Publications

This paper is based on a fireside chat which took in Technological University Dublin, on June 1st at a conference entitled ‘ICH Q9(R1):The Next Frontier’. Given the brevity of an interview format the author expanded on the topics discussed at the conference, and shares in this paper more detail lessons he has learned for his experience in facilitating or leading Risk-Based DecisionMaking.

Essentially any concept of risk is built on fundamental concepts of chance, likelihood or probability. Although risk is generally a probability of loss of something of value, given that a risk-generating event will occur or has occurred, it is ironic that the quality risk management literature and guidelines on quality risk management tools are relatively silent on the meaning and uses of “probability.” The probability concept is typically applied by risk managers as a combination of data-based, “frequentist” measures of probability and a Bayesian, “degree of belief” meaning of probability. Probability as a concept that is crucial for understanding and managing risk is discussed though examples from the most general, scenario-defining and ranking tools that use probability implicitly to more specific probabilistic tools in risk management. A rich history of probability in risk management applied to other fields suggests that high-quality risk management decisions benefit from the implementation of probability concepts in both risk modeling and risk management.

Subjectivity is a necessary part of decision-making under uncertainty and quality risk management. Nevertheless, subjectivity can be controlled through the use of more objectively-based tools of multicriteria decision analysis.

The objective of this study is to use near-infrared spectroscopy (NIRS) coupled with multivariate chemometric models to monitor granule and tablet quality attributes in the formulation development and manufacturing of ciprofloxacin hydrochloride (CIP) immediate release tablets. Critical roller compaction process parameters, compression force (CFt), and formulation variables identified from our earlier studies were evaluated in more detail. Multivariate principal component analysis (PCA) and partial least square (PLS) models were developed during the development stage and used as a control tool to predict the quality of granules and tablets. Validated models were used to monitor and control batches manufactured at different sites to assess their robustness to change. The results showed that roll pressure (RP) and CFt played a critical role in the quality of the granules and the finished product within the range tested. Replacing binder source did not statistically influence the quality attributes of the granules and tablets. However, lubricant type has significantly impacted the granule size. Blend uniformity, crushing force, disintegration time during the manufacturing was predicted using validated PLS regression models with acceptable standard error of prediction (SEP) values, whereas the models resulted in higher SEP for batches obtained from different manufacturing site. From this study, we were able to identify critical factors which could impact the quality attributes of the CIP IR tablets. In summary, we demonstrated the ability of near-infrared spectroscopy coupled with chemometrics as a powerful tool to monitor critical quality attributes (CQA) identified during formulation development.

The goal of this study was to utilize risk assessment techniques and statistical design of experiments (DoE) to gain process understanding and to identify critical process parameters for the manufacture of controlled release multiparticulate beads using a novel disk-jet fluid bed technology. The material attributes and process parameters were systematically assessed using the Ishikawa fish bone diagram and failure mode and effect analysis (FMEA) risk assessment methods. The high risk attributes identified by the FMEA analysis were further explored using resolution V fractional factorial design. To gain an understanding of the processing parameters, a resolution V fractional factorial study was conducted. Using knowledge gained from the resolution V study, a resolution IV fractional factorial study was conducted; the purpose of this IV study was to identify the critical process parameters (CPP) that impact the critical quality attributes and understand the influence of these parameters on film formation. For both studies, the microclimate, atomization pressure, inlet air volume, product temperature (during spraying and curing), curing time, and percent solids in the coating solutions were studied. The responses evaluated were percent agglomeration, percent fines, percent yield, bead aspect ratio, median particle size diameter (d50), assay, and drug release rate. Pyrobuttons® were used to record real-time temperature and humidity changes in the fluid bed. The risk assessment methods and process analytical tools helped to understand the novel disk-jet technology and to systematically develop models of the coating process parameters like process efficiency and the extent of curing during the coating process.

Qualitative risk assessment methods are often used as the first step to determining design space boundaries; however, quantitative assessments of risk with respect to the design space, i.e., calculating the probability of failure for a given severity, are needed to fully characterize design space boundaries. Quantitative risk assessment methods in design and operational spaces are a significant aid to evaluating proposed design space boundaries. The goal of this paper is to demonstrate a relatively simple strategy for design space definition using a simplified Bayesian Monte Carlo simulation. This paper builds on a previous paper that used failure mode and effects analysis (FMEA) qualitative risk assessment and Plackett-Burman design of experiments to identity the critical quality attributes. The results show that the sequential use of qualitative and quantitative risk assessments can focus the design of experiments on a reduced set of critical material and process parameters that determine a robust design space under conditions of limited laboratory experimentation. This approach provides a strategy by which the degree of risk associated with each known parameter can be calculated and allocates resources in a manner that manages risk to an acceptable level.

