Introduction
In today's world, the issue of quality has become of vital importance for any organization that seeks to satisfy the needs of its customers and the international consensus points to the application of methodologies that allow managing quality to obtain the expected results.
statistical-control-variability-manufacturing-gel-hydroxide-aluminumThe specific conditions in which the international markets for products and services operate today demand that companies, with increasing rigor, transform their resources and productive efforts into competitive results, based on high effectiveness and efficiency indicators such that provide consistency and stability to those results.
At the international level and within our country, one of the most important sectors is the production of medicines due to their link with health, which is why the design and implementation of programs that guarantee the highest quality is the central axis of the system. Thus, good quality must be built from within, during the entire manufacturing process.
The Pharmaceutical Laboratory Company has established, documented, implemented and maintains a QMS based on ISO 9001: 2000, regulation 16/2006 of the CECMED Good Manufacturing Practices for Pharmaceutical Products and Regulation 37/2004 of the CECMED Good Laboratory Practices, and the continuous improvement of its effectiveness according to the requirements of said rule and regulations. However, it has not implemented Statistical Process Control. If you do, there will be greater control over quality costs and less variability due to the existence of preventive, predictive and corrective actions during the production process, guaranteeing high levels of effectiveness, efficiency and good customer service.
The pharmaceutical industry in our country has managed to take significant steps in this area, perfecting the use of statistical tools efficiently in the study of process variability.
Problem:
How to use robust CEP methods and Six Sigma tools dynamically; as part of the decision-making process.
Development
To carry out the variability study, the methodology in the development phase of the Quality Research group was used.
Experimental methodological procedure
The methodological procedure bases its work logic on the basis that, the work with the CEP tools are not always applied in a harmonious and coherent way, so that not only do they provide the mathematical element, but also the improvement of the analyzes carried out, and the way in which the best and refined incident information on quality costs can be obtained. So the first phase presents in its structure a diagnostic element based on the needs of the allegorical decision-making action to the process, together with a second stage of deployment of the tools and without excluding the impact on quality costs and the CEP implementation proposal.
Phase I. Phase of process perception, reception of data needs.
Objective: To determine the variables under investigation, their analysis, estimation of the elements that could hinder the collection of this data and its transformation into information, according to the needs of the decision-making process.
Step I. Characterization of the manufacturing process.
a) Reception of the needs of the decision-making process.
b) Selection and description of the variables (description of the process that gives rise to the variable, possible sources of difference in the process, importance of the selected variables, behavior impact on the final product, need for control).
c) Data collection: sampling and registration system.
Tool: Systematic sampling (Gutiérrez, 1996); acceptance sampling for variables Military Standard 414 (Doty, 1991).
Analysis of the data using the Statistical Package Minitab 16.
d) Work with the primary data (determination of primary analysis of the data, analysis of errors in its behavior).
Step II. Description of the organization's perception capacity.
Internal customer satisfaction elements for the use of CEP tools.
Analysis of the elements that affect the perception capacity.
a) Factors to consider in Perception (Mazorra Lopetey, 2009).
Existence of standardized formats.
Existence of standardized procedures.
Existence of management review procedures for the point in question.
Competition for the position.
Improvements arising from the point in question.
b) Studies of possible causes of the elements that characterize data collection.
Phase II: Analysis of the variables generated by the process under study.
Objective: To carry out the analysis of the variables under investigation, and the estimation of the organizational methodological elements that could hinder the transformation of this data and its transformation into information, according to the needs of the decision-making process.
Step I. Standardize the manufacturing conditions to evaluate the process by applying the control chart.
a) Selection of the analysis graph for each variable and degree of implementation and operation of the CEP.
Tools:
Shewhart control charts: Means, standard deviation and range.
CUSUM and EWMA charts.
B) Determination of Control Limits (LC) and study of process stability.
To determine the stability study, a total of 84 samples were taken for fill volume and pH and 56 samples for concentration, these grouped into 14 subgroups of 6 samples for the first two variables mentioned above and 4 for the last (see Appendix 2). The charts were constructed using the Minitab 16 and Statgraphics 5.1 statistical packages, the latter being used only for CUSUM charts.
