Issue link: https://www.e-digitaleditions.com/i/1171020
Pharmaceutical Technology REGULATORY SOURCEBOOK SEPTEMBER 2019 13 that the effects of random events on the precision of the analytical procedure should be established while also stating the factors such as days/analysts/ equipment need not be studied individually. The validation of analytical procedures is often treated as a 'check box' exercise (2) where generic validation studies are often used to assess each validation characteristic. This has led to the use of simple/generic IP studies (3) being common prac- tice, which are deemed acceptable by most regula- tory agencies. The Japanese National Institute of Health Sciences (NIHS) proposed a generic design in 2002 (4) that involves the use of six independent analytical runs (5) (five degrees of freedom) with sufficient combination of factors (days, analysts, and equipment) as an acceptable approach for studying intermediate precision. Other approaches in the lit- erature include various designs based on two opera- tors analyzing on two days using two instruments (e.g., eight runs with three samples for each run [6]; four runs using a half fraction of the eight-run fac- torial design with nine tablets for each run [7]) or designs run over a number of days (e.g., two analy- ses of several batches from two matrices on seven different days where it is not stated whether other factors such as analyst are held constant [8]). The NIHS design (4) is illustrated in Table I (not in the test date order as in the NIHS paper [4]). The intent of the six analytical runs (5) shown in Table I is to simulate longer term variation by deliberately varying and setting up the equipment between runs. While Kojima advocated that fac- tors should be determined based on the conduct of the test in the applicant's laboratory and the labo- ratory environment, etc, the authors' experience is that the design shown in Table I is typically used as a generic design. Following the advent of quality by design (QbD) and the development of QbD-related ICH guide- lines for drug substances and drug products (ICH Q8, Q9, Q10, and Q11) (9–12), the same science and risk-based principles have been proposed and applied to analytical procedures (13,14). Instead of using a generic design to study intermediate preci- sion, a risk-based approach can be used to design these studies instead and/or incorporate interme- diate precision in a ruggedness assessment (15,16) that covers all three precision elements: repeatabil- ity, intermediate precision, and/or reproducibility together. Such an approach ensures the factors selected for an intermediate precision study are the most likely (present the highest risk) to impact the performance of the analytical procedure while retaining the Japanese requirement of performing at least six runs. Use of such designs provides an increased level of confidence in the expected vari- Table I: Design for studying intermediate precision proposed as proposed by NIHS (4,5). Condition Factor: Analyst 1 1 1 2 2 2 Equipment A B A B A B Column Used New New New Used Used Test Date 1 2 5 4 3 6 Independent Run 1 2 3 4 5 6 Within Run Two test results for each condition