CONVENTIONAL WISDOM HOLDS that scientific experiments must include controls. Conventional manufacturing wisdom is that materials and processes have to meet or exceed specified standards. Most studies ought to be controlled; not all of them are, because it is all-too convenient to justify performance by asserting that a particular process meets or passes Standard “X,” and leave it at that.
Both are desirable
It is not a case of “either/or.” Both controls and standards are critical for achieving and maintaining cleaning processes, surface quality, andcontamination control.
Most standards reflect a group understanding of acceptable performance. As such, one might reasonably argue that all standards reflect the knowledge, prejudices, intellectual environment, technological era, and corporate culture of the individuals involved in establishing those standards. Standards serve a valuable purpose in allowing individual industries to demonstrate how they compare with others in the similar fields.
In contrast, a control tends to be customized to a specific application or a specific study. A control often has an historical vector relative to the process history of the particular company. A control is what allows an individual facility to say “all other things being equal” when reporting results.
Benefits of experimental controls
There are many types of experimental controls. An experimental control can mean “no treatment.” In clinical trials, the group receivingthe placebo is a control group.
When we do process modification studies, we often ask clients to apply representative soils to a group of components, clean some components using their current method, and then ship the cleaned parts as controls to the cleaning equipment vendor along with soiled parts. All the parts are wrapped and shipped in the same manner. The equipment vendor is asked to open and observe all of the parts but to clean only the soiled ones using a process that we have all agreed sounds promising, not the ones marked “control.” All of the parts are then returned to the clientfor evaluation.
Such controls serve several purposes. If there is inadvertent, overall contamination during shipping and handling, the controls will be soiled; and we will know that a potentially promising process should not be abandoned. In addition, such controls let the vendor know what a clean part ought to look like; this is not always obvious. The controls also serve to assure that the proposed process is being tested under plausible manufacturing conditions. It is all too tempting to set up an overly stringent cleaning test by applying an atypical soil, baking it on, or applying many times the normal amount of soil, and then expecting the hapless cleaning equipment or cleaning agent supplier to work miracles. If a well-functioning process is being replaced, and the client cannot clean the control successfully, the cleaning hurdle is probably unrealisticallystringent.
In interpreting the results of analytical testing, controls provide a practical sanity check. The lab report may indicate that the level of a contaminant is below the detection limit when using ASTM method XYZ. However, if a positive control (a sample with a known level of contamination) is analyzed at the same time and it also comes back with the same reassuring lab report, we have an indication that something was amiss in the laboratory. Conversely, a report indicating a significant level of contamination is more convincing if the results for a negative control are appropriate. Often, the analytical lab includes such controls as a matter of course. It is wise to find out if they do so. Particularly for critical applications, it is also prudent to periodically re-submit a retained sample where the contamination level has been well-established. This is important because sometimes standardized samples do not reflect all of the variables present in your process; some of those variables can interfere with test results. In addition, periodic reanalysis of historical samples can provide an early warning sign of problems related to laboratory personnel changes or changes that are not covered by the specification. As an example in clinical testing, a problem with a clinical laboratory test was suspected, although the standard QC samples indicated nothing was amiss. Historical samples were submitted and provided the first clue that the test methodology required reassessment. Specifically, the results indicated hormonal levels consistent with pregnancy. Because thesamples had been obtained from Carl and John, pregnancy seemed unlikely.
Conclusions
Meeting standards is important, but not to the exclusion of controls. Designinga controlled study and setting up in-house controls require thought and awareness of the specific potential problems of your process. The investment in time and effort is worthwhile. Selecting and instituting appropriate manufacturing controls yields a robust, well-defined, superior process. Controls also allowyour company to continuously monitor and improve the process and the product.
Barbara Kanegsberg and Ed Kanegsberg are independent consultants in critical and precision cleaning,surface prepara-tion,and contamination control.They are the editors of “Handbook for Critical Cleaning,”CRC Press.Contact them at BFK Solutions LLC.,310-459-3614;[email protected]; www.bfksolutions.com.
