Analysis of the imputed datasets suggests that accounting for the missing data explained much of the difference seen between each of the regions and South England, although the point estimate for the Midlands/Wales still suggests lower CVD incidence compared to South England (HR 0.85 (95% CI 0.64 to 1.12)).
The low amount of missing ASAT data explains why pmix could not be estimated in the MIXconc method: due to the low amount of information in the remaining concentration data on ASAT (n = 6 patients), this was not sufficient to discriminate between patients with increased and normal ASAT.
In addition, authors will be expected to provide supplementary appendices in which the details of the method for analysis of missing data are clearly explained.
The steps required in preparing this complex dataset for analysis (transform, account for death, impute missing data – "tdi") are explained by example here, and in more detail in Appendix 1.
The retrospective design explains missing data for ECG, MRI and measurements of bone mineral density, as those examinations were not part of a prospective study protocol.
Participants can choose to leave earlier than this with possible reasons ranging from lack of success and weight regain to weight loss success and the feeling they can 'go it alone' and no longer need the service; this makes the issue of missing data very difficult and explains why we and other studies get large differences in results when using baseline or LOCF.
NSSI = non-suicidal self-injury, xthere are three individuals with missing data on gender, which explains why the number of girls and boys do not add up to the total, †together or every- other week, *not specifically for self-injury, +including alcohol or drug abuse, mental health problems, in jail, prolonged handicap or illness, separated from parent(s), divorce, quarreling during divorce.
This large amount of missing data for percentages was explained by the qualitative data.
These missing data points can be explained in part by the presence of heterozygous regions, because data of ambiguous intensity were converted into missing values.
Throughout the quantitative data collection interim analyses will be conducted on the immature data sets to help identify and explain missing data, and shift the analysis burden.
