Search

Summary Background Pediatric thoracoscopic lobectomy is a technically challenging procedure that may result in better pain control, better cosmetic results, and shorter hospital stay. However, data describing the learning curve of total thoracoscopic lobectomy (TTL) have yet to be obtained. To evaluate our learning curve for TTL in children, we reviewed the safety and efficiency of our initial experiences with TTL in pediatric patients with congenital lung malformation. Methods This was a retrospective study of all pediatric patients undergoing TTL between March 2011 and January 2017. Cumulative summation (CUSUM) analysis of operative time (OT) was used. Results One hundred patients were retrospectively analyzed and chronologically divided into three phases: the ascending (A), plateau (B), and descending (C) phases of CUSUM of OT. Phases A, B, and C comprised 35, 22, and 43 cases, respectively. OT decreased significantly from phases A to B (P = 0.035) and B to C (P = 0.019). Age and weight of patients both reduced significantly from phase A to B (p = 0.017 and p = 0.012, respectively), while the two measures did not vary from phase B to C (p = 0.987 and p = 0.874, respectively). Chest tube duration and length of hospital stay had similar trend. All complications occurred in five cases in phase A (5/35). Six cases were converted to open surgery (6%). Four conversions occurred within phase A and two in phase C (4/35 vs 2/43, p = 0.490). There were no mortalities. Conclusions Repeated standardized training plays a role in overcoming the learning curve for thoracoscopic lobectomy in children, and CUSUMOT indicates that a learning curve of approximately 57 cases is required in our institute.

Let $\boldsymbol{X}=(X_t)_{t\ge 0}$ be a process behaving as a general increasing Lévy process (subordinator) prior to hitting a given unknown level $m_0$, then behaving as another different subordinator once this threshold is crossed. This paper addresses the detection of this unknown threshold $m_0\in [0,+\infty]$ from an observed trajectory of the process. These kind of model and issue are encountered in many areas such as reliability and quality control in degradation problems. More precisely, we construct, from a sample path and for each $\epsilon >0$, a so-called detection level $L_\epsilon$ by considering a CUSUM inspired procedure. Under mild assumptions, this level is such that, while $m_0$ is infinite (i.e. when no changes occur), its expectation $ \mathbb{E}_{\infty}(L_{\epsilon})$ tends to $+\infty$ as $\epsilon$ tends to $0$, and the expected overshoot $ \mathbb{E}_{m_0}([L_{\epsilon} - m_0]^+)$, while the threshold $m_0$ is finite, is negligible compared to $ \mathbb{E}_{\infty}(L_{\epsilon})$ as $\epsilon$ tends to $0$. Numerical illustrations are provided when the Lévy processes are gamma processes with different shape parameters.

Summary Aim of the study To analyze the collective learning curve in the performance of safe liver resections, using the decrease of severe postoperative complications (SPC) as a proxy for overall safety competency. Material and Methods This was a retrospective analysis of a prospective database in the setting of a liver surgery program implementation in a tertiary center in Morocco. The 100 first consecutive cases of elective liver resections starting from January 1st, 2018 were included in the analysis. SPC were defined as CD > IIIa during the first 90 postoperative days. We used a cumulative sum (CUSUM) technique to determine the number of cases required to achieve safety competency. We then compared case characteristics before and after the learning curve completion. Results SPC occurred in 15 cases (15%), including 5 deaths (5%). The CUSUM chart revealed a learning curve completion at the 49th case, marked by an inflection point towards the decrease in SPC (24.5% vs 5.9%; P = 0.009). In period 2 (after), cases were associated with less diabetes, less synchronous digestive resection, more cirrhosis, and more prolonged preoperative chemotherapy. The rates of major resection (30.6% vs 29.9%; P = 0.89) and biliary reconstruction were comparable, as were the operating time, and estimated blood loss. Conclusion Approximately 50 cases were required to complete the learning curve and improve the overall safety of liver resection. In our setting, the learning curve chronology was consistent with collective measures, including team stabilization and protocol development.

Objectives To evaluate the time course and caseload required to achieve proficiency by plotting the learning curve of video-assisted thoracoscopic sleeve lobectomy. Methods We reviewed 127 cases of video-assisted thoracoscopic sleeve lobectomy by a single surgeon at Shanghai Pulmonary Hospital to evaluate its learning curve using the cumulative sum (CUSUM) analysis. The changes of perioperative outcomes were assessed. Results The inflection points of the CUSUM curve were around case 30 and 90, according to which 3 phases were identified: Phase I, Phase II, and Phase III. Significant downtrends were observed regarding operative time (Phase I, 194 [173-233 minutes] vs Phase II, 172 [142-215 minutes] vs Phase III, 138 [117-164 minutes], P, Graphical abstract 30 cases are needed to lay the technical foundation and 90 cases are required to achieve proficiency. Significant improvements concerning the perioperative outcomes were observed as the learning curve developed. CUSUM, Cumulative sum; OT, operative time.

A family of self-normalized CUSUM tests for structural change under long memory is proposed. The test statistics apply non-parametric kernel-based long-run variance estimators and have well-defined limiting distributions that only depend on the long-memory parameter. A Monte Carlo simulation shows that these tests provide finite sample size control while outperforming competing procedures in terms of power.

Introduction Robotic surgery has become a safe and effective approach for the treatment of pulmonary surgical pathology. However, the adoption of new surgical techniques requires the evaluation of the learning curve. The objective of this study is to analyze the learning curve of robotic anatomical lung resections. Methods Retrospective analysis of all robotic anatomical lung resections performed by the same surgeon between June 2018 and March 2020. The learning curve was evaluated using CUSUM charts to estimate trend changes in surgical time, surgical failure and the occurrence of post-operative cardiorespiratory complications throughout the sequence of cases. Results The study included a total of 73 cases. The median duration of all complications was 120 min (interquartile range: 90−150 min), the prevalence of surgical failure was 23.29%, while 4/73 patients had any postoperative cardiorespiratory complication. Based on the CUSUM analysis, the learning curve was divided into 3 different phases: phase i (from the first to the 14th intervention), phase ii (between the 15th and 30th intervention) and phase iii (from the 31st intervention). Conclusions The learning curve for robotic anatomical lung resections can be divided into 3 phases. The technical competence that guarantees satisfactory perioperative outcomes was achived in phase iii from the 31st intervention.

