Background
Methods
Protocol and registration
Eligibility criteria
Information sources
Search
Study selection
Data collection process and data items
Risk of bias in individual studies
Summary measures
Results
Study selection
Study characteristics
Authors (Date) | Measure | Reliability (Cronbach’s Alpha) | Internal consistency (Inter-Item Correlations) | Test-test Reliability | Measurement error |
---|---|---|---|---|---|
Moran & Mohr (2005) [39] | BDI | BDI SEM = 3.26; HRSD SEM = 3.20 | |||
Aikens et al. (1999) [24] | BDI-II | Full BDI α = .86, Cognitive / Affective α = .85 Somatic α = .64 BDI-18 α = .84 | |||
Amtmann et al. (2014) [33] | PHQ-9; CESD-10; PROMIS-D-8 | PHQ-9 = .35-.67; CESD-10 = .33-.67; PROMIS-D-8 = .75-.84 | |||
Sjonnesen et al. (2012) [46] | PHQ-9 |
α = .82 | Item-total correlations (PwMS): anhedonia (.71); depressed mood (.65); fatigue (.57) and concentration (.55). | ||
Chang et al. (2003) [37] | CMDI | Sleep Disturbance α = .64 Cognitive Inefficiency α = .86 Fatigue α = .94 Evaluative α = .82 Mood α = .95 | |||
Solari et al. (2003) [49] | CMDI | MS Vegetative α = .88 Evaluative α = .7 Mood α = .93 | Item-subscale: mood = .58-.80, evaluative = .37-.70, vegetative = .32-.57 | ICC = 0.78 (95 % CI 0.62–0.89); Subscales range 0.71–0.79 (2 weeks) | |
Patten et al. (2010) [45] | CES-D |
r = .69 over 1 year; r = .70 over 2 years; r = .65 over 3 years; r = .73 year 1-year 2; r = .73 year 2-year 3 | |||
Verdier-Taillefer et al. (2001) [48] | CES-D |
α = .90 |
Authors (Date) | Measure | Structural Validity (Factor Analysis) | Content Validity Item Inclusion |
---|---|---|---|
Mohr et al. (1997) [70]a
| BDI | Fatigue, work difficulty, and concerns about health (ps < .001) within MS population | |
Moran & Mohr (2005) [39] | BDI | All items showed significant reductions following treatment (p < .05) | |
Aikens et al. (1999) [24] | BDI-II | MS had higher scores than healthy controls on work difficulty, F = 8.05, p < .001 and sexual disinterest, F = 9.99, p < .005. | |
Strober & Arnett (2010) [41] | mBDI | MS-NON-DEP: fatigue, indecision, loss of libido, work difficulty, irritability, loss of interest, crying, dissatisfaction, self-criticism. DEP-PwMS: Irritability, loss of interest, crying, dissatisfaction, self-criticism, sadness, pessimism, failure, guilt, appetite, disappointment, weight loss. MS (more severe in DEP-MS) or related to depressive symptoms: irritability, loss of interest, crying, dissatisfaction, self-criticism. | |
Amtmann et al. (2014) [33] | PHQ-9; CESD-10; PROMIS-D-8 | Fit indices from a one-factor CFA were acceptable. CFI for all models ≥ of 0.95. TLIs for PHQ-9 and CESD-10 < 0.95 (0.94 and 0.93, respectively). TLI > 0.95 for PROMIS-D-8. RMSEAs > .05 for any of the measures. | |
Sjonnesen et al. (2012) [46] | PHQ-9 | Exclusion of fatigue and concentration items no change in prevalence estimates (p > .05). Fatigue item contributed the most to the total score in both groups (MS 35.1 % (95%CI 30.9–39.3), control 34.8 % (95%CI 33.5–36.1). Item contribution for fatigue and concentration items between groups (p > .05). Anhedonia was lower in MS than controls (OR 0.47, p = .03), PwMS positively endorsed guilt (OR 2.17, p = .03) and fatigue (OR 1.51, p = .05); frequency of endorsement between groups (p > .05) | |
