Introduction
Patient-reported outcomes are defined as outcomes that are reported directly by the patient and are often collected using questionnaires called patient‐reported outcomes measures (PROMs).277 PROMs provide a useful way to obtain standardised measurements of a patient’s health status and/or quality of life and can be used to assess and compare treatment outcomes from the patient’s perspective.277 Generic PROMs have been developed for use across different health conditions, while disease-specific PROMs measure aspects of health that are relevant to a particular health condition.
Individual PROMs can be made up of numerous items and scales that should assess outcomes that matter to patients. However, the diversity of items and scales between different PROMs can result in the combining and aggregation of dissimilar items and scales that, on face value, appear to be measuring similar concepts or result in measuring different aspects of the outcome of interest. It is, therefore, important to establish whether items contained in PROMs are comparable across different measures to determine whether PROMs data can be meaningfully pooled. There are a variety of ways to measure PROMs depending on whether the intention is to allow generic or disease-specific comparisons and depending on differences in how PROMs developers have considered the relevant concepts of interest when developing question items.278 For instance, of two PROMs that measure quality of life, one might have items that consider concepts such as pain, sleep quality, fatigue and return to normal activities, while the other might have items that consider pain and physical aspects such as the inability to undertake activities of daily living, but not sleep. Without conducting an item analysis, it would be difficult to establish their comparability. Variability in individual items included in these tools and how they contribute to the overall ‘quality of life’ assessment can raise questions about the legitimacy of combining these measures in meta-analyses. For example, an analysis of PROMs used in studies of radical treatments for oesophageal cancer identified 32 different health domains from 21 PROMs that included 116 scales and 32 single items.279 The authors noted that 94 different verbatim names were used to describe the PROM scales and single items and, while many of these names were similar, content analysis showed that the component questions did not always measure comparable issues. Similarly, a synthesis of PROMs in colorectal cancer surgery identified 51 health domains from 50 PROMs that included over 900 question items.278 The authors noted that none of the PROMs measured all 51 domains and found major heterogeneity of PROM measurement and variation in the assessed content. The findings from both these analyses highlight how the variation in measuring PROMs in clinical trials can hinder comparisons between studies.
Methods to determine the validity and methodological quality of outcome measurements have been developed through the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN), which aim to improve the selection of health measurement instruments.280 Climent-Sanz and colleagues (2020) recently published a systematic review to identify and describe existing sleep outcome PROMs used in studies that recruited adults diagnosed with fibromyalgia.35 The review identified seven reports of five PROMs: the FMSD,39 the JSS,281 the MOS-SS,256,282,283 the PSQI284 and the SQ-NRS.40,256 With the exception of the FMSD, the PROMs were not developed with a fibromyalgia patient population and can be used to measure sleep outcomes in different health conditions. Climent-Sanz and colleagues (2020) assessed the five PROMs as valid and reliable tools for assessing sleep quality in the context of fibromyalgia, although none met the full COSMIN criteria.35 While the review examined the psychometric properties of the PROMs, the items contained within the different measures were not compared to each other. It is, therefore, unclear if important sleep domains are missing and whether the PROMs are similar enough to be combined in evaluations of interventions to treat fibromyalgia-related sleep problems. We therefore sought to examine the item variability of PROMs of sleep outcomes for people with fibromyalgia. Since the Climent-Sanz (2020)35 review focused exclusively on studies that assessed adults with fibromyalgia published in English or Spanish, we also updated and extended the scope of the review to identify any additional PROMs in adults and children without applying any language restriction.
Methods
Search methods for identification of studies
The search strategy of the Climent-Sanz et al. synthesis was developed based on the COSMIN ‘search filters for finding studies on measurement properties’ provided as an additional tool in the COSMIN website285 and was performed on 6 March 2020. We repeated the Climent-Sanz et al. searches to identify relevant reports published from 6 March 2020 to 5 November 2021. Additionally, their searches were repeated with text terms to identify studies including children that were excluded from their original synthesis; these searches covered all years to 17 November 2021.
Inclusion criteria
Studies reporting sleep measures validated in people with fibromyalgia were eligible for inclusion in our review. All measures identified (as opposed to included) in the Climent-Sanz et al. synthesis were included in our PROMs analysis, as our purpose was to critique item coverage, not psychometric properties, which are already addressed by the existing evidence synthesis. A summary of eligibility criteria is shown in Table 13.
| Type of studies | Type of participants |
|---|---|
| Validation or cross-cultural adaptation studies.Studies that report the item development of PROMs originally developed in the context of fibromyalgia. | Validation or cross-cultural adaptation studies involving adult and child participants (as defined by the studies’ authors) diagnosed with fibromyalgia. |
Where eligible studies included PROMs that were originally developed in a non-fibromyalgia patient population, we also consulted the original development study to inform our analysis where this was possible.
Exclusion criteria
Studies of sleep measures validated in people with clinical conditions other than fibromyalgia were excluded.
Study selection and data extraction
Literature search results were screened by two reviewers (CR, MI) to identify relevant studies reporting PROMs. To ensure consistency, the reviewers independently screened 10% of all citations identified by the search strategies and compared screening decisions at the beginning of the study selection process. To ensure accuracy, each reviewer checked the data extracted by the other review author for 10% of all included studies.
Data were extracted on:
- the name of the PROM(s)
- the reported PROMs scales
- individual verbatim items, and
- whether and how patients were involved in tool development.
Data analysis
The analysis was informed by previous research that has analysed PROMs into individual outcome domains.286
The individual verbatim items from each PROM were analysed by using an inductive content analysis approach. All PROM items were examined and systematically categorised into conceptual health domains according to the aspect which they aimed to capture; however, items were coded to more than one domain where appropriate. Domains were generated inductively from the identified individual items and were informed by terms and definitions contained in the Sleep Foundation Dictionary287 and the World Health Organization (WHO) International Classification of Functioning, Disability and Health (ICF).288 Domain mapping was conducted independently by the two review authors with assistance from a senior member of the research team (KG) with experience in PROM coding. Any disagreements were resolved through discussion or referred to another member of the team with research and/or clinical expertise in sleep (DW and NT) and our patient representative (DD) as appropriate. The patients’ input in the item inception and development phases of the PROMs measures was also recorded. The PROM items and their associated domain categories were summarised narratively.
