FREE ELECTRONIC LIBRARY - Dissertations, online materials

Pages:   || 2 | 3 | 4 | 5 |   ...   | 13 |

«A Thesis Presented to the faculty of the School of Engineering and Applied Science University of Virginia In Partial Fulfillment of the requirements ...»

-- [ Page 1 ] --

Are Patients at Higher Risk if Having Operations

at Nights, on Weekends or Holidays?

A Thesis Presented to

the faculty of the School of Engineering and Applied Science

University of Virginia

In Partial Fulfillment of the requirements for the Degree of

Master of Science in Systems and Information Engineering


Hui Guo

July 2003

Approval Sheet

This thesis is submitted in partial fulfillment of the requirements for the degree of Master

of Science in Systems and Information Engineering.

Hui Guo, Author

This thesis has been read and approved by the Examining Committee:

Dr. Donald E. Brown, Committee Chair Dr. Stephanie Guerlain, Advisor Dr. Frank E. Harrell Jr.

Dr. Florence E. Turrentine Accepted for the School of Engineering and Applied Science Dean, School of Engineering and Applied Science July, 2000 i Abstract To improve patient safety is one of the ultimate goals for nearly all medical centers. However, the influence of intrinsic variability associated with different medical situations on patient outcomes makes it difficult to identify key factors to improve. One question related to patient safety is whether the patients having surgery at nights, on weekends or holidays are at higher risk than those having surgery during the regular operating room hours, e.g., due to reduced staffing. Previous studies have suggested that this is true, but did not take into account many of the other factors that might cause such a difference.

The records from 1996 to 2000 for all patients who underwent surgery at the University of Virginia Health System were gathered. Two risk indexes were employed to evaluate surgical outcomes: 30-days mortality and the patient’s length of stay in the hospital. In addition to the service time (Night, Weekend and Holiday), patient age, gender, ASA score, co-morbidity, case type, diagnosis and surgeon’s experience were used in the model.

Logistic regression analysis for patient mortality did show a significantly higher death rate for patients having surgeries at night; however, the surgeries performed on weekends or holidays did not show such difference from those performed during the regular operating hours. Patient age, co-morbidity, ASA score and case type were all significant factors affecting surgery outcome. Further analysis within each case type showed the night factor was only significantly associated with emergent cases, but not with elective and late posted cases. The weekend and holiday stayed insignificant for

–  –  –

to patient’s sickness, health status and case type rather than patient’s age or surgeon’s experience.

The linear regression analysis for patient length of stay found both night and weekend were statistically significant. Holiday was not significant for length of stay.

A strong interaction was found between night with ASA score and case type, which would even strengthen the night effectiveness on length of stay. Again, patient age, comorbidity, ASA score, and case type were significantly associated with length of stay.

Besides patients’ sickness and health status, surgeon’s experience contributed to the negative outcome on weekends, which was another significant factor associated with length of stay.

However, Cox proportional hazards model for length of stay after surgery censored by discharge status failed to find the significance of Night, neither to Holiday and Weekend. Age, ASA, case type, co-morbidities, the attending’s experience and diagnosis are significant predictors.

In summary, night is one significant factor affecting surgical outcomes. Surgery at night is associated with a higher mortality and longer length of stay; weekend is only significantly associated with length of stay; holiday is an insignificant factor for surgical outcomes. Patients’ sickness and health status are major reasons of the negative surgery outcome at night or on weekends. Another reason affecting length of stay can be attributed to surgeon’s experience. According to these results, suggestions to improve patient safety for surgeries at night and on weekends are discussed. For future research, surgical outcome can be analyzed within surgical procedure and disease which can

–  –  –

Writing a thesis for me is a process of self-identification with the profession;

without many people’s support and assistance, I would not have come this far. For many reasons, the first one I would like to thank is my advisor: Dr. Stephanie Guerlain. She is knowledgeable and considerate. As an advisor, she is always there to support her students.

What I have learned from her is not only the knowledge, but the way to carry on a research project and how to put an idea into reality. Thanks to Dr. Turrentine, without her help, I could not have finished this project. She provided a lot of help, especially when I met difficulties in the medical area. I could always feel her encouragement during the whole procedure. I also should give thanks to Dr. Harrell. He was always so patient to answer my questions and teach me how to go further on my research. Thanks for Dr.

Brown, who led me into this area from his class and gave me a great deal of important advice for this project.

I would also like to thank the following people: Debbie Lamb, Ken Scully, Paige Williams, Jin Jun, Jeff Young, David Bogdonoff and Stuart Lowson who assisted me in accomplishing this research.

