Statistical Process Control Methodology Research Articles

Multivariate Statistical Process Control Using LASSO

This article develops a new multivariate statistical process control (SPC) methodology based on adapting the LASSO variable selection method to the SPC problem. The LASSO method has the sparsity property of being able to select exactly the set of nonzero regression coefficients in multivariate regression modeling, which is especially useful in cases where the number of nonzero coefficients is small. In multivariate SPC applications, process mean vectors often shift in a small number of components. Our primary goals are to detect such a shift as soon as it occurs and to identify the shifted mean components. Using this connection between the two problems, we propose a LASSO-based multivariate test statistic, and then integrate this statistic into the multivariate EWMA charting scheme for Phase II multivariate process monitoring. We show that this approach balances protection against various shift levels and shift directions, and thus provides an effective tool for multivariate SPC applications. This article has supplementary material online.

Control charts, Cusum techniques and funnel plots. A review of methods for monitoring performance in healthcare

Quality control in medicine is generating more and more interest. Industrial concepts of quality control have been refined and transformed to be useful in healthcare monitoring. Whereas medical practitioners first reaction to this new concept of quality control was negative 'we're treating patients, we're not a part of an industrial process', some dramatic cases of inferior medical performance urged the need to adequately monitor healthcare outcomes. To date, several methods have been described, and more and more reports deal with the subject. Most of us, however, are overwhelmed by the new and different tools in use such as Shewhart control charts, cumulative sum charts and funnel plots. This paper will review the methodology of statistical process control and its application in medical practice.

What can we learn from a decade of database audits? The Duke Clinical Research Institute experience, 1997—2006

Despite a pressing and well-documented need for better sharing of information on clinical trials data quality assurance methods, many research organizations remain reluctant to publish descriptions of and results from their internal auditing and quality assessment methods. We present findings from a review of a decade of internal data quality audits performed at the Duke Clinical Research Institute, a large academic research organization that conducts data management for a diverse array of clinical studies, both academic and industry-sponsored. In so doing, we hope to stimulate discussions that could benefit the wider clinical research enterprise by providing insight into methods of optimizing data collection and cleaning, ultimately helping patients and furthering essential research. We present our audit methodologies, including sampling methods, audit logistics, sample sizes, counting rules used for error rate calculations, and characteristics of audited trials. We also present database error rates as computed according to two analytical methods, which we address in detail, and discuss the advantages and drawbacks of two auditing methods used during this 10-year period. Our review of the DCRI audit program indicates that higher data quality may be achieved from a series of small audits throughout the trial rather than through a single large database audit at database lock. We found that error rates trended upward from year to year in the period characterized by traditional audits performed at database lock (1997-2000), but consistently trended downward after periodic statistical process control type audits were instituted (2001-2006). These increases in data quality were also associated with cost savings in auditing, estimated at 1000 h per year, or the efforts of one-half of a full time equivalent (FTE). Our findings are drawn from retrospective analyses and are not the result of controlled experiments, and may therefore be subject to unanticipated confounding. In addition, the scope and type of audits we examine here are specific to our institution, and our results may not be broadly generalizable. Use of statistical process control methodologies may afford advantages over more traditional auditing methods, and further research will be necessary to confirm the reliability and usability of such techniques. We believe that open and candid discussion of data quality assurance issues among academic and clinical research organizations will ultimately benefit the entire research community in the coming era of increased data sharing and re-use.

Monitoring service quality with residuals control charts

Purpose – The purpose of this paper is to devise a robust statistical process control methodology that will enable service managers to better monitor the performance of correlated service measures.Design/methodology/approach – A residuals control chart methodology based on least absolute value regression (LAV) is developed and its performance is compared to a traditional control chart methodology that is based on ordinary least squares (OLS) regression. Sensitivity analysis from the goal programming formulation of the LAV model is also performed. The methodology is applied in an actual service setting.Findings – The LAV based residuals control chart outperformed the OLS based residuals control chart in identifying out of control observations. The LAV methodology was also less sensitive to outliers than the OLS approach.Research limitations/implications – The findings from this study suggest that the proposed LAV based approach is a more robust statistical process control method than the OLS approach. In a...