Objectives Failure to rescue (FTR) is defined by the number of deaths among patients experiencing major complications after surgery. In this report we analyze FTR and apply a cumulative sum control chart (CUSUM) methodology for monitoring performance in a large series of operated lung carcinoma patients. Methods Prospectively stored records of cases undergoing anatomical lung resection in one center were reviewed. Postoperative adverse events were coded and included as a binary variable (major, or minor complications). The occurrence of 30-day mortality was also recorded. Patients dying after suffering major complications were considered as FTR. Risk-adjusted CUSUM graphs using EuroLung1 and 2 variables were constructed for major complications and FTR. Points of plateauing or trend inversion were checked to detect intentional or non-adverted changes in the process of care. Results 2237 cases included. 9.1% cases suffered major complications. The number of cases considered as failures to rescuing was 46 (2.1% of the total series and 22.5% of cases having major complications). The predictive performance of EuroLung1 and 2 models was as follows: EuroLung1 (major morbidity) C-index 0.70 (95%CI: 0.66–0.73); EuroLung2 (applied to FTR) C-index 0.81 (95%CI: 0.750.87). CUSUM graphs depicted improvement in rescuing complicated patients after case 330 but no improvement in the rate of non-complicated cases until case 720. Conclusions FTR offers a complementary view to classical outcomes for quality assessment in Thoracic Surgery. Our study also shows how the analysis of FTR on time series can be applied to evaluate changes in team performance along time.

The complexity of modern chemical and petrochemical plants is becoming increasingly problematic in the recent years. At the same time, the demands to ensure safety and reliability of process operations rise. Early detection of abnormal event in complex real systems decrease maintenance cost and lead to guarantee the safety of human operators and environment. In the present work, a fault detection (FD) method which exploits the advantages of black-box modeling and statistical measure for fault detection in real chemical process as a distillation column is proposed. This technique is developed by applying the Nonlinear Auto-Regressive Moving Average with eXogenous input (NARMAX) model and Bhattacharyya distance (BD). In order to determine the NARMAX model, a real data set recorded during normal operations is used. Then, the BD is used to quantify on-line the dissimilarity between the current and reference probability distributions of the residual obtained from the NARMAX model for fault detection purposes. The ability of the proposed FD approach is demonstrated using real fault of separation unit. The obtained results indicate that the developed technique produces favorable performance compared to the conventional Cumulative Sum (CUSUM) test.

Process monitoring techniques are of paramount importance in the chemical industry to improve both the product quality and plant safety. Small or incipient irregularities may lead to severe degradation in complex chemical processes, and the conventional process monitoring techniques cannot detect these irregularities. In this study to improve the performance of monitoring, an online multiscale fault detection approach is proposed by integrating multiscale principal component analysis (MSPCA) with cumulative sum (CUSUM) and exponentially weighted moving average (EWMA) control charts. The new Hotelling's T2 and square prediction error (SPE) based fault detection indices are proposed to detect the incipient irregularities in the process data. The performance of the proposed fault detection methods was tested for simulated data obtained from the CSTR system and compared to that of conventional PCA and MSPCA based methods. The results demonstrate that the proposed EWMA based MSPCA fault detection method was successful in detecting the faults. Moreover, a comparative study shows that the SPE-EWMA monitoring index exhibits a better performance with lower values of missed detections ranging from 0% to 0.80% and false alarms ranging from 0% to 21.20%.

Background Thoracoscopic repair of esophageal atresia with tracheoesophageal fistula (EA/TEF) remains technically challenging due to the rarity of these procedures. The aim of this study is to report our experience with thoracoscopic repair of type C EA/TEF and to evaluate the learning curve based upon the surgeon's skill level. Methods We retrospectively reviewed data of thoracoscopic EA/TEF repair performed in our center between October 2008 and May 2019. The learning curve was evaluated using the cumulative sum (CUSUM) method based on operative time. Results Of the 50 consecutive cases evaluated, the mean birth weight was 2634 ± 608 g and the median age at operation was 3 days (range, 1–29 days). The mean operation time was 144 ± 65 min. Anastomosis leakage occurred in 3 cases (6%) and strictures requiring balloon dilatations occurred in 16 cases (32%). The CUSUM analysis evaluated a learning curve of approximately 10 cases of thoracoscopic type C EA/TEF repair. A lower gestational age was associated with longer operation time. Conclusions Thoracoscopic repair of type C EA/TEF is a feasible and safe procedure. The number of procedures required to achieve a stable learning curve was 10. The learning phase may be shortened by adequate set-up under the supervision of an expert endoscopic surgeon. Type of study Retrospective Comparative Treatment Study. Level of evidence III

In electrical discharge machining (EDM), accurate and timely detection of the breakthrough of the tool electrode is rather intricate when drilling though-holes due to considerable longitudinal tool wear. This paper proposes a novel method for detecting the completion of the drilling process using the cumulative sum (CUSUM) control chart. In this method, variations in the amount of normal discharge pulses per unit time are monitored for detecting the completion of the drilling process. The actual process conditions are sensed at the start of the drilling process an the control limit for detecting the hole completion is accordingly self-tuned. The robustness of the detection method to different process conditions is supported by a design of experiments study. Experimental validation has been carried out by drilling through micro holes at relative angles between the tool and workpiece of 30°, 45°, and 60°. The proposed method has been shown to guarantee the hole completion, while in the same time avoiding machining a surface situated 1 mm below the outlet of the drilled hole. This is critical for EDM drilling applications in the aerospace and automotive sectors.

