Multidimensional daily diary of fatigue–fibromyalgia–17 items (MDF-fibro-17). part 1: development and content validity

The study objective was to further evaluate qualitative and quantitative data previously gathered in order to gain a full understanding of the patient experience of FM-related fatigue, and to develop a multidimensional PRO measure of FM-related fatigue for use in clinical studies. The study aims were to a) confirm the multidimensional components of FM-related fatigue, b) explore the understanding, interpretation, and relevance of the 23 item pool, derived from earlier elicitation interviews using the patients’ language to describe the complex experience of FM-related fatigue, c) explore item redundancy and performance in order to optimize scale sensitivity and concept saturation, and d) to propose a list of items and potential themes or domains, that would be suitable for inclusion within a multidimensional measure of FM-related fatigue.

The present research was based on data gathered over the three studies that had previously been conducted for the development of an earlier measure of fatigue in FM. These studies included a concept elicitation study, a cognitive debriefing study, and a cross-sectional methodology study. The outcomes from these studies included a 23-item pool describing the experience of fibromyalgia-fatigue, and a five-item global fatigue measure. In these three studies, all participants were adults who met the American College of Rheumatology criteria for FM [2], with no significant co-morbidities that could confound reporting of FM. The concept elicitation study was conducted with 40 individuals with FM in the US (n = 20), Germany (n = 10) and France (n = 10). Key themes and concepts were identified using qualitative analysis methods based upon grounded theory. The cognitive debriefing and pilot testing of the 23-item pool was conducted in an additional 20 individuals with FM in the US only. The cross-sectional methodology study included a total of 145 patients with FM in the US, who ranked their symptom severity using the 23 items administered on an electronic daily diary for two weeks. The data was then used to evaluate the psychometric properties of the 23 item-pool. Further details of the methodology, original analyses, and results of these three studies involved have been published elsewhere [6, 14].

In the current study, the original concepts identified in the concept elicitation interviews, the feedback that was obtained on items from the cognitive debriefing, and the item level quantitative data from the original studies were reviewed to inform the development of a multidimensional tool.

The qualitative re-analysis involved a review of the original transcripts and analytical coding to confirm that the patient data supported the importance of all aspects of fatigue, and that the 23 items reflected the qualitative findings and language used by the patients. The qualitative and quantitative data on each item was also reviewed to confirm patient understanding and item performance as expected. Based on this work, a 17-item, multidimensional measure was proposed. A confirmatory factor analysis was then conducted to explore the fit of items to the multidimensional structure proposed in the initial conceptual model. This re-analysis was conducted by SM and YL, and findings were reviewed with those involved in the original analysis (CB, TS) and other experts in the area (JS, SD).

A single domain of “global fatigue experience” was identified to capture global and general descriptors of FM-related fatigue that appeared to transcend the other, more specific, domains. Patients used a variety of general terms to describe their fatigue. For example, patients would describe their fatigue in the following ways:

“When tiredness happens, there’s no relief. It’s like an overwhelming, overarching, penetrating, consuming tiredness. There’s nothing you can do about it.” (003-005, US, Female) [10]

“The absolute tiredness just, for no reason that can strike at any time. It’s like you’re totally exhausted mentally and physically.” (003-004, US, Female) [10]

“I would feel tired, or exhausted, or just run down, those kind of terms, which is all connected.” (001-001, US, Female) [10]

“And now to do the simplest little thing, I can be where I’m just totally worn out. Like I’ve been doing something all day long very strenuous.” (003-004, US, Female) [10]

No single term was identified that adequately captured the global experience of FM-related fatigue in the same way for all. Therefore, a domain containing multiple items using global terms was felt to be an important element of a comprehensive PRO measure to ensure that it captured the global experience of fatigue using patient derived language. All patients reported experiencing fatigue as part of their FM.

Patients asserted that FM-related fatigue affected them both physically and mentally. Cognitive limitations were described as an integral part of FM-related fatigue. In the concept elicitation interviews, patients often differentiated between FM-related fatigue and normal tiredness by the impact of the former on concentration, thinking clearly and remembering, and used this to explain the severity of their fatigue.

“It can be difficult, in the sense that the fatigue makes it hard to concentrate. That’s for sure.” (004-007, France, Female) [10]

“It’s like if something happens while I’m fatigued, I have brain fog. I forget about it. But if I do something when I’m clearheaded, then I’ll remember that as a memory.” (006-004, US, Female) [10]

In the concept elicitation interviews, 21 patients (52.5%) talked about a mental or cognitive component to their fatigue.

