Reducing variability in apparent operative inclination during total hip arthroplasty: findings of a randomised controlled trial