service, Council of Coaches (COUCH) [29 ] (see Fig. 2), is a web-based service designed for adults with Diabetes Mellitus Type 2 or Chronic Pain, and older adults who are dealing with age-related impairments. The goal of COUCH is to encourage a healthy lifestyle via conversations with virtual coaches. Within COUCH six different coaches are available: a physical activity coach, a nutrition coach, a social coach, a cognitive coach, a chronic pain coach (only available for users with chronic pain), and a diabetes coach (only available for users with diabetes). During the summative evaluation of COUCH, participants could use the eHealth service for four weeks. The last eHealth service, the selfBACK app [30 (link)–32 (link)] (see Fig. 3), is a mobile self-management application for adults with neck and/or low back pain. The selfBACK app provides users with a weekly tailored plan to self-manage this pain. The weekly plain focusses on three aspects: Physical activity (i.e., daily step data), physical exercises to strengthen the muscles and increase flexibility, and educational messages to motivate users and to give them advice. This study with the selfBACK app lasted for six weeks.Fig. 1

Screenshot of eHealth service Stranded

Fig. 2

Screenshot of eHealth service Council of Coaches. (Names of the virtual coaches f.l.t.r.: Carlos (peer), Olivia (physical activity coach), Emma (social coach), Katarzyna (diabetes coach), Helen (cognitive coach), Coda (helpdesk robot), François (nutrition coach))

Fig. 3

Screenshot of eHealth service selfBACK app (showing weekly self-management plan)

A prospective, single-blinded, parallel-group pilot RCT design was adopted (Figure 2). The family caregivers of frail older adults were randomized into either the intervention group receiving the social media–based MBI embedded with MM and SA or the control group receiving brief education on older adult care and usual care.