Patient-report as an option for describing muscle weakness: An integrative review

Abstract

BACKGROUND:

Muscle strength, although usually measured by performance, can be measured by patient-report.

OBJECTIVE:

Review the utility and clinimetric properties of muscle strength assessed by patient-report.

METHODS:

PubMed and hand searches were used to identify relevant literature. Findings were systematically summarized.

RESULTS:

Most patient-report measures identified individuals with muscle weakness, however, the clinimetric properties of measures were highly limited. Particularly missing was information on reliability and responsiveness.

CONCLUSIONS:

There is a place for the inclusion of patient-reported muscle strength, but clinimetric support for its use is still limited.

1. Introduction

Adequate muscle strength is a body function fundamental to the performance of everyday activities. Clinically, muscle strength is usually assessed as a performance outcome determined by manual muscle testing, field tests or dynamometry. However, muscle strength/weakness can also be assessed as a patient-reported outcome. The legitimacy of doing so is dependent on the utility and clinimetric properties of muscle strength measurements so obtained. As a review of the literature revealed no summary of such information, this integrative review was undertaken to identify patient-reported measures of muscle weakness and summarize what is known about their utility, reliability, validity, and responsiveness.

2. Methods

Relevant literature was identified via an electronic search of PubMed and a hand search of references cited in articles identified in the electronic search. The electronic search, which was conducted on November 10, 2018, went back to 1977 and used the search string: (“patient report” OR “self report”) AND (“strength” OR “weakness”). The titles and abstracts of 1500 potentially relevant articles were reviewed and the full texts were examined by the author to determine compliance with inclusion and exclusion criteria. Inclusion required that an article addressed a patient-reported weakness measure. Studies were excluded if they involved single cases. Studies were also excluded if patient-reported weakness was expressed in combination with another symptom (e.g., “weakness and tiredness”) or described using the Strength component of the Stroke Impact Scale [1]. The Strength component of the Stroke Impact Scale is the possible focus of a future paper.

Retained studies were examined for specific information on utility and clinimetric properties. Utility was defined as the ability of a patient-report measure to identify weakness in a study population. Reliability was defined a priori to include internal consistency, test-retest reliability, and inter-tester reliability. Validity was evinced by the convergent validity of patient reported measures with other measures of strength or physical status; known groups validity was supported by logical differences in patient-reported weakness between groups. Responsiveness was characterized by expected changes in status over time. Such changes could be documented using descriptive statistics, inferential statistics of difference (e.g., paired $t$ -test), or summary statistics focused on differences over time (e.g., effect size, standardized response mean, minimal detectable change, or minimally clinically important difference). This information was then tabulated.

3. Results

Table 1 summarizes 8 studies identified by my search in which patient-reported measures other than the Strength component of the Stroke Impact Scale were used to document weakness [2, 3, 4, 5, 6, 7, 8, 9]. Four of the studies included American individuals [2, 6, 7, 8], 2 focused on Dutch participants [3, 4], and 2 incorporated Scandinavians [5, 9]. Studies included between 28 [2] and 1189 [6] participants. The study populations for each study were quite different (e.g., survivors of intensive care admission versus adults with Parkinson’s disease) [2, 8].