Mathematical basis of Shewhart Graphs.
Mathematical basis of the CUSUM and EWMA charts respectively.
CUSUM letter.
EWMA letter.
c) Evaluate the process capacity under the conditions established in the previous step, by interpreting the capacity indices.
Tools: Determination of process capacity through the use of basic tools such as: frequency histogram and control charts; use of the indices: Potential capacity (Cp), Real capacity (Cpk) and Taguchi index (Cpm) and the decentration index (k) according to (Gutiérrez Pulido 2007).
Reference values for decision making (Gutiérrez Puido, 2007)
Phase III. Estimation of the impact of process capacity on quality costs.
Evaluation of the economic impact for non-conformity.
Evaluation of its impact on total sales.
Application Experimental methodological procedure
For the application, one of the study variables was chosen in the process under study.
Phase I: Analysis of process perception.
Step I. Characterization of the manufacturing process.
e) Reception of the needs of the decision-making process, for the production of hydroxide gel from aluminum hydroxide gel.
The Decision-Making Process (PTD) of a drug-producing organization, as previously mentioned, requires the control of strict process parameters that result in quality, as required under the license granted by Quimefa.
For what is required, the control of all the parameters associated with the intermediate processes and the final product according to the USP30 Pharmacopoeia (US Pharmacopeia), code: PNO2.2.1.039. As the need to establish trends in the processes, and its ability to achieve these characteristics in a stable way. Such as the detection of systematic errors in the production and control processes. For all the aforementioned needs, data collection formats were generated.
f) Selection and description of the variables (description of the process that gives rise to the variable, possible sources of difference in the process, importance of the selected variables, behavior impact on the final product, need for control)
For the product under study, they are valued by the incidence on the quality of the medicine and the safety components that these generate the variables: PH of the final product, its concentration and due to the impact that the volume of filling has on the client, this will also be under analysis because it is in the first impression of the product and its stability is directly perceived in the market. All the variables under study turn out to be quantitative, and you continue. Being necessary its study because falling out of specifications would lead to adjustments, reprocessing, or destruction of the product in the worst case.
g) Data collection: sampling and registration system.
The data was taken by defining the batches as: number of units produced from the preparation of the raw material to the packaging of all the units that render this initial preparation, taking 1% (Systematic sampling (Gutiérrez, 1996); sampling of acceptance for Military Standard 414 variables (Doty, 1991).), these, produced and distributed in subgroups K of 14 samples, with 6 observations for the variables PH and filling volume and with 4 for the concentration variable, for a confidence level 95%, so the variables were achieved.
PH Aluminum Hydroxide Gel with 84 observations.
h) Work with the primary data (determination of primary analysis of the data, analysis of errors in their behavior).
One of the most important elements in the use of the CEP is the work of the primary data with which the subsequent analyzes of the variables are carried out. In this case, a descriptive analysis was performed:
Determining:
Normality
study Descriptive
study PH study: this variable presents elements of USP30 Pharmacopoeia (US Pharmacopeia), code: PNO2.2.1.039 with Value parameters.
Normality analysis.
Proving that the data show normal behavior, due to the P-Value> 0.15, making this analysis useful for subsequent analyzes. Having as output of Mach: Normal Prob Plot: C2 PH, considering its approach to the normal distribution.
Descriptive study: Descriptive Statistics: C2
• A comparison is made between average and objective value (assuming the latter as the average of the specifications, since an optimal value does not appear with which to make comparisons), this shows us the degree of decentralization that the process presents, with respect to the desired value 7.63.
• When establishing a comparison of the maximum value with the LSC, we obtain that since the maximum value 7.9 is less than the LSC 8.0, then the points would be within the LSC, so the process is within the specifications..
• Analyzing the value Q3 = 7.7 this indicates that 75% of the values are below Q3.
• Since the tricked mean (TrMean = 7.6237) is below the mean (Mean = 7.6274), it removes the values that are above and then leans to the left.
Data analysis summary:
The results of the maximums and minimums allow the analysis with the mean to be useful due to the non-existence in any of the variables of values that distort the use of the mean.
