|
|
October 16-20, 2010 San Diego, CA Home Abstract Archive Search Abstracts 2010 Session Grid 2010 Meeting Website Policy Statements ASA Website Feedback 
Visit Anesthesiology.org |
| Previous Abstract | Next Abstract |
| Printable Version |
|
A1682
October 22, 2008 9:00 AM - 11:00 AM Room Hall E2-Area G, |
||||||||||||||
| A Novel Integrated Pulmonary Index (IPI) Quantifies Heart Rate, Etco2, Respiratory Rate and Spo2%. | ||||||||||||||
|
Arthur Taft, Ph.D., Michal Ronen, Ph.D., Chad Epps, M.D., Jonathan Waugh, Ph.D., Richard Wales, B.S. School of Allied Health, Respiratory Therapy, Medical College of Georgia, Augusta, Georgia |
||||||||||||||
|
Introduction An Integrated Pulmonary Index (IPI) aimed at providing clinicians with information regarding a patient's pulmonary status is presented. The index reflects knowledge of clinicians skilled in respiratory areas, plus commonly accepted theories from medical literature. It is generated by mathematical analysis of simulated data, end-tidal carbon dioxide (EtCO2), respiration rate (RR), oxygen saturation percent (SpO2%) and heart rate (HR), integrated into a single number. The IPI ranges 1-10, where 10 is healthy and 1 requires immediate intervention. With this index, the clinician is able to capture IPI status of a patient via measured parameters and their interrelations. Methods: To determine IPI value based on EtCO2, RR, HR and SpO2%, data sets were scored by 18 medical experts (nurses, respiratory therapists, physiologists and anesthesiologists). The data sets included 85 cases with different parameter values. The experts were asked to give an IPI value according to a predefined code.[table1]A fuzzy logic inference was built using experts' know-how and fine-tuned using the consensus of experts' data. The design was carried out using MATLAB® fuzzy logic toolbox (Mathworks Inc.). Initially, membership functions were assigned for each variable, determining the ranges for: Normal, High, Very High, Low and Very Low. Then a rule set was determined relating the inputs to the output, using the verbal descriptor of the membership functions. The model developed is nonlinear, yet it is smooth.[figure1]Figure 1: IPI for different EtCO2 and Respiration Rate (RR) values. HR=70, SpO2=100 (Adult model). There are differences in normal ranges of HR and RR between adults and children; hence IPI-models were designed for distinct ages. Results The IPI rated by the experts are subjective and depend on their medical background and experience.[figure2]Figure 2: IPI for different cases sorted by similarity. FL model= fuzzy Logic inference, Avg= average. Color range: Blue- IPI=10 to red- IPI=1 Our model results agree with the experts' average (mean absolute differences =0.64±0.5). Comparing to all experts and cases, the average absolute difference between experts and model is 1±0.35. The agreement between the model and experts views is apparent by the distribution of differences (abs diff< 2 for 70% of the data tested and < 3 for 92% of the data). Conclusion: The IPI incorporated expert's knowledge into a fuzzy-logic inference. The IPI presents one value demonstrating real time respiratory status based on RR, HR, SpO2%, and EtCO2. Anesthesiology 2008; 109 A1682 |
||||||||||||||
|
Figure 1
 Figure 2  |