Background
Methods
Search strategy
Eligibility criteria
Data extraction and analysis
Results
Literature search yield
First Author | Reasons for or against using the SWD | Variations of the SWD | Settings where the SWD can be applied | Method for estimating sample size | Method for statistical analysis |
---|---|---|---|---|---|
Bellan [93] | Yes | No | Yes | No | No |
Beard/Lewis [94] | Yes | No (see Copas et al. instead) | Yes | No (see Baio et al. instead) | No (see Davey et al. instead) |
Baio [38] | No | No | No | Yes (By simulation) | Yes (GLMM) |
Brown [9] | Yes | No | Yes | No | No |
Brown [19] | Yes | No | Yes | No | No |
Brown [36] | Yes | No | Yes | Yes | No |
Copas [43] | No | Yes | No | No | No |
Davey [35] | No (see Bear/Lewis instead) | No | No | No | Yes (GLMM, Cox Proportional Hazards in cohort SWD setting) |
de Hoop [41] | Yes | No | No | No | No |
Fatemi [95] | Yes | Yes | Yes | No | No |
Fok [27] | Yes | No | No | No | Yes (GLMM) |
Gruber [28] | No | No | No | No | Yes (Causal Modelling) |
Haines | Yes | Yes | Yes | No | No |
Handley [11] | Yes | Yes | No | No | No |
Hargreaves [96] | Yes | No (see Copas et al. instead) | Yes | No (see Baio et al. instead) | No (see Davey et al. instead) |
Hargreaves [97] | No | Yes (Randomisation vs Non-randomisation | No | No | No |
Hemming [17] | Yes | Yes | No | No | No |
Hemming [42] | Yes | No | No | No | No |
Hemming [24] | Yes | No | Yes | No | Yes (GLMM, GEE) |
Hemming [37] | No | Yes | No | Yes (Extension of Hussey) | No |
Hussey [29] | No | No | No | Yes (Wald Test) | Yes (GLMM, GEE with Jack-knife SE, Cluster Summaries) |
Keriel-Gascou [98] | Yes | No | No | No | No |
Kotz [99] | Yes | No | No | No | No |
Kotz [90] | Yes | No | No | No | No |
Kotz [44] | No | No | No | No | No |
Medge [10] | Yes | No | Yes | No | No |
Medge [100] | Yes | Yes | Yes | No | No |
Medge [101] | Yes | No | No | No | No |
Moulton [26] | No | No | No | Yes | Yes (GLMM, GEE, Cox Proportional Hazards) |
Pater [102] | No | No | Yes | No | No |
Prost [103] | Yes | No | Yes | No | No |
Rhoda [15] | Yes | No | No | Yes (Based on Hussey) | No |
Sanson-Fisher [104] | Yes | No | No | No | No |
Schelvis [105] | Yes | No | Yes | No | No |
Scott [30] | No | No | No | No | Yes (GEE with few clusters adjustment) |
Van den Heuvel [31] | Yes | Yes | No | No | Yes (GLMM) |
van der Tweel [106] | Yes | No | Yes | No | No |
Viechtbauer [107] | Yes | No | No | No | No |
Woertman [16] | No | No | No | Yes (Design Effect based on Hussey) | No |
Wolkewitz [108] | No | No | Yes | No | No |
Wyman [32] | Yes | No | No | No | Yes (GLMM) |
Zhan [109] | Yes | No | Yes | No | No |
First Author | Type of SWD | n clusters (n steps) | Statistical analysis method used for primary outcome (Distribution of outcome) | Adjusted for time | Sample size estimation | Publication phase |
---|---|---|---|---|---|---|
Aoun [110] Aoun [111] Aoun [112] Aoun [113] Austin [114] Grande [115] | Cohort | 3 (3) | GLMM (Normal) | Yes | Yes: Cluster Design Effect Adjustment | Results |
Lo [116] | Cross sectional | 54 (3) | GLMM (Binomial) | Unknown | Unknown | Results |
Bacchieri [47] | Cohort | 42 (2) | GLM with RVE (Poisson) | Yes | Yes: No Clustering adjustment | Results |
Badenbroek [76] | Cohort | 40 (4) | Not Mentioned | Unknown | Yes: Cluster Design Effect Adjustment | Protocol |
Bailet [117] | Cross sectional | 174 (2) | GLMM (Normal) | Yes | No | Results |
Bailey [77] | Cohort | 4 (4) | Unknown | Unknown | Unknown | Results |