Aims To illustrate the construction of statistical control charts and show their potential application to analysis of outcomes in children's surgery. Patients and methods Two datasets recording outcomes following esophageal atresia repair and intestinal resection for Crohn's disease maintained by the author were used to construct four types of charts. The effects of altering the target signal, the alarm signal and the limits are illustrated. The dilemmas in choice of target rate are described. Simulated data illustrate the advantages over hypothesis testing. Results The charts show the author’s institutional leak rate for esophageal atresia repair may be within acceptable limits, but that this is dependent on the target set. The desirable target is contentious. The leak rate for anastomoses following intestinal resection for Crohn's disease leak is also within acceptable limits when compared to published experience, but may be deteriorating. The charts are able to detect deteriorating levels of performance well before hypothesis testing would suggest a systematic problem with outcomes. Conclusions Statistical process control charts can provide surgeons with early warning of systematic poor performance. They are robust to volume–outcome influences, since the outcome is tested sequentially after each procedure or patient. They have application in a specialty with low frequencies of operations such as children's surgery. Type of study Diagnostic test. Level of evidence Level II.

Background Recently, robot-assisted kidney transplantation (RAKT) was recently introduced as renal replacement mini-invasive surgery. Objective To report surgical technique, including tips and tricks, and the learning curve for RAKT. Design, setting, and participants All consecutive RAKTs performed in the five highest-volume centers of the European Robotic Urological Society RAKT group were reviewed, and a step-by-step description of the technique was compiled. Surgical procedure Surgeries were performed with Da Vinci Si/Xi. The patient was placed in the lithotomy position. The Trendelenburg position was set at 20–30° and the robot was docked between the legs. Measurements Shewhart control charts and cumulative summation (CUSUM) graphs and trifecta were generated to assess the learning curve according to rewarming time (RWT), intra/postoperative complications, and renal graft function (glomerular filtration rate) on days 7 and 30, and at 1 yr. Linear regressions were performed to compare the learning curves of each surgeon. Results and limitations Arterial anastomosis time was below the alarm/alert line in 93.3%/88.9% of RAKTs, while venous anastomosis time was below the alarm/alert line in 88.9%/73.9%. The nonanastomotic RWT exceeded +3 standard deviation (SD) in 24.7% of procedures and +2SD in 37.1%. In only 46% cases, the RWT was below the alert line. The ureteroneocystostomy time was below +2SD and +3SD in 87.9% and 90.2% of cases, respectively. CUSUM showed that the learning curve for arterial anastomosis required up to 35 (mean = 16) cases. Complications and delayed graft function rates decreased significantly and reached a plateau after the first 20 cases. Trifecta was achieved in 75% (24/32) of the cases after the first 34 RAKTs in each center. Conclusions A minimum of 35 cases are necessary to reach reproducibility in terms of RWT, complications, and functional results. Patient summary Robot-assisted kidney transplantation requires a learning curve of 35 cases to achieve reproducibility in terms of timing, complications, and functional results. Synergy between the surgeon and the assistant is crucial to reduce rewarming time. High-grade complications and delayed graft function are rare after ten surgeries. Hands-on training and proctorship are highly recommended.

We aim to assess linear relationships between the non-constant variances of economic variables. A two-step methodology is proposed to solve this problem. First, the conditional mean is filtered by mean of a vector autoregressive (VAR) model. Then, a bootstrap cumulative sum (CUSUM) test is applied to the residuals. Simulations suggest a good behavior of the test, for sample sizes commonly encountered in practice. The tool we provide is intended to highlight relations, or draw common patterns between economic variables, through their non-constant variances. The outputs of this paper are illustrated considering U.S. regional data.

We develop a CUSUM-type monitoring procedure based on the ordinary least squares residuals for detecting structural changes in models with a trend. A proper boundary function is designed to control the size. We derive the limiting null distribution and the consistency of the procedure under the alternative. In addition, we derive the asymptotic distribution of the delay time for the CUSUM procedure as well as the fluctuation procedure proposed by Qi e t a l . (2016). The simulation and empirical results indicate that although neither procedure is uniformly superior to the other, the CUSUM test is more suitable for an early break.

Post-operative pancreatic fistula (POPF) is a major cause of morbidity following pancreatoduodenectomy. The risk of POPF varies between individuals and thus assessment without risk adjustment is crude. However, despite the availability of numerous scores to determine risk, no study has provided a risk adjusted assessment of POPF outcomes.The observed and risk adjusted occurrence of POPF from consecutive patients operated upon by eight surgeons were recorded. Surgeons varied in experience from newly appointed (n = 5) to established (n = 3). CUSUM (cumulative sum) analysis was used to assess performance.104 POPF occurred among 519 patients (20.0%). The occurrence of POPF was significantly lower among experienced surgeons (20/186, 10.7% vs 84/333, 25.2%; p 0.001). Following risk adjustment surgeons observed 16.6 fewer to 6.5 excess POPF per 100 patients than predicted. Analysis of the CUSUM plots demonstrated the experienced surgeons performed steadily with a gradual reduction in observed POPF compared to what was predicted. The new surgeon's performance was less consistent and evidence of a learning curve was observed with steady improvement occurring after 50-70 patients.Risk adjusted assessment of POPF demonstrates differences between experienced and less experienced surgeons. This method could be used to audit practice and observe effects of changes to technique.

