10.17037/DATA.272
This dataset contains information extracted from studies published in the NIHR Health Technology Assessment journal between the 1st January 2013 and 31st December 2015. The study team subsequently used it as a basis to critically reviews the extent of and methods used to address missing data in recently published trial-based CEA. Detailed methodology and variables definition are reported in the accompanying Health Economics paper, “Missing data in trial-based cost-effectiveness analysis: an incomplete journey”.
The following variables are contained in the missing_data_review dataset.
Variable Name | Variable Label | Answer Label | Answer Code | Variable Type |
author | Author | Open ended | String | |
year | Year | Open ended | Integer | |
futime | FU time (months) | Open ended | double | |
fucat | Follow-up design | Numeric | ||
Continuous | 1 | |||
One follow-up | 2 | |||
Repeated assesments | 3 | |||
nfu | Number of assesments, including baseline | Open ended | Numeric | |
effectcat | Primary effectiveness measure type | Numeric | ||
QALY | 1 | |||
Binary | 2 | |||
Clinical scale | 3 | |||
Time to recovery | 4 | |||
pcc | Proportion of complete cases | Open ended | double | |
n | Number complete cases | Open ended | Integer | |
n_ | Number of participants | Open ended | Integer | |
reportcc | Report exact CC | Numeric | ||
No | 0 | |||
Yes | 1 | |||
arm1 | Proportion CC arm1 | Open ended | double | |
arm2 | Proportion CC arm2 | Open ended | double | |
diffarm | Missingness differs between arms | Numeric | ||
Yes | 1 | |||
No | 2 | |||
No missing | 3 | |||
Unclear | 8 | |||
cceffect | CC effect | Open ended | double | |
cccost | CC cost | Open ended | double | |
missinfo2 | Missingness cost vs. effect | Numeric | ||
None | 1 | |||
Both missing | 2 | |||
Cost more missing | 3 | |||
Effect more missing | 4 | |||
Unclear | 8 | |||
md | Included - missing data >5% | Open ended | Numeric | |
primcat | Primary analysis category | Numeric | ||
CC | 1 | |||
MI | 2 | |||
Other | 3 | |||
Mixed | 4 | |||
Unclear | 5 | |||
primcat3 | Primary analysis | Numeric | ||
CC | 1 | |||
MI | 2 | |||
Other | 3 | |||
mi | Conducted multiple imputation | Numeric | ||
No | 0 | |||
Yes | 1 | |||
cc | Conducted complete-case analysis | Numeric | ||
No | 0 | |||
Yes | 1 | |||
sa | Conducted sensitivity analysis | Numeric | ||
No | 0 | |||
Yes | 1 | |||
other1 | Secondary approach to MD 1 | Open ended | String | |
other2 | Secondary approach to MD 2 | Open ended | String | |
other3 | Secondary approach to MD 3 | Open ended | String | |
mnarsa | MNAR SA | Numeric | ||
No | 0 | |||
Yes | 1 | |||
a1 | Recommendations indicator A1.Maximise response rate | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
a2 | A2.Alternative data sources | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
a3 | A3.Monitor completeness | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
b1 | B1.Assumption for primary analysis | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
b2 | B2.Appropriate primary method | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
c1 | C1.Discuss departures from the primary assumption | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
c2 | C2.Consider broad range of assumptions | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
c3 | C3.Method valid under these assumptions | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
d1 | D1.Missing data by endpoint, arm and time-point | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
d2 | D2.Discuss reasons for missing data | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
d3 | D3.Describe methods used and assumptions | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||
d4 | D4.Conclusions in light of missing data | Numeric | ||
Yes | 1 | |||
No | 2 | |||
Unclear | 3 | |||