What Is Probabilistic Data Sampling In Environmental Mointoring

7-October-2019

Axel Schroeder of Concept Heidelberg and Raphael Bar of BR Consulting analyse the steps to compliant EM monitoring alee of the course EM Monitoring Data Management on xx-21 November hosted by ECA University in Barcelona

The manufacture of pharmaceutical products must be performed nether controlled conditions. The production rooms of pharmaceutical facilities should have an appropriate blueprint, which depends on the requirements of the products to exist manufactured. Rooms, including walls, ceilings and floors, technical facilities like heating, ventilation, and air conditioning (HVAC), and equipment, including machinery and piece of furniture, must be designed in such a way that defined environmental weather condition for production can be met.

Constant environmental conditions are essential to guarantee a consistent quality of the manufactured products. Of form, this too applies to the staff working in these rooms. That is why the relevant directives crave monitoring the environmental conditions and the personnel.

Under both non-sterile and aseptic manufacturing weather condition, monitoring is the method of choice for detecting deviations from the qualified or validated status.

Monitoring is, therefore, an important office of proving that the manufacturing process is nether control, and in hygienic production it plays a peculiarly of import role.

Monitoring the environmental conditions in non-sterile production areas has besides gained in relevance recently. It has become credible that microbiological monitoring in these areas serves as a basis for necessary decisions and procedures, both in the analysis of deviations and in quality take a chance assessment.

At that place are relevant guidelines and regulations for microbiological monitoring in sterile and not-sterile production, and these are, in practice, supplemented by the definition of warning and action limits.

The post-obit list shows some relevant guidelines and regulations for microbiological monitoring in sterile and not-sterile production areas:

EU GMP guideline, Annex 1: Industry of Sterile Medicinal Products
ISO 13408-1 Aseptic processing of health care products ISO 14644-1 Cleanrooms and associated controlled environments: Classification of air cleanliness by particle concentration
ISO 14698-1 Cleanrooms and associated controlled environments: Biocontamination command
FDA, Guidance for Manufacture: Sterile Drug Products Produced by Aseptic Processing - Electric current Good Manufacturing Exercise • USP Chapter <1115> Bioburden Control of Nonsterile Drug Substances and Products
USP Affiliate <1116> Microbiological Command and Monitoring of Hygienic Processing Environments
PDA Technical Report No. 67 - Exclusion of Objectionable Microorganisms from Nonsterile Pharmaceuticals, Medical Devices, and Cosmetics. 2014: Parenteral Drug Clan, Inc.
PDA Technical Written report No. xiii (Revised) - Fundamentals of an Environmental Monitoring Programme. Parenteral Drug Association, Inc., 2014
ANVISA Quality Guide for Air Handling and Environmental Monitoring Systems in the Pharmaceutical Industry (2013)
Pic/Due south PE 009-13 Guide to Good Manufacturing Practice for Medicinal Products Annexes These guidelines also lay downwards the permitted limits, both for particles and for microbiological contamination

Action limit: every bit mentioned in the requirement, the regulatory limit can be used as an activeness limit, or alternatively, a statistically derived lower limit can be adopted; for instance, at 99% percentile of the historical information. Obviously, an intensive investigation, including cosmetic and possibly preventive measures, is expected when exceeding an activity limit.

Similar this story? Subscribe to Cleanroom Technology mag for incisive analysis of the latest news and developments in hi-tech industries manufacturing in controlled environments.

Warning limit: If the warning limit is exceeded, an investigation should take place. This limit tin exist somewhat lower than when the action limit is exceeded. In this example, corrective and preventive measures may not exist necessary.

The monitoring programme should, therefore, be set out in writing (i.due east. a respective SOP should be available) and comprise the post-obit:

Sampling points in sufficient item
Sampling frequencies that provide meaningful results
Sampling time (due east.one thousand. during filling or at the terminate)
Duration of sampling
Sampling size (e.k. area of the imprint, air volume)
Sampling equipment
Sampling techniques
Alarm and action levels
Corrective and preventive actions if warning and activeness levels are exceeded

Large amounts of data are also generated by the broad-based monitoring of air, surfaces and, if necessary, water, which in disquisitional areas may fifty-fifty exist carried out every bit permanent monitoring or with a loftier sampling frequency. Effigy one shows an case of microbial counts recorded in six locations in a controlled room Grade C.

Figure 1: Example of microbial counts monitored weekly in half dozen locations within a Class C room

Dealing with "big data"

Today, the growing technology of rapid microbiological methods (RRM), aimed at real-fourth dimension continuous monitoring of both microbes and particles in water and air, introduces an even college level of information complexity. Instruments based on low-cal-induced fluorescence (LIF) are capable of instantaneously detecting and counting particles and microorganisms in air or water samples every two seconds (R. Bar, Charting and evaluation of real-time continuous monitoring water bioburden, PDA J Pharm Sci Technol, in print, 2019). Thus, continuous monitoring leads to a vast amount of data recorded every day at a much larger book than that obtained with traditional sampling commonly performed on a daily, weekly or monthly ground in controlled rooms.

The following instance makes the case more evident: a single day of continuous monitoring at a charge per unit of one bespeak per 2 seconds generates about 43,200 data points, leading to a plot of a forest of crowded information that overshadow the underlying pattern of the monitoring process. The following process challenges the responsible person in quality. While conducting an ongoing take a chance analysis she/he has to carry out the following steps:

Nautical chart and analyse data
Ready activeness and alert limits
Distinguish between statistically derived procedure performance limits and the regulatory action limits
Interpret the overall monitoring process behaviour
Detect a trend or shift in contagion levels
Monitor excursion rates

Wait once again at Figure 1, which type of command chart would you depict? How would you ready control limits? Would yous average the data? Would you unmarried out location 6?

A practical approach to information analysis

Rooms of Grade C and D, which are typically characterised with sizeable numbers of microbial counts, are amenable to standard control charting. Rooms graded B and A, which experience infrequent contamination, tin can too be monitored with contagion recovery rates as suggested in USP <1116>.

Figure 2 shows an example of a record of monthly and cumulative percentage of contamination recovery rates calculated from the microbial counts of passive air samples in a production room Grade B.

According to the contempo revision of the EU Annex 1 (Dec. 2017), results from monitoring should be considered when reviewing batch documentation for finished product release. For this purpose, a good statistical simply practical evaluation of the nerveless information is indispensable.

Figure 2: Monthly and cumulative percent contamination recovery rates (per USP <1116>) calculated from the microbial counts of passive air samples recorded during i year in a room Grade B.
Does this conform with the USP <1116> suggested limit?

The application of the traditional Shewhart command charts is based on a fundamental assumption that the process data is statistically contained, and that the process operates with a constant hateful. Even so, the data is not always independent, and additionally, may be auto-correlated. Therefore, data assay should be conducted with a applied approach aimed at:

Minimising false alarm signals
Setting applied alarm and action limits
Revealing the monitoring process behaviour
Evaluating whether the monitoring process is operating in an adequate state of control

In practice, the possibilities offered by optimal statistical data assay are frequently non fully exploited. However, by optimising the handling of the results, the identification of trends and patterns in environmental data offering many opportunities to facilitate risk assessments, to amend processes and lead times in the laboratory, and to reduce sampling plans.

In this context, the Ecology Monitoring Data Management form by ECA Academy (xx-21 November in Barcelona, Spain) will present the bones methodology of evaluating the data using elementary statistical process control tools equally well equally the empirical approaches to prepare microbial control limits for cleanrooms.

N.B. This commodity is featured in the October 2019 issue of Cleanroom Technology. Subscribe today and go your print copy!