The limits of compliance can be inferred by the parametric method if the normality of the differences is indicated. or the use of non-parametric percentiles, if these assumptions are not included. In particular, 10,000 iteration simulation studies were conducted to calculate the probability of simulated coverage of accurate and approximate confidence intervals for percentiles in a standard distribution N (0, 1). The sample size is six different sizes: No. 10, 20, 30, 50, 100 and 200. In addition, eight probabilities of percentiles are studied: p – 0.025, 0.05, 0.10, 0.20, 0.80, 0.90, 0.95 and 0.975. For each replication, the lower and upper confidence limits () (“widehat” (Uptheta) L , “uptheta” U, DIE A/ (“widehat” (“breithat” – “uptheta”) AL, “widehat” and “uptheta” and “breithat”) were calculated to establish unilateral confidence intervals of 95 and 97.5%, as well as reciprocal confidence intervals of 90 and 95%. The probability of simulated coverage was the proportion of 10,000 replications whose confidence interval contained the normal percentile of the population. Second, the adequacy of one- and two-side interval procedures is determined by error – simulated coverage probability – nominal probability of coverage. The results are summarized in Tables 1, 2, 3 and 4 for precise and approximate confidence intervals with two-sided confidence coefficient 1 – α – 0.90 and 0.95.

Choudhary PK, Nagaraja HN. Measuring compliance in method comparison studies – an audit. In: Balakrishnan N, Kannan N, Nagaraja HN, editor. Investment and selection progress, multiple comparisons and reliability. Boston: Birkhauser; 2004. 215-44. Lin LI, Hedayat AS, Sinha B, et al. Statistical methods for evaluating the agreement: models, problems and instruments. J Am Stat Assoc. 2002;97:257–70. Researchers studied the agreement between primary care and ambulatory monitoring during the day for blood pressure measurement. The study subjects were patients with newly diagnosed high blood pressure or borderline blood pressure, or patients receiving treatment for hypertension but having poor control.

A total of 179 patients were recruited from three general practices and eight physicians participated in blood pressure measurement. Daily outpatient monitoring was conducted between 0700 and 2300 hours.1 Although the practical implementation of the exact interval procedure at Carkeet [19] is well illustrated, the explanation of the differences between the exact and approximate methods focused mainly on the relative sizes and symmetrical/asymmetric limits resulting from the confidence limits. On the other hand, Bland-Altman`s 95% agreement limits are generally considered to be related to the measurement of compliance in comparing methods. Carkeet [19] and Carkeet and Goh [20] therefore focused on comparing approximate confidence intervals for the upper and lower limits of torque chords and tolerance intervals on both sides for normal distribution.