Muscle quantity and function measurements are acceptable to older adults during and post- hospitalisation: results of a questionnaire-based study

To evaluate the acceptability of handgrip strength, gait speed, quadriceps ultrasonography, and BIA to patients, when measured as part of an observational study during and post-hospitalisation. The aim of the main study was to characterise acute sarcopenia in hospitalised older patients.

The main protocol for this study has been published elsewhere [6]. Our reporting is consistent with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Patients were recruited to one of three cohorts from the Queen Elizabeth Hospital Birmingham (QEHB) – general medical patients with infections, elective colorectal surgery, or emergency abdominal surgery. Inclusion criteria for each cohort were aged 70 years and older and hospitalised (or expected to be hospitalised for the elective cohort) for an acute bacterial infection, major colorectal surgery procedure, or emergency abdominal surgery procedure. Exclusion criteria were the inability to understand verbal and written English, or imminently dying. Informed consent or personal consultee declaration was obtained for all participants. Medical patients were recruited within 48 h of admission, emergency surgery patients were recruited pre-operatively or within 48 h post-operatively, and elective surgery patients were recruited in pre-operative assessment clinic.

Quadriceps ultrasound, BIA, handgrip strength, and physical performance (either SPPB or gait speed alone depending on cohort and timing of assessment) were measured serially as part of this study. These were performed within 48 h of admission/surgery, within one week of admission/surgery, and three months after admission/surgery. In the elective cohort, measurements were also performed prior to admission.

Quadriceps ultrasound was performed anteriorly over both thighs at the midpoint between the greater trochanter, and the joint line of the knee. Participants were positioned with their knees in natural relaxation, with a firm wedge below the knees, and the upper body reclined to 45^o. Contact gel was applied to the skin and measurements were taken using a linear probe using a Venue 50 device (GE Healthcare). A minimum of three measurements were taken on each side; a fourth was taken if rectus femoris (RF), vastus intermedius (VI), or subcutaneous (SC) measures varied by more than 10% between each other. These measures were used to calculate the Bilateral Anterior Thigh Thickness (BATT – right RF + left RF + right VI + left VI) [4]. BIA was performed in the same position by applying electrodes to the right hand and foot and recording measures using a Bodystat Quadscan 4000 as per the manufacturer’s instructions. BIA was not performed in participants with implanted cardiac devices. Handgrip strength was measured using a Jamar dynamometer; participants sat in a chair with their elbow flexed at 90^o and advised to squeeze as hard as they could [7]. Two readings were taken on each side. Gait speed was measured over a 4 m course; participants were advised to walk at a normal comfortable pace, using walking aids if necessary.

Frailty was defined dichotomously (frail vs. non-frail) according to the phenotype definition [8] at the point of the completion of acceptability questionnaire. Frailty was defined as scoring three or greater of weight loss (recorded or self-report), low handgrip strength, low walking speed, self-reported exhaustion, or low physical activity, as detailed in the main study protocol [6].

An acceptability questionnaire was developed, as described in the main study protocol [6] and online supplement, which asked participants to state how highly they agreed with positive statements about seven different aspects of acceptability for each of handgrip strength, 4 m gait speed, ultrasound quadriceps, and BIA (Table 1) [1]. The questionnaire was completed by the same researcher who administered the muscle quantity and function assessments. We evaluated gait speed alone rather than SPPB to ensure consistency across cohorts, and prevent burden to participants from the acceptability evaluation. Responses were given using a Likert scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, 5 = strongly agree). Participants were also able to provide additional comments related to the study in general or any study-related procedures. This questionnaire was administered to all participants at the point of their three-month follow-up, in either their own home, or the Inflammation Research Facility, QEHB. Recruitment was paused due to the Coronavirus 2019 (COVID-19) pandemic and the protocol was later amended to remove in-person follow-up at three months, to reduce unnecessary contact with vulnerable participants. This sub-study includes participants who were recruited prior to this amendment.

Data were imported into IBM SPSS Version 26. Counts for each Likert score were derived and presented visually with horizontal bar charts. For each outcome measure, a total score was derived for all acceptability domains for each participant (minimum possible score 7, maximum possible score 35). Median total acceptability scores were calculated for each outcome. We used one-sample Kolmogorov-Smirnov normal tests to evaluate distributions of total acceptability scores for each outcome. We used the Friedman test to assess differences in total acceptability scores between outcome measures and Mann-Whitney U tests to assess for differences in individual total scores between those with and without frailty. The main study was powered for a different primary outcome. A post-hoc power calculation showed that a sample size of 40 was able to detect a difference in total acceptability score of 1.35, with 80% power and 5% alpha, assuming a null hypothesis median score of 20 (i.e. neither agree nor disagree selected for all answers) and an expected normal distribution.