Implementing statistical process control: an organizational perspective

PurposeThe purpose of this paper is to contribute to the understanding of how statistical process control (SPC) methodology can be implemented and used in organizational settings.Design/methodology/approachAn action research model was used. Data were collected through formal meeting protocols, interviews and participant observation.FindingsBased on the results of an action research project, the paper emphasizes the need for: top management support with respect to roles such as infrastructural assistance, mentor, critic, financer; creating system validity through the involvement of people with experiential knowledge about the “world” in which SPC should be applied; keeping a small, highly knowledgeable development team with appropriate expertise together during the whole process from beginning to end; keeping the various end‐users in focus but separate and prioritising between their different needs; and working with iterative design methodology.Research limitations/implicationsThe paper provides the research field with a unique case of implementing SPC using a computerized administrative data system.Practical implicationsOrganizations are given guidelines to use when implementing SPC.Originality/valueThe paper contributes knowledge in an underdeveloped field of research. It may provide a basis for further research and scholarly analysis.

A process capability/customer satisfaction approach to short‐run processes

AbstractShort‐run processes pose many challenges to standard statistical process control methodologies. Classical short‐run approaches attempt to standardize the measurements before applying traditional control procedures. A technique is proposed that standardizes the measurements from a customer satisfaction perspective and then uses small sample process capability techniques to provide analytical control mechanisms for a class of short‐run processes. Examples motivated from a print shop are used to illustrate the procedures. Copyright © 2008 John Wiley & Sons, Ltd.

Real-Time Checking of Electronic Anesthesia Records for Documentation Errors and Automatically Text Messaging Clinicians Improves Quality of Documentation

The quality of electronic anesthesia documentation is important for downstream communication and to demonstrate appropriate diligence to care. Documentation quality will also impact the success of reimbursement contracts that require timely and complete documentation of specific interventions. We implemented a system to improve completeness of clinical documentation and evaluated the results over time. We used custom software to continuously scan for missing clinical documentation during anesthesia. We used patient allergies as a test case, taking advantage of a unique requirement in our system that allergies be manually entered into the electronic record. If no allergy information was entered within 15 min of the "start of anesthesia care" event, a one-time prompt was sent via pager to the person performing the anesthetic. We tabulated the daily fraction of cases missing allergy data for the 6 mo before activating the alert system. We then obtained the same data for the subsequent 9 mo. We tested for systematic performance changes using statistical process control methodologies. Before initiating the alert system, the fraction of charts without an allergy comment was slightly more than 30%. This decreased to about 8% after initiating the alerts, and was significantly different from baseline within 5 days. Improvement lasted for the duration of the trial. Paging was suspended on nights, weekends, and holidays, yet weekend documentation performance also improved, indicating that weekday reminders had far-reaching effects. Electronic anesthesia documentation performance can be rapidly managed and improved by using an automatic process monitoring and alerting system.

Detection of Low Volatility Organic Analytes on Soils Using Infrared Reflection Spectroscopy

Previous work on the detection of low-volatility liquid organic (and organophosphorus) analytes on soil indicated a strong signal in Fourier transform infrared spectra in the 2966–2855 cm−1 range attributed to C-H vibrational stretching modes. This range is the focus of detection strategies examined here as an anticipated prelude to subsequent detection and classification approaches utilising additional spectral bands. This laboratory study is being used to test detection algorithms that may be useful for rapid stand-off detection of organic compounds on soil. Six signal processing methods (designed to minimise irrelevant variability in the recorded soil reflection spectra while enhancing signal from analyte compounds possibly present on soil) were studied. These included Savitzky-Golay second derivatives, extended multiplicative scatter correction (EMSC) and a novel alternative to piece-wise EMSC. The pre-processed signal was then used for detection. Previous work utilised principal components analysis (PCA)-based multivariate statistical process control methodologies for detection. Six alternative anomaly detection statistics were examined here that were based on the pre-processed signal, weighted signal and generalised weighted signal. The latter statistic can be considered easier than PCA with many of the same benefits. Two tests corresponding to different monitoring strategies were examined: test data “local” and “non-local” to the calibration data. Although the results are expected to be optimistic, the generalised weighted approach worked extremely well for the “local” detection test, but was less successful for the “non-local” test. Results suggest that good characterisation of analyte-free soil prior to monitoring will lead to the best detection performance. Second derivative pre-processing showed the best results on receiver operator characteristic curves and its ease of application is a distinct advantage. However, it may not be universally successful for all bands to be considered in the future. Both EMSC and the novel “piece-wise” EMSC methodology showed promise, but some improvements may be necessary.

Data-based process monitoring, process control, and quality improvement: Recent developments and applications in steel industry

The issue of how to improve product quality and product yield in a brief period of time becomes more critical in many industries. Even though industrial processes are totally different in appearance, the problems to solve are highly similar: how to build a reliable model from a limited data, how to analyze the model and relate it to first principles, how to optimize operating condition, and how to realize an on-line monitoring and control system and maintain it. In this paper, statistical process monitoring and control methodologies are briefly surveyed, and our application results in steel facilities are presented. The achievements of the present work are as follows: (1) the development of a new method that can cope with qualitative quality information and relate operating conditions to product quality or product yield, (2) the simultaneous analysis of multiple processing units including a converter, a continuous caster, a blooming process, and rolling processes, and (3) the successful application results in the steel industry.