Results
Descriptive characteristics: included studies
Our updated literature searches identified 292 titles and abstracts. Following title and abstract screening, 19 reports were selected for full-text eligibility assessment. After full-text assessment, one report associated with a study already identified by the original search strategies developed by Climent-Sanz et al.35 was considered eligible for inclusion. This report is a validation study of a Turkish version of the JSS.289 Thus, the current evidence base in terms of PROMs consists of a total of eight reports related to five studies. We did not identify any new PROMs tools validated for sleep outcomes in people with fibromyalgia. All included reports are published in English and focus only on adult participants. A visual summary of the study screening process is provided as Figure 9. The bibliographic details of the included studies are provided in Report Supplementary Material 13.
FIGURE 9 Preferred reporting items for systematic reviews and meta-analyses flow diagram for identification of the PROM studies.50 (Figure not reproduced here; see the original article.)
Across both the Climent-Sanz (2020) review and our current review, the studies were published between 2009 and 2020 and were conducted in the USA,39,40,256,281–283 Spain284 and Turkey.289 Three of the PROMs are associated with development studies that were conducted in non-fibromyalgia patient populations. The JSS was developed for use in the general population by Jenkins and colleagues (1988)38 in a population of air traffic controllers290,291 and cardiac surgery patients.292,293 The JSS was validated with fibromyalgia patients in the included studies conducted by Crawford and colleagues (2010)281 and Unal-Ulutatar and Ozsoy-Unubol (2020).289 The MOS-SS was developed in individuals with chronic illnesses37 and was subsequently validated with fibromyalgia patients in the studies by Cappelleri and colleagues (2009),283 Martin and colleagues (2009)256 and Sadosky and colleagues (2009).281 The PSQI was developed with people with major depressive disorder and healthy controls.36 The Spanish version of the PSQI was validated with fibromyalgia patients by Hita-Contreras and colleagues (2014).284 A summary of the study and participant characteristics of the development studies is presented in Appendix 7, Table 23.
The baseline participant and study characteristics of the included studies are summarised in Table 14. The included studies report data for 3226 people with fibromyalgia. The two reports conducted by Cappelleri and colleagues (2009) detailing the evaluation of the MOS-SS283 and the SQ-NRs40 were both based on data from fibromyalgia participants recruited to two randomised, double-blind, placebo-controlled trials of pregabalin: studies 1056 (n = 748) and 1077 (n = 745). The report by Martin and colleagues (2009) evaluated both the MOS-SS and SQ-NRS; therefore, the same participant sample (n = 20) was used for both analyses in this report.256 Across studies the majority (92.8%) of participants were middle-aged women; the youngest and oldest reported mean ages were 44.3 years289 and 52.8 years,284 respectively. Race and ethnicity were NR by three studies.282,284,289 In the two studies that reported race and ethnicity, the majority of participants were white [2604/2878 (90.5%)]. The studies often did not report baseline socioeconomic data such as educational attainment level or employment status. Of the three studies that did report education, half of participants were educated to high-school/college level [120/239 (50.2%)].256,284,289 Of the three studies that reported participant employment status most participants [180/288 (62.5%)] were unemployed or retired.39,256,284
| PROM, Country, Author ID | PROM general characteristics | Sample size | FM participant characteristics | Methods of the study |
|---|---|---|---|---|
| FMSDUSAKleinman 201439 | The tool was created to measure sleep disturbance in FM patients | Total FM patients: n = 49Focus group participants n = 34Cognitive interview participants n = 15The sample also included therapeutic area experts in FM n = 4 | Focus group (n = 34)Age (years), mean (SD) range: 47.8 (11.9); 22–70Gender, n (%): male: 4 (11.8%), female: 30 (88.2%)Race/ethnicity, n (%):white: 25 (73.5%)Hispanic or Latino: 8 (23.5%)black or African American: 1 (2.9%)Asian: 2 (5.8%)Native Hawaiian or other Pacific: 4 (11.7%)American Indian or Alaska Native: 1 (2.9%)other: 6 (17.6%) | (1) A literature review identified key concepts associated with FM patients’ experience of sleep and relevant PROMs; (2) Qualitative interviews with therapeutic experts; (3) Focus groups with FM patients; (4) Development of a conceptual framework and the FMSD and 5) Cognitive interviews with FM patients to explore the content validity of the FMSD. |
| Marital status: NRSociodemographic status:Employment status, n (%):employed, full-time: 7 (20.5%)employed, part-time: 9 (26.4%),homemaker: 2 (5.8%)student: 1 (2.9%)unemployed: 7 (20.5%)retired: 3 (8.8%)disabled: 9 (26.4%)other: 2 (2.8%)unemployed or disabled due to FM (n, %): yes 10 (62.5), no 6 (37.5%)Cognitive interview participants (n = 15)Age (years), mean (SD) range: 51.4 (10.1); 27–64Gender, n (%): male: 1 (6.7%), female: 14 (93.3%)Race/ethnicity, n (%)white: 11 (73.3%)Hispanic or Latino: 0 (0%)black or African American: 3 (20.0%)Asian: 1 (6.7%)Native Hawaiian or other Pacific: 0 (0%)American Indian or Alaska Native: 0 (0%)other: 0 (0%) | ||||
| Marital status: NRSociodemographic status:Employment status, n (%):employed, full-time: 4 (26.7%)employed, part-time: 1 (6.7%)homemaker: 1 (6.7%)student: 0 (0%)unemployed: 3 (20.0%)retired: 2 (13.3%)disabled: 6 (40.0%)other 0 (0%)unemployed or disabled due to FM (n, %): yes 7 (77.8%), no 2 (22.2%) | ||||
