Table 3.
Reported influences on accrual predictions during trial design stratified by final accrual sufficiency status
| Potential factor influencing patient accrual predictions | Median score and percent citing factor | |||||
|---|---|---|---|---|---|---|
| Study Chairs | Lead Statisticians | |||||
| Percent citing factor | Accrual status | Percent citing factor | Accrual status | |||
| Sufficient | Insufficient | Sufficient | Insufficient | |||
| SC's experience in disease, disease stage, or intervention | 95% | 2 | 3 | 94% | 3 | 3 |
| Cooperative group's experience in disease, disease stage, or intervention | 99% | 1 | 1 | 97% | 2 | 2 |
| Accrual estimates by experts on disease sub-committee | 94% | 3 | 3 | 87% | 3 | 3 |
| Accrual estimates by local investigators at participating institutions | 93% | 3 | 3 | 81% | 3 | 3 |
| Survey or focus group of prospective participating clinicians | 58% | 4 | 4 | 65% | 4 | 4 |
| Survey or focus group of prospective participating patients | 41% | 5 | 5 | 58% | 5 | 5 |
| Published accounts of accrual experience in similar trials | 70% | 3 | 3 | 76% | 3 | 3 |
Results presented by median score on five-point Likert score for respondents considering factor relevant to trial and by percent factor not used for respondents considering factor not relevant to trial.
Likert scale from 1-5 used to rate strength of influence
1=Strong influence; 3=Some influence; 5=No influence