The existence of errors is detected in taking measurements, such as in data manipulation:
Identical measurements
Trends in measurements
Not having instruments timed by powered microcomputers, the number of operations and the level of expertise required of the operators. In this circumstance, the temptation to consider that the results of the instruments are perfect is irresistible, even when these instruments are subject to systematic errors (unless they possess sufficient intelligence to self-calibrate).
This type of error also comes from human bias, some chemists suffer from some degree of astigmatism or color blindness (this is more frequent in men than in women), which could cause errors in instrument reading and other observations. Different types of bias in numbers have been registered, for example: tendency to favor similar numbers by round or appreciation, or by a supposed historical behavior, which is sometimes in the mind. The analyst should carefully consider each stage of the experiment he is about to carry out, the apparatus he uses, the sampling and the analytical procedure to be adopted, such as the incidence of considering the sensitivity of the equipment to be used, with respect to the necessary sensitivity of the variable for the requirements of the contracts,which reveals problems in the necessary accuracy.
Step II. Description of the organization's perception capacity.
In the present investigation due to the discipline in the collection of information, which may be increasing in the organization in a perennial way, in the collection of data no differentiation elements are established such as:
Brigades.
Lots of materials against batch of product.
Work shift against produced batch.
Analysis of the elements that affect the perception capacity.
a) Factors to consider in Perception.
Existence of standardized formats: It is revealed that the organization does not use data collection formats that allow a chronological analysis of these variables, as well as the non-use of the formats proposed in this research, which undermines the in-depth study of causes of variation, such as standardization and product traceability, allowing analysis through statistical tools.
Existence of standardized procedures: Although the procedures are more entrenched, for the execution of the measurements, they present deficiencies in the handling of the data until reaching their deeper analysis, with tools that require specific behavior of the data at the time of analysis. Compliance with assumptions and descriptive analyzes that are basic information for further studies.
Existence of management review procedures for the point in question: The points where some of these variables are measured are not reviewed with frequency and systematicity, in a logical way contained in a methodological procedure to know if the discipline For the collection, recording or handling of primary data, it is carried out according to registered procedures, so that these do not become a source of variation in the measurements, introducing for this reason trends that are not precisely the behavior of the variable.
Competition for the position: In the positions for both selection and training, only the purely technical elements, the measurements and the chemical nature are assessed, so there are deficits of competence regarding the handling of primary data.
Improvements arising from the point in question: The integration of the elements where these variables are measured, as they are not subject to chronological analysis or to the detection of causes that can be assigned by the study of statistical quality tools, does not develop an improvement process in these, developing only an operational activity of measurement and recording, being left out of the process of continuous improvement of the organization.
b) Study of possible causes of the elements that characterize data collection.
One of the main causes of the non-conformities detected for the use of the CEP is due to the fact that:
The continuous improvement process is not grounded at all levels and functions of the organization, and this is far from being based on planned studies. or the chronological analysis of the variables.
The principle of decision-making based on facts, presents difficulties in the organization, since the statistical, probabilistic and operations analysis regarding the handling of information is somewhat low.
Phase II: Analysis of the processes that generate the variables under study
Step I. Standardize the manufacturing conditions to evaluate the process by applying the control chart.
a) For the study of the variable the chart or -S chart is used, and with the chart the behavior of the means will be analyzed to detect changes in the central tendency of the process and with the chart S the standard deviations of the subgroups will be guaranteed To detect changes in the magnitude of the dispersion of the process, based on Gutiérrez Pulido (2007). The use of the CUSUM and EWMA charts will be done to perceive the minimum changes with a size less than 2.5 times the value of the standard deviation, which are not perceived by traditional graphs.
Interpretation:
Despite the fact that these results comply with the specification limits, they are not capable of meeting the statistical limits of the tool due to its variability, this instability is due to the existence of patterns such as the cyclical behavior of the points, which may be given by the regular rotation of operators.
