Understanding the practical application of statistics?

The application of statistical methods is specified all through 21 CFR and guidance documents for the pharmaceutical, biopharmaceutical, and medical device industries for these activities:

o  Setting validation criteria and specifications

o  Performing Measurement Systems Analysis (MSA)

o  Conducting stability analysis

o  Using Design of Experiment (DOE) for process development and validation

o  Developing process control charts

o  Determining process capability indices.

The Quality System Regulation (QSR) for medical devices states that manufacturers should take steps, where appropriate, for establishing and maintaining procedures with which to identify valid statistical techniques needed to establish, control, and verify that the process capability and product characteristics are acceptable.

Methods are specified in 21 CFR and guidance documents

Both 21 CFR and guidance documents emphasize the need for statistical methods from discovery through product discontinuation. While 21 CFR specifies the suitable statistical procedures needed to establish both in-process and final specifications; the guidance documents require applying statistical methods for development and validation of measurement systems, process understanding using Quality by Design (QbD) principles, process validation, as well as ensuring that the manufacturing process is in a state of control and is capable.

Even though many statistical methods may be applied to fulfil this part of the QSR; there exist a few commonly accepted methods that all companies can and preferably should use for:

o  Developing acceptance criteria

o  Ensuring accurate and precise measurement systems

o  Fully characterizing manufacturing processes

o  Monitoring and controlling process results, and

o  Selecting an appropriate number of samples.

Learn the dynamics of statistical methods

The nitty gritty of application of statistical methods will be the teaching a two-day seminar from GlobalCompliancePanel, a leading provider of professional trainings for the regulatory compliance areas, will be offering. At this seminar, Heath Rushing, who is the cofounder of Adsurgo and author of the book Design and Analysis of Experiments by Douglas Montgomery: A Supplement for using JMP; will be the Director.

This course has been pre-approved by RAPS as eligible for up to 12 credits towards a participant’s RAC recertification upon full completion. To understand the areas of implementation of statistical methods that comply with the requirements set out in CFR 21 and guidance documents, please enroll for this session by visiting “Understanding the practical application of statistics”

A full description of statistical processes

This course provides instruction on the statistical methods for data analysis of applications related to the pharmaceutical, biopharmaceutical, and medical device industries. The Director will explain the appropriate statistical approaches: descriptive statistics, data intervals, hypothesis testing, ANOVA, regression, ANCOVA, and model building need to be applied. Upon establishment of competence in each of these areas; he will present the way these need to be applied in an industry-specific manner.

During the course of these two days, Heath will cover the following areas:

·        Describe and analyze the distribution of data

·        Develop summary statistics

·        Generate and analyze statistical intervals and hypothesis tests to make data-driven decisions

·        Describe the relationship between and among two or more factors or responses

·        Understand issues related to sampling and calculate appropriate sample sizes

·        Use statistical intervals to setting specifications/develop acceptance criteria

·        Use Measurement Systems Analysis (MSA) to estimate variance associated with: repeatability, intermediate precision, and reproducibility

·        Ensure your process is in (statistical) control and capable.

Article on “Statistical Sampling Plans for Medical Devices”

One of the important aspects of design control of medical devices is statistical sampling plans for medical devices. To gain an understanding of the idea of statistical sampling plans for medical devices, one needs to understand the process of medical device design controls.

Statistical sampling plans for medical devices needs to be seen in this background: Under Sec. 820.30 of Title 21 of Code of Federal Regulations (CRF) the FDA sets out requirements from medical device manufacturers -which want to market certain categories of medical devices in the US -for establishing and maintaining procedures to implements design controls into the device.

An understanding of design controls is necessary first

First of all, what are design controls? Design controls are linearly and logically described and recommended steps that manufacturers have to take for ensuring that they have developed what they meant to develop. In addition, design controls have to also be implemented to ensure that the final product is in line with the expectations and needs from the customer’s perspective.

Statistical sampling plans for medical devices come at a slightly later stage. Design controls pave way for the validation processes of design verification and design, which are done to ensure that the device design has met critical specifications or outputs and fulfill the requirements for the safety requirements, intended use, or specified application.

Verification and validation

This stage makes way for the next, which is fulfillment of design verification and validation (V&V), as required under Sec. 820.50 of 21 CFR. A core part of this code is the requirement from manufacturers for establishing and maintaining procedures to locate valid and proper statistical techniques for the process capability and product characteristics to be considered applicable. These are what constitute statistical sampling plans for medical devices

More of rule of thumb

Statistical sampling plans for medical devices need to be written and based on a well-established statistical foundation. However, the FDA does not prescribe a formal plan for writing down statistical sampling plans for medical devices. These are to be based more on rule of thumb. In other words, there are no acceptable limits violations under statistical sampling plans for medical devices. The statistical sampling plans for medical devices need to be implemented on a case-to-case basis, based on the device’s characteristics and features.

In arriving at statistical sampling plans for medical devices, the FDA sets out the following rule:

Table 1

Binomial Staged Sampling Plans

Binomial Confidence Levels

able 2

Binomial Staged Sampling Plans

Binomial Confidence Levels

ucl = Upper Confidence Level

These constitute the core guidance for statistical sampling plans for medical devices.

Learn more on this topic by visiting: Article on “Statistical Sampling Plans for Medical Devices”