Finally, I dedicate this dissertation to Min Zheng, my wife. I genuinely appreciate her love and support.

–  –  –



Chapter 1. Introduction

1.1 Patient Safety in the Operating Room

1.2 Thesis purpose and research question

1.3 Thesis Outline

Chapter 2 Literature Review

2.1 Patient Safety Evaluation

2.2 Mortality Variation in Time

2.3 Previous Research for Mortality during the Nights and Holidays/Weekends

Chapter 3 Methodology

3.1 Data Source

3.2 Data pre-processing

3.2.1 Data cleaning

3.2.2 Variable definition

3.2.3 Data Error

3.3 Modeling Strategy

3.3.1 Variable selection

3.3.2 Missing Values

3.3.3 Model Validation

3.3.4 Predictor Complexity

3.3.5 Modeling strategy for hypothesis testing

3.4 Software

Chapter 4 Statistical Analysis and Result

4.1 Pre-Analysis

4.1.1 Descriptive statistics

4.1.2 Predictor Pre-analysis

4.2 Mortality analysis

4.2.1 Full model

4.2.2 Model within CaseType

4.2.3 Model Within Diagnosis

4.2.4 Conclusion for Mortality analysis:

4.3 LOS (Length of Staying) Analysis

4.3.1 Response Transformation

4.3.2 Full Model Analysis

4.3.3 Model within diagnostic category

4.3.4 Conclusion for LOS analysis

4.4 LOS Analysis by Cox Proportional Hazards Model

4.4.1 Model without Stratification

4.4.2 Model with Stratification

4.4.3 Assumption Test

4.4.4 Conclusion for Cox PH model

4.5 Other Analysis


4.6 Summary

Chapter 5 Conclusion and Future Work

5.1 Conclusion

5.2 Future work

5.3 Summary

Appendix A: ICD-9 Disease Categories

Appendix B: Co-morbidity

Appendix: C. Procedure deleted

Appendix D. Statistical Methodology

D.1 Multiple linear regression

D.1.1 Model and assumption

D.1.2 Significance test

D.1.3 Goodness-of-Fit

D.1.4 Predictive accuracy

D.2 Logistic Regression

D.2.1 Model

D.2.2 Parameter interpretation

D.2.3 Significance test

D.2.4 Goodness-of-Fit

D.2.5 Predictive Accuracy

D.3 Cox Proportional Hazards Model

D.3.1 Model

D.3.2 Parameter Interpretation and Assumption

D.3.3 Test Statistics

Appendix E: Descriptive Statistics

Appendix F: Mortality Analysis

Appendix G: LOS Analysis

Appendix H: Cox Proportional Hazards Model

Appendix I: Splus Code


List of Table

Table 3- 1 ASA code description

Table 3- 2 Death Summary

Table 3- 3 Case type error summary

Table 3- 4 ASA score summary

Table 3- 5 Audit result of ASA 6

Table 3- 6 Audit result of missing ASA

Table 3- 7 Audit results for selected sample

Table 3- 8 IPOP code and discharge date summary

Table 4- 1 Descriptive statistics of continuous variable in SOR dataset

Table 4- 2 Descriptive statistics of categorical and ordinal variable

Table 4- 3 Test Result for Status and LOS grouped by Gender, IPOP

Table 4- 4 Test for variables of interest grouped by Night, Weekend and Holiday......... 38 Table 4- 5 Frequencies of surgeries by time period and case type

Table 4- 6 Wald Statistics for Holiday, Night and Weekend

Table 4- 7 Test Statistics- Full Model without interaction

Table 4- 8 Validation Result for the Full model

Table 4- 9 Odds Ratio of the important variables

Table 4- 10 Test Statistics- Full model within EL CaseType

Table 4- 11 Odds Ratio for variables of interest (EL cases only)

Table 4- 12 Test Statistics- Full model within EM CaseType

Table 4- 13 Odds Ratio for variables of interest (EM Cases only)

Table 4- 14 Test Statistics- Full model within LP CaseType

Table 4- 15 Odds Ratio for variables of interest (LP cases only)

Table 4- 16 Death summary for combined diagnostic groups

Table 4- 17 Test Statistics in diagnosis 2

Table 4- 18 Odds Ratio for variables of interest (diagnosis 2)

Table 4- 19 Test Statistics in diagnosis 7

Table 4- 20 Odds Ratio for variables of interest (diagnosis 7)

Table 4- 21 Test Statistics in diagnosis 9

Table 4- 22 Odds Ratio for some important variables (diagnosis 9)

Table 4- 23 Test Statistics in diagnosis 17

Table 4- 24 Odds Ratio for some important variables (diagnosis 17)