Statistical Process Control for Hospitals

The health care industry is slowly embracing the use of statistical process control (SPC) to monitor and study causes of variation in health care processes. While the statistics and principles underlying the use of SPC are relatively straightforward, there is a need to be cognizant of the perils that await the user who is not well versed in the key concepts of SPC. This article introduces the theory behind SPC methodology, describes successful tactics for educating users, and discusses the challenges associated with encouraging adoption of SPC among health care professionals. To illustrate these benefits and challenges, this article references the National Hospital Quality Measures, presents critical elements of SPC curricula, and draws examples from hospitals that have successfully embedded SPC into their overall approach to performance assessment and improvement.

SPC Procedures for Monitoring Autocorrelated Processes

The inference about the statistical properties of quality control methodologies is based on the assumptions of normality and independence. In real industrial environments though process data is often correlated or exhibits some serial dependence affecting the efficiency of Statistical Process Control (SPC) methodologies. New technology gives managers the option of using more sophisticated SPC models which more accurately reflect the process being monitored, by relaxing some of the assumptions. The aim of this paper is to present, to apply and to evaluate control charts that are designed to account for autocorrelation.

Application of the generalized lambda distributions in a statistical process control methodology

In statistical process control (SPC) methodology, quantitative standard control charts are often based on the assumption that the observations are normally distributed. In practice, normality can fail and consequently the determination of assignable causes may result in error. After pointing out the limitations of hypothesis testing methodology commonly used for discriminating between Gaussian and non-Gaussian populations, a very flexible family of statistical distributions is presented in this paper and proposed to be introduced in SPC methodology: the generalized lambda distributions (GLD). It is shown that the control limits usually considered in SPC are accurately predicted when modelling usual statistical laws by means of these distributions. Besides, simulation results reveal that an acceptable accuracy is obtained even for a rather reduced number of initial observations (approximately a hundred). Finally, a specific user-friendly software have been used to process, using the SPC Western Electric rules, experimental data originating from an industrial production line. This example and the fact that it enables us to avoid choosing an a priori statistical law emphasize the relevance of using the GLD in SPC.

Statistical Process Control Charts Applied to Steelmaking Quality Improvement

The complex nature for steelmaking processes makes the classical Statistical Process Control (SPC) methodologies non-optimal when used to monitor and control steam boiler generation used to supply the required steam for vacuum degassing process. These processes include a large number of variables that need to be monitored and controlled, while classical SPC requires a control chart for each variable. Thus the effect of one variable can be confounded with effects of other correlated variables. Such a situation can lead to false alarm signals. Univariate control charts are also difficult to manage and analyze because of the large numbers of control charts of each process. An alternative approach is to construct a single multivariate control T2 chart that minimizes the occurrence of false process alarms, monitors the relationship between the variables, and identifies real process changes not detectable using univariate charts. It is necessary to simultaneously monitor and control these variables to achieve optimal vacuum degassing process performance to remove harmful gases from the molten steel before casting. This application represents the main focus of the presented paper. This paper also studies the application of univariate and multivariate control charts in the field of steel industry. Performance analysis for each charting method is studied using the Average Run Length (ARL). A comparison of the univariate out-of-control signals with the multivariate out-of-control signals is also made to illustrate the efficiency of the Hotelling’s T2 statistic.

How firms grow: clustering as a dynamic model of internationalization

This paper provides a methodology for identifying the dynamics of international firm expansion, demonstrating systematic evidence of both ‘toe in the water’ and concentrated bursts of internationalization by US multinational enterprises. We use the methodology of statistical process control to analyze initial investments in host countries and develop a numerical measure of temporal clustering or bunching. Using this measure, we identify two distinct patterns of growth; whereas some firms cluster their investments over time, others do not cluster their investments. For firms that cluster, three types of clustering strategy are identified: concentrating, hibernating, and hybrid. Arguments based on internal firm architecture, experience, and learning are advanced to explain these findings.