| JSSUSACrawford 2010281JSS-TRTurkeyUnal-Ulutatar 2020289 | The tool assesses common sleep symptoms during the previous month. | Crawford 2010281Total n = 1316One Phase 2 validation trial (n = 195) and twoPhase 3, RCTs (n = 1121) in patients with fibromyalgiaUnal-Ulutatar 2020289Total n = 81 | Crawford 2010281Phase 2 Validation trial (n = 195)Age (years), mean (SD) (range): 46.5 (11.35) [20–83]Gender, n (%): male: 11 (5.6%) female: 184 (94.4%)Race/ethnicity, n (%):white/Caucasian: 180 (92.3%)black/African American: 11 (5.6%)Asian: 1 (0.5%)American Indian/Alaska Native: 0 (0%)other: 3 (1.5%) | Crawford 2010281Analysis began with the Phase 2 trial and was then followed by the two Phase 3 trials. The number of missing responses for each item over time, the. internal consistency reliability, test–retest reliability and responsiveness to change were evaluated. Construct validity was measured by examining correlations of the rescored JSS with other patient-reported outcome (PRO) scales: the Pain VAS, Fatigue VAS, FOSQ, SF–36, FIQ, and ESS.Unal-Ulutatar 2020289The JSS-TR was performed twice with a fortnightly interval. The intraclass correlation coefficient was used to evaluate test–retest reliability. The internal consistency was assessed by Cronbach’s alpha coefficient. Convergent validity was measured via the correlation between the JSS-TR and PSQI, FIQ, FSS, EQ-5D-3 L and BDI. |
| Marital status: NRSociodemographic status: NRTime since first fibromyalgia symptoms (years), mean (SD) [range]: 10.6 (8.40) [0–41]Time since first fibromyalgia diagnosis (years), mean (SD) [range]: 6.0 (6.02) [0–41]Trial 1 (n = 548)Age (years), mean (SD) [range]: 47.0 (11.26) [18–79]Gender, n (%): male: 48 (8.8%); female: 500 (91.2%)Race/ethnicity, n (%):white/Caucasian: 498 (90.9%)black/African American: 33 (6.0%)Asian: 7 (1.3%)American Indian/Alaska Native: 6 (1.1%)other 4 (0.7%) | ||||
| Marital status: NRSociodemographic status: NRTime since first fibromyalgia symptoms (years), mean (SD) [range]: 9.7 (8.49) [0–47]Time since first fibromyalgia diagnosis (years), mean (SD) [range]: 5.9 (6.75) [0–39]Trial 2 (n = 573)Age, (years), Mean (SD) [range]: 46.6 (10.72) [19–80]Gender, n (%): M: 60 (10.5%); F: 513 (89.5%)Race/ethnicity, n (%):white/Caucasian: 524 (91.4%)black/African American: 39 (6.8%)Asian: 4 (0.7%)American Indian/Alaska Native: 3 (0.5%)other: 3 (0.5%) | ||||
| Marital status: NRSociodemographic status: NRTime since first fibromyalgia symptoms (years), mean (SD) [range]: 9.7 (8.75) [0–51]Time since first fibromyalgia diagnosis (years), mean (SD) [range]: 4.9 (5.61) [0–48]Unal-Ulutatar 2020289Age (years), mean (SD) minimum–maximum: 44.28 (10.6) 19–70Gender, n (%): male; 10 (12.3%), female: 71 (87.7%)Race/ethnicity: NR | ||||
| Marital status, n (%):married: 63 (77.8%)single: 18 (22.2%)Sociodemographic status:Education, n (%):primary-secondary school: 55 (67.9%)high school: 12 (14.8%)university: 14 (17.3%)Work status, n (%):employed: 24 (29.6%)unemployed: 55 (67.9%)retired: 2 (2.5%)Symptom duration (months), mean (SD) minimum-maximum: 47.88 (69.83) 3–480Disease duration (months), mean (SD) minimum-maximum: 5.31 (12.89) 0–96 | ||||
| MOS-SSUSACappelleri 2009283Martin 2009256Sadosky 2009282 | The MOS-SS tool assesses sleep quality | Cappelleri 2009283n = 1493FM patients were recruited from two RCTs (trial 1077 and trial 1056) comparing pregabalin vs. placeboMartin 2009256n = 20Sadosky 2009282n = 129Total n = 1642 | Cappelleri 2009283Trial 1056 (n = 748)Age (years), mean (SD) 95% CI: 48.8 (10.9) 48.0 to 49.6Gender, n (%): male: 42 (5.6), female: 706 (94.4)Race/ethnicity, n (%): white 675 (90.2%), black 35 (4.7%), other 38 (5.1%)Marital status: NRSociodemographic status: NRDuration of FM prior to study start (months), mean (SD) 95% CI: 111.7 (95.0) 104.9 to 118.5 (n = 747) | Cappelleri 2009:283 The MOS-SS was completed by FM patients at baseline and end of treatment The patients’ scores were compared with scores obtained from a nationally representative sample in the USA.Martin 2009:256 Twenty adults with FM were asked to complete each question using a ‘think-aloud’ process and answer follow-up probes from the interviewers related to the content, wording and comprehension of the item and gave feedback on their relevance to FM and suggest improvements.Sadosky 2009:282 The MOS-SS was completed by FM patients with a current pain rating of > 2 on a 0–10 point NRS. All patients completed the 4-week and 1-week recall period versions of the MOS-SS. The test and retest of the 4- week recall version was separated by a time interval ranging between 1 and 3 days. The test–retest period of the 1-week recall version was 7 days. |
| Trial 1077 (n = 745)Age (years), mean (SD) 95% CI: 50.1 (11.4) 49.3, 50.9Gender, n (%): male: 41 (5.5), female: 704 (94.5)Race/ethnicity, n (%):white: 678 (91.0%),black: 33 (4.4%),Other: 34 (4.6%)Marital status: NRSociodemographic status: NRDuration of FM prior to study start, (months) mean (SD) 95% CI: 120.2 (96.2)113.3 to 127.1 | ||||
| Martin 2009256Age (years), mean (range): 50.3 (29–64)Gender, n (%): male: 4 (20%) female: 16 (80%)Race/ethnicity, n (%):white: 13 (65%)black: 7 (35%)Marital status: NRSociodemographic status:Education, n (%):advanced degree: 3 (15%)college graduate: 2 (10%)associates degree: 2 (10%)some college: 9 (45%)high-school degree or equivalent: 4 (20%)Employment status, n (%):employed full-time: 5 (25%)employed part-time: 1 (5%)unemployed: 1 (5%)disabled: 8 (40%)retired: 3 (15%)other: 2 (10%) | ||||
| Years since diagnosis of FM, mean (range): 8.9 (< 1–18)Average (SD) of reported pain level, 0–10-point scale where 0 indicates no pain and 10 indicates worst possible pain: 6.0 (1.6)Sadosky 2009282Age (years), mean (SD): 49.4 (11.0)Gender, %: male: 8.7%, female: 91.3%Race/ethnicity: NRMarital status: NRSociodemographic status: NRSeverity of FM, moderate, %: 88.1%≥ 2 years since diagnosis, %: 88.3% | ||||