The physical aspect of FM-related fatigue was also described as an important part of FM-related fatigue throughout the interviews. In the concept elicitation interviews, 16 patients (40%) talked about their body feeling heavy, nine patients (22.5%) talked about their muscles feeling weak, and 32 patients (80%) described experiencing tiredness in specific areas of their body.:

“My body can feel very heavy. Like I have to drag to move, because I’m so tired.” (002-002, US, Female) [10]

“Interviewer: Could you describe the feeling of fatigue?” Patient response: “Just tired and heavy and draggy and slow.” (001-005, US, Female) [10]

“I would think about the energy it takes, and I would feel – I just feel like weighted down. That’s what the fatigue feels like. It’s almost a weak feeling, or a heavy feeling.” (006-005, US, Female) [10]

Patients consistently described motivation as a direct manifestation, and integral part, of FM-related fatigue, and appeared to link this to severity of fatigue on a daily basis. Difficulties with motivation were described as difficulty “getting going,” or as an increase in effort to conduct activities. Although such comments may reflect a lack of motivation, or an underlying mood disorder such as depression, it is important to note that patients with comorbid depression were excluded from the study to avoid any confounding of symptoms in the patient experience.

“But the other thing is, the energy, to me, also is connected to motivation…it decreases, at least for me, the motivation to do things. And that’s what I find very frustrating.” (001-001, US, Female) [10]

“I can’t do everything because I’m extremely tired.” (004-002, France, Female) [10]

“Yes, well, the lack of enthusiasm which I have, that I have to overcome to motivate myself into doing something.” (005-004, Germany, Female) [10]

In the concept elicitation interviews, 19 patients (48%) across all three countries discussed the impact of fatigue using general terms such as describing an “increased effort to do things,” “forcing oneself to do things,” and having trouble “getting motivated to do things.” More specifically, for 16 patients (40%), the concept of motivation was also discussed in relation to getting out of bed in the morning or having to “get going” in the morning.

A clear message from the interviews was that patients with FM often communicate their experience of FM by describing how fatigue negatively impacts their functional capacity, including their ability to conduct daily activities. Patients often used general statements referring to “things” as an overarching evaluation of impact, rather than specific activities. When specific activities were mentioned during a probed discussion, these were numerous, with 17 categories of activity identified, and no individual activity was reported by more than 7.5% of the overall population across the three countries. Therefore, an Impact of Function domain, using “doing things” as a general descriptor, was felt to be important in a comprehensive measure of FM-related fatigue and included in the final conceptual model.

“I mean, you just have no energy to get up and go any place. Like to work or any activities or to make dinner or just do anything.” (001-002, US, Male) [10]

“It’s a slowing down in being able to do things. Like where it would normally take me a minute, two minutes to wash a plate…the first one may start out that way, but as I start going through the dishes, I start to slow down.” (002-002, US, Female) [10]

Overall, 27 patients (67.5%) in the concept elicitation interviews spontaneously talked about their fatigue/tiredness affecting general things they wanted to do, and 24 patients (60%) talked about their fatigue/tiredness specifically affecting their daily activities.

The data from the cognitive debriefing of the original pool of 23 items, capturing all elements of fatigue that emerged from the qualitative work, was then re-evaluated in light of this multidimensional framework. The patient feedback confirmed that most items were clearly understood and considered to be relevant, and overall supported using 17 of the original pool of 23 items to capture these five dimensions of FM-related fatigue. The six items that were flagged for exclusion from the item pool, along with the rationale for their exclusion, are provided in Table 1. The decision to remove the items was based on the qualitative and quantitative data. Findings from the cognitive debriefing revealed a lack of conceptual clarity, and a tendency to over-represent severe symptoms. The item level analysis data from the cross-sectional validation study confirmed the poor performance of these items that had been flagged for removal. One item (energy) had low correlations with all other items (0.177 to 0.339), and the others had high correlations (all above 0.84) with other items within their target domain, indicating potential item redundancy. These items also showed a tendency towards a “floor” effect. Taken together, these findings lead to the removal of the six weakest items from the item pool.

Based on these results, a 17-item multidimensional measure of FM-related fatigue has been developed, capturing five domains: Global Fatigue Experience (four items); Cognitive Fatigue (four items); Physical Fatigue (three items); Motivation (three items); and Impact on Function (three items).

The revised multidimensional conceptual model of FM-related fatigue, and framework for the MDF-Fibro-17 is presented in Fig. 1. Amongst these 17 items, there were no floor or ceiling responses observed (1.4 – 11.3% selecting zero, and 2.1 – 9.5% selecting ten), and a relatively even spread of responses across the 0 to 10 scale (mean scores ranged from 4.2 to 6.3, with a minimum of zero and a maximum of ten on all items). Inter-item correlations were high (all above 0.720), and when evaluating correlations between items within a domain, all were above 0.889. Many inter-item correlations were above 0.9, suggesting potential item redundancy, however the qualitative data confirmed the relevance of the multidimensional aspects of FM-fatigue from the patient perspective. Therefore, items that were otherwise supported through the qualitative work were retained despite the high inter-item correlation to maintain the comprehensive nature of the proposed measure. An initial confirmatory factor analysis was conducted using the dataset obtained from the cross-sectional validation study confirmed that the 17-item model performed better than all 23 items. The 17 item, multi-dimensional model resulted in an appropriate fit on all fit measures, including comparative fit index (CFI) = 0.96 and a root mean square error of approximation 0.08. These results are shown in Table 2.