TRENDING PERSPECTIVE IN EVALUATION OF INSPECTION CHARACTERISTICS OF PHARMACEUTICAL COMPOUND: COMPARATIVE STUDY OF CONTROL CHARTS

Mostafa Essam Eissa 

Independent Researcher, Pharmaceutical Research Facility, Cairo, Egypt.

ABSTRACT

Background and objectives: The quality and efficacy of the pharmaceutical medicinal product are of paramount importance to the patients. Reproducibility and consistency of the properties of the dosage form start with the constituting ingredients either active or inactive. The present work provides an approach for the trending of one of the crucial inspection characteristics of a well-known excipient that is commonly used as a disintegrant of a water-insoluble synthetic cross-linked Polyvinyl Pyrrolidone (PVP). Establishing manufacturing criteria and quality limits should be assessed from the beginning level of the project using exploratory process-behavior plots.

Materials and Methods: This study involved an assay of the initial 16 batches of the manufactured 1-ethenyl pyrrolidine-2-one expressed as a dried substance using a standard analysis method according to British Pharmacopoeia (BP). Exploratory Shewhart charts were plotted after preliminary distribution identification using Statistical Process Control (SPC) software. Based on the goodness of fit test results, the most appropriate assumed distribution was implemented in drawing the Individual/Moving Range (I/MR) control charts.

Results: The goodness of fit study and probability plot showed that the most close-fitting to the data dispersion pattern was Johnson transformation to the normal spreading, smallest extreme value and Weibull distribution, respectively. While there were no alarms that could be detected from the process mean and the variability charts, further improvement is required from the supplier side to meet the benchmark minimum limit threshold, in addition, the process mean was not centered but shifted towards the Upper Control Limit (UCL).

Conclusion: Implementation of SPC techniques is crucial in the modern competitive healthcare industry. Trending charts play a pivotal role in this perspective. They provide a quantitative estimation for the current situation and the actions needed for the future improvement of the monitored characteristic.

Keywords: Capability plot, control charts, Johnson transformation, probability plot, smallest extreme value, Weibull.

 

 

INTRODUCTION

 

In the world of competitive healthcare and the pharmaceutical industry, many companies and firms compete in the drug and medicinal products market1-3. However, safety, quality and efficacy come as the first-place priority of the patient Health4-6. Medicinal ingredients either active or inactive should be the cornerstone for the standard quality before the analysis of the inspection properties of the finished medicinal dosage forms.

Implementation of the Statistical Process Control (SPC) techniques are now widely used and considered essential in all pharmaceutical firms to achieve a high level of predictable and acceptable quality7-9. One of the most pivotal SPC tools is the Shewhart plot10. It has vast applications in many industrial and non-industrial areas to assess and control processes and inspection characteristics11-15. The manufacturers of the raw chemicals of pharmaceutical grades have expanded worldwide making them easy to be attained in the retail markets and brokers anywhere16. Nevertheless, sustainable quality assurance of the expected chemical and physical properties is of paramount necessity to guarantee the current and the future merit of the pharmaceutical products.

There are growing trends in the number of chemical manufacturing facilities, notably in the developing nations. The Good practices in various fields (GxP) status of them including the medicinal and healthcare sectors are questionable17,18. Good Manufacturing Practice (GMP) could be projected into the quality of the final product19. Hence, an organization with the right quality concept in mind throughout the whole firm would yield products with acceptable, stable and predictable properties with minimal risk of failure.

In a time of crisis, there is a great possibility of witnessing degradation in the quality of goods that are available from the brokers, wholesalers and market retailers to meet the demand of the customers with low prices at the expense of the essential quality inspection characteristics. Due to the emphasized challenges, the current study sought to assess the excellence and goodness of a selected excipient that is commonly incorporated in pharmaceutical preparations from chemical manufacturing companies. The study will focus on an important test that is officially considered one of the critical inspection properties of inactive material.  