| PSQISpanish versionSpainHita-Contreras 2014284 | The PSQI was developed to provide a reliable, valid and standardised measure of sleep quality; to discriminate between ‘good’ and ‘poor’ sleepers; and to provide a brief, clinically useful assessment of a variety of sleep disturbances that might affect sleep quality. | Total n = 138 | Age (years), mean (SD): 52.83 (9.32)Gender, %: male: 0 (0%), female: 138 (100%)Race/ethnicity: NRMarital status: NRSociodemographic status:Education n (%):primary education: 62 (44.9%)secondary education: 38 (27.5%)tertiary education: 14 (10.1%)Occupation n (%):retired: 58 (42.0%)working: 34 (24.6%)unemployed: 46 (33.3%)Years since FM was diagnosed, mean (SD)8.27 (5.65)Years since fibromyalgia pain onset, mean (SD): 15.77 (9.76) | The English version of the PSQI was independently translated into Spanish by two bilingual experts. The Spanish version was then completed by 20 women with FM to verify their comprehension. The bilingual experts then compared the translated version with the original to check semantic and linguistic equivalence. The Spanish version was then completed by 96 participants. These results were compared with those obtained from the same participants two weeks earlier to evaluate test–retest reliability. Some of the Spanish PSQI items were modified because of cultural differences. |
| SQ-NRSUSACappelleri 200940Martin 2009256 | The SQ-NRS was developed to allow a quick-to-complete daily record of the quality of sleep and to provide a generic approach to the global impact of sleep problems in patients with FM. | Total n = 1513Cappelleri 2009:40 n = 1493 FM patients recruited from two RCTs (trial 1077124 and trial 1056126) comparing pregabalin vs. placeboMartin 2009:256 n = 20 FM patients | See descriptions for the participants included in the MOS-SS validation studies conducted by Cappelleri 2009 and Martin 2009 | Cappelleri 2009:40Test–retest reliability was evaluated by computing intraclass correlation coefficients. Responsiveness to treatment was evaluated by standardised effect sizes (difference between least squares mean changes in Sleep Quality scores in the two treatment groups divided by the SD of Sleep Quality scores across all patients at baseline).Martin 2009:256 Twenty adults with FM were asked to complete each question using a ‘think-aloud’ process and answer follow-up probes from the interviewers related to the content, wording and comprehension of the item. Participants also provided feedback on, the item’s relevance to FM and offered suggestions for improvement. |
Descriptive characteristics: patient-reported outcome measures from included studies
Information on the characteristics of the PROMs identified in the included studies is presented in Table 15. Only one study reported including patients in the conceptual item identification stage of the PROMs tool.39 In this study, focus groups were conducted with 34 fibromyalgia patients who experienced sleep disturbance to inform the development of a conceptual framework for the FMSD. Cognitive interviews were then conducted with a further 15 fibromyalgia patients to explore the content validity of the FMSD. Cognitive interviewing is a method that is used in the development of PROMs. Cognitive interviews are used to evaluate whether the tool sufficiently captures the concept of interest and checks whether participants’ understanding of how to complete the PROM is as intended by the PROM developers, including the participants’ interpretation of the meaning of the PROMs items, recall period and how to complete the response scales.294 Cognitive interview results showed that the concepts were relevant to fibromyalgia patients and content was interpreted as intended.
| Tool, Country, Reference ID | Dimension (number of items) | Response options and scoring | Mode of administration | Tool recall period | Time to complete | Were patients/participants involved in developing the tool items? | Content validity for FM patients explored? | Sample items |
|---|---|---|---|---|---|---|---|---|
| FMSDUSAKleinman 201439 | 8 items | 11-point numerical scale ranging from 0 to 10. Alternative response options included using the ‘worst possible symptom and best possible symptom’ as anchor points as well as ‘no difficulty’ and ‘worst possible difficulty’. | Self-completion | Daily | NR | YesQualitative interviews with therapeutic area experts (n = 4) and focus groups with FM patients (n = 34) who experienced sleep disturbance were conducted to inform the development of a conceptual framework and the FMSD. | Yes | How difficult was it to fall asleep last night? |
| JSSJenkins 198838Crawford 2010281JSS-TRUnal-Ulutatar 2020289 | 4 items | Six-point Likert scalewhere 0 indicates no sleep problems and 5 indicates frequent sleep problems. The JSS is scored as the sum of the four items, resulting in a score from 0 to 20. Higher scores indicate greater sleep problems (0 indicates no sleep problems and 20 points indicates significant sleep problems). | Self-completion | 4 weeks | NR | Jenkins 198838UnclearCrawford 2010281NoUnal-Ulutatar 2020289No | No | During the past month did you have trouble falling asleep? |
| MOS-SSUSAHays 199237Cappelleri 2009283Martin 2009256Sadosky 2009282 | 12 items with 6 sleep domains derived subscales. Answers are based on a retrospective assessment over the past 4 weeksThere is a nine-item version of the MOS-SS, named Sleep Problems Index II and a 6-item versiondefined as the Sleep Problems Index I. Neither of the scales excludes any item from the original scale,but rather groups them in unique items so that the only difference is that these two scales are shorter. | One open numerical response option (item 2), one five-point Likert scale (item 1) and 10 six-point Likert scales (items 3 to 12).Quantity of sleep is scored as the average hours slept per night. The other scales and problems index are scored on a 0–100 possible range, except the ‘quantity of sleep’ item (ranges between 0 and 24) and the ‘adequacy of sleep’ item (ranges between 0 and 1).High scores indicate worse sleep problems, except for the ‘sufficiency of sleep’ and ‘sleep quantity’ items, where lower scores indicate worse sleep problems. | Self-completion | 4 weeks | 3–5 minutes | Hays 199237UnclearThe 12 MOS-SS items were identified based on two pilot studies that tested 17 items, but it is NR whether patients had any input to the development of the items beyond piloting the tool to assess whether the tool items could be scored separately and to identify the smallest number of items necessary to assess the independent sleep dimensions.Cappelleri 2009,283 Martin 2009,256 Sadosky 2009282No | Yes | How often during the past 4 weeks did you feel that your sleep was not quiet (moving restlessly, feeling tense, speaking, etc.) while sleeping?- All of the time- Most of the time- A good bit of the time- Some of the time- A little of the time- None of the time |
