Multi-parameter data collection example J. Zaleski October 4th, 2009

Discontinuation from mechanical ventilation Parameter Threshold Value/Range Vital Capacity, Vc > 10mL/kg Positive End-Expiratory Pressure, PEEP 5 cm H2O Negative Inspiratory Force, NIF -20 cm H2O Inspired Oxygen Fraction, FiO2 < 0.6 Spontaneous Tidal Volume, Vt > 5 mL/kg Spontaneous Respirations, fR 8 < Resp Rate < 30 Blood Alkalinity/Acidity 7.32 < pH < 7.48 Partial Pressure of Oxygen, SPO2 > 80 mmHg Partial Pressure of Carbon Dioxide, PCO2 30 mmHg < PCO2 < 50 mmHg Normal Body Temperature, Tcore ~37 C Ventilation Mode CPAP Clinical decision making is a multi-parameter action Behavior, trending of several parameters coincidentally influences the recommended ordering of clinicians

Discontinuation from mechanical ventilation Parameter Threshold Value/Range Vital Capacity, Vc > 10mL/kg Positive End-Expiratory Pressure, PEEP 5 cm H2O Negative Inspiratory Force, NIF -20 cm H2O Inspired Oxygen Fraction, FiO2 < 0.6 Spontaneous Tidal Volume, Vt > 5 mL/kg Spontaneous Respirations, fR 8 < Resp Rate < 30 Blood Alkalinity/Acidity 7.32 < pH < 7.48 Partial Pressure of Oxygen, SPO2 > 80 mmHg Partial Pressure of Carbon Dioxide, PCO2 30 mmHg < PCO2 < 50 mmHg Normal Body Temperature, Tcore ~37 C Ventilation Mode CPAP Clinical decision making is a multi-parameter action Behavior, trending of several parameters coincidentally influences the recommended ordering of clinicians Parameters, Pi Value Thresholds, Vpthi

Clinical Decision Making is Based Upon Multiple Data Sources LIS RIS CIS Devices EHR … Data from multiple sources (enterprise and departmental health systems, devices, etc.) P1 P2 P3 … Key Parameters Used to Determine Action Clinical Decision Support Systems Vpt1 < Vpth1 Vpt2 < Vpth2 … Vpti < Vpthi Action (Note: NOT simply a linear sum of elements, but a system of mutually-satisified constraints)

Examples of Multi-Parameter Medical Device Data Spontaneous Mandatory 12 fR (br/min) VT (l/br) 0.5 CO (l/min) 3 Tc(C) 37 HR (/min) 88 Nitroprusside (u/kg/min) 0.5 1 2 3 4 … Time, in hours

Synchronous and Asynchronous Data Collection Spontaneous Mandatory x 12 x x x x x x x x x x x x x x x x x x x x x x x x x x x fR (br/min) x x x x x x x x x x x x x VT (l/br) x x x x x x x x 0.5 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x CO (l/min) x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 3 Tc(C) x 37 x x x x x x HR (/min) x x x x x x x x x x x x x x x x x 88 x x x x x x x x Nitroprusside (u/kg/min) x x x 0.5 x x x x x x x 1 2 3 4 … Time, in hours

Synchronous and Asynchronous Data Collection Spontaneous Mandatory 12 x x x x x x x x x x x x x x x x x x x x x x fR (br/min) x x x x x x x x x x x x x x x x x x x VT (l/br) x x x x x x x x x x x x x x x x x x 0.5 x x x x x x x x x x x x x x x x x x x x x x x x x CO (l/min) x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 3 Tc(C) x 37 x x x x x x HR (/min) x x x x x x x x x x x x x x 88 x x x x x x x x x x x Nitroprusside (u/kg/min) 0.5 x x x x x x x x x x Time, in hours Data availability and staleness in relation to decision making can influence decisions

Synchronous and Asynchronous Data Collection Spontaneous Mandatory 12 x x x x x x x x x x x x x x x x x x x x x x fR (br/min) x x x x x x x x x x x x x x x x x x x VT (l/br) x x x x x x x x x x x x x x x x x x 0.5 x x x x x x x x x x x x x x x x x x x x x x x x x CO (l/min) x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 3 Tc(C) x 37 x x x x x x HR (/min) x x x x x x x x x x x x x x 88 x x x x x x x x x x x Nitroprusside (u/kg/min) 0.5 x x x x x x x x x x Data availability and staleness in relation to decision making can influence decisions Decision support systems may require higher-fidelity data or more frequent collection

Synchronous and Asynchronous Data Collection Spontaneous Mandatory x 12 x x x x x x x x x x x x x x x x x x x x x x x x x x x fR (br/min) x x x x x x x x x x x x x VT (l/br) x x x x x x x x 0.5 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x CO (l/min) x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 3 Tc(C) x 37 x x x x x x x HR (/min) x x x x x x x x x x x x x 88 x x x x x x x x x x x x x Nitroprusside (u/kg/min) x x x x 0.5 x x x x x x x Trigger queries for data to fill missing gaps (that is RECEIVING system requests missing data) OR To update parameters requested of rules algorithms based on gaps Data availability and staleness in relation to decision making can influence decisions

Objective Aim is not on communication (transport), but on method for allowing triggering of new results based on functional needs of algorithms Can operate using existing or new messages, according to current standards Independent of real-time messaging protocols

Initiate query for missing data to appropriate target device Example Hub Synchronous Data Communication CIS / Departmental / CDSS / EMR Gateway Interface Engine CDSS evaluates multiple parameters, determines relevance and importance of each with respect to specific condition being tested (example: spontaneous breathing trials) Gateway Hub Support query for parameter that does not meet certain criteria on currency, or for which data are lacking. CIS / Departmental / CDSS / EMR Gateway Interface Engine Initiate query for missing data to appropriate target device Gateway

Motivating Needs Cardiac Output changes with respect to nitroprusside levels Early reduction in vasodilator levels can lead to cardiac output reduction and distress; Associating changes in infusion of specific drips with cardiovascular performance requires timely knowledge of both infusion and CO, heart rate (changes in CO can occur in just a few seconds) Decision support algorithms require latest information on both to recommend action Spontaneous breathing rates and rapid shallow breathing with respect to changes in mandatory support levels Sudden increases in RSBI after changes in support levels are indicators to notify staff of need for intervention Need for suctioning and events surrounding intubation can occur suddenly and decision support algorithms require synchronous information on spontaneous parameters and work of breathing together with temperature Decision to reduce support levels driven by multiple parameters Change in pH and end-tidal CO2 levels influence respiratory support levels. Determination of changes in support require latest information on blood gases and (from Swan) invasive cardiac and real-time blood gas parameters Need for updated blood gas data (pH, PCO2, B.E., HCT, etc.) can be triggered by the need for a CDSS to operate on new data (i.e., recommend spontaneous breathing trials subject to result of latest blood gas) Changes in systemic vascular resistance (SVR) leading to variation in cardiac output can occur due to sympathetic and parasympathetic tone changes (and through infused drug) in less than 1 second. Patient responses cannot be interpolated from past data—discontinuities and sudden changes from moment to moment can occur