Patten et al. (2005) [44] | CES-D | Agreement between full and modified scales, r = .99, 1.00, .98; k = .93, .96, .90. Prevalence for full scale (>15) was 32.8 % (95%CI 18.9–36.8), exc. fatigue 30.0 % (95%CI 26.3–34.0), exc. cognitive 31.0 % (95%CI 27.3–35.0), exc. both 30.4 % (95%CI 26.6–24.4). | |
Verdier-Taillefer et al. (2001) [48] | CES-D | Factor 1 = depressed affect; Factor 4 = interpersonal relationships. Factor 2 = positive affect items (MS group), but contained somatic complaints (other groups); Factor 3 = somatic complaints (MS group) but positive affect (other groups) | |
Chang et al. (2003) [37] | CMDI | CFA confirmed the subscale/factor structure | Vegetative subscale showed misfitting in PwMS. ‘Fatigue’ and ‘useless’ items were endorsed more by MS than controls. |
Mohr et al. (2007) [38] | Two-item measure | The sensitivity and NPV of using either question alone significantly better than either question alone; the sensitivity of both questions was significantly lower than other methods |
Authors (Date) | Measure | Discriminant | Convergent |
---|---|---|---|
Aikens et al. (1999) [24] | BDI-II | BDI totals differed by group, F = 30.99, p < .001; controls less depressed than MS, p < .05, and MS less depressed than chronic pain patients, F = 12.9 p < .001. On the somatic subscale, MS scored higher than all other groups (ps < .001), except diabetes patients (ns) | BDI totals similar for MS and diabetes F = 3.84, p = ns |
Moran and Mohr (2005) [39] | BDI | BDI total score reduced from 23.7 (SD = 6.9) to 10.5 (SD = 6.5) following treatment, t = 10.91, p < .001 | |
Sullivan et al. (1995) [47] | BDI | Participants with major depression scored significantly higher on negative attitude towards self and performance components, F(2,41) = 3.98, p < .05, F(2,41) = 3.42, p < .05, but not on the somatic component, F(2,41) = 1.30, p = ns. | |
Benedict et al. (2003) [36] | BDI-FS | Groups differed in scores between treated and untreated groups (anti-depressants), F = 13.26, p < .01, persisting after controlling for physical disability. Treatment effect size was largest for BDI-FS than other measures (d = 1.5). | |
Strober and Arnett (2010) [41] | mBDI | PwMS (DEP, NON-DEP) endorsed fatigue (p < .001), work difficulty (p < .001), indecision (p < .01), irritability (p < .01), loss of libido (p < .05), loss of interest (p < .01), crying (p < .05), dissatisfaction (p < .05 and self-criticism (p < .05) on the BDI | BDI and depression proneness rating scale, r = .52, p < .01; BDI and CMDI, r = .76, p < .01 |
Amtmann et al. (2014) [33] | PHQ-9; CESD-10; PROMIS-D-8 | PHQ-9 with CESD-10 (.85); PROMIS-D-8 (.73). CESD-10 with PROMIS-D-8 (.80). | Fatigue scores correlations: PHQ 9 = .73; CESD-10 = .71; PROMIS-D-8 = .55. Sleep disturbance: PHQ-9 = .57; CESD-10 = .56; PROMIS-D-8 = .39. Pain interference: PHQ-9 = .60; CESD-10 = .55; PROMIS-D-8 = .47. |
Pandya et al. (2005) [43]a
| CES-D | MS with a major depressive disorder diagnosis had CES-D score (36.5) than those who did not (26.0), Mann-Whitney = 12.1, df = 1, p < .001 | |
Patten et al. (2005) [44] | CES-D | CES-D scores did not correlate as highly with physical QoL, as with mental QoL, Spearman's r = -.64. CES-D scores and the physical, Spearman's r = .61, or the cognitive, Spearman's r = .66, subscales of the fatigue impact scale. | CES-D scores correlated highly with mental QoL, Spearman's r = -.80. CES-D scores and the Fatigue Impact Scale social subscale, Spearman's r = .74. |