Bailey [82] | Cross sectional | 6 (6) | GEE Few clusters adjustment (Binomial) | Yes | Yes: Simulated | Results |
Banga [78] | Cross sectional | 4 (4) | Not Mentioned | Unknown | Yes: Woertman method | Protocol |
Barton [59] Somerville [58] | Cohort | 119 (2) | Mann–Whitney U/ANCOVA (Normal) | Yes | Yes: No Clustering adjustment | Results |
Bashour [118] | Cross sectional | 4 (4) | GLMM (Normal) | Yes | Yes: Hussey/Hughes method | Results |
Bennett [119] | Cohort | 15 (3) | GLMM (Normal) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Bernabe-Ortiz [120] | Cross sectional | 6 (6) | GEE (Normal) | No | Yes: Hussey/Hughes method | Protocol |
Bouwsma [51] | Cohort | 9 (9) | Cox PH models with gamma random effects (Time to Event) | Yes | Yes: Hussey/Hughes method | Protocol |
Brimblecombe [80] | Cross sectional | 20 (4) | GLMM (Normal) | Yes | Yes: Simulated | Protocol |
Brown [121] | Cross sectional | 9 (5) | GEE (Binomial) | Unknown | Yes: Cluster Design Effect Adjustment | Protocol |
Brownell [122] | Cohort | 20 (5) | GLMM (Binomial) | Yes | No | Results |
Chavane [123] | Cross sectional | 10 (10) | GLMM (Binomial) | Unknown | Yes: Cluster Design Effect Adjustment | Protocol |
Chinbuah [124] Chinbuah [125] | Cross sectional | 114 (2) | GLMM (Poisson/Binomial) | No | Yes: Cluster Design Effect Adjustment | Results |
Ciliberto [71] Patel [73] | Cohort | 7 (6) | GLM (Binomial)/Unspecified Survival Analysis (Time to Event) | No | Yes: No Clustering adjustment | Results |
Cowan [65] | Cross sectional | 6 (3) | GLM fixed effect for cluster (Binomial) | Yes | Yes: Hussey/Hughes method | Protocol |
Crain [126] Solberg [127] | Cohort | 96 (5) | GLMM (Not specified) | Yes | No | Protocol |
Craine [66] | Cross sectional | 5 (5) | GLM fixed effect for cluster (Binomial) | Yes | Yes: Hussey/Hughes method | Results |
Dainty [128] Morrison [129] | Cross sectional | 32 (4) | GEE (Binomial) | Yes | Yes: Hussey/Hughes method | Results |
De Allegri [130] Gnawali [131] | Cohort | 33 (3) | GLMM (Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Results |
Dilworth [61] Turner [132] | Cross sectional | 5 (5) | GEE/Discourse Mapping (Normal) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Doherty [55] | Cross sectional | 55 (2) | χ2 test for trend (Multinomial)/Kruskal-Wallis (Nonparametric) | No | Yes: Details not clear | Results |
Dreischulte [133] | Cohort | 40 (10) | GLMM (Binomial) | Yes | Yes: Details not clear | Protocol |
Dryden-Peterson [134] | Cross sectional | 20 (10) | GEE (Binomial) | Yes | Yes: Details not clear | Results |
Due [53] | Cohort | 189 (2) | T-test (Normal) | No | Yes: Details not clear | Results |
Duijster [135] Monse [136] | Cohort | 13 (2) | GLMM (Normal) | Yes | No | Results |
Durovni [49] Golub [50] Moulton [26] | Cross sectional | 29 (14) | GLM with Scale Parameter (Poisson)/Cox PH with either Bootstrap SE, RVE or gamma random effect (Time to event) | Yes | Yes: Moulton | Results |
Durovni [137] Trajman [138] | Cross sectional | 14 (7) | GLMM (Poisson and Binomial) | Yes | Yes: Moulton | Results |
Emond [139] | Cross sectional | 9 (3) | GLMM (Normal and Binomial | Yes | Yes: Hussey/Hughes method | Protocol |
Enns [140] | Cross sectional | 4 (4) | GEE (Poisson) | Yes | No | Results |
Etchells [141] | Cross sectional | 2 (2) | GLMM (Normal and Binomial | No | No | Results |
Fernald [142] | Cohort | 506 (2) | GLM adjusting for sampling design and clustering (Normal and Binomial) | No | Yes: Cluster Design Effect Adjustment | Results |
Franklin [143] | Cross sectional | 15 (5) | GLMM (Binomial) | Yes | Yes: Details not clear | Results |