Stealthy false data injection attacks on cyber-physical systems introduce erroneous measurements onto sensors with the intent to degrade system performance. An intelligent attacker can design stealthy attacks with knowledge of the system model and noise characteristics to evade detection from state-of-the-art fault detectors by remaining within detection thresholds. However, during these hidden attacks, an attacker with the intention of hijacking a system will leave traces of non-random behavior that contradict with the expectation of the system model. Given these premises, in this paper we propose a run-time monitor called Cumulative Sign (CUSIGN) detector, for identifying stealthy falsified measurements by identifying if measurements are no longer behaving in a random manner. Specifically, our proposed CUSIGN monitor considers the changes in sign of the measurement residuals and their expected occurrence in order to detect if a sensor could be compromised. Moreover, our detector is designed to be a memoryless procedure, eliminating the need to store large sequences of data for attack detection. We characterize the detection capabilities of the proposed CUSIGN technique following the well-known $\chi^2$ failure detection scheme. Additionally, we show the advantage of augmenting CUSIGN to the model-based Cumulative Sum (CUSUM) detector, which provides magnitude bounds on attacks, for enhanced detection of sensor spoofing attacks. Our approach is validated with simulations on an unmanned ground vehicle (UGV) during a navigation case study., Comment: 7 pages, 8 figures (one of which is in the Appendix), 4 tables (two of which are in the Appendix), to be published in the International Federation of Automatic Control (IFAC) 2020 World Congress Proceedings

Background Laparoscopic liver resection (LLR) requires training in both hepatobiliary surgery and advanced laparoscopy. Available data on LLR learning curves are derived from pioneer surgeons. The aims of this study were to evaluate the LLR learning curve for second generation surgeons, and to compare different CUSUM methodology with and without risk adjustment. Methods Retrospective analysis of a prospective database of 111 consecutive patients who underwent LLR by two surgeons at a single centre between 2011 and 2016. The LLR learning curve for minor hepatectomy (MH) was evaluated for each surgeon using standard CUSUM before and after risk-adjusting for operative difficulty using the Iwate index, and compared with Learning Curve (LC) CUSUM. The end points were operative time and conversion rate. Results Standard CUSUM analysis identified a learning curve of 50–60 MH procedures. The corresponding learning curve reduced to 25–30 after risk-adjusting for operative difficulty, whilst LC-CUSUM identified a learning curve of 17–25 procedures. Conclusions The learning curve for laparoscopic minor liver resection by second generation surgeons is shorter than that for pioneer surgeons. Laparoscopic HPB fellowship programmes may further shorten the learning curve, facilitating safe expansion of LLR. The LC-CUSUM method is an alternative technique that warrants further study.

Random causes of variation are part of every process and are harmless to product quality characteristics, while special cause variations appear due to a fault in the process and need special attention. The adaptive exponential weighted moving average (EWMA) control chart based on the cumulative sum (CUSUM) accumulation error is proposed in this study. The aim of this study is detecting imbalanced mixed range (different sizes) shifts (special cause) in process dispersion. The Huber and Tukey bi-square functions are used to enhance the proposed chart efficiency to simultaneously detect these small to large shifts. To check the sensitivity of the proposed chart, numerical results are provided through Monto Carlo simulations. The average run length (ARL) for single shift performance and extra quadratic loss (EQL), relative average run length (RARL) and performance comparison index (PCI) measures are used for overall performance evaluation of the proposed chart. The proposed scheme is compared to existing charts in the literature to determine superiority. For practical purposes, the guidelines are provided using real-life data for their practical implementation.

A stable money demand function is essential when using monetary aggregate as a monetary policy. Thus, there is need to examine the stability of the money demand function in Nigeria after the deregulation of the financial sector. To achieve this, the study employed CUSUM (cumulative sum) and CUSUMSQ (CUSUM of square) tests after using autoregressive distributive lag bounds test to determine the existence of a long run relationship between monetary aggregates and their determinants. Results of the study show that a long-run relationship holds and that the demand for money is stable in Nigeria. In addition, the inflation rate is found to be a better proxy for an opportunity variable when compared to interest rate. The main implication of the study is that interest rate is ineffective as a monetary policy instrument in Nigeria.

Infant failures are the natural enemy of customer satisfaction of electronic product. The change point is a precondition to optimize the burn-in time, warranty period, and root causes in production reliability assurance. However, the method for modeling and detecting the change point is unavailable due to the deficient research on the mechanism of infant failure, thereby hindering the continuous optimization of reliability in this field. Therefore, a quantitative method for modeling and detecting the change point of infant failure rate by using the Weibull distribution is proposed in this paper. Firstly, the mechanism of infant failure and the connotation of the change point of infant failure rate are explained. Then a statistic index is proposed to quantitatively model the trend of change point with the non-constant Weibull shape parameter. Also, the proposed statistic index is adopted to detect the trend of change point by CUSUM chart. Finally, an analysis example of a control board is presented to verify the effectiveness of the proposed approach in detecting the changing trend of infant failure rate for electronic products.

Pressure transients have been identified as one of the major contributing factors in many pipe failures in water distribution systems (WDSs). The behavior of these pressure transients is largely unknown and cannot be fully assessed by numerical simulation or modeling. This study investigates the behavior of pressure transients in WDSs as measured by high-frequency pressure sensors. A Time Series Data Mining (TSDM) approach is proposed to detect and cluster pressure transients to reveal recurrent and consistent patterns. The proposed technique, based on a modified two-sided cumulative sum (CUSUM) algorithm, is used to detect pressure transients. Dynamic Time Warping (DTW) is adopted to measure the similarity between the detected pressure transients, and k-means clustering algorithm is used to discover the characteristic patterns. Several performance scores are suggested to evaluate the quality of the clustering results, including sum of squared error, Silhouette index, and Calinski-Harabaz index. Results demonstrate that the proposed approach is able to reveal consistent and unique patterns across multiple sensing locations. The proposed approach provides a fast and efficient way to discover the hidden information in WDSs by analyzing high-frequency pressure signals from distributed sensors.

Background On July 1st, 2015, Hong Kong became the first city in Asia to implement a policy regulating sulfur dioxide (SO2) in shipping emissions. We conducted an accountability study assessing the improvement in ambient air quality and estimating the effect on health outcomes of the policy.MethodWe used interrupted time series (ITS) with segmented regression to identify any change in ambient concentrations of SO2 in contrast to other ambient pollutants (particulate matter

The coefficient of variation (CV), a measure of relative variability, is an important quality control issue worthy of consideration in some manufacturing and service-oriented companies when the process mean is not constant and/or the process variance is a function of the process mean. In this paper, we propose two adaptive EWMA (AEWMA) charts for monitoring the infrequent changes in the CV and multivariate CV (MCV) when sampling from univariate and multivariate normally distributed processes, named the AEWMA CV and AEWMA MCV charts, respectively. With extensive Monte Carlo simulations, the run length characteristics of the proposed control charts are computed. It is found that the AEWMA CV chart performs substantially and uniformly better than the existing optimal EWMA and CUSUM CV charts when detecting moderate-to-large shifts in the process CV. Moreover, the AEWMA MCV chart also performs substantially and uniformly better than the existing Shewhart MCV chart. The proposed control charts are implemented on real datasets to support the theory.

