Çok değişkenli istatistiksel süreç kontrolü: Bir hastane uygulaması
Multivariate statistical process control: A hospital application
- Tez No: 323750
- Danışmanlar: PROF. DR. CENGİZ KAHRAMAN
- Tez Türü: Yüksek Lisans
- Konular: Endüstri ve Endüstri Mühendisliği, Industrial and Industrial Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2012
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Endüstri Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Endüstri Mühendisliği Bilim Dalı
- Sayfa Sayısı: 137
Özet
Küreselleşen dünyada kalite, işletmelerin müşteri ihtiyaçlarına cevap vererek yoğun rekabet ortamında ayakta kalmalarına yardımcı olan en önemli faktörlerden biri olarak görülmektedir. Bununla birlikte kalite anlayışı, bugüne kadar genelde üretim işletmeleri için uygulanan bir çalışma iken günümüzde, giderek artan bir hızla hizmet işletmelerinde de kalite bilinci yerleşmeye ve işletmeler, hizmet kalitelerini arttırmanın yollarını aramaya başlamışlardır. Kalite, işletmelerin maliyetlerini düşürerek müşteri beklentilerini karşılamalarına yardımcı olmaktadır.Sağlık sektörü için kalite, işletme (operasyon) maliyetlerin düşürülmesinin yanında sağlık hizmetinin doğası gereği insan hayatı ile doğrudan ilişkili olması bakımından özel bir öneme sahip olan sağlık sistemlerinin iyileştirilmesini de kapsamaktadır. Sağlık sektörünün en önemli kurumu olan hastanelerde hizmet kalitesinin arttırılması, önemi yeni anlaşılmaya başlanmış bir konudur. Kuşkusuz bu önemin nedeni, yapılacak bir hatanın veya süreçlerdeki bozulmaların doğrudan insan yaşamını tehlikeye sokabilecek olması ve aynı zamanda bir hizmet işletmesi olarak insanlarla sürekli etkileşim ve iletişim halinde olunmasıdır.Son dönemde, hastanelerdeki süreçleri görüntülemek amacıyla yapılan tek değişkenli istatistiksel süreç kontrolü çalışmaları artış göstermektedir. Bununla birlikte, tek değişkenli istatistiksel süreç kontrolü için belirlenen kalite karakteristikleri, genellikle bu yöntemle dikkate alınmayan diğer bazı karakteristiklerle de ilişki halindedir ve bu ilişkiler ihmal edilerek yapılan analizler yanıltıcı sonuçlara neden olabilmektedir.Bu araştırmanın amacı, hastaneler için belirlenen kalite karakteristiklerinin birbiri ile olan ilişkilerini göz önüne alarak hastanelerdeki hizmet performansını görüntülemek, diğer bir deyişle sürecin istatistiksel olarak kontrol altında olup olmadığını denetlemek ve kontrol dışında ise buna neden olan kalite karakteristiğini veya karakteristiklerini belirlemektir.Bu çalışmada, bir devlet hastanesindeki müşteri memnuniyeti düzeyini ve yoğun bakım bölümünün performansını görüntülemek için çok değişkenli istatistiksel süreç kontrolü tekniklerinden yararlanılmaktadır. Bu amaçla, Hotelling T2 kontrol grafiği ve kontrol dışında olduğu belirlenen gözlemlerin nedenlerini ortaya çıkarmak için ise Mason-Young-Tracy (MYT) Ayrıştırması tekniği kullanılmaktadır.Çalışma sonuçları incelendiğinde, çok değişkenli Hotelling T2 kontrol grafiğinin hastane performansını görüntülemede başarılı bir yöntem olduğu belirlenmiş ve MYT Ayrıştırması tekniği ile de kontrol dışı duruma neden olan değişkenler ortaya çıkarılmıştır. Bu çalışma, çok değişkenli kontrol grafiklerinin hizmet sektöründe kullanılması açısından diğer çalışmalardan ayrılmakta ve alanında öncülük etmektedir.