Fuller [81] | Cross sectional | 60 (5) | GLMM (Binomial) | Yes | Yes: Simulated | Results |
Gambia Hepatitis Study Group [14] | Cross sectional | 17 (16) | Not Mentioned | Unknown | Yes: No Clustering adjustment | Protocol |
Gerritsen [144] Leontjevas [145] Leontjevas [146] Leontjevas [147] | Cohort | 5 (5) | GLMM (Normal and Binomial) | Yes | Yes: Hussey/Hughes method | Results |
Golden [34] | Cross sectional | 23 (4) | GLMM/GEE (Binomial/Poisson) | Yes | Yes: Hussey/Hughes method | Results |
Groshaus [67] | Cohort | 6 (6) | GLM fixed effect for cluster (Normal and Binomial) | No | No | Results |
Gruber [148] | Cohort | 24 (6) | GEE with RVE (Binomial) | Yes | Yes: Hussey/Hughes method | Results |
Grunewaldt [54] | Cohort | 20 (2) | χ2 tests (Binomial)/t-tests (Normal) | No | No | Results |
Gucciardi [149] | Cross sectional | 4 (3) | GLMM/GEE (Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Haines [150] | Cross sectional | 6 (6) | GLMM (Normal) | Yes | Yes: Hussey/Hughes method | Protocol |
Haugen [74] | Cross sectional | 5 (5) | GLM (Binomial) | Yes | Yes: Details not clear | Results |
Haukka [151] | Cross sectional | 9 (4) | Unspecified Survival Analysis (Time to event)/GEE (Normal) | No | No | Protocol |
Hayden [152] | Cohort | 4 (4) | GLMM (Binomial) | Yes | Yes: Details not clear | Results |
Hill [153] Hill | Cross sectional | 8 (4) | GLM RVE (Negative Binomial and Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Results |
Horner [154] | Cohort | 65 (3) | GLMM (Binomial) | Yes | No | Results |
Howlin [155] | Cohort | 18 (2) | GLMM (Ordinal) | No | Yes: Cluster Design Effect Adjustment | Results |
Hunter [156] | Cross sectional | 8 (2) | GLMM (Binomial) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Husaini [157] | Cohort | 45 (2) | GLMM (Binomial) | No | No | Results |
Kelly [68] | Cross sectional | 4 (4) | GLM fixed effect for cluster or Bootstrap SE (Normal) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Keriel-Gascou [98] | Cross sectional | 8 (4) | GLMM (Binomial) | Yes | Yes: Hussey/Hughes method | Protocol |
Killam [158] | Cross sectional | 8 (8) | GEE (Binomial) | Yes | No | Results |
Kitson [159] Schultz [160] | Cross sectional | 25 (4) | GLMM (Normal) | Yes | Yes: Hussey/Hughes method | Results |
Kjeken [161] | Cohort | 6 (6) | GLMM (Normal) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Larsen [46] | Cross sectional | 46 (18) | GLM with RVE (Binomial) | Yes | Yes: Details not clear | Results |
Li [162] | Cross sectional | 104 (4) | GEE (Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Liddy [163] | Cohort | 83 (3) | GLMM (Normal) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Lilly [87] | Cross sectional | 7 (7) | GLMM (Binomial and Normal) | Yes | Yes: No Clustering adjustment | Results |
Maheu-Giroux [164] | Cross sectional | 6 (3) | GLMM (Binomial) | Yes | No | Results |
Marrin [165] | Cross sectional | 6 (3) | GLMM (Normal) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Marshall [166] | Cross sectional | 32 (10) | GEE (Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Mhurchu [167] Mhurchu [168] | Cohort | 14 (4) | GLMM (Binomial and Normal) | Yes | Yes: Hussey/Hughes method | Results |
Mosha [72] | Cross sectional | 10 (10) | GLM (Binomial) | No | No | Results |
Mouchoux [62] | Cross sectional | 3 (3) | GLM fixed effect for cluster (Binomial) | Yes | Yes: Hussey/Hughes method | Protocol |