Despite significant advances made on anomaly detection systems, few reports are found documenting their practical integration into the industrial realm. Furthermore, the literature reports a wide range of complex detection strategies, which may require hardware changes/updates in order to be supported by critical industrial equipment such as industrial controllers (e.g., Programmable Logic Controllers). To address these issues, this paper documents a systematic methodology for the practical integration of lightweight anomaly detection algorithms into industrial control applications. It shows that industrial controllers, and in particular the scheduling rate of user programs, are sensitive to network traffic-based disturbances. Therefore, the methodology embraces the task scheduling rates found in control applications, and their deviation from the “normal” behavior. It designs a “monitoring” task, and an innovative algorithm for detecting abnormal task scheduling rates by leveraging the cumulative sum model (CUSUM) and a regression strategy applied on a specific time interval. Essentially, the approach enhances the industrial controller with a “security module” that can trigger alerts to identify early cyber attacks. The approach is extensively analyzed in the context of two industrial controllers: a Phoenix Contact ILC 350-PN controller, and a Siemens SIMATIC S7-1200 Programmable controller.

The absence of continuous long term meteorological dataset has led to limited knowledge of glaciers’ response to climate change over Himalayas. This study presents an open source long term temperature dataset Climatic Research Unit (CRU) available since 1901 to study trend analysis of temperature (Tmax, Tmin and Tmean) for Gangotri basin in Himalayas. The study first establishes close agreement between CRU time series data and observed temperature dataset available from National Institute of Hydrology (NIH), Roorkee for a period of 11 years from 2005 to 2015 using standard anomaly, Wilcoxon Signed-Rank (WSR) and correlation tests. The close agreement of CRU with NIH data validate the use of CRU time series to study variation in meteorological parameter for hilly terrain of Himalayas. The second part includes application of different statistical tests such as Mann-Kendall (MK), Sen’s slope and CUSUM technique on CRU data to detect existence of any possible trends and identification of change points in Tmax, Tmin and Tmean on long term scale. On annual scale, significant increasing trends for Tmean and Tmin were observed with no significant trend for Tmax. On seasonal and monthly scale, Tmax showed significant decreasing trend for monsoon season and increasing trend for winters while Tmin show significant increasing trend for all months (except May) and seasons. CUSUM technique identified 8 change points from 3 annual time series with 2 for Tmean (1974 and 1999), 3 each for Tmax (1941, 1975 and 1999) and Tmin (1941, 1965 and 1999) respectively. Overall, significant increase in Tmin with no significant trend for Tmax has been identified over the study area.

Cumulative sum graphs are quality control charts that are possibly the most frequently used for monitoring clinical-care processes. One of their main advantages is the use in rare cases and in events with low incidence, where it would be necessary to obtain a large sample and a long follow-up time with conventional statistical methods, which is impossible in certain cases. This is also why they are useful for studying learning curves, the introduction of new technologies and, in general, for assessing the quality of care outcomes themselves, because their profile is sensitive to very subtle changes in trends (positive or negative), which would not be observed with other methods. On the other hand, their use can be expanded beyond quality control or monitoring, which is a new aspect in clinical research.

For change point problems with Gaussian distributions, the CUSUM method is most efficient for detecting mean shifts. In contrast, it is not so efficient for heavy-tailed or contaminated data because of its sensitivity to outliers. To address this issue, Csorgő and Horvath (1988) introduced the Wilcoxon–Mann–Whitney test based on two-sample U -statistics. In practice, however, the tail structure of distributions is typically unknown. For example, Barndorff-Nielsen and Shephard (2001) showed that with higher frequency, stock returns’ tails become heavier. To our knowledge, there are no uniformly most powerful testing methods for both heavy and light-tailed distributions. To deal with this issue, we construct a new family of test statistics and combine them to adapt to different tails. As the final test statistic is complex, we design a low-cost bootstrap procedure to approximate its limiting distribution. To capture temporal data dependence, we assume that the data follow a near epoch dependent process (Borovkova et al., 2001), which includes ARMA and GARCH processes, among others. We explore the validity of our method both theoretically and through simulation. We also illustrate its use with data on the S&P 500 index.

Background Parameters defining attainment and maintenance of proficiency in thoracoscopic video-assisted thoracic surgery (VATS) lobectomy remain unknown. To address this knowledge gap, this study investigated the institutional performance curve for VATS lobectomy by using risk-adjusted cumulative sum (Cusum) analysis. Methods Using The Society of Thoracic Surgeons General Thoracic Surgery Database, the study investigators identified centers that had performed a total of 30 or more VATS lobectomies. Major morbidity, mortality, and blood transfusion were deemed primary outcomes, with expected incidence derived from risk-adjusted regression models. Acceptable and unacceptable failure rates for outcomes were set a priori according to clinical relevance and informed by regression model output. Results Between 2001 and 2016, 24,196 patients underwent VATS lobectomy at 159 centers with a median volume of 103 (range, 30 to 760). Overall rates of operative mortality, major morbidity, and transfusion were 1% (244 of 24,189), 17.1% (4,145 of 24,196), and 4% (975 of 24,196), respectively. Of the highest-volume centers (≥100 cases), 84% (65 of 77) and 82 % (63 of 77) (p = 0.48) were proficient by major morbidity standards by their 50th and 100th cases, respectively. Similarly, 92% (71 of 77) and 90% (69 of 77) (p = 0.41) of centers showed proficiency by transfusion standards by their 50th and 100th cases, respectively. Three performance patterns were observed: (1) initial and sustained proficiency, (2) crossing unacceptability thresholds with subsequent improved performance; and (3) crossing unacceptability thresholds without subsequent improved performance. Conclusions VATS lobectomy outcomes have improved with lower mortality and transfusion rates. The majority of high-volume centers demonstrated proficiency after 50 cases; however, maintenance of proficiency is not ensured. Cusum provides a simple yet powerful tool that can trigger internal audits and performance improvement initiatives.

