Application of statistical process control to the use of a peristaltic pump for batch production of hospital sterile injectable preparations.
4 October 2018Lucie Estrade1, Guillaume Bouguéon1,2, Aude Berroneau1, Sylvie Crauste-Manciet1,2 Pharmaceutical Technology Department, Bordeaux university hospital (CHU de Bordeaux), France
ARNA ChemBioPharm U1212 INSERM - UMR 5320 CNRS Bordeaux University, France
Statistical Process Control (SPC) is a defect prevention method conventionally used in industrial production that makes the production process more reliable and stable over time by maintaining the required level of quality. A process capability index is a measure relating the actual performance of a process to its specified performance. It’s calculated from a process characteristic.
3 lots of hospital preparations were produced from a “mother preparation” divided using a qualified peristaltic pump (Vantage 3000, VERDER) into 45 preparations of 60 mL each. The net mass (mass full bag- mass empty bag) is the characteristic to monitor.
The 3 capability long term parameters of the masses obtained (Pp, PpK et PpM) correspond respectively to intrinsic index (comparison of total variation with tolerances), ‘potential capability’ (centering of process), and global capability (cost of no-quality) were determined.
If all three parameters are superior in 1.33, the level of confidence in the process capability is acceptable. Application inertial tolerancing (inertia) is an alternative to the traditional method, based on the value repartition around target (Cpi >1) used to estimate the cost of no-quality.
In parallel, control charts of individual values was subjected to a posteriori analysis.
The intrinsic capability (dispersion) were respectively 1.86, 2.28 and 1.55 for the batch 1, 2 and 3. Those figures were satisfying. As for the centering of process, results were respectively for the batch 1 and 3 (0.37 and 0.38) and insufficient for the batch 2 (1.15).
The global capability (dispersion and centering) is insufficient for all the lots (0.41, 0.65 or 0.43 for the lots 1 to 3). Only lot 2 has a good inertia (Cpi = 0,94) and hence has an “ideal” cost.
The reading of control charts allows to visualize the unsatisfactory centering of masses and so of the volumes with respect to the target and the deviation of the process characteristics at the time.
Also, no individual values are included in this interval [LCL*; UCL**] (Lot 1 [59.16; 61.02]; lot 2 [59.61 ; 60.57] ; lot 3 [59.24 ; 60.94])mg revealing the existence of special causes (loos of calibration for example…) responsible of the process variability .
* Lower Control Limit
** Upper Control Limit
According to classical reasoning, the percentage of volumes out of tolerance (accuracy [95-105]%) remains weak (Lot 1 : 13.3%, Lot 2 : 0%, Lot 3 : 10.6%) about 5 bags per lot, the pump would not need to make adjustments. However, capability and inertia analysis showed the need to adjust the pump by performing a new calibration for the 60 mL volume to ensure long-term process stability.
The application of the MSP is important in the monitoring of hospital production processes allowing anticipation of non-compliance and the realization of savings in human resources and unnecessary production costs (cost of no-quality).