 

MATERIALS AND METHODS

 

A chemical manufacturing plant of raw materials of pharmaceutical grade was assessed for the quality of the manufacturing output16,17. Sixteen batches of one of the common excipient materials were investigated for the assay result trend16,17.The inactive material - which is commonly known to be used as a thickening agent, solubility enhancer, disintegrant and emulsifier – is an insoluble polyvinyl pyrrolidinone polymer that is crosslinked or Polyvinylpyrrolidone (Crospovidone) with the chemical formula (C6H9NO)n

According to the International Union of Pure and Applied Chemistry (IUPAC) nomenclature, it is named 1-Ethenylpyrrolidin-2-one20. Each manufactured batch was subjected to the analysis using the standard official method of the British Pharmacopoeia (BP)21. The assay limit is 11.0-12.8 % of nitrogen based on dried substance21. The preliminary assay data was analyzed using Minitab version 17.0 to screen for the distribution fitting and to construct an appropriate SPC examination profile22-25. Based on the output results, the initial state of control could be determined from the control charts, capability analyses and histograms.

The assay was conducted as follows: 0.100 g of the substance to be tested (in milligrams) was introduced to a combustion flask along with five g of a mixture that included one g each of copper sulfate pentahydrate R, titanium dioxide R, and di-potassium sulfate R, as well as three glass beads21. To rinse away any particles that stuck to the neck and into the flask, a very small amount of water R was employed. Sulfuric acid R (seven milliliters) was added and allowed to drip down the inner wall of the flask21. The solution was gradually heated in the flask until it took on a clear, yellowish/green hue and the interior wall of the flask was free of carbonized material. Then, the flask was heated for an additional 45 minutes.

The contents of the flask must be cautiously poured into 20 mL of water R after it has cooled. Attaching the flask to the distillation apparatus—which has previously been cleaned by passing steam through it is necessary21. To submerge the bottom end of the condenser tube in the absorption flask, enough water (R), 0.15 mL of a bromocresol green/methyl red solution (R), and thirty milliliters of a 40 g/L boric acid solution (R) must be added21. A funnel should be filled with thirty milliliters of strong sodium hydroxide solution R. The funnel should then be gently rinsed with ten milliliters of water R. Immediately after, the rubber tube clamp needs to be closed, and the steam distillation process needs to be initiated in order to extract 80-100 milliliters of distillate.

Following the removal of the absorption flask from the lower end of the condenser tube and a quick wash with some water R, 0.025 M sulfuric acid was added to the distillate until the color of the solution changed from green to pale greyish blue to pale greyish red-purple21. Any necessary corrections were to be made after performing a blank measurement.

Statistical analysis

Distribution identification was done at 95 % Confidence Interval (CI) and α 0.05; best fitting spreading plot was verified using the Anderson/Darling (AD) test. Raw data that fails to follow any definite spreading was subjected to transformation by the Johnson family algorithm for normalization and if it passes the normality test then it can be subjected to further analysis. The most applicable types of distributions were applied for drawing the variable process-behavior charts with the associated capability plots and capability histograms. All the computations and graphing were done using statistical software.

 

RESULTS AND DISCUSSION

 

This project is part of a total examination of the organization to achieve the Total Quality Management (TQM) goals of the chemical plant26,27. The technique of total quality management (TQM) continuously detects, reduces, or eliminates production errors27. It ensures that employees get the newest training possible, speeds up supply chain coordination, and raises customer satisfaction27. To achieve complete quality management, all parties involved in the manufacturing process must be held accountable for the overall quality of the final product or service.

Detailed dispersion pattern of the dataset using probability plots are shown in Figure 1 and Figure 2. The goodness of fit test demonstrated that only three distributions are valid in the following descending order: Normal (after Johnson transformation), smallest extreme value and Weibull. The selection was based on Anderson-Darling (AD)the goodness-of-fit figure probability charting and the p-values.

A probability called the p-value assesses the strength of the evidence against the null hypothesis28. The null hypothesis for an AD test is that the data are distributed as expected29. Lower p-values thus offer more proof that the data do not fit the distribution. The preliminary exploratory Individual/Moving Range (I/MR) process behavior charts in Figure 3, Figure 4, and Figure 5 showed stable variations and process means according to the selected types of the best-fitting distributions.