| PSQIUSABuysse 198936SpainHita-Contreras 2014284 | 18 items, which combine to form 7 component scores to generate a global score that is used to determine cut-offs for good or poor sleep.Six additional questions rated by the bedpartner or roommate are used for clinical information only and are not tabulated in the scoring of the PSQI. | Mixture of four-point Likert scales and numerical coding of the following items: usual bedtime, number of minutes taken to fall asleep, usual getting up time and number of hours sleep per night.The 19 self-rated items are combined to form seven ‘component’ scores, each of which has a range of 0–3 points. A score of ‘0’ indicates no difficulty, while a score of ‘3’ indicates severe difficulty. The seven component scores are then added to give one ‘global’ score, with a range of 0–21 points, ‘0’ indicating no difficulty and ‘21’ indicating severe difficulties in all areas. | Self-completion | 4 weeks | 5–10 minutes | Buysse 198936NoItems in the PSQI were derived from three sources: clinical intuition and experience with sleep disorder patients; a review of previous sleep quality questionnaires reported in the literature; and clinical experience with the instrument during 18 months of field testing.Hita-Contreras 2014284NoSome of the translated Spanish PSQI items were modified because of cultural differences but the translations were conducted by two bilingual experts and not FM patients | Hita-Contreras 2014284NoSome of the translated Spanish PSQI items were modified because of cultural differences, but content validity for FM not explored. | During the past month, how often have you had trouble sleeping because you…Cannot get to sleep within 30 minutes- Not during the past month – Less than once a week– Once or twice a week– Three or more times a week |
| SQ-NRSUSACappelleri 200940Martin 2009256 | 1 item | 11-point numeric scale ranging from 0 (‘best possible sleep’) to 10 (‘worst possible sleep’). | Self-completion | 24 hours | NR | Cappelleri 200940UnclearMartin 2009256No | Yes | Please complete the following question upon awakening. Select the number that best describes the quality of your sleep during the past 24 hours. (Circle one number only.)0 1 2 3 4 5 6 7 8 9 10 |
In addition to the FMSD, two reports explored the content validity of two further PROMs for fibromyalgia patients: the MOS-SS256 and the SQ-NRS.256 For the MOS-SS, Martin and colleagues (2009) held qualitative cognitive debriefing with 20 adult fibromyalgia participants using a ‘think-aloud’ process.256 In general, participants found the measure to be relevant to their sleep symptoms. However, participants indicated that several items in the MOS-SS could be re-worded, split, or deleted to improve relevance for fibromyalgia patients, and that some important concepts were not captured, such as ‘trouble waking up’ or ‘getting out of bed, and other non-sleep specific symptoms or impacts of fibromyalgia’. The authors state that further research to address these modifications could improve the psychometric performance of the MOS-SS for fibromyalgia patients. To our knowledge, the MOS-SS currently remains a generic instrument and has not been adapted for fibromyalgia patients. Martin and colleagues (2009) used the same method to explore the content validity of the SQ-NRS with the same sample of fibromyalgia patients.256 The authors reported that most participants responded positively when asked how well the SQ-NRS captured the effect of fibromyalgia on their sleep. As with the MOS-SS, some participants stated that specific details regarding their sleep were not captured, such as the number of awakenings through the night, the ability to go to sleep, reasons for poor sleep quality and non-sleep-related fibromyalgia symptoms; however, the authors note that the brevity and ease of use of the SQ-NRS support its use in research and clinical practice.
The JSS was evaluated for construct validity, reliability and ability to detect a change in fibromyalgia patients.281 The Spanish version of the PSQI concentrated on validating the translation of items with fibromyalgia patients.284 It is unclear whether the authors also explored the content validity of the PSQI for fibromyalgia patients.
The total number of items varied across PROMs, ranging from 1 item in the SQ-NRS to 24 items in the PSQI. However, one of the items contained in the PSQI concerns information about whether the respondent shares a bed with a partner or has a roommate. Five additional items are answered by the bed partner or roommate rather than the patient. These six items are used for clinical information only and are not used for calculating the global PSQI score. The PSQI is the only one of the five tools to include questions addressing roommates. These six items were not included in our analysis. Our analysis, therefore, included a cumulative total of 43 individual items across the five PROMs (median = 8).
Two PROMs included 11-point NRS response options (the FMSD and SQ-NRS) and the remaining PROMs included a range of four- to six-point Likert scales and numerical response options. The PROMs varied in terms of their cut-offs to determine what constitutes significant sleep problems. Except for the five items directed at bed or roommates contained in the PSQI, all items were self-completed by the patient. The recall period of the PROMs ranged from 24 hours (FMSD and SQ-NRS) to 4 weeks (JSS, MOS-SS and PSQI). The MOS-SS was reported to take 3–5 minutes and the PSQI was reported to take 5–10 minutes for the participant to complete. The studies did not report the time taken to complete the remaining PROMs.
