Pn P DSS Pregnancy Parturition Process DSS Himanshu

Pn. P DSS Pregnancy & Parturition Process DSS Himanshu Chavda Lemuel Dizon Ateeqth. Jaffer March 5 , 2009 1

Introduction VBAC - Vaginal Birth After Cesarean. 1996 28% VBAC Healthy People 2010 § Target VBAC 37% 2006 8% VBAC ACOG – American College of Obstetricians and Gynecologists. 9 out of 10 “Once a Cesarean, always a Cesarean” 2

Facts about Cesarean delivery rate (CDR) in the US in 1970 was 5. 5 per 100 delivery and increased to 24. 7 in 1988. By 2007 the rate has increased to 33%. U. S. health objective to reduce CDR to 15% World Health Organization (WHO) recommends 5 – 10%. Above 15% seem to do more harm than good. Cause of Increasing CDR Low priority of enhancing women's own abilities to give birth. Side effects of common labor interventions. Refusal to offer the informed choice of vaginal birth. Casual attitudes about surgery and cesarean sections in particular. Limited awareness of harms that are more likely with cesarean section. Providers' fears of malpractice claims and lawsuits. Incentives to practice in a manner that is efficient for providers. 3

Pn. P DSS Overview MODEL § Decision Tree § Knowledge Engineering SYSTEM § Architectural Diagram § Technology § User Interface EVALUATION § Sensitivity Analysis § Pilot Programs 4

Pn. P Model Design This CDSS will allow clinicians and expected mother on the informed choice of delivery mode. Decision tree models various studies for uncertainties in terms of probabilities, values/utilities and attributes. The terminal values are utility scores from a certain study that was attributed to a natural ranking between 0 & 1. A better outcome meant a higher score. The core logic utilize Bayesian inference to compute the probability of each uncertainty from [1, 2] research studies. The model analyzes some the common pregnancy uncertainties such as: §Maternal Death §Cerebral Palsy §Still Birth §Hysterectomy §Pregnancy §No Pregnancy §Rescue CS §Normal delivery §Morbidity §Healthy Birth §Mortality/Death The core objective is to help expected mothers with history of prior CS to reach a best possible decision during the pregnancy. 5

Pn. P Model – Decision Tree 6

Knowledge Engineering Tree. Age Software Ver. 3. 5 Published study by Sadan et al in Arch Gynecol Obstet. Published study by Vandenbussche et al in BIRTH. TOL Vs Repeat CS. 7

Pn. P System Design Architectural Diagram 8

Pn. P System Design - Technology Method A – Using the Pn. P Website to Query the data directly from Pn. P DSS stationed at the Vendor Site The Pn. P DSS will have an interactive website which the authorized personnel can access for their needs. The login page on the Pn. P website would ask for the user credentials. The Patient registration form on the Pn. P website should have the following Fields to capture pertinent data (Patient data, Medical History, Comments) The clinician registration form on the Pn. P website should have the following Fields to capture pertinent data (Clinician data, Area of Expertise, CSI #, Comments) The application will be driven using Apache Tomcat, which will implement the servlets and JSP or struts (framework) Here’s an excerpt of the Data Interchange process between Pn. P Website & Pn. P DSS The Pn. P application server captures the data submitted from the Pn. P website by the end-users, and will communicate the data request to the web-server in the Pn. P DSS. The data will be transmitted via HTTP/HTTPS protocol. The Pn. P DSS system will query its database with the pertinent data (requested thru the website) & render it back to the Pn. P website. 9

Pn. P System Design - Technology Method B – Using the Pn. P DSS driven by feeds from Pn. P ODS (Operational Data Store) In this option, the Pn. P DSS residing at the Clinical Institution gets a feed from the Pn. P ODS on a predetermined basis. The Pn. P DSS is utilized to serve the needs of the patients & Clinicians in making an informed & educated decision. The Pn. P ODS will serve as a Knowledge repository which contains the data pertaining to Pn. P process as well as a ‘Data Collection & Analysis’ repository to acquire the latest information pertaining to the Pn. P process & reinforce the Knowledge respository with the updated/refined information. Here’s an excerpt of the Ongoing Data Interchange process between Pn. P DSS & Pn. P ODS The Pn. P application server from the Pn. P DSS will communicate the data to the webserver in the Pn. P ODS. The data will be transmitted via HTTP/HTTPS protocol. The Pn. P ODS system will update its database with the pertinent data (Delta, Changes) & incorporate it in the program logic Once the validity of the new findings is confirmed, the Pn. P ODS application server will communicate that to the Pn. P DSS webserver. The Pn. P DSS can now be queried for the latest information pertaining to the Pn. P process. Here’s an excerpt of the Data retrieval process between Pn. P Website & Pn. P DSS The Pn. P application server captures the data submitted will communicate the data request to the webserver in the Pn. P DSS. The Pn. P DSS system will query its database with the pertinent data (requested thru the website) & render it back to the Pn. P website. 10