Each corresponding dispersion was acceptable according to the probability plot30,31. Moreover, the capability histograms illustrated that the spreading of data was confined within the Lower Specification Limit (LSL) and the Upper Specification Limit (USL) 32. In addition, the bins' dispersion frequency is close to the hypothetical presumed underlying distribution assumption33. It is evident that the process average is not centered in the middle of the specification window, instead it showed that the preliminary results are all above the centerline risking an excursion above the upper threshold. The sample mean± the overall standard deviation was approximately ≈ 12.5±0.19. In the capability analysis, the width of the inspection characteristic variation i.e., the Voice of Process (VoP) should be less than the Voice of Customer (VoC) 34. Ppk is the index for performance centering35. It gauges the degree to which the data is centered inside the specified bounds. The resultant values in Figures 3, 4 and 5 lie within 1.00 and 1.3336.

This is in the middle of the roadand more improvement should be made to achieve the benchmark target value. On the other hand, Pp is the index of performance35,36. It gauges the degree to which the data could fit inside the specified bounds (USL, LSL)35,36. Nevertheless, it doesn't matter if it is centered within the boundaries window. In this case, it is obvious that the inspection characteristic of the assay results yielded satisfactory results in terms of the overall capability examination study. Z. bench (overall) is the percentile that converts the predicted probability of process defects to an upper tail probability on a standard underlying assumed distribution37. It is computed using the overall standard deviation and is dependent on the overall performance of the process37. While Z.LSL was acceptable, Z.USL showed values below Z. bench.

The predicted percentage of parts outside of the specified limitations is the percent total for expected overall performance. Calculating the expected value involves utilizing the total standard deviation. The probability that the measurement of a randomly chosen portion from the overall process distribution is beyond the specified bounds is the percentage total for predicted overall performance.

Based on the overall variation of the monitored process, the percentage above USL is used for predicted overall performance to estimate the percentage of nonconforming products that might be expected to be above the upper specification limit38. The actual process performance that the inspection property experiences over time is indicated by the overall performance values38. In comparison to the maximum specification limit, lower values of the percentage over USL indicate more process capability. Few or no parts should have measurements that are higher than the highest limit of the specification. The same could be extrapolated for the percentage below LSL.

It should be emphasized that establishing statistical quality monitoring from the beginning of the manufacturing process for chemical production is a mandatory activity that should not underestimated to ensure obtaining raw materials with stable, predictable and conforming inspection properties indicating a Good Manufacturing Practice (GMP) during the production processes39. It is important to implement GxP for any reputable company to take the lead in any competitive market where the common acronym for quality standards and recommendations that are considered to be "good practice" is GxP40. The different areas it can be applied to are represented by the "x". GxP is an abbreviation that is frequently used to describe a group of quality standards. The quality of the single components would be reflected in the functioning values of the final dosage forms. This is crucial when considering the health and life of the patients in the world of ever-growing populations of sick individuals.

Limitations of the study   

While monitoring and control of the inspection properties could be ensured by the above-implemented technique, transformation might change the original shape of the time series plot which could make the extrapolation during investigation more difficult and challenging. Also, more data (and hence more batches) are required to establish solid criteria through the control charts rather than being exploratory in the initial stages.

 

 

CONCLUSION

 

Process-behavior plots based on the non-normally distributed datasets showed the quality of the manufactured raw material did not demonstrate out-of-control states between the successive batches. However, the process capability monitoring showed that the performance needs improvements and tightening in the variations of the inspection characteristic (assay) between consecutive products is essential to achieve better performance index level. 

Accordingly, the mean of the process should be brought close to the center to avoid any risk of out-of-control results above the upper limiting threshold in the future.

Recommendations 

Further research studies with other inspection characteristics of the pharmaceutical raw material should be conducted and other medicinal ingredients produced by the manufacturer should be included. 

