19045 Farmington Road
Livonia, Michigan 48152 USA
Machine Vision Systems and Turnkey Inspection Solutions
Use of the MIB-50 & MIB-100 Manual Inspection Booths
Harmonization of in-process measurements, announced as a target area by the FDA, requires accurately transferable data. The high priority for this requirement stems from the USP affirmation of the manual clinical inspection at the injection site as the reference performance standard that must be matched or exceeded. This benchmark inspection is a single container inspection which uses manual manipulation to inspect the injectable dose for freedom from visible particle contamination. The USP affirmation imposes specific requirements on the validation of any alternative visible particle inspection system. The first requirement is an accurate assessment of the human inspection capability for visible contaminating particles. Following this initial determination, in accordance with cGMP, any alternative inspection method must be shown to be at least as effective in the elimination of particle contaminated containers from a batch of injectable products. Accurate measurement of the single container human inspection capability for visual contamination in parenteral products is an essential prerequisite. It supplies the benchmark performance against which any alternative inspection method or device is validated. To achieve replicable measurements from human performance, the conditions and action sequence of the inspection must be accurately defined and reproduced.
Knapp’s 1980 PDA Journal Paper established the fact that the primary data describing the inspection for contaminating particles in injectable products is the probability of particle detection. This determination has, since first publication, been widely repeated and accepted. The direct use of the procedure described by Knapp requires multiple inspections of each container in a sampled set to determine the containers experimental rejection probability. Following this determination the containers are sorted into two groups. The Reject Zone group consists of those with a rejection probability equal to or greater than 0.7071 and those containers in the Accept and Gray whose rejection probability is less than this limit. Only the number of Reject Zone containers are counted as ISO Table rejects. This procedure results in "simply repeatable data" compatible with ISO sampling procedures. The direct use of raw, probabilistic, visible particle inspection data results in false batch acceptance of undesirable quality and false batch rejections of acceptable quality.
A more convenient evaluation of a sampled group of containers is available with the use of a calibration curve, which relates particle size to particle detection probability. Such a curve can be prepared with a set of durable NIST traceable microspheres. The rejection probability for each container in this set is evaluated within 95% confidence limits by trained inspectors using standard methodology in standard conditions. The particle size detected with a probability of 0.7071, approximately 100µm from initial data, defines the onset of the Reject Zone. From the data on hand a 50µm particle is detected with 10% probability and a 33µm particle is detected with 4% probability With the availability of this calibration curve, evaluation of a sampled aliquot of containers to determine acceptability of the batch is reduced to a count of the number of containers whose contaminating particle is equal or larger than the particle detected with a probability of 0.7071. Conversion of this research procedure, which requires multiple inspections of a sample to determine the experimental rejection probability of each container in a sampled set, can now be converted to the measurement of the size of any contaminating particles in each container. This measurement can be accomplished with a low power stereo microscope with a horizontal container axis, with a comparator or with current technology capable of automated NIST traceable particle size measurements such as the Phoenix Imaging Automatic Visual Inspection System, AVIS.
The Phoenix Imaging Manual inspection Booth has been designed to provide an optimum test environment for critical visible particle inspections. Earlier standardization efforts were based on measured light intensity at a specific inspection point under a light source. The exponential variation of light intensity with distance from the light source limited the utility and replicability of such specifications. When manual inspection point variability is considered, a spatial volume in which the light intensity is maintained constant is required for practical inspection applications. A recently patented design achieves such a test environment. The variation of light intensity in a 1.4 cubic foot volume was reduced to ± 5%, less than ½ the light intensity variation that is manually detectable. When inspections for contaminating particles are conducted within the controlled light intensity volume, data variability due to light intensity variation is eliminated.
The Manual inspection Booth, following the current recommendations of the Illumination Engineering Society, operates with 550 foot-candles in the light intensity stabilized inspection volume. NOTE: This illumination intensity is in the range recommended for critical inspections for long periods. The particular value has been selected to provide separation between sub-visible and visible contaminating particle measurements. This higher intensity also minimizes the effect of the reduction in contrast sensitivity that accompanies the aging process.