Pn. P System Design - Technology Method B – Using the Pn. P DSS driven by feeds from Pn. P ODS (Operational Data Store) Data Interchange Options Option 1 - Web services can be employed to effect data Interchange between Pn. P DSS & Pn. P ODS Web services are Web based applications that use open, XML-based standards and transport protocols to exchange data with clients. 11

Pn. P System Design - Technology Method B – Using the Pn. P DSS driven by feeds from Pn. P ODS (Operational Data Store) Data Interchange Options Option 2 – Data Interchange between Pn. P DSS & Pn. P ODS via HTTP/S Forms Data can be interchanged between Pn. P DSS & Pn. P ODS through the respective application servers using the http/https protocol. This is kind of a tight coupling. 12

User Interface 13

Implementation & Evaluation §Sensitivity Analysis §Pilot Programs 14

Feedback: Expectation and Knowledge Scale 15

Pn. P DSS Product Evaluation 16

References : 17 1. Sadan, O. , Leshno, M. , Gottreich, A. , Golan, A. , Lurie, S. , “Once a cesarean always a cesarean? A computer-assisted decision analysis. ” Arch Gynecol Obstet (2007) 276: 517– 521. Pub. Med. Gulter Health Sciences Library, Chicago. 2 Feb. 2009. http: //www. galter. northwestern. edu/. 2. Frank P. H. A. Vandenbussche, MD, Lieke C. De Jong-Potjer, MD, Anne M. Stiggelbout, Ph. D, Saskia Le Cessie, Ph. D, and Marc J. N. C. Keirse, MD, DPhil, FRACOG, . FRCOG, “Differences in the Valuation of Birth Outcomes Among Pregnant Women, Mothers, and Obstetricians. ” BIRTH 26: 3 September 1999. Pub. Med. Gulter Health Sciences Library, Chicago. 2 Feb. 2009. http: //www. galter. northwestern. edu/. 3. Caroline Signore, MD, MPH, Anusha Hemachandra, MD, MPH, and Mark Klebanoff, MD, MPH. , “Neonatal Mortality and Morbidity After Elective Cesarean Delivery Versus Routine Expectant Management: A Decision Analysis. ” Seminars in Perinatology 2006. 07. 010. Pub. Med. Gulter Health Sciences Library, Chicago. 4 Feb. 2009. 4. Naomi E. Stotland, MD, a, b Lisa S. Lipschitz, MD, a and Aaron B. Caughey, MD, MPP, MPHa. , “Delivery strategies for women with a previous classic cesarean delivery: A decision analysis. ” 2002, Mosby, Inc. mob. 2002. 127123. Pub. Med. Gulter Health Sciences Library, Chicago. 4 Feb. 2009. 5. Cynthia J. Sims, MD, a Leslie Meyn, BS, a Rich Caruana, Ph. D, b, c R. Bharat Rao, Ph. D, d Tom Mitchell, Ph. D, b and Marijane Krohn, Ph. Da. “Predicting cesarean delivery with decision tree models. ” 2000, Mosby, Inc. mob. 2000. 108891. Pub. Med. Gulter Health Sciences Library, Chicago. 4 Feb. 2009. 6. Alan A Montgomery, Clare L Emmett, Tom Fahey, Claire Jones, Ian Ricketts, Roshni R Patel, Tim J Peters, Deirdre J Murphy and Di. AMOND Study Group. “Two decision aids for mode of delivery among women with previous caesarean section: randomised controlled trial. ” BMJ, doi: 10. 1136/bmj. 3921 7. 67101955 (published 04 April 2007). Pub. Med. Gulter Health Sciences Library, Chicago. 31 Jan. 2009. 7. http: //www. time. com/time/magazine/article/0, 9171, 1880665, 00. html