Free text comments were transcribed by the researcher (CW) linked against their identifiable study number. The researcher (CW) familiarised themselves with the comments and identified emergent themes. Thematic analysis was conducted using an inductive approach, with no prespecified hypotheses of what data may arise from these comments. The participant details were linked to text after identification of emergent themes. Consensus agreements of themes was reached by researchers not involved in initial transcription and data reduction.

Patients and members of the public were extensively involved in the planning and development of the main study. The questionnaire used in this study was developed with direct involvement of healthy older adults. The results of this study itself will be of direct relevance to future studies and clinical practice involving the measures described.

Figure 2a-d shows the distribution of response scores for each acceptability domain for each outcome measure. Overall, domains rated highly for all outcome measures, with the majority of participants stating that they agreed or strongly agreed with each positive statement for each outcome. The domain with the least agreeability was burden for both handgrip strength and gait speed; some participants disagreed with the statement that minimal effort was required to complete these tests. The domain with the highest agreeability was self-efficacy, particularly for ultrasound and BIA; participants agreed or strongly agreed that they enjoyed participating in these tests.

Table 3 shows the median overall scores for each outcome measure, separated by phenotypic frailty. All total score distributions were individually significant. However, total scores did not significantly differ between outcome measures. Additionally, scores did not significantly differ for outcomes between those with and without frailty.

Some participants expressed that other aspects of the SPPB, the acceptability of which were not formally assessed in this study, were more burdensome.

“Apart from ‘getting up from the chair’ [chair stands] it was no effort”

“The only one that really got me was ‘the chair’ [chair stands]”

“The only thing was the balance thing [tandem stand]”

Although not the primary focus of this study, participants expressed comments relating to their reasons for participating in research. Common themes that emerged were around the desire to help others and feeling that they had been able to provide a service.

“If it helps anyone else to get better then so be it”

“I do them because I know that I'm helping to improve things”

“If it's gonna be useful to you and to someone else that's good enough for me”

“I'm glad to be of service to someone – whatever helps you and your research”

“I'm glad that I was able to help …”

Other participants expressed that they felt they had been able to learn things through participating in research, which had benefitted them personally.

“I just find I learn something and you learn something. My motivation is I want to see the boundaries pushed back”

“We found out how these things work”

“I enjoyed it all – interesting and educational”

“I've just been really interested in what you've done”

Providing the option of being able to have follow-up conducted in the participants’ own homes was also considered very positively.

“Grateful to visitors - we enjoyed”

“I've enjoyed you coming and seeing you … it gives you an insight into what's going on”

“I'm pleased that you're able to come to me and I've not got to travel anywhere …”

This is the first study to specifically evaluate the acceptability of the measures described to patients during and post- hospitalisation. The questionnaire devised for this study was multi-faceted and developed from recognised domains within acceptability [1]. The simplicity of the survey ensured high completion rates, enabling gathering of both quantitative and qualitative results. Additionally, obtaining feedback at the end of study completion enabled participants to have appropriate time to really consider their feedback on participation, and to be able to provide this in a comfortable environment (either their own home or a quiet clinic room). At this stage participants had also completed the assessments multiple times so were familiar with the tests. This ensured higher completion/response rates.

We acknowledge that there are a number of limitations to this study. There is no agreed standard way of assessing acceptability of a medical test. Firstly, the questionnaire itself was devised by the study team. It is unknown how these results would compare against other tests that are commonly used in clinical practice i.e. we do not know whether these results represent “above average” acceptability. Additionally, perceived acceptability of tests may be biased by the agreement of participants to participate in the study in the first place and to complete follow-up; patients who refused to participate and those who did not complete follow-up may have responded differently. Unfortunately, this is an inevitable bias of any study that aims to assess acceptability via participant responses; it would not be possible to assess acceptability of a study procedure in a participant who had not agreed to participate. Results may also be biased by the fact that the questionnaires were administered by the same researcher who conducted the muscle quantity and function assessments; participants may have wished to ingratiate themselves with the research team [11]. We also acknowledge that religious or cultural differences may affect the results of this study. The majority of participants were White British or Irish and we did not collect personal information about religious beliefs. Acceptability of tests may be viewed differently in other groups e.g. individuals of some religious backgrounds may consider quadriceps ultrasonography to be more personally obtrusive [12]. As described, feedback was obtained after the participants’ final follow-up assessments, although we consider this a strength, this can also be considered a limitation. Participants may have responded differently if they had been asked to complete feedback in hospital. It is important to consider that some participants had cognitive impairment and were unable to recall the initial tests, which did not obviously affect responses.