Verdier-Taillefer et al. (2001) [48] | CES-D | PwMS endorsed more depressed affect and fatigue items than GP patients; MS and GP patients endorsed more depressive symptoms than healthy workers (ps < .05) | |
Beeney and Arnett (2008) [35] | CMDI | Physical disability was correlated with vegetative subscale, r(94) = .36, p < .05; mood, r(94) = .16, p = ns, or evaluative, r(94) = .20, p = ns, subscales. | History of depression correlated with CMDI-Mood, r(94) = .38, p < .001; CMDI-Evaluative, r(94) = .31, p < .001; CMDI-Vegetative r(94) = .10, p = ns. Depression proneness correlated with mood, r(94) = .53, p < .001; evaluative, r(94) = .48, p < .001, subscales; vegetative subscale, r(94) = .15, p = ns. |
Chang et al. (2003) [37] | CMDI | PwMS scored higher on all dimensions than controls (CMDI-Mood t = 6.10, p < .001, CMDI-Evaluative t = 6.61, p < .001, CMDI-Vegetative t = 8.04, p < .001). | |
Nyenhuis et al (1995) [40] | MDI | PwMS endorsed more vegetative symptoms than mood and evaluative, than the depressed group, controlling for total BDI score F(2,322) = 5.7, p < .01; PwMS endorsed less mood items than the depressed group, t(161) = 2.5, p < .05, with no difference in evaluative, t(163) = 0.34, p > .10, or vegetative, t(163) = -0.58, p > .10 items. | BDI (30.5 %) estimated depression significantly greater than MDI-mood (17.7 %), p < .05; MDI-Vegetative (34.6 %) estimated significantly higher prevalence than MDI-Evaluative (22.2 %), p < .05, and MDI-Mood, p < .01; MDI-Mood estimated lower prevalence than MDI-Total (26.6 %), p < .05. |
Solari et al. (2003) [49] | CMDI | Significant differences in all subscale scores, p < .001. Adjusted odds ratios for MS vs. healthy controls for depression was 2.72 (95%CI 1.14–6.50, p = .02) total scale; 2.79 (95 %CI 1.19–6.54, p = .02); mood scale, 2.00 (95%CI 0.88–4.56, p = .10); evaluative scale; 6.49 (95%CI 1.92–21.93, p = .001) vegetative scale. | |
Nicholl et al. (2001) [31] | HADS-D | HADS-D correlated with both the GHQ-12 (r = .49, p < .01), and the GHQ-28 (r = .48). Kappa between HADS-D and BDI 0.12, p = .22. |
Authors (Date) | Measure | Criterion Measure | Correlation | Sensitivity | Specificity | ROC Curve Analysis | PPV | NPV |
---|---|---|---|---|---|---|---|---|
Benedict et al (2003) [36] | BDI-FS | BDI, CES-D, neuropsychiatric inventory (NPI) | BDI r = .85, p < .001; CES-D r = .86, p < .001; NPI subscales range r = .32-.50, ps < .05 | |||||
Sullivan et al. (1995) [47] | BDI | Structured clinical interview (DSM) | 88 % (≥9), 71 % (≥13) | 46 % (≥9), 79 % (≥13) | 54 % (≥9), 70 % (≥13) | 84 % (≥9), 79 % (≥13) | ||
Honarmand and Feinstein (2009) [42] | HADS-D (≥8) | SCID-I (DSM) | 90 %, (95%CI 73–98)* 86.7 % (95%CI 68–96)* fatigue item excluded (≥6) | 87.3 %, (95%CI 81–92)* 86.7 % (95%CI 80–92)* fatigue item excluded (≥6) | AUC = 0.94, 0.93 fatigue item excluded (≥6) | 90.6 %, 90.6 % fatigue item excluded (≥6) | ||
Nicholl et al. (2001) [31] | HADS-D | BDI (Predictive) |
r = .58, p < .01 | 25 % - When using cut-off of 5/6 (identified by ROC curve), 75 % | 86 % - When using cut-off of 5/6 (identified by ROC curve), 69 % | Identified the optimum cut-off of 5/6 when compared to the BDI. | ||