Özet (Çeviri)
In globalising world in quality is considered as one of the most important focus that helps responding customer requirements in order to survive in competitive environment. The perception of quality is usually applied to manufacturing enterprises, however nowadays service enterprises strart to be aware of the quality and enterprices arelooking for options to improve service quality. Quality helps businesses by reducing costs to meet customer expectations.Quality in healthcare systems has a special importance in terms of being directly related to human life. Improving the service quality in hospitals which are the most important institutions in the health care systems, is an issue with increasing importance all over the world. Undoubtedly, this is because of, an error caused directly endanger human life, and also that it did not intertwined with the people as a service organization.In recent years, studies with multivariate statistical process control in health care systems are extended. However, quality characteristics for univariate statistical process control are usually associated with other characteristics and analysis without taking into account these relationships can provide misleading results. This study aims to monitor hospital performance taking into consideration corelations among the quality characteristics. In other words, whether or not under the control of the process to ensure statistically controlled, to determine the quality characteristics who caused the out of control signal.In this study, multivariate statistical process control techniques were applied for monitoring the level of patient satisfaction and hospital intensive care department of a state hospital. Hotelling's T2 control chart is used for monitoring the process and Mason-Young-Tracy (MYT) decomposition technique are also used for out-of control signal interpretation.In chapter 1, the main goal of the study and the importance of monitoring hospital processes are emphasized.Chapter 2 contains detailed explanations regarding to purpose and definition of statistical process control and gives information about the tools of statistical process control including flow chart, cause-effect diagram, check sheets, pareto diagram, histogram, scatter chart and control charts.Chapter 3 is based on univariate statistical process control. The definition of univariate statistical process control and its usage for process improvement are explained. Univariate control charts which are the most important statistical process control tool are classified according to the type of application and data type. Afterwards, the univariate control charts are examined individually and given information on construction and usage of these control charts.In chapter 4, multivariate statistical process control is emphasized. Firstly, the importance and requirement of multivariate statistical process control are explained and differences of multivariate and univariate process control are emphasized. Multivariate data is defined and then the multivariate control chart types are examined one by one in more detail. Additionally assumptions for constructing multivariate Hotelling T2 control charts including multivariate normality, multicollinearity and autocorrelation are explained. The decomposition and interpretation of an out-of-control signal in multivariate statistical process control are a larger area of research and there are various methods and approaches in its application. Therefore, out of control signal interpretation methods are explained one by one in more detail.In chapter 5, literature review on academic studies about multivariate statistical process control and statistical process control applications in hospitals is introduced. Basically studies are seperated into two parts as application and theoretical studies. Initially studies providing information on the history of multivariate process control, new approaches developing and application studies are examined respectively. Last part of the literature review includes statistical process control studies in health care systems.In chapter 6, proposed method for monitoring hospital processes and interpretation of out of control signals is introduced below systematically. At the end of the chapter, the findings as the result of the application are given. Proposed method has four phases.The initial step of the multivariate statistical process control is the decision concerning the variables that will be used in the control procedures. Hovewer, assigning hospital performance is a complicated process. Therefore, monitoring hospital procesesses are divided into subgroups by taking into consideration investigated academic studeis and selected two subgroups (patient satisfactory level and hospital intensive care unit performance) are examined. Patient satisfactory level scores and length of stay variables are chosen for monitoring patient satisfactory level process. Patient satisfactory level is the average of the scores gathered from patient satisfactory surveys and it has an obvious relation with the length of stay. Intensive care mortalite rates, infection rates and re-hospitalization rate in intensive care surgery infection rate are choosen as variables for intensive care processes.In this study, the multivariate statisticial