Design and validation of a risk matrix for the assessment of sterile compounded preparations
Pharmacy Compounding from the Spanish Society of Hospital Pharmacy
The resolution CM/ResAP(2011)1 established the need of undertaking an appropriate risk assessment when making a pharmacy preparation in order to determine the level of the quality system which should be applied. Sterile preparations include compounded formulations and reconstitution of medicinal products and both have a potential risk of contamination, chemical instability and non-sterility that should be mitigated.
The Spanish Group of Pharmaceutical Compounding worked in the design of a quality tool that allows the classification of sterile preparations following the resolution premises. The design process included three stages: literature review, identification of risks associated to the elaboration process by means of the Failure Mode and Effect Analysis methodology, and estimation of the severity associated to the risks detected. The final model includes 6 different dimensions of risk: compounding process, route of administration, drug’s safety profile, amount prepared, distribution and susceptibility for microbiological contamination.
After initially trying a model of numeric scoring which had several problems of concordance the classification matrix was changed to an alphabetical classification, grading each criterion from A to D. Each preparation assessed is given a 6-letter combination with three possible risk levels: low, intermediate, and high. This model was easier for risk assignment, and more reproducible. An attached table proposes a microbiological beyond-use date based on risk-level, preparation environment and storing conditions, always considering physical and chemical stability
Once the risk matrix was designed, the tool was validated in order to assure its validity and reliability. The analysis included construct validity, as well as inter and intrarater reliability, assessed by unweighted Kappa coefficient (Ligh’ts Kappa). Qualitative instruments are considered reliable if overall agreement is >95% and Kappa≥0.6. A sample of 15 representative sterile preparations usually compounded at hospital settings was used to do this qualitative study. These were evaluated by 10 hospital pharmacists working in the compounding area. The final risk matrix showed an overall percentage of agreement of 96.7%, with Light’s kappa values between 0.68-1(lower limit of confidence interval >0.4) in dimensions 1-5. Intrarate reliability obtained also a kappa coefficient ≥0.6 for dimensions 1-5. Dimension 6, related to distribution of the preparation, showed a high homogeneity in the answers; so, kappa was not calculated.
In conclusion, the risk matrix designed is a reproducible tool adapted to daily practice in hospital settings that may increase patient safety and allows a better use of resources in sterile preparations.