Table 1 Data recoding
Variable | Original responses | Recoded responses | Rationale |
|---|---|---|---|
Country of Origin | Australia New Zealand South Africa England Phillipines Fiji India China | Australia Asia-Pacific Other | Prevent quasi-separation in between-group comparisons |
Comorbidities | Free text | None Single Multiple | Include as a model predictor |
Secondary diagnosis | Free text | Medial meniscus pathology (Yes; No) Lateral meniscus pathology (Yes; No) | Include as a model predictor |
Quadriceps measurement | 1RM Not recorded HHD MMT | Yes No | Prevent quasi-separation in between-group comparisons; address missing data |
Hamstrings measurement | Not recorded HHD MMT | Yes No | Prevent quasi-separation in between-group comparisons; address missing data |
Balance measurement | SL Balance SEBT Not recorded | Yes No | Prevent quasi-separation in between-group comparisons; address missing data |
Rehab Status | Completed (Discharged) Failed to attend Withdrew | Completed Did not complete | Prevent quasi-separation in between-group comparisons |
Number of staff involved | Integer 1 - 9 | 1-3 4-6 7+ | Prevent quasi-separation in between-group comparisons |
Complications | Free text | Yes No | Include as a model predictor |
Knee extension angle | Continuous (Degrees) | > 5 < 5 | Relate to a clinically meaningful threshold |
Reason for failed to attend | Free text | Too hard to attend Discharge belief Happy with knee Changed service | Summarise into key themes |
IKDC - subjective knee function score | Continuous, 0 - 100 | > PASS < PASS | Relate to a clinically meaningful threshold |