The adaptive multivariate EWMA (AMEWMA) and adaptive multivariate CUSUM (AMCUSUM) charts are recently proposed as they provide an overall good detection over a range of mean shift sizes than their non-adaptive conventional counterparts. In this paper, we propose an adaptive MEWMA (AMEWMA) chart for monitoring the infrequent changes in the mean of a multivariate normally distributed process. The idea is to consider an unbiased estimator of the mean shift using an MEWMA statistic, and then use it to adaptively update the smoothing constant of the MEWMA chart. By using Monte Carlo simulations, the run length characteristics of the AMEWMA chart are computed. The AMEWMA chart is comprehensively compared with the existing AMEWMA and AMCUSUM charts in terms of the average run length when detecting mean shifts in different ranges. It is found that the proposed chart is able to perform uniformly and substantially better than the existing charts. Both real and simulated datasets are considered to support the theory.

Due to several applications in applied statistics, there is an increasing attention to the coefficient of variation (CV) in quality control. In this paper, we propose investigating the effect of measurement errors on the performance of one-sided cumulative sum (CUSUM) control charts monitoring the CV. According to the simulated results, the precision and accuracy errors have considerable influence on the performance of the CUSUM control charts. It is also shown that increasing the number of times measuring an item in the linear covariate error model insignificantly improves the proposed chart performance.

In this paper we described a new tool used for frequent growth monitoring in pig production. The aim of the constructed tool was to alert farmers about deterioration in pigs’ growth and to provide current and historical growth statistics for consecutive batches. The tool was built as a dynamic linear model (DLM) for systems with identified and unidentified pigs. The growth of pigs was described by parameters representing an initial body weight (BW), average daily gain and daily fluctuations in BW of pigs. Moreover, the constructed tool was built to account for increasing variation in BW of pigs over time as well as autocorrelation between BW measurements of observed individuals. The forecast errors obtained from the DLM were standardized and monitored with a tabular cusum. Alarms obtained from the cusum were registered both on batch (identified and unidentified pigs) and individual level (identified pigs). The constructed tool was tested on data gathered between February and September 2016. Altogether, data from 3 batches, 1058 individuals and 146,926 BW observations were collected and analyzed. For all 3 batches, filtered and smoothened growth means were obtained and plotted. Using the model for identified and unidentified pigs, the growth problems were detected on certain dates in Batch 2 and Batch 3. In both identification scenarios, similar batch alarms were obtained. The example on growth alarms for individual pigs (using version of the model for identified pigs) was also provided. The filtered herd mean, obtained from the constructed DLM, was calculated based on all pigs present at the start of fattening. Therefore, the obtained batch BW mean was accounting also for the heaviest individuals sold after exceeding a defined target weight from the observed batches. Hence, the presented tool can be useful in obtaining growth statistics for systems where pigs are sold continuously. Frequent BW information can be useful in informing farmers about unexpected events influencing growth e.g. outbreaks of diseases or management problems. Moreover, the historical information on growth might be valuable in making optimal decisions regarding management. In the future the economic value of information from frequent BW monitoring should be assessed. In this study, we have demonstrated that frequent observations of unidentified pigs can provide meaningful alarms on growth deterioration at the batch level.

The Poisson cumulative sum (CUSUM) control chart is a method to monitor count data, which are commonly modeled by a Poisson distribution. However, the assumption of a Poisson distribution may not be valid in practice and the robustness of the Poisson CUSUM chart is shown to be poor when the volatility is greater than expected, a phenomenon known as overdispersion. Motivated by this, we propose an improved weighted Poisson CUSUM control chart based on the weighted log-likelihood ratio. For the purpose of discounting the unexplained volatility, smaller weights are assigned to the larger shifts. The continuity, monotonicity and convergence of the cumulative statistic proved in this paper make the design more practical. Compared with the conventional Poisson CUSUM and the Winsorized Poisson CUSUM control charts, the new weighted Poisson CUSUM scheme performs most robustly under overdispersion. In addition, the fast initial response (FIR) feature makes the chart more sensitive to start-up cases. A real-data example of monitoring daily orders on a start-up E-commerce company illustrates the superiorities of the proposed scheme.

BACKGROUND The Ross procedure offers several advantages in patients requiring aortic valve replacement (AVR). However, it is a more complex procedure than a standard AVR. Whether there is a risk associated with expanding its eligibility and having multiple surgeons performing this operation remains a matter of debate. METHODS AND RESULTS From 2011-2020, 568 Ross procedures were performed by 5 surgeons in 2 Canadian institutions. The cohort was divided in 5 periods of 100 patients (P1 through P5) per center to assess efficacy and safety throughout time. The efficacy endpoints were cross-clamp time, bypass time and aortic regurgitation (AR)>1 at discharge. Safety was assessed with a cumulative sum analysis (CUSUM) adjusted using the STS score. The mean age increased throughout the study period (P1: 45±13 years, P2: 47±12 years, P3: 47±12 years, P4: 48±11 years, P5: 53±10 years; p 1 (1 in P2 and 1 in P5; p=0.76). The CUSUM analysis showed that the risk of complications decreased after 100 cases (Figure 1). Three patients had a perioperative myocardial infarction (1 in P4 and 2 in P5; p=0.25) and 3 patients presented a transient ischemic attack (2 in P1 and 1 in P2; p=0.70). There was a non-statistically significant decrease in the need for temporary dialysis (P1: 5 patients, P2: 3 patients, P3: 0 patient, P4: 1 patient, P5: 1 patient; p=0.08). A similar trend was observed in terms of reintervention for bleeding (P1: 5 patients, P2: 1 patient, P3: 0 patient, P4: 2 patients, P5: 0 patient; p=0.10). The perioperative outcomes did not differ between centers. Two patients died within 30-days, both during P1 (p=0.20). There was no perioperative death in center 2. The overall mortality was 0.4%, (O/E ratio=0.24). CONCLUSION In dedicated programs, the learning curve for the Ross procedure is ≈70 cases. Following this initial phase, nor the addition of other experienced surgeon or increased patient comorbidities had a significant impact on safety and efficacy. Overall mortality remained lower than predicted.