Supplementary notes can be added here, including code and math.

As outlined in the ICH Q8(R2) guidance, identifying the critical quality attributes (CQA) is a crucial part of dosage form development; however, the number of possible formulation and processing factors that could influence the manufacturing of a pharmaceutical dosage form is enormous obviating formal study of all possible parameters and their interactions. Thus, the objective of this study is to examine how quality risk management can be used to prioritize the number of experiments needed to identify the CQA, while still maintaining an acceptable product risk profile. To conduct the study, immediate-release ciprofloxacin tablets manufactured via roller compaction were used as a prototype system. Granules were manufactured using an Alexanderwerk WP120 roller compactor and tablets were compressed on a Stokes B2 tablet press. In the early stages of development, prior knowledge was systematically incorporated into the risk assessment using failure mode and effect analysis (FMEA). The factors identified using FMEA were then followed by a quantitative assessed using a Plackett–Burman screening design. Results show that by using prior experience, literature data, and preformulation data the number of experiments could be reduced to an acceptable level, and the use of FMEA and screening designs such as the Plackett Burman can rationally guide the process of reducing the number experiments to a manageable level.

Quality risk management for the pharmaceutical industry was recently defined in internationally harmonized guidance as a systematic process for the assessment, control, communication, and review of risks to the quality of the drug product across the product life cycle. Two overarching principles for quality risk management are that evaluations of risk should be scientifically based and ultimately linked to risk to the patient, and the level of effort and documentation of quality risk management processes should be commensurate with the level of risk. Numerous tools for risk management come from other applied sciences and manufacturing having a longer history of risk management. The degree of quantitative sophistication among tools varies from generalized and qualitative, “high-level” tools to computationally rigorous, “low-level,” quantitative tools. Risk is described in recent guidance as a combination of the probability of occurrence of harm and the severity of that harm. Risk management always comes with uncertainty, given that it calls for projections of the likelihood of adverse events for given severities. As a relatively new application of risk management, quality risk management can benefit from adapting existing theory and practice to pharmaceutical manufacturing.

Supplementary notes can be added here, including code and math.

This article describes development, implementation, and results of an expert elicitation survey about risks associated with pharmaceutical manufacturing processes, and discusses potential application of this data collection methodology to a broader range of experts.

Microbiological risk assessments generally focus on estimating adverse human health risks from exposures to human pathogenic microbes. The assessment of potential human health risks posed by pathogens that have acquired resistance to antimicrobial drugs is a new application of risk assessment that is closely related to microbiological risk assessment. Antimicrobial resistance risk assessment is a risk analytical process that focuses on resistance determinants as hazardous agents that might lead to drug-resistant microbial infections in humans exposed to bacteria carrying the determinants. Antimicrobialresistant infections could occur directly from actively invading or opportunistic pathogens or indirectly from the transfer of resistance genes to other bacteria. Here, we discuss risk assessment models that might be employed to estimate risks from drug-resistant bacteria in the animal food pathway and the types of models and data that may be used for microbiological risk assessments or antimicrobial resistance risk assessments.

The DNA base adduct, 8-hydroxyguanine (8-OHGua), has been reported to be a key biomarker relevant to carcinogenesis and cellular oxidative stress important in tumor promotion. Although investigators often report artificially high levels of 8-OHGua in DNA samples that have been exposed to phenol solutions and/or air during processing, few quantitative results are available. We show that routine phenol-based DNA purification procedures can increase 8-hydroxydeoxyguanosine (8-OHdG) levels 20-fold in samples that are exposed to air after the phenol is removed from the solutions. Surprisingly, air exposure alone accounts for a significant portion of this increase (4-fold) when compared to dG or DNA samples that have been solubilized in buffers purged with nitrogen. Most importantly, phenol treatments of DNA are shown to sensitize DNA to 8-OHdG formation by subsequent exposures to air. The sensitization of DNA occurs even though extensive dialysis is used between phenol treatment and enzymatic DNA digestion. Alternate procedures, including chlorofornrisoamyl-alcohol extractions, also yield air-sensitive DNA samples. Other artifacts of organic extraction prior to air exposure include alterations in DNA base ratios after nuclease digestions. Overall, these results strongly suggest that studies of 8-OHdG in carcinogenesis should avoid dry conditions, such as lyophilization followed by exposure to air, and that all four of the bases should be monitored before 8-OHdG concentrations are normalized by undamaged deoxynucleoside concentrations. Failure to heed these precautions can lead to 2- to 20-fold overestimates of 8-OHdG in target tissues or in vitro models.'