Table 1

Summary of studies included in review

Study	Population	Patient reported strength measure	Findings
Choi et al. (2014) ${}^2$	American patients surviving admission to intensive care unit ( $n=$ 28)	Component of Modified Given Symptom Assessment Tool in which presence or absence of weakness was reported. Rating from 0 (none) to 10 (worst)	Utility: Weakness reported by 84.8% ( $⩽$ 2 weeks post ICU discharge). Reliability: Not reported. Validity: Correlation with fatigue ( $<$ 2 weeks post ICU discharge) $=$ 0.28, (2 months post ICU discharge) $=$ 0.42 ( $p<$ 0.05), (4 months post ICU discharge) $=$ 0.53 ( $p<$ 0.05). Responsiveness: Mean weakness rating $\downarrow$ from 7.1 ( $⩽$ 2 weeks post ICU discharge), to 5.3 (2 months post ICU discharge) & to 6.4 (4 months post ICU discharge).
Op het Veld et al. (2018) ${}^3$	Dutch community dwelling older adults ( $n=$ 135)	Custom instrument incorporating 10 questions relevant to hand-grip strength (e.g., “Do you encounter problems in daily life due to lack of hand strength?”) Yes or no reported	Utility: Weakness (according to at least 1 question) reported by 30.4%. Reliability: Internally consistent (Cronbach’s $\alpha =$ 0.88). Validity: Relative to dynamo-metrically measured grip strength cut-off values of the Fried frailty criteria, 2 hand-grip questions together provided the best indication of weakness: sensitivity (73.2%), specificity (71.3%). Responsiveness: Not reported.
Voermans et al. (2011) ${}^4$	Dutch adults with Ehlers-Danlos syndrome ( $n=$ 300)	Custom instrument with a single question: “Do you suffer from muscle weakness (i.e., reduction of muscle strength in one or more muscles?” Yes or no reported	Utility: Weakness reported by 73%. Reliability: Not reported. Validity: Correlation with fatigue (0.60, $p=$ 0.038), Medical Research Council sum score ( $-$ 0.79, $p=$ 0.004), number of muscles demonstrating weakness by dynamometry (0.58, $p=$ 0.048). Responsiveness: Not reported.
Grøvle et al. (2013) ${}^5$	Norwegians with disc-related sciatica ( $n=$ 380)	Component of Sciatica Bothersome Index asking about “Weakness in leg or foot” rated on 0 to 6 scale where 0 represents “not bothersome” & 6 represents “extremely bothersome”	Utility: Weakness reported by 33.5% at baseline. Reliability: Not reported. Validity: Weakness score significantly ( $p<$ 0.01) greater for patients determined to be weak by clinical examination. Responsiveness: Weakness score $\downarrow$ significantly ( $p<$ 0.01) over 2 years.
Sarkisian et al. (2008) ${}^6$	American older adults without disability ( $n=$ 1189)	Component of Hopkins Symptom Checklist with question “During the past week, how much have you been distressed by weakness in parts of your body?” Reported as: not at all, a little, quite a bit, extremely	Utility: Not reported. Reliability: Not reported. Validity: Weakness loaded highly (0.73) on a component of the Ten Expanded MacArthur Study of Successful Aging Frailty instrument. Responsiveness: Not reported.
Stiffler et al. (2013) ${}^7$	American older adults admitted to emergency department ( $n=$ 90)	Modification of component of Hopkins Symptom Checklist with statement “Before the illness or injury that brought me to the emergency department I felt weak.” Reported as: 0 (none of the time), 1 (some of the time: 1–2 days/week), 2 (a moderate amount of time: 3–4 days/week), 3 (most of the time)	Utility: Weakness reported by 12.2%. Reliability: Not reported. Validity: Weakness was significantly more prevalent in patients with functional limitations (54.5%) than in patients without functional limitations (25.3%). Relative to dynamometrically measured grip strength cut-off values of the Fried frailty criteria, sensitivity was poor (18%) but specificity was good (90%). Responsiveness: Not reported.
Friedman & Abrantes (2012) ${}^8$	American adults with Parkinson’s disease not weak on physical examination ( $n=$ 113)	Custom instrument with a question to patients regarding “whether they suffered from a sense of muscular weakness resulting from their Parkinson’s disease.” Reported as yes or no. Elaboration was sought as to body parts involved.	Utility: Weakness reported by 43.8%. Reliability: Not addressed. Validity: Weakness was significantly ( $p=$ 0.001) more common in patients who also reported fatigue than in patients who did not. Weakness was not related to disease duration or severity. Responsiveness: Not reported.
Brogårdh & Lexell (2015) ${}^9$	Swedish adults with late effects of polio ( $n=$ 325)	Component of Self-reported Impairments in Persons with Late Effects of Polio with a question “during the past 2 weeks how much have you been bothered by various post-polio related …muscle weakness. Reported as not at all, little, quite a bit, and extremely.	Utility: Weakness reported by 94%. Reliability: Not reported. Validity: Weakness correlated significantly (0.47–0.64, $p<$ 0.01) with all 12 items of the Walk 12. Responsiveness: Not reported.

Options for describing self-reported weakness varied considerably. Some were custom instruments [3, 4, 8] whereas others were components of other established instruments (e.g., Hopkins Symptom Checklist) [6, 7]. The instruments ranged from having a single question (e.g., “Do you suffer from weakness [i.e., reduction of muscle strength in one or more muscles?]) with a simple yes or no response, (4) to a single question with multiple ordinal choices (i.e., 4 to 11) for rating bothersomeness [5, 9], distress [6], or frequency [7], to having multiple questions (i.e., 10) with a simple yes or no response [3].

All but one study [6] provided some indication of utility – the percentage of patients self-identifying as weak in the 7 other studies ranged from 12.2% [7] to 94.0% [9].

Only one study addressed reliability. That study showed high internal consistency in patient responses to 10 diverse questions focused on hand grip strength [3]. All 8 studies dealt with some aspect of validity of patient reported weakness. Four studies reported a relationship between self-reported weakness and strength determined by some other means. Two compared self-reported strength to that measured with a grip dynamometer [3, 7]. Both showed self-reported strength to be highly specific relative to grip dynamometry, however only 1 of the 2 studies showed self-reported strength to be adequately sensitive relative to dynamometry. Grøvle et al. found self-reported weakness to be greater in patients judged to be weak by clinical examination [5]. Voermans et al. demonstrated a good correlation between self-reported strength and strength measured using the Medical Research Council sum score and a fair correlation between patient-reported strength and the number of muscles determined to be weak by dynamometry [4]. Patient-reported weakness has also been shown to be related to fatigue [2, 4, 8], albeit not necessarily significantly. Self-reported strength is higher among older adults who are without functional limitations [7], who report better walking performance [9], and who are successfully aging [6]. In a sample of adults with Parkinson’s disease but no weakness on physical examination, self-reported weakness was not related to disease duration or severity [8].

Two studies provided information relevant to responsiveness. They simply reported changes in self-reported weakness over time [2, 5]. No study used statistics such as effect size, minimal detectable change, or minimal clinically important difference to formally describe the responsiveness of self-reported measures.

4. Discussion

This review clearly demonstrates that muscle strength/weakness can be characterized using a number of different patient-report measures. Few patient-report measures, however, are used in multiple studies. So it is not possible to compare the utility of published measures. There is virtually no evidence for the reliability of any patient-report measures. Most particularly needed is information on between session test-retest reliability of the patient-report measures as it can serve as a basis for responsiveness. Some evidence for validity exists. This resides inconsistently in correlations with other measures of strength and function. There is virtually no evidence for the responsiveness of any patient-report measures of strength. Such evidence will require repeated measures over time in patients with known trajectories of decreasing (e.g., amyotrophic lateral sclerosis) or increasing (e.g., stroke) strength.

In conclusion, there may be a place for using of patient-reported measures for describing strength deficits in specific populations. However, considerable more work is required before any one published measure can be justified on the basis of its clinimetrics.