Other drawbacks associated with earlier generations of fluorescent lights used for inspection purposes have also been eliminated. These drawbacks include light intensity variation as the lamps aged and as the line voltage varied. The current sources employ phosphors with improved color rendering capability. The variation of light intensity as the line voltage varied has also been eliminated. The lamps are feedback stabilized to maintain set point light intensity until end-of-life. Predicted end-of-life is now 5 times longer than any available in 1980. Also eliminated was the inspector fatigue resulting from the stroboscopic images that result for any movement in a volume illuminated with fluorescent lamps excited with 60 cycle line voltage. [This effect is increased in regions in which 50 cycle line voltage is standard. The specially designed negative feedback stabilized 50 kHz fluorescent excitation source reduces the flicker intensity of the fluorescent light sources below human detection limits].
Inspector fatigue has been shown to adversely affect the accuracy of the inspection for visible contaminating particles. The Phoenix Imaging Manual inspection Booth has been designed so that the sequence of motions required from an inspector during the inspection are smoothly continuous and are adjustable to the body dimensions of the inspector. These adjustments are made tool-free.
This adjustment concept normalizes the onset of fatigue for all inspectors. It departs in an important way from the design concept of commercially available test booths in which inspectors of all physical dimensions are required to perform a sequence of motions between spatially identical positions. In presently available manual inspection stations the movements required during the inspection are the same for all inspectors. A consequence of this arbitrary requirement is that shorter inspectors must stretch between the required movement points and tall inspectors are cramped in performing the required movements. Both extremes impose additional fatigue elements. An alternative to this one-size-fits-all approach is to provide the flexible adjustment capability of the Phoenix Imaging Manual inspection Booth and a smooth optimized inspection sequence.
In the Phoenix Imaging Manual inspection Booth all adjustments to conform inspection motions to individual physical dimensions are performed at most once per shift. The required adjustment sequence starts with the adjustment of the vertical eye position of the inspector to the marked vertical center-line between the two light sources. The primary adjustment brings the eyes of the inspector to the marked center-point between the two vertically disposed inspection light sources. This adjustment places the nominal inspection point position at the center of the stabilized light volume. This adjustment is made with a task designed chair.
The next two adjustments place the operating position of the inspector within the tactile center of the inspection position. The vertical position of the inspectors elbow is measured and recorded. The center points of both trays are adjusted to this y height. The trays are then adjusted to provide full access with forearm and wrist motions. This is accomplished with adjustments in x, y and θ. The lever functions for all movements are identified with labels. The position of the adjustments (as determined from the scales mounted in each dimension) are logged against the inspectors name to achieve simple resets when required.
Although the light stabilized inspection volume is ~1.0 cubic feet and is 6 inches deep. The inspectors are trained to target the center point vertically and horizontally at which the movement sequence of the inspection commences. The next adjustments place the center of an inclined tray of vials to be inspected in a position accessible with the left forearm and wrist motions (this assumes a right handed person). A similar inclined tray receives the accepted and rejected containers in 2 compartments that are accessible with right forearm and wrist motions. (All dimensional adjustments are standardized to tighten with clockwise motions and release with counter-clockwise motions.)
The inclination of the input tray avoids the random incidence of particles on or near the spin axis, which reduces or eliminates particle movement following container agitation. The inclination of the tray and the use of a delay time after its initial placement is used to allow gravity to position contaminating particles in the container heel or in the highest position of the meniscus. Both positions are remote from the spin axis of the container thus assuring successful approximately uniform particle velocity transfer. Particle movement with a stationary vial during the inspection period is a prerequisite for a successful inspection.
The inspection is performed in a smooth, continuous sequence. Sample pick-up and disposal are accomplished with tactile stimuli; the head and torso remain upright in a comfortable, stable position. For a right-handed person, the sequence commences with an entry of the last accept/reject decision with the left hand. This is followed by grasp and pick-up of a fresh container, which is placed in the inclined transfer position.
The sequence of right hand inspection movements commences with pickup from the container transfer position using peripheral vision and tactile sensory. The container is transferred to the nominal inspection point and then rotated 1-¼ turns to an abrupt stop, which transfers velocity to the contaminating particle. Inspection commences as soon as the container agitation sequence is finished. A more detailed discussion of the inspection sequence and operational controls follows.