Use Statistical Process Control Analysis to Improve Quality
ART practitioners constantly ask themselves: "Am I producing the best possible outcomes for my patients". Sadly, it is difficult to provide this assurance to ourselves and our patients. Quality assurance programs can confirm that steps in the ART program are being performed as intended. But quality assurance cannot inform the practitioner whether these steps, when assembled in a system, are producing the best possible quality and patient success.
Process stability is a hallmark of a successful enterprise. Whether you are making autos or embryos, the best quality results from a optimized system that consistently performs within tight specifications. Conversely, a chaotic "out of control" process yields highly variable results - both good and bad - and inferior quality on a long term basis.
Industrial engineers have developed methods known as Statistical Process Control (SPC) to monitor variation within a system and to determine whether the process is stable and "in control".
The mathematics of SPC make use of the normal curve in a manner that is somewhat different than we are accustomed to in medical research. Traditional use of the normal curve compares two or more populations to determine whether the populations are significantly different. SPC, in contrast, uses the normal distribution to determine whether the products of the process are all alike - that is statistically consistent - or if the process produces products that are so different from the population that they could not arise by chance. In the first case, a process that produces statistically consistent products is said be "in statistical control". The second case, wherein the process has so much variation that some of the products lie outside the normal distribution, is said to be "out of statistical control".
SPC employs "Shewhart Control Charting" to determine whether a process is in control. The method is simple. Samples of a product (y-axis) are taken over time (x-axis).
With a sufficient sample size, the mean and standard deviation of the sample are calculated. Upper and lower "Control Limits" for the outcome of the process are typically set at three standard deviations from the mean.
Any variation of the products produced by the process that falls within the Control Limits is said to arise from "Common Causes of Variation" - the amount of variation that results from the normal differences within the capabilities of the machines, material, and people that comprise the system.
Common variation is both normal and unavoidable. Just as "no two snow flakes are alike", no two widgets or any other products of a process are identical. Rather they vary within the limits of the ability of the process. Indeed, the normal curve provides that points within the upper and lower control limits have 997 chances out of a 1,000 of not being the result of some "special" factor impacting the process. In the IVF lab "Common Variation" might arise from the lot to lot variation in reagents or gas mixtures or the ability of the incubators to maintain a constant temperature.
A more inefficient process will have wide control limits while a tight and efficient process will have narrow control limits.
In contrast, a process that is "out of statistical control" will have data points that reside outside the upper and lower control limits.
Points outside the Control Limits result from "Special Causes of Variation" - outside influences that disrupt a stable system. Examples in the IVF lab include air filters failing before their rated time, the presence of VOCs from an outside source, or a unskilled technician.
Alternatively, a Control Chart of an out of control process may contain points which fall within the control limits but form a pattern that could not have arisen by chance.
In either instance, Special Cause Variation results in process instability and reduced quality.
The purpose of SPC is to first ensure that the process is stable and to detect any Special Cause Variation affecting outcome. Next, SPC is used to improve process efficiency with a continuous cycle of identifying issues, implementing solutions, and assessing outcome. The ultimate outcome is the evolution of a process that both highly stable and highly efficient, resulting in the highest possible quality.
© Copyright 2011. Artisan Medical Solutions LLC. All Rights Reserved.