Background: Laparoscopic surgical approaches for gastric cancer are increasing, yet limited evidence between the learning curve and long-term outcomes. To analyze survival in GC patients at a single center over 10 years and characterize the relationship between the learning curve and prognosis. Methods: This retrospective cohort study studied 3674 patients who underwent laparoscopic radical gastrectomy for gastric cancer. Cusum and Cox regression were used to assess the association between the surgeon’s experience and the 3 years overall survival (OS). Results: The 3-year OS of all patients was 71.8%. This increase of 3-year OS was associated with laparoscopic cases (r=0.638, p=0.047). Analysis of the CUSUM curve showed a significant change in the 3-year OS of 1400 cases. Further propensity score matching of patients during and after the learning curve (

As an example for the fast calculation of distributional parameters of Gaussian processes, a new Monte Carlo algorithm for the computation of quantiles of the supremum norm of weighted Brownian bridges is proposed. As it is known, the corresponding distributions arise asymptotically for weighted CUSUM statistics for change-point detection. The new algorithm employs an adaptive (sequential) time discretization for the trajectories of the Brownian bridge. A simulation study shows that the new algorithm by far outperforms the standard approach, which employs a uniform time discretization.

Structural breaks are pervasive among macroeconomic and financial time series; consequently, forecasts may lose accuracy out of sample, which renders monitoring structural breaks a critical practice. We develop a structural break monitoring scheme, Bayesian Sequential Probability Test (BSPT), for dynamic regression models, which consists of two components: probabilistic detecting statistics, and a sequential stopping procedure. We demonstrate the finite sample property and effectiveness of the BSPT by comparing its performance with that of CUSUM under a variety of DGPs and in a few economic applications.

This article presents a simple algorithm for online multivariate changepoint detection of a mean in rare changepoint settings. The algorithm is based on a modified cusum statistic and guarantees control of the type I error on any false discoveries, while featuring O(1) time and O(1) memory updates per series as well as a proven detection delay.

For sequentially observed functional data exhibiting multiple change points in the mean function, we establish consistency results for the estimated number and locations of the change points based on the norm of functional CUSUM processes and standard binary segmentation. Our main results are established under conditions allowing for general serial dependence structures in the observations, and low moments.

This paper presents a new algorithm to identify and diagnose stochastic faults in Tennessee Eastman (TE) process. The algorithm combines Ensemble Empirical Mode Decomposition (EEMD) with Principal Component Analysis (PCA) and Cumulative Sum (CUSUM) to diagnose a group of faults that could not be properly detected and/or diagnosed with previously reported techniques. This algorithm includes three steps: measurements pre-filtering, fault detection, and fault diagnosis. Measured variables are first decomposed into different scales using the EEMD-based PCA, from which fault signatures can be extracted for fault detection and diagnosis (FDD). The T2 and Q statistics-based CUSUMs are further applied to improve fault detection, where a set of PCA models are developed from historical data to characterize anomalous fingerprints that are correlated with each fault for accurate fault diagnosis. The algorithm developed in this paper can successfully identify and diagnose both individual and simultaneous occurrences of stochastic faults.

Objectives To analyze the learning curve for the management of tyrosine kinase inhibitors as the first line of treatment for patients with metastatic renal cancer. Material and methods We evaluated 32 consecutive patients treated in our department for metastatic renal cancer with tyrosine kinase inhibitors (pazopanib or sunitinib) as first-line treatment between September 2012 and November 2015. We retrospectively analyzed this sample. We measured the time to the withdrawal of the first-line treatment, the time to progression and overall survival using Kaplan–Meier curves. The learning curve was analyzed with the cumulative sum (CUSUM) methodology. Results In our series, the median time to the withdrawal of the first-line treatment was 11 months (95% CI 4.9–17.1). The mean time to progression was 30.4 months (95% CI 22.7–38.1), and the mean overall survival was 34.9 months (95% CI 27.8–42). By applying the CUSUM methodology, we obtained a graph for the CUSUM value of the time to withdrawal of the first-line treatment (CUSUM TW), observing 3 well-differentiated phases: phase 1 or initial learning phase (1–15), phase 2 (16–26) in which the management of the drug progressively improved and phase 3 (27–32) of maximum experience or mastery of the management of these drugs. The number of treated patients needed to achieve the proper management of these patients was estimated at 15. Conclusions Despite the limitations of the sample size and follow-up time, we estimated (in 15 patients) the number needed to reach the necessary experience in the management of these patients with tyrosine kinase inhibitors. We observed no relationship between the time to the withdrawal of the first-line treatment for any cause and progression.

Change point detection methods signal the occurrence of abrupt changes in a time series. Non-parametric approaches, such as the Gaussian kernel based change point (KCP) detection (Arlot et al., 2012), are especially attractive because they impose less assumptions on the data. Yet, a drawback of these methods is that most of them are sensitive to changes in the mean, the variance, etc., making them less sensitive to specific kinds of changes. We show that KCP can be adapted to detect a particular type of change only. We focus here on correlation change, which has been put forward by different theories but proved hard to trace in multivariate time series. We propose KCP-corr, which boils down to applying KCP on the running correlations. To confirm that KCP-corr is more sensitive than merely applying KCP on the raw data (KCP-raw), a simulation study was conducted in which the number of (noise) variables and the size of the correlation change were varied. KCP-corr emerged as the better method especially in the more difficult but realistic settings where the correlation change is minimal and/or noise variables are present. KCP-corr also outperforms Cusum, a non-parametric method that specifically targets the detection of correlation changes.