Statistical process control via context modeling of finite-state processes: an application to production monitoring

Conventional Statistical Process Control (SPC) schemes fail to monitor nonlinear and finite-state processes that often result from feedback-controlled processes. SPC methods that are designed to monitor autocorrelated processes usually assume a known model (often an ARIMA) that might poorly describe the real process. In this paper, we present a novel SPC methodology based on context modeling of finite-state processes. The method utilizes a series of context-tree models to estimate the conditional distribution of the process output given the context of previous observations. The Kullback-Leibler divergence statistic is derived to indicate significant changes in the trees along the process. The method is implemented in a simulated flexible manufacturing system in order to detect significant changes in its production mix ratio output.

Context-Based Statistical Process Control

Most statistical process control (SPC) methods are not suitable for monitoring nonlinear and state-dependent processes. This article introduces the context-based SPC (CSPC) methodology for state-dependent data generated by a finite-memory source. The key idea of the CSPC is to monitor the statistical attributes of a process by comparing two context trees at any monitoring period of time. The first is a reference tree that represents the “in control” reference behavior of the process; the second is a monitored tree, generated periodically from a sample of sequenced observations, that represents the behavior of the process at that period. The Kullback–Leibler (KL) statistic is used to measure the relative “distance” between these two trees, and an analytic distribution of this statistic is derived. Monitoring the KL statistic indicates whether there has been any significant change in the process that requires intervention. An example of buffer-level monitoring in a production system demonstrates the viability of the new method with respect to conventional methods.

Silicon Oxynitride Film Thickness Measurements Using HRTEM and Grazing Incidence X-Ray Photoelectron Spectroscopy (GIXPS)

Abstract The ability to accurately and precisely measure the thickness of ultrathin (∼ 3 nm) dielectric films, used as gate dielectrics in integrated circuits, is critical to the continued success of the semiconductor manufacturing industry. Many metrology tools have been applied to this problem in the past, but recent research has focussed on ellipsometry, x-ray reflectivity, secondary-ion mass spectrometry, capacitance-voltage curves, medium energy ion scattering, high resolution transmission electron microscopy (HRTEM), and grazing incidence x-ray photoelectron spectroscopy (GIXPS). Unfortunately, these techniques disagree about a given film's thickness by amounts larger than their individual precisions. To support the statistical process control methodologies used in production wafer fabrication, these disagreements need to be investigated and the true accuracy and precision of the tools need to be determined. This work compares the ability of two techniques, HRTEM and GIXPS, to measure the thickness of silicon oxynitride films on silicon substrates.

Statistical control charts for monitoring the mean of a stationary process

Recently statistical process control (SPC) methodologies have been developed to accommodate autocorrelated data. A primary method to deal with autocorrelated data is the use of residual charts. Although this methodology has the advantage that it can be applied to any autocorrelated data it needs time series modeling efforts. In addition for a X residual chart the detection capability is sometimes small compared to the X chart and EWMA chart. Zhang (1998) proposed the EWMAST chart which is constructed by charting the EWMA statistic for stationary processes to monitor the process mean. The performance of the EWMAST chart the X chart the X residual chart and other charts were compared in Zhang (1998). In this paper comparisons are made among the EWMAST chart the CUSUM residual chart and EWMA residual chart as well as the X residual chart and X chart via the average run length.

Using statistical process control to demonstrate the effect of operational interventions on quality indicators in the emergency department.

When our emergency department (ED) initiated a continuous quality improvement (CQI) program, we selected as a quality indicator the percentage of patients leaving without being seen (LWBS) by a physician. Because the primary reason for LWBS patients was determined to be dissatisfaction with waiting time, we devised four interventions in clinical operations to decrease delays in patient flow through the ED. Statistical process control (SPC) methodology was then used to assess the effect of these interventions. Because baseline data were available, we constructed control charts of the percentage of LWBS patients versus consecutive months beginning in January 1990 with the mean percentage of LWBS patients and upper and lower control limits. Postintervention data, plotted using control statistics from the baseline period, demonstrated sustained special-cause variation, indicating a fundamental change in the overall system. A new control chart was then constructed using postintervention data. A significantly lowered mean percentage LWBS and a narrowed control limit range were observed, leading to the conclusion that the interventions improved the quality of care as measured by a reduction in percentage LWBS.

Multivariate Statistical Process Control in Chemicals Manufacturing

Abstract The close monitoring of the operational performance of process plants and their associated instrumentation and control is of increasing strategic importance. Failures can lead to increased costs, reduced product quality, consistency and production, plant shutdowns and increased environmental impact. Multivariate Statistical Process Control methodologies can provide a diagnostic tool for the comprehensive on-line monitoring and detection of process malfunctions with applications to both continuous and batch manufacturing. This paper reviews the development of a Multivariate Statistical Process Performance Monitoring system for a complex industrial reactor process.