Bias due to interval censored outcomes in a study of flare risk after hydroxychloroquine taper/cessation in systemic lupus erythematosus