 

CONFLICT OF INTERESTS 

 

The author does not have any conflict of interest. 

 

AUTHOR’S CONTRIBUTION

 

All contributions were done in the study by the corresponding author only.

 

 

ACKNOWLEDGEMENTS 

 

None to declare. 

 

REFERENCES

 

  1. EİSSA M. Implementation of modified Q-control chart in monitoring of inspection characteristics with finite quantification sensitivity limits: A case study of bioburden enumeration in capsule shell. El-Cezeri 2021; 8(3):1093-107.https://doi.org/10.31202/ecjse.871179
  1. Eissa M. Bioburden control in the biopharmaceutical industry. BioPharm Int 2017; 30(9):24-7.
  2. Eissa M, Mahmoud A, Nouby A. Evaluation and failure risk of microbiological air quality in production area of pharmaceutical plant. RGUHS J Pharm Sci 2016; 5:155-66.https://doi.org/10.5530/rjps.2015.4.5
  1. EİSSA M. Establishment of biocidal activity evaluation study protocol in healthcare facility for routine monitoring of antibacterial activity of disinfectants. J Exp Clin Med 2022;39(4):939-47.https://doi.org/10.52142/omujecm.39.4.4
  1. Essam Eissa M. Assessing the antimicrobial efficacy of eye drop products. European Pharm Rev 2019.
  2. Eissa ME. Conventional culture media: An outdated microbiological tool but still useful. Int J Drug Disc 2018; 2(011).
  3. Essam M. Pharmaceutical Component Kinetics
  4. Mostafa Eissa MEA. Quality criteria establishment for dissolution of ascorbic acid from sustained release pellets. Novel Tech Nutrition Food Sci 2018;2(2).http://doi.org/10.31031/ntnf.2018.02.000531
  1. Eissa ME. Statistical process control and quantitative microbiological risk assessment combination in healthcare industry. Biomed Engg Biotech 2018;2(1):14-17.
  2. Eissa ME. Shewhart control chart in microbiological quality control of purified water and its use in quantitative risk evaluation. UK J Pharm Biosc 2016; 4(1):45-51.https://doi.org/10.20510/ukjpb/4/i1/87845
  1. Eissa M, Rashed E, Eissa DE. Study of tellurium-129M (129MTE) ground deposition following fukushima nuclear disaster: descriptive analysis of unscear database using statistical process techniques. Mugla J Sci Tech 2021; 7(2):67-72.https://doi.org/10.22531/muglajsci.955946
  1. Eissa ME. Studies on Morbidities and Mortalities from COVID-19: Novel public health practice during pandemic periods. Asian J App Sci 2023; 16(3):84-94.https://doi.org/10.3923/ajaps.2023.84.94
  1. Eissa D, Rashed E, Eissa M. Measuring public health effect of coronavirus disease 2019: A novel perspective in healthcare in pandemic times. BatıKaradeniz Tıp Dergisi 2023;7(2):266-8. https://doi.org/10.29058/mjwbs.1257163
  2. Eissa M, Rashed E, Eissa DE. Quality improvement in routine inspection and control of healthcare products using statistical intervention of long-term data trend. Dicle Üniversitesi Fen Bilimleri Enstitüsü Dergisi 2021;10(2):163-84.
  3. Eissa ME. Food outbreak: An overview on selected cases over long-term web-based monitoring. SM J Nutr Metab 2019;5(1):1029.
  4. Eissa M, Mahmoud A. Evaluation of microbial recovery from raw materials for pharmaceutical use. J Food Pharm Sci 2016; 4(1):6-11. https://doi.org/10.14499/jfps
  5. Eissa ME, Abid AM. Application of statistical process control for spotting compliance to good pharmaceutical practice. Brazilian J Pharm Sci 2018; 54.https://doi.org/10.1590/s2175-97902018000217499
  1. Eissa ME. Adulterated pharmaceutical product detection using statistical process control. Bangladesh Pharm J 2018 Aug 15;21(1):7-15.
