Table 3.
Approaches to missing data, by year, number of follow‐ups, and extent of missing data (n = 47)
| Primary analysis method | Reported a sensitivity analysis | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| CCA | MI | Other | Yes | No | ||||||
| n | % | n | % | n | % | n | % | n | % | |
| Publication year | ||||||||||
| 2013 (n = 13) | 6 | 46 | 3 | 23 | 4 | 31 | 5 | 38 | 8 | 62 |
| 2014 (n = 15) | 9 | 60 | 1 | 7 | 5 | 33 | 6 | 40 | 9 | 60 |
| 2015 (n = 19) | 5 | 26 | 10 | 53 | 4 | 21 | 11 | 58 | 8 | 42 |
| Number of follow‐up assessmentsa | ||||||||||
| 1 (n = 10) | 7 | 70 | 1 | 10 | 2 | 20 | 3 | 30 | 7 | 70 |
| ≥2 (n = 36) | 13 | 36 | 13 | 36 | 10 | 28 | 18 | 50 | 18 | 50 |
| Proportion of complete casesb | ||||||||||
| <50% (n = 15) | 4 | 27 | 6 | 40 | 5 | 33 | 8 | 53 | 7 | 47 |
| 50–75% (n = 18) | 10 | 56 | 4 | 22 | 4 | 22 | 9 | 50 | 9 | 50 |
| 75%–95% (n = 14) | 6 | 43 | 4 | 29 | 4 | 29 | 5 | 36 | 9 | 64 |
| Information missingc | ||||||||||
| Similar (n = 22) | 13 | 59 | 6 | 27 | 3 | 14 | 10 | 45 | 12 | 55 |
| More cost missing (n = 3) | 1 | 33 | 2 | 67 | 0 | 0 | 2 | 67 | 1 | 33 |
| More effect missing (n = 10) | 4 | 40 | 2 | 20 | 4 | 40 | 6 | 60 | 4 | 40 |
Note. % = row percentages. CCA = complete‐case analysis; MI = multiple imputation.
a
Excluding one study with continuous follow‐up (n = 46).
b
For the five studies with less than 5% of incomplete cases, four used CCA and one an ad hoc hybrid method for their primary analysis. One of the five studies conducted a sensitivity analysis to missing data.
c
Excluding 12 studies where this was unclear (n = 35).