Table 4- 25 Test Statistics for Missing Diagnosis code

Table 4- 26 Odds Ratio for some important variables (missing code group)

Table 4- 27 Test Statistics for other groups

Table 4- 28 Odds Ratio for some important variables (other groups)

Table 4- 29 Test statistics for Holiday, Weekend and Night

Table 4- 30 Test Statistics for each variable (except co-morbidities)

Table 4- 31 Predictor effects on TLOS

Table 4- 32 Result of model validation

vii Table 4- 33 Insignificant Variable and relation within each diagnostic regression.......... 73 Table 4- 34 Test statistics Cox PH model stratified by Gender

Table 4- 35 Hazard Ratio for Variables of interest

Table 4- 36 Summary for the number of surgery for each hospitalization

Table 4- 37 Model performance for two datasets


List of Figure

Figure 4- 1 Variable clustering (a)

Figure 4- 2 Variable clustering (b)

Figure 4- 3 Spearman’s ρ 2 rank correlation (response: Status)

Figure 4- 4 Spearman’s ρ 2 rank correlation (response: LOS)

Figure 4- 5 Nonparametric regression estimation for continuous variable against Status 43 Figure 4- 6 Nonparametric regression estimation for continuous variable against LOS.. 43 Figure 4- 7 Log odds of the continuous predictors to the response Status

Figure 4- 8 Power Transformation using Box-Cox Method

Figure 4- 9 Histogram of the LOS distribution

Figure 4- 10 Variable Effects on TLOS

Figure 4- 11 Residual Analysis (with 95% confidence level)

Figure 4- 12 Q-Q plot of residual

Figure 4- 13 Log relative prob. Of discharge alive vs. continuous variable

1 Chapter 1. Introduction

1.1 Patient Safety in the Operating Room The U.S Institute of Medicine estimates that there are about 44,000~98,000 Americans who die each year from medical mistakes, which costs the nation about $17~29 billion each year. However, those numbers do not include the much larger number of medical mistakes that lead to sickness or injury but do not result in death, about 770,000 per year (Kohn, et al., 2000) The operating room (OR) is a high-risk environment and one of the largest consumers of hospital resources. Patients undergoing surgeries usually have a higher risk of dying than those not undergoing surgery. Surgeons, anesthesiologists and nurses may experience under stress and fatigue. Intrinsic variability is associated with different surgical situations and complex modern operative technology. All these make Adverse Events (AEs) more likely to happen in the OR.

Research has found that more than two-thirds of AEs are preventable, where AEs are defined as unintended injuries that are caused by medical management and result in measurable disability (Leape, et al., 1993). (If an AE could not be prevented based on the current state of knowledge, then it is classified as unpreventable; on the other hand, AE is preventable if no error is identified, but a high incidence of this type of complication results from low standards of care or technical service.) It was also noticed that surgical AEs are more likely to be prevented than those resulting from non-surgical management (Leape, et al. 1993). Unfortunately, the complexity of the surgical environment makes

–  –  –

1.2 Thesis purpose and research question The first step for patient safety improvement would be identifying the system failures underlying the errors that occurred. In other words, it is necessary to recognize each contributor to the AEs. Recognition of these factors will allow a better management of them, which in turn will minimize the probability of AE’s happening.

Pages:   || 2 | 3 | 4 | 5 |   ...   | 13 |

Similar works:

«Les moteurs du développement de la métropole Nantes Saint-Nazaire Par Laurent DAVEZIES Stratégie métropolitaine L’analyse par Laurent Davezies des moteurs du développement économique de la métropole Nantes Saint-Nazaire Dans le cadre de la mission qui lui a été confiée par le Syndicat Mixte du SCOT, la SAMOA conduit une réflexion sur la stratégie de la métropole Nantes Saint-Nazaire en privilégiant une démarche de projets qui doit concourir à la constitution d’une culture...»

«STANDARD CLAUSES FOR EXPLORATION CONTRACT Section 1 Definitions 1.1 In the following clauses: (a) “exploration area” means that part of the Area allocated to the Contractor for exploration, described in schedule 1 hereto, as the same may be reduced from time to time in accordance with this contract and the Regulations; (b) “programme of activities” means the programme of activities which is set out in schedule 2 hereto as the same may be adjusted from time to time in accordance with...»

«   Interdependent   Discourses  of   Disability   A  Critical  Analysis  of  the  Social/Medical   Model  Dichotomy       Jan  Grue     Thesis  submitted  for  the   degree  of  Ph.D.,  to  the   Faculty  of  Humanities,   University  of  Oslo,  2011             Contents   Preface  and  acknowledgements   Part  I:  Introduction   Chapter  1:  Preliminaries   Disability:  A  complex  phenomenon  ...»