Interpretation:
In this individual chart, it is evident that the variable is not in statistical control, since it does not comply with the chart's control limits due to its instability. In this chart, where a large part of the points outside the limits of statistical control are visualized, patterns of variation such as trends, sudden cyclical behaviors, with dangers of incurring costs, are shown, making a comparison of real limits (7,711-7,544) with the control limits (7,668-7,587), we realize that there are differences between the limits, indicating that the process is unstable, therefore it is inferred that the process is leaning towards the lower end of the specification, although there are points out of statistical control both by the lower limit and by the upper limit.
CUSUM graph for the study of PH.
The control chart was constructed under the assumption of normality with a mean equal to 7.62738 and a standard deviation equal to 0.0281881. These parameters were estimated from the data. Of the 14 non-excluded points shown in the graphs, 10 are outside the control limits at the top of the graph while 4 are outside the limits at the bottom, which shows that the variable is out of statistical control., because it does not comply with the control limits of the card due to its instability.
EWMA chart for PH.
Test results for PH EWMA graph
PROOF 1. One point more than 2.50 standard deviations from the center line. The test failed on points: 1; two; 3; 4; 6; 8, so it is inferred that the process is out of statistical control.
Process capacity index of the PH variable:
Interpretation:
According to the data obtained from the PH variable, it can be inferred that the process is adequate or potentially capable of meeting the specifications, since the indicator that determines this is Cp = 10, 40 and exceeds the stipulated, which is Cp> 1.33.
The process is capable of meeting technical limits since Cpk = 3.10 which is greater than 1, indicating that the products meet the specifications.
The decentration index has a positive sign because the process mean is higher than the nominal value, moving away from it by 70.16% within the specifications, exceeding the 20% allowable value, so it is inferred that the process is decentralized.
Phase III. Description of the impact of process capacity on quality costs.
The existence of deviation in the process leads to generating production units that do not meet specifications, which translates into the risk of not being accepted by the client and in the event that these units come into the hands of the client, the risk is even greater, with respect to this situation being verified by the client, which could translate into claims and possible loss of the same, with direct influence on the level of sales, that is, less income. With respect to all the variables studied, the following behavior is chosen.
It should be noted that the cost of production of the product under study added to energy expenditure is $ 9,761.02 (see Annex 3). Since the company does not control compliance costs, it is unsafe to analyze to what extent these affect profits.
Analysis of the energy expenditure of the equipment per batch. (1 kw / h = 0.42 cents)
However, according to the process capacity study, for a confidence level of 95% under operating conditions, 124944 units would be obtained out of specification for every million units produced.
The lot size being 11387 units, then there would be 1422.7 units out of specifications for each lot produced.
Taking as a base that the unit cost of production is $ 0.85, then the cost of non-conformity for each batch produced under the conditions of this process would be $ 1209.3.
Having a sales revenue for each batch of $ 8,619.96, it is inferred that the cost of non-compliance in relation to this process means 14% of total sales for each batch of the product.
Phase IV: Improvement for the implementation of the CEP, control of the processes
Proposal of a plan of measures for the implementation of the statistical control of the process and increase of the organization's perception and analysis capacity.
Based on the research, the lines of work to be taken to improve the organization's perception and analysis capacity through the use of the CEP would be given by:
Redesign and comply with the data collection formats, that these are connected to the analysis tools and these in turn with the reports that drive the decision-making process and continuous improvement.
Work on the basis of accreditation of laboratories and obtaining test sheets and documentation of these.
They must determine an equilibrium point between the value of the information and the cost of the analyzes, to determine the amount of the most useful analysis and optimal costs.
Transmit to human resources the need for change and increase the culture regarding the CEP.
Conclusions:
1. In the present investigation the weaknesses of the process of control of the variables and the limitations of the organization to establish useful information for the decision-making process of them are exposed.
2. By applying the statistical control tools, the possible deficiency regarding accuracy and precision could be inferred, showing the non-existence of statistical control in the variables studied. It was also concluded that the continuous improvement process does not present the principle of “Decision-making based on facts” as a basis.
3. During the investigation it became clear that compliance costs are not kept under control, even if non-compliance costs exceed 14% of the total sales per batch of the product.
4. The production process does not have sufficient real capacity to guarantee the efficiency of the organization in order to achieve its vision.