Muntinga [169] | Cohort | 35 (4) | GLMM (Normal) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Ononge [75] | Cross sectional | 6 (Unclear) | GLM (Normal and Binomial) | No | No | Results |
Palmay [170] | Cross sectional | 6 (6) | GLMM (Negative binomial) | Yes | No | Results |
Palmer [83] | Cohort | 11 (3) | GLMM (Normal) | Yes | Yes: Simulated | Protocol |
Pearse [52] | Cross sectional | 90 (15) | GLMM (Binomial)/Cox PH (Time to Event) | Yes | Yes: Hussey/Hughes method | Protocol |
Pickering [171] | Cohort | 4 (4) | GLMM (Normal) | Yes | Yes: Details not clear | Results |
Poldervaart [172] | Cross sectional | 10 (10) | GLMM (Binomial) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Priestley [88] | Cross sectional | 16 (7) | GLMM (Binomial) | Unknown | Yes: No Clustering adjustment | Results |
Proeschold-Bell [79] | Cohort | 3 (3) | Unknown | Unknown | No | Protocol |
Rasmussen [173] | Cohort | 21 (4) | GLMM (Normal) | No | Yes: Woertman method | Results |
Reuther [174] | Cohort | 12 (5) | GLMM (Binomial) | Yes | Yes: Hussey/Hughes method | Protocol |
Roy [175] | Cross sectional | 27 (5) | GLMM (Binomial) | No | Yes: Cluster Design Effect Adjustment | Results |
Schnelle [60] | Cohort | 2 (2) | Repeated Measures ANOVA (Normal) | Yes | Unknown | Results |
Skrovseth [57] | Cohort | 2 (2) | GLM (Normal)/Wilcoxon Rank sum (Non parametric) | No | Yes: Details not clear | Results |
Solomon [176] Solomon [177] | Cross sectional | 128 (4) | GLMM/GEE (Normal/Binomial) | Yes | Yes: Hussey/Hughes method | Results |
Stern [84] | Cohort | 12 (11) | GLMM (Normal) | Yes | Yes: Simulated | Results |
Strijbos [89] | Cross sectional | 8 (4) | GLMM (Binomial) | No | Yes: No Clustering adjustment | Protocol |
Suman [178] | Cohort | 4 (4) | GLMM (Binomial) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Tielsch [85] | Cohort | 12 (12) | GLMM (Normal and Binomial) | Yes | Yes: Simulated | Protocol |
Tiono [179] | Cohort (Cross sectional subgroup) | 40 (8) | GLMM (Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Protocol |
Tirlea [180] Tirlea [181] | Cohort | 12 (3) | GLMM (Normal) | Yes | Yes: Cluster Design Effect Adjustment | Results |
Toftegaard [182] | Cohort (GP level) Cross Sectional (patient level) | 8 (8) | GLMM (Binomial) | No | Yes: Woertman method | Protocol |
Viera [56] | Cohort | 3 (2) | McNemar’s test (Binomial) | No | Retrospectively: No clustering adjustment | Results |
Ward [183] | Cross sectional | 24 (3) | GLMM/GEE (Binomial) | No | Yes: Cluster Design Effect Adjustment | Protocol |
Weiner [69] | Cross sectional | 36 (36) | GLM fixed effect for cluster (Binomial) | No | No | Results |
Williams [184] | Cohort | 12 (3) | GLMM (Normal) | Yes | Yes: Hussey/Hughes method | Protocol |
Williamson [70] | Cohort | 6 (3) | GLM fixed effect for cluster (Binomial and Normal) | Yes | Yes: Hussey/Hughes method | Protocol |
Wilrycx [185] | Cohort | 2 (2) | GLMM (Normal) | Unknown | No | Results |
Zwijsen [186] Zwijsen [187] Zwijsen [188] Zwijsen [189] | Cohort | 17 (5) | GLMM (Binomial) | No | Yes: Hussey/Hughes method | Results |
van Daalen [86] | Cross sectional | 9 (4) | GLMM/GEE (Normal/Binomial) | No | Yes: Simulated | Protocol |
van de Steeg [190] van de Steeg [191] | Cross sectional | 18 (10) | GLMM (Binomial) | Yes | Yes: Cluster Design Effect Adjustment | Results |
van den Broek [63] van den Broek [64] | Cross sectional | 190 (3) | GLM fixed effect for cluster (Binomial) | No | Yes: Simulated | Results |
van Holland [192] | Cohort | 5 (5) | GLMM (Normal and Binomial) | Yes | Yes: Hussey/Hughes method | Protocol |