«THEMATIC REVIEW ON ADULT LEARNING GERMANY COUNTRY NOTE Visit: December 2003 Final Version: May 2005 Germany has granted the OECD permission to include this document on the OECD Internet Home Page. The views expressed in the document are those of the author(s) and not necessarily those of Germany, the OECD or its Member countries. The copyright conditions governing access to information on the OECD Home Page are provided at http://www.oecd.org.rights TABLE OF CONTENTS 1. INTRODUCTION 1.1....»

«����� ���������� ���� ����� Barrie/Midland/ Orillia Final Report ��������� 2007 Commissioned by The Ontario Federation of Indian Friendship Centres The Ontario Metis Aboriginal Association The Ontario Native Women’s Association Urban Aboriginal task force Barrie / Midland / Orillia Final Report Urban Aboriginal task force Barrie/Midland/ Orillia Final Report september 2007 Commissioned by The Ontario Federation of Indian...»

«DESIGN AS.: THINKING OF WHAT DESIGN MIGHT BE Filippo A. Salustri, Ryerson University Nathan L. Eng, University of Cambridge NOTE: This is a preprint of a paper submitted for review to a journal. It is NOT an official publication. Introduction “Basic research is what I’m doing when I don’t know what I’m doing.” – Wernher von Braun (1912-1977) One question that persists in design research, despite many vigorous and on-going research efforts, is What is design? All of us who practice,...»

«Oil & Gas Bulletin 2 1.0 9.2 0 1 2 ► Natural gas reserve found on Turkey’s Black Sea coast Today’s Zaman, 16.09.2012 Drillers looking for natural gas deposits on the western Black Sea coast announced the discovery of one of Turkey’s first natural gas reserves over the weekend, a development that confirmed earlier reports of a large natural gas deposit there. An exploratory drilling operation headed by the state-run Turkish Petroleum Corporation (TPAO) has reached a natural gas deposit...»

«The CoSMo Company Tel +33(0)4 37 66 71 57 5, Passage du Vercors contact@thecosmocompany.com 69007 LYON FRANCE www.thecosmocompany.com PRESS RELEASE Lyon, 31 March 2014 The CoSMo Company raises €3.8 million ($5.2 million) during an initial round of financing and opens an office in the United States. The Lyon based startup, created in 2010 and specialized in complex systems modeling and simulation has successfully completed its first fund-raising campaign. Total subscriptions amounted to €3.8...»

«The internal reliability of some City & Guilds tests A project report prepared by City & Guilds and its research partners for the Office of Qualifications and Examinations Regulation Andrew Boyle The City and Guilds of London Institute Zeeshan Rahman The City and Guilds of London Institute Ofqual/13/5257 January 2013 This report has been commissioned by the Office of Qualifications and Examinations Regulation. Contents Executive summary Introduction Review Method Findings Discussion Conclusions...»

«THE NEW TRANSLATION OF THE HOLY MASS: AN INTRODUCTION By Brian MacMichael On the First Sunday of Advent next year (Nov. 27, 2011), English-speaking Roman Catholic communities around the world will begin using a new English translation of the Roman Missal. The Vatican’s Congregation for Divine Worship and the Discipline of the Sacraments gave its “recognitio,” or approval, to the new translation in late April. However, the delivery of the final texts and the announcement of an...»

«SWC Walk DRAFT Kemble circular: the source of the River Thames. The main feature of this tranquil walk is a visit to the source of the Thames (the Thames head) which is in a field just 3 km from Kemble station Beyond the source the route continues along a section of the disused Severn and Thames canal which is not a typical towpath walk but more like a trail through dense woodland as the canal is completely overgrown and has become an important wildlife corridor, particularly for butterflies....»

«Nº 240 MARTES, 20 DE OCTUBRE DE 2015 XUNTA DE GALICIA CONSELLERÍA DE MEDIO AMBIENTE, TERRITORIO E INFRAESTRUTURAS. AUGAS DE GALICIA. SERVIZO TERRITORIAL DE LUGO Anuncio INFORMACIÓN PÚBLICA O Concello de Muras, con CIF P2703300J, solicita de Augas de Galicia autorización de obras en zona de policía. O presente anuncio ten por obxecto abrir a regulamentaria información pública que prescribe o art. 52 do Regulamento do dominio público hidráulico. Expediente: DH.W27.50236 As obras...»

<<  HOME   |    CONTACTS
2016 www.dissertation.xlibx.info - Dissertations, online materials

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.