  2. Eissa ME. Novel rapid method in ecological risk assessment of air-borne bacteria in pharmaceutical facility. Mahidol Univ J Pharm Sci 2016; 43(3):115-26.https://doi.org/10.14456/mujps.2016.14
  1. Kappacher C, Neurauter M, Rainer M, Bonn GK, Huck CW. Innovative combination of dispersive solid phase extraction followed by NIR-detection and multivariate data analysis for prediction of total polyphenolic content. Molecules 2021 Aug 9; 26(16):4807.https://doi.org/10.3390/molecules26164807
  1. British Pharmacopoeia Commission. British Pharmacopoeia 2023. London: TSO; 2023.
  2. Eissa M. Evaluation of microbiological purified water trend using two types of control chart. European Pharm Rev 2018; 23(5):36-8.
  3. Eissa ME, Seif M, Fares M. Assessment of purified water quality in pharmaceutical facility using six sigma tools. Int J Qual Assur 2015;6(2):54-72.https://doi.org/10.25258/ijpqa.6.2.4
  1. Eissa ME, Mahmoud AM, Nouby AS. Control chart in microbiological cleaning efficacy of pharmaceutical facility. Dhaka University J Pharm Sci 2016 Jun 28;14(2):133-8.
  2. Eissa ME. Variable and attribute control charts in trend analysis of active pharmaceutical components: Process efficiency monitoring and comparative study. Exp Med (EM) 2018;1(1):32-44.
  3. Eissa M. Implementation of trending charts in analysis of inventory pattern. European Pharm Rev 2021;26(2):30-3.
  4. Barone A. What Is Total Quality Management (TQM) and Why Is It Important? Investopedia. 2023.
  5. Rashed ER, Eissa ME. Long-term monitoring of cancer mortality rates in USA: A descriptive analysis using statistical process control tools. Iberoamerican J Med 2020;2(2):55-60. https://doi.org/10.53986/ibjm.2020.0012
  6. EİSSA M. Statistical comparison of parallel-line symmetrical microbiological models: Analysis of agar diffusion assay in 8 x 8 large rectangular plates. İstatistikve Uygulamalı Bilimler Dergisi 2021 Dec 12;2(2):48-64.https://doi.org/10.52693/jsas.989584
  1. Eissa ME, Rashed ER. Analysis and modeling of morbidity and mortality from coronavirus outbreak in Brazil: A case study of public health challenge using unique statistical tools. J Health Sci Res; 8(2):57-61.https://doi.org/10.25259/JHSR_3_2023
  1. Eissa ME. Global health quality assessment using statistical control monitoring tools based on WHO database record: A descriptive analysis. health research 2019;3:8-18.https://doi.org/10.31058/j.hr.2019.32001
  1. Eissa M. SPC and evaluation of analyst performance| pharma articles [Internet]. Pharmafocusasia.com. 2020 [cited 9 August 2020].
  2. Eissa ME, Hamed HS. Application of statistical quality control tools for monitoring of pharmaceutical product properties: SISPQ, GxP, pharmaceutical product, drug properties. Biol Sci-PJSIR. 2019; 62(1):39-48.
  3. Staffi Six Sigma. Voice of the process vs. voice of the customer: What’s the difference?com. isixsigma.com. 2023.
  4. Capability indices: Ppk | Data collection tools | Quality
  5. Singla M. What is Process Capability Analysis? -. 2021.
  6. bench as an estimate of sigma capability - Minitab. (C) Minitab, LLC. All rights Reserved 2023.
  7. Expected overall performance for Normal Capability Analysis - Minitab. (C) Minitab, LLC. All rights Reserved. 2023.
  8. Eissa M, Rashed E. Application of statistical process optimization tools in inventory management of goods quality: Suppliers evaluation in healthcare facility. J Turkish Oper Manag 2020 Jun 6; 4(1):388-408.
  9. Grumbach C, Czermak P. Process analytical technology for the production of parenteral lipid emulsions according to good manufacturing practices. Processes 2022 Jun 10; 10(6):1174. https://doi.org/10.3390/pr10061174