Item domain classification
Our review identified 21 relevant sleep domains representing the 43 items contained in the 5 PROMs. During analysis it was determined that question item 3 contained in the FMSD was considered to measure two domains: sleep maintenance and degree of sleep disturbance. The domains are, therefore, represented by 44 items. Table 16 presents a summary of the domain labels, definitions and example items from the PROMs. The domains are conceptually distinguishable, but many are closely related. We often found it challenging to separate items measuring sleep maintenance and sleep disturbance as the items in these domains share a high degree of relatedness. It also important to note that we used the term non-restorative sleep as our domain name for coding question item 4 of the JSS: ‘During the past month did you wake up after your normal amount of sleep feeling tired and worn out?’. Our coding and definition for this domain were informed by current published literature and advice from our clinical expert (NT).295 A summary table of the domain coding for each of the verbatim PROMs items is presented in Report Supplementary Material 14.
| WHO ICF category | Selected WHO ICF terms | Selected Sleep Foundation Dictionary terms | Definition of terms and domainsa | Agreed domain name | Example PROMs item |
|---|---|---|---|---|---|
| b134Sleep functions | b1340 Amount of sleepMental functions involved in the time spent in the state of sleep in the diurnal cycle or circadian rhythm. | Sleep duration | The quantity of time that a person sleeps. Sleep duration may be measured for just one sleep period or over the course of a 24-hour day. | Sleep duration | PSQI Q4. During the past month, how many hours of actual sleep did you get at night? (This may be different than the number of hours you spend in bed.)This question item is used to calculate sleep efficiency in the PSQI PROM tool. |
| Time in bed | The total amount of time that a person spends in bed regardless of whether or not they are sleeping during that time. This term is most often used in sleep studies to calculate sleep efficiency or to assess adherence to specific sleep rescheduling instructions during treatment. | Bedtime | PSQI Q1. During the past month, when have you usually gone to bed at night?This question item is used to calculate time in bed and sleep efficiency in the PSQI PROM tool. | ||
| Wake-up time | PSQI Q3. During the past month, when have you usually gotten up in the morning?This question item is used by the authors in the calculation of time in bed and sleep efficiency. | ||||
| Sleep efficiency | The proportion of time during a sleep episode that is actually spent sleeping. It is calculated by dividing total sleep time by total time in bed × 100%. | N/A | N/A | ||
| b1341 Onset of sleepMental functions that produce the transition between wakefulness and sleep | Sleep onset | Falling asleep or initiating a sleep period. | Sleep onset | FMSD Q1. How difficult was it to fall asleep last night?[0–10 scale] | |
| Sleep latency | The amount of time from ‘lights out’, or bedtime, to actually falling asleep. | Sleep latency | PSQI Q2. During the past month, how long (in minutes) has it usually taken you to fall asleep each night?Number of minutes | ||
| b1342 Maintenance of sleepMental functions that sustain the state of being asleep | Sleep maintenance | Staying asleep for the desired or planned amount of time after initially falling asleep. | Sleep maintenance | JSS Q3. During the past month did you have trouble staying asleep, including waking up far too early? | |
| Insufficient sleep | A condition in which the mind or body do not function properly because of short sleep duration or excessive sleep fragmentation. | Insufficient sleep | MOS-SS Q12. How often during the past 4 weeks did you get the amount of sleep you needed? All of the timeMost of the timeA good bit of the timeSome of the timeA little of the timeNone of the time | ||
| Sleep disturbance | A disruption in sleep that causes arousal or awakening. | Degree of sleep disturbance | MOSS-SS Q3. How often during the past 4 weeks did you feel that your sleep was not quiet (moving restlessly, feeling tense, speaking, etc.) while sleeping? All of the timeMost of the timeA good bit of the timeSome of the timeA little of the timeNone of the time | ||
| Bad dream | A dream period that involves negative or bothersome content but does not cause a person to wake up from the dream. | Sleep disturbance due to bad dream | PSQI Q5h. During the past month, how often have you had trouble sleeping because you … had bad dreams? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week“ | ||
| Nightmare | A dream with negative content that causes a person to wake up from sleep. Immediately after waking up, a person normally remembers the content of the nightmare. | ||||
| Nocturia | Frequent urination at night. Most technical definitions consider nocturia to be awakening one or more times to urinate, but some studies focus on the effects of multiple bathroom trips. | Sleep disturbance due to toilet visit | PSQI Q5c. During the past month, how often have you had trouble sleeping because you … have to get up to use the bathroom? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week | ||
| Sleep apnoea | A type of sleep disorder marked by disordered or abnormal breathing. The two main types are obstructive sleep apnoea and central sleep apnoea. | Sleep disturbance due to breathing problem or headache | MOS-SS Q5. How often during the past 4 weeks did you awaken short of breath or with a headache? All of the timeMost of the timeA good bit of the timeSome of the timeA little of the timeNone of the time | ||
| Sleep environment | The setting where a person sleeps. Usually a bedroom, the sleep environment includes elements such as the mattress and bedding as well as ambient light, sound, smell and temperature. | Sleep disturbance due to sleep environment | PSQI Q5f. During the past month, how often have you had trouble sleeping because you … feel too cold? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week | ||
| Snoring | An audible vibration of tissue at the back of the throat. When frequent, snoring may be known as chronic snoring or primary snoring. | Sleep disturbance due to coughing or snoring | PSQI Q5e. During the past month, how often have you had trouble sleeping because you … cough or snore loudly? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week | ||
| No corresponding dictionary term | N/A | Sleep disturbance due to pain | PSQI Q5i. During the past month, how often have you had trouble sleeping because you … have pain? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week | ||
| No corresponding dictionary term | N/A | Sleep disturbance due to other reason | PSQI Q5j. During the past month, how often have you had trouble sleeping because you … other reason(s), please describe: | ||
| How often during the past month have you had trouble sleeping because of this? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week | |||||