Background In medical oncology, changes in practices are almost always based on randomized trials but medical history shows that it is different in surgical oncology. In the past, many surgical procedures were routinely performed without a rigorous evaluation of the risk-benefit. To highlight the complexity of developing randomized surgical trials, disquisitions on methodology presented in the medical literature. This is particularly true when we consider breast reconstruction after surgical treatment for breast cancer. It is illusory to perform and conduct a randomized clinical trial (RCT) when a surgical procedure is routinely used by most surgeons. Methods As a case study, we present the scientific rationale and the design of the MAPAM01 trial which evaluates the security of the nipple sparing mastectomy. Other alternative approaches, such as propensity score and CUSUM, are presented. Results In this situation, to design surgical trials using alternative methodological approaches present a particularly important challenge both for surgeons and methodologists. Alternative approach to randomized trials can be useful to evaluate surgical procedures routinely used. Conclusion Close collaboration between surgeons and methodologists is needed to propose appropriate and well-designed surgical trials.

The first step in cleft lip repair is the precise positioning of anatomical landmarks and tracing of the incisions on the patient's lip at the beginning of the procedure. The aim of this study was to evaluate progress made in learning cleft lip repair tracing using a quantitative assessment of learning curves: LC-CUSUM (learning curve - cumulative sum). Eight surgical residents were enrolled and asked to trace lip repair incisions on five cases of unilateral left cleft lip over 5 consecutive weeks. Results were compared to a reference tracing based on the positioning of nine anatomical landmarks and assessed using LC-CUSUM. Competence was defined as the accurate positioning of the nine landmarks (less than 1.4mm deviation from the reference positions, with an accepted 15% failure rate). After five tracing sessions, competence was not achieved evenly for all trainees, or for all landmarks, underlining differences in inter-individual learning ability even with similar training. However, despite an initial marked lack of theoretical and practical training in lip repair techniques, repeated drawings of cleft lip incisions allowed a satisfactory level of competence to be reached for most landmarks and most trainees. Nevertheless it was found that not all landmarks are understood by students with similar ease, and that landmark positioning reveals significant inter-individual differences. This approach allowed a global assessment of the teaching of cleft repair and will help to focus training on specific problematic points for which competence was not obtained according to the LC-CUSUM test.

Background Perceived excess morbidity during the early learning curve of minimally-invasive pancreaticoduodenectomy (MIPD) has limited widespread adoption. It was hypothesized that robot-assisted reconstruction (RA) after MIPD allows anastomotic outcomes equivalent to open pancreaticoduodenectomy (PD). Methods Intent to treat analysis of centrally audited data accrued during early adoption of RA-MIPD at five centers. Results CUSUM analysis of operating times at each center identified 92 RA-MIPD during the early learning curve. Mean age was 65 ± 12 years with body mass index 25.8 ± 5.0. Surgical indications included malignant (60%) and premalignant (38%) lesions. Median operating time was 504 min (interquartile range 133) with 242 ml median estimated blood loss (IQR 398) and twelve (13%) conversions to open PD. Major complication rate (Clavien-Dindo III/IV) was 24% with 2 (2.2%) deaths and ten (10.9%) reoperations. Nine (9.9%) clinically significant pancreatic fistulae were observed (4 grade B; 5 grade C). Margin negative resection rate for malignancy was 90% (75% for PDA) with mean harvest of 16 ± 8 lymph nodes. Conclusions These multicenter data during the early learning curve for RA-MIPD do not demonstrate excess anastomotic morbidity compared to open. Further studies are required to determine whether surgeon proficiency and evolving technique improve anastomotic outcomes compared to open.

Background and Aims Endoscopic sleeve gastroplasty (ESG) is a novel, incisionless technique for gastric volume reduction to promote weight loss. Our aim was to describe the learning curve for performing ESG using a prospective case series. Methods Using a prospective case series design, we analyzed the first 128 consecutive patients at a tertiary care academic medical center who underwent ESG performed by a single operator from August 2013 to December 2016. Efficiency (refining performance to decrease procedure time) and mastery (absence of outliers) for performing ESG was evaluated by using a penalized basis-spline regression and cumulative sum analysis. Results Efficiency for ESG was attained after 38 ESGs, with mastery after 55 procedures. At 12 months, the mean percent total body weight loss was 15.8% (standard deviation, 9.47%). A total of 71.7% of patients achieved successful weight loss. When using multiple linear regression analysis, both number of sutures and baseline weight were significantly associated with procedure time. Other patient characteristics, such as age, sex, and race, did not significantly affect procedure time. In addition, number of sutures, baseline weight, and endoscopist achieving efficiency were not correlated with successful weight loss at 12 months. Conclusions Mastery of ESG by a single operator is suggested after sufficient endoscopic experience and may help guide widespread clinical adaptability.

This paper presents a method for the detection of faults in electrical pumps. The method relies on the computation of two features, the pump efficiency and the hydraulic balance, that present reference values during healthy pump operation and change when the pump is affected by a fault. In this paper, the CUSUM Change Detection Test and the Mann-Whitney Change Point Method are proposed as change detection algorithms to process both features. The method has been applied to a real installation with several pumps and the results are reported in the paper., This work has been partially funded by the Spanish State Research Agency (AEI) and the European Regional Development Fund (ERFD) through the projects DEOCS (ref. MINECO DPI2016-76493), SCAV (ref. MINECO DPI2017-88403-R) and through the grant IJCI-2014-20801. This work has also been partially funded by AGAUR of Generalitat de Catalunya through the Advanced Control Systems (SAC) group grant (2017 SGR 482) and by AGAUR and the European Social Fund (ESF) and the Secretary of University and Research of the Department of Companies and Knowledge of the Government of Catalonia through the grant FI-DGR 2015 (ref. 2015 FI_B 00591).