analysis performed using Qualstat software package. Although there are already a few software packages for multivariate statistical process control, Qualstat make a difference with wide multivariate statistical process control procedures based on Hotelling T2 control chart, decomposition technique of out of control signals and it does not need individual programming.The next step in the multivariate statistical process control is the construction of multivariate control chart. Hotelling T2 control chart which is one of the most popular control charts is constructed for monitoring said multivariate processes. Initially, an historical data set, an in control preliminary data set for the future stages of the procedure formed. This step is called Phase I. The step of formation a historical data set include the collection of preliminary data set for selected quality characteristics (variables), detecting potential data quality problems such as multicollinearity and autocorrelation, and detection and removal of outliers. Principal component analysis are conducted and condition index of these components are computed for testing multicollinearity. Trend tests are applied to all quality characteristics for detecting autocorrelation. Hotelling T2 control charts are constructed to detect and removal of outliers at the end of Phase I and remaining data are used for computing Phase II Hotelling T2 control chart parameters. The multivariate normality is also checked during charting of Phase I T2 control chart by Q-Q plots checks beta distribution of T2 values computed while T2 control chart construction.The last step of the multivariate statistical process control is constructing Hotelling T2 control charts for the new observations taken on the same process variables from the process after the end of Phase I and assessing Phase II Hotelling T2 control chart parameters. It is discovered that there are out of control observations for both processes which means the monitoring processes are not under the statistical control.The detected signals by Phase II Hotelling T2 control chart are decomposed by using the signal decomposition tool of Qualstat. The decomposed out of control signals showed that the relationship between patient satisfactory level and length of stay causes out of control signal for patient satisfactory level process while infection rates and re-hospitalization rate in intensive care causes out of control signals for intensive care processes. The assigned signaling variables must be investigated and the improvement efforts must be conducted to eliminate special variability. Hovewer re-hospitalization rate in intensive care indicates greater variability in the intensive care processes. Therefore it is proposed to initiate improvements with this characteristic.Study results showed that Hotelling's T2 control chart is a successful method for monitoring hospital performance. MYT decomposition technique is also revelaed variables caused the out of control signal. The study is unique with the approach that using multivariate control charts for monitoring service sector process.This study is completed by specifying the improvement possibilities of proposed method for future studies. Multivariate T2 control charts can be constructed for different departments of the hospital (departmet of new-born, etc.) in the future studies. Besides that decomposition and interpretation of the out of control signals might be accomplished by the different methods provided in the literature and results may be compared.
Benzer Tezler
- Pamukkale Üniversitesi Sağlık Araştırma ve Eğitim Merkezi yataklı servislerinde yatmakta olan hastalarda gelişen acinetobacter kan dolaşımı infeksiyonlarının prospektif izlemi ve risk faktörlerinin analizi
Prospective follow-up and analysis of risk factors for acinetobacter bloodstream infections in Pamukkale University Faculty of Medicine, health research and application center
MEHMET UÇAR
Tıpta Uzmanlık
Türkçe
2014
Klinik Bakteriyoloji ve Enfeksiyon HastalıklarıPamukkale ÜniversitesiEnfeksiyon Hastalıkları ve Klinik Mikrobiyoloji Ana Bilim Dalı
YRD. DOÇ. DR. MURAT KUTLU
- Usability of mobile applications: A conceptualization and instrument development study based on Apple human interface guidelines
Mobil uygulamaların kullanılabilirliği: Apple insan arayüzü yönergelerine dayalı bir kavramsallaştırma ve enstrüman geliştirme çalışması
KÜBRA ÇETİN YILDIZ
Yüksek Lisans
İngilizce
2023
Endüstri ve Endüstri Mühendisliğiİstanbul Teknik ÜniversitesiEndüstri Mühendisliği Ana Bilim Dalı
DOÇ. DR. ÇİĞDEM ALTIN GÜMÜŞSOY
- Hastanelerde yalın yönetim ve dijital dönüşüm sinerjisi
Lean management and digital transformation synergy in hospitals
DUDU BİLGİÇ GEDİK
Doktora
Türkçe
2021
Mühendislik Bilimleriİstanbul Teknik Üniversitesiİşletme Mühendisliği Ana Bilim Dalı
PROF. DR. HATİCE CAMGÖZ AKDAĞ
- Vankomisine dirençli enterokok (VRE) kolonizasyonu olan hastalarda, primer VRE bakteriyemisi gelişmesinde etkili risk faktörlerinin incelenmesi
Analysis of risk factors for primary vancomycin resistant enterococci (VRE) bacteremia among VRE colonized patients
ZEYNEP MEMİŞ
Tıpta Uzmanlık
Türkçe
2014
Klinik Bakteriyoloji ve Enfeksiyon Hastalıklarıİstanbul ÜniversitesiEnfeksiyon Hastalıkları ve Klinik Mikrobiyoloji Ana Bilim Dalı
DOÇ. SERAP ŞİMŞEK YAVUZ