Avasarala et al. (2003) [34] | YSQ | BDI (Cut-off of ≥13) | 65.3 % (95%CI 50–78) | 87.3 % (95%CI 77–94) | 78.0 % (95%CI 62–89) | 78.5 % (95%CI 68–87) | ||
Mohr et al. (2007) [38] | Two-item measure | SCID (DSM) | Both questions: 51 % (95%CI 38–63), either: 99 % (95%CI 91–100), 1 only: 75 % (95%CI 62–84), 2 only: 75 % (95%CI 62–84) | Both: 98 % (95%CI 94–99), either: 87 % (95%CI 81–91), 1 only: 94 % (95%CI 89–97), 2 only: 94 % (95%CI 89–97) | Both: 90 % (95%CI 74–97), either: 72 % (95%CI 61–80), 1 only: 73 % (95%CI 61–83), 2 only: 81 % (95%CI 68–89) | Both: 85 % (95%CI 80–89), either: 99 % (95%CI 96–100), 1 only: 91 % (95%CI 86–95), 2 only: 91 % (95%CI 86–95) | ||
Pandya et al. (2005) [43] | CES-D | DSM-IV psychiatric interview | 59.6 % (95%CI 45.0–74.1) for major depression; 74.5 % (95%CI 61.5–87.4) for any depressive disorder. | |||||
Solari et al. (2003) [49] | CMDI | Previously diagnosed according to DSM-IV | Total scale: 96.8 %; mood: 96.9 %; evaluative and vegetative: 75.0 % | |||||
Vahter et al (2007) [50] | One-item measure | BDI (>10 cut-off) and structured clinical interview | 81%
a
| 89%
a
| 93.5%
a
|
Risk of bias within studies (COSMIN)
PAPERS | Measure | Internal Consistency | Reliability | Measurement Error | Content Validity (Face Validity Incl.) | Structural Validity | Hypotheses testing | Cross Cultural Validity | Criterion Validity | Responsiveness |
---|---|---|---|---|---|---|---|---|---|---|
Mohr et al. (1997) [70] | BDI | - | - | - | Excellent | - | Fair | - | - | - |
Moran & Mohr (2005) [39] | BDI | - | - | Fair | Excellent | - | Fair | - | - | Fair |
Sullivan et al. (1995) [47] | BDI | - | - | - | Fair | - | Fair | - | Fair | - |
Aikens et al. (1999) [24] | BDI-II | Fair | - | - | Excellent | - | Fair | - | - | - |
Benedict et al. (2003) [36] | BDI-FS | - | - | - | Fair | - | Fair | - | Fair | - |
Strober & Arnett (2010) [41] | mBDI | - | - | - | Excellent | - | Fair | - | - | - |
Pandya et al. (2005) [43] | CES-D | - | - | - | - | - | Fair | - | Poord
| - |
Patten et al. (2005) [44] | CES-D | - | - | - | Excellent | - | Fair | - | - | - |
Patten et al. (2010) [45] | CES-D | - | Fair | Fair | - | - | Good | - | - | Fair |
Verdier-Taillefer et al. (2001) [48] | CES-D | Good | - | - | Excellent | Good | Good | - | - | - |
Amtmann et al. (2014) [33] | CESD-10; PHQ-9; PROMIS-D-8 | Poora
| - | - | Fair | Fair | Fair | - | - | - |
Sjonnesen et al. (2012) [46] | PHQ-9 | Poorb
| - | - | Fair | - | Fair | - | - | - |
Beeney & Arnett (2008) [35] | CMDI | - | - | - | Excellent | - | Fair | - | - | Fair |
Chang et al. (2003) [37] | CMDI | Fair | - | - | Excellent | Fair | Fair | - | - | - |
Nyenhuis et al. (1995) [40] | MDI | - | - | - | Fair | - | Fair | - | - | - |
Solari et al. (2003) [49] | CMDI | Fair | Fair | - | Excellent | - | Fair | Poorc
| - | - |
Honarmand & Feinstein (2009) [42] | HADS | - | - | - | - | - | Fair | - | Fair | - |
Nicholl et al. (2001) [31] | HADS | - | - | - | - | - | Good | - | Good | - |
Avasarala et al. (2003) [34] | YSQ | - | - | - | - | - | Fair | - | Fair | - |
Mohr et al. (2007) [38] | Two-item measure | - | - | - | Excellent | - | Fair | - | Fair | - |
Vahter et al (2007) [50] | One-item measure | - | - | - | - | - | Fair | - | Fair | - |