| Nap | A short sleep period, usually taken during the day, apart from a person’s primary sleep period. A nap may also be referred to as a siesta, its name in Spanish. | Nap | MOS-SS Q11. How often during the past 4 weeks did you take naps (5 minutes or longer) during the day? All of the timeMost of the timeA good bit of the timeSome of the timeA little of the timeNone of the time | ||
| Somnolence | The state of feeling sleepy or drowsy. | ||||
| Sleep inertia | A drowsy or groggy feeling that occurs shortly after waking up from sleep. | Sleep inertia | FMSD Q7. How difficult was it to begin your day?[0–10 scale] | ||
| b1343 Quality of sleepMental functions that produce the natural sleep leading to optimal physical and mental rest and relaxation. | Sleep quality | An individual’s satisfaction with their sleep, integrating aspects of sleep initiation, sleep maintenance, sleep quantity and feeling refreshed upon awakening. Sleep quality is not always defined in the same way and often relies on subjective assessments by individuals of how they slept. | Sleep quality | FMSD Q6. How rested were you when you woke up for the day?[0–10 scale] | |
| No corresponding dictionary term | The feeling that sleep is restless, light, or of poor quality even though the duration may appear normal.296 | Non-restorative sleep | JSS Q4. During the past month did you wake up after your normal amount of sleep feeling tired and worn out? | ||
| b1349 Sleep functions, unspecified | Sleep aid (use of) | A substance or medication used to try to improve sleep. Most sleep aids are either prescription drugs, over-the-counter medications, or dietary supplements. Other approaches, such as aromatherapy, may also be considered sleep aids. | Use of sleeping medication | PSQI Q7. During the past month, how often have you taken medicine (prescribed or ‘over the counter’) to help you sleep? Not during the past monthLess than once a weekOnce or twice a weekThree or more times a week | |
| Sedative (use of) | A substance or medication that induces drowsiness. | ||||
| b130 Energy and drive functions | b1300 Energy level | Fatigue | A feeling of a lack of mental or physical energy. Fatigue frequently overlaps with excessive daytime sleepiness, cognitive impairment and other symptoms of sleep problems. | Daytime impairment | MOS-SS Q6. How often during the past 4 weeks did you feel drowsy or sleepy during the day? All of the timeMost of the timeA good bit of the timeSome of the timeA little of the timeNone of the time |
| b1301 Motivation | Cognitive impairment | Difficulty related to mental functions such as thinking, attention, reaction time, memory, learning and judgement. | |||
| Daytime impairment | Negative effects of sleeping problems that occur during waking hours. Daytime impairments can include cognitive deficits, physical problems, and emotional or mood disturbances. |
All PROMs (FMSD, JSS, MOS-SS and the PSQI) can be considered as being multi-dimensional (i.e. capturing more than one domain) based on their coded domains, apart from the SQ-NRS, which can be considered as being unidimensional (i.e. only capturing one domain) as this tool contains only one item: sleep quality. The PSQI is considered the most comprehensive tool as this includes 15 of the 21 identified domains, followed by the MOS-SS, which included 11 domains. Table 17 presents the domains identified across the included PROMs. We identified 12 items that reported measuring the broad concept of sleep disturbance across two PROMs tools (the MOS-SS and PSQI); however, these individual items measure seven different types of sleep disturbance, as presented in Table 17. Given that clinical treatments and patient management strategies might differ depending on the underlying causes of sleep disturbance, we chose to code these items as distinct domains according to the type of sleep disturbance being measured, rather than group the items under a broad sleep-disturbance domain. Two items from the FMSD and the MOS-SS were coded as measuring the degree of sleep disturbance, and this was considered as a further distinct domain associated with sleep disturbance. The most frequently reported domains across the PROMs were sleep maintenance, which consists of six items, representing 13.6% of the total number of items included in the domains across four PROMs, followed by sleep quality, consisting of five items (11.4%) across four PROMs, sleep onset, consisting of four items (9.1%) across three PROMs, and daytime impairment, consisting of four items (9.1%) across two PROMs. Two PSQI items that were categorised as measuring the domains bedtime (PSQI question item 1) and wake-up time (PSQI question item 3) are used in the PSQI tool in the calculation of time in bed and sleep efficiency. One further PSQI item (PSQI question item 4) – which was categorised as measuring sleep disturbance in our analysis – is also used in the calculation of sleep efficiency in the PSQI tool. As the individual PSQI question items do not capture time in bed or sleep efficiency in isolation, sleep efficiency and time in bed were not identified as distinct domains in our analysis. Other domains that were reported by only one PROM include: non-restorative sleep (one question item in the JSS), sleep inertia (one question item in the FMSD), nap (one question item in the MOS-SS) and use of sleeping medication (one question item in the PSQI).
| Domain (n = 21) | Number of items in the domain | FMSD | JSS | MOS-SS | PSQI | SQ-NRS | Number of PROMs included in the domain |
|---|---|---|---|---|---|---|---|
| Sleep onset | 4 | 2 | 1 | 1 | 3 | ||
| Sleep latency | 3 | 1 | 2 | 2 | |||
| Bedtime | 1 | 1 | 1 | ||||
| Wake up time | 1 | 1 | 1 | ||||
| Sleep duration | 2 | 1 | 1 | 2 | |||
| Insufficient sleep | 2 | 1 | 1 | 2 | |||
| Sleep maintenance | 6 | 2 | 2 | 1 | 1 | 4 | |
| Sleep disturbance due to bad dream | 1 | 1 | 1 | ||||
| Sleep disturbance due to breathing problem or headache | 2 | 1 | 1 | 2 | |||
| Sleep disturbance due to coughing or snoring | 2 | 1 | 1 | 2 | |||
| Sleep disturbance due to pain | 1 | 1 | 1 | ||||
| Sleep disturbance due to sleep environment | 2 | 2 | 1 | ||||
| Sleep disturbance due to toilet visit | 1 | 1 | 1 | ||||
| Sleep disturbance due to other reason(s) | 1 | 1 | 1 | ||||
| Degree of sleep disturbance | 2 | 1 | 1 | 2 | |||
| Sleep quality | 5 | 2 | 1 | 1 | 1 | 4 | |
| Non-restorative sleep | 1 | 1 | 1 | ||||
| Sleep inertia | 1 | 1 | 1 | ||||
| Nap | 1 | 1 | 1 | ||||
| Daytime impairment | 4 | 2 | 2 | 2 | |||
| Use of sleeping medication | 1 | 1 | 1 | ||||
| Total number of items | 44 | 9a | 4 | 12 | 18 | 1 | N/A |
| Total number of domains included in PROM | N/A | 6 | 3 | 11 | 15 | 1 | N/A |
Discussion
This systematic review of PROMs for people with fibromyalgia-related sleep problems is the first to systematically characterise the item content across these measures into individual outcome domains. It extends the findings of the Climent-Sanz (2020)35 review to highlight the heterogeneity in domains measured across the sleep PROMs. Although we identified one additional relevant report, we did not identify any new sleep-related PROMs validated for use with fibromyalgia patients.
The identified PROMs varied in terms of the number of individual items and the domains covered by such items. None of the PROMs covered all 21 domains identified by our analysis. The most comprehensive identified PROM, in terms of the number of domains measured, is the PSQI, which captures 15 domains and includes the highest number of items across the five PROMs. Unsurprisingly, the least comprehensive PROM is the SQ-NRS as this is a single-item tool that only measures the sleep quality domain. Sleep quality is also measured by the PSQI and MOS-SS (one item each), and two items contained in the FMSD, but is not covered by the JSS. The SQ-NRS can, therefore, be considered to provide an overall assessment of sleep quality that is in keeping with most of the other PROMs but might lack the depth of the other tools. The sleep maintenance domain contained the highest number of items (six items) and is covered by all PROMs except the SQ-NRS; however, it should be noted that the dominance of the sleep maintenance domain is partly due to our decision to code items measuring the different underlying reasons for sleep disturbance as individual domains. Ten items across three PROMs (PSQI, MOS-SS and FMSD) measure different types of sleep disturbance and two further items contained in the MOS-SS and FMSD measure the degree of sleep disturbance. The large number of items associated with different types of sleep disturbance reflects the concentration of PSQI items measuring these domains. Just under half [8/18 (44.4%)] of the items contained in the PSQI measure different types of sleep disturbance. Sleep maintenance and sleep disturbance can be seen as interconnected and at times overlapping domains, as both capture concepts associated with the interruption of the state of being asleep or problems with sleep continuity. Sleep disturbance has a broader focus and could include sleep-disordered breathing, excessive limb movement, nightmare and nocturia. Interestingly, as discussed in Chapter 3, Climent-Sanz et al. noted that participants across the studies included in their qualitative meta-synthesis reported that the most common sleep problems associated with fibromyalgia were those related to the maintenance of sleep.33 This was also associated with the feeling of being continuously sleep-deprived, and that constantly waking up is the worst thing about their health condition – aspects that would be captured by these existing PROMs.
Four tools included items measuring the domain daytime impairment. It should be noted that while we have chosen the term ‘daytime’ to correspond with the term used by the Sleep Foundation, we believe that this domain captures the negative effects of sleeping problems that occur during waking hours and could also measure impairment during the evening or nighttime, depending on the usual or desired waking hours of the respondent.
While we found heterogeneity among the item content of the PROMs, it is worth noting that our clinical expert and patient advisors (see Acknowledgements for further details) expressed the opinion that all identified PROMs capture constructs associated with sleep quality and, conceptually, are similar enough to be combined in a synthesis. This underlines our rationale for conducting separate network meta-analyses in our quantitative evidence synthesis, where we conducted an NMA that pooled data for all PROMs in our main ‘sleep’ outcome as well as performing analyses for each individual PROM (see Chapter 2).
It is noteworthy that we did not identify any studies that reported exploring the content validity of the JSS and PSQI items for use with fibromyalgia patients. Of the three PROMs that were explored for content validity, participants indicated that both the MOS-SS and SQ-NRS missed items that are of relevance for fibromyalgia patients. Only the FMSD was developed with fibromyalgia patients, highlighting a lack of input from this patient group in the item development of the PROMs. This is not unexpected as the remaining PROMs did not include fibromyalgia patients during item conception. It should be noted that there is a possibility that the item coverage of the PROMs may be inadequate in terms of outcomes that matter to people who have fibromyalgia-related sleep problems. For example, our qualitative synthesis (reported in Chapter 3) demonstrated that aspects of the relationship between poor sleep quality and other symptoms of fibromyalgia, as well as feelings of frustration or failure, and fear of going to bed are highly relevant to fibromyalgia patients. It is our opinion that these concepts are not covered by the items contained in the five identified PROMs.
Due to the poor reporting of demographic data, such as race, ethnicity and sociodemographic status, it is unclear whether the participants in the validation studies are fully representative of the wider population of fibromyalgia patients. We did not identify any studies that evaluated the PROMs for children with fibromyalgia. It is, therefore, uncertain whether the identified PROMs measure sleep outcomes that are most relevant for children with fibromyalgia. Given our concerns about the inadequacy of item coverage in the existing PROMs for adults with fibromyalgia, and the lack of any such PROM for children, we recommend the development of separate core outcome sets (COSs) for measuring sleep outcomes in adults and children with fibromyalgia. Any future COS could then inform the development of new standardised PROMs for measuring sleep outcomes in children and adults with fibromyalgia. It is recommended that further evaluation of the current PROMs, and the development of any future PROMs, is conducted in accordance with guidance from recent relevant initiatives such as the National Institute for Health and Care Research INnovations in Clinical trial design and delivery for the UnDEr-served framework, published in 2020, to enhance the inclusion of diverse participant populations from historically underserved groups, and ensure that outcomes represent what is important and that they are measured in ways that are acceptable to the wider fibromyalgia patient community.298
Strengths and limitations
This systematic review of PROMs for people with fibromyalgia-related sleep problems extends the findings of the previously published review conducted by Climent-Sanz and colleagues (2020)35 by characterising the item content of the PROMs and establishing their comparability. We conducted highly sensitive and comprehensive literature searches as part of our wider mixed-methods complex evidence synthesis to identify PROMs for assessing sleep outcomes that have been validated among fibromyalgia patients. While we employed hand-searching during our screening process, we did not conduct a systematic search of the grey literature. It is, therefore, possible that some relevant articles were not detected, although our clinical expert and patient advisors were unaware of any relevant missing reports. Other established sleep PROMs that have not currently been validated for the fibromyalgia patient group, such as the Insomnia Severity Index and the Sleep diary, were not eligible for inclusion in this review. Their exclusion does not indicate that these PROMs are unsuitable for use with fibromyalgia patients, and it is possible that these PROMs capture additional outcomes that are important to people with fibromyalgia. Scientifically rigorous methods were adopted to identify and code the relevant domains across the included PROMs. Content analysis was conducted by two reviewers, who worked independently before reaching a consensus agreement. While the analysis was informed by the WHO ICF category definitions and terms contained in the Sleep Foundation dictionary, along with advice based on the experiences and reflections of our clinical expert and patient advisors, the content analysis is a subjective qualitative interpretation of the PROMs items. Hence, we cannot exclude with certainty that the same analysis conducted by different researchers with different perspectives and lenses may produce different overall findings.