Upload
mdpnpuiuc
View
124
Download
1
Embed Size (px)
Citation preview
A Treatment Validation Protocol for Medical Cyber-Physical-Human Systems
Po-Liang Wu, Dhashrath Raguraman, Lui Sha (Computer Science, UIUC)
Richard B. Berlin Jr. (College of Medicine/ Computer Science, UIUC)
Julian M. Goldman (Massachusetts General Hospital/ CIMIT) To appear In EUROMICRO Software Engineering and Advanced
Applications (SEAA), 2014.
1
Outline § Introduction § System Architecture and Models § Treatment Validation Protocol § UPPAAL Model and Verification § Conclusion and Future Work
2
Introduction § Statistics indicates that preventable medical error rate is highest in
Intensive Care Unit (ICU) as compared to other hospital units. § Many preventable medical errors are result from unintended deviation
from the best practice medical guidelines § Timely and correctly perform treatments while monitoring potential side
effects of performed treatments is critical for patient safety. § Treatment Validation:
• Preconditions » If any precondition is not satisfied, a corrective treatment should be performed.
• Potential side effects » The side effects of a treatment may adversely affect other treatments.
• Expected physiological responses » The effectiveness of a treatment is non-deterministic.
§ The purpose of this work is to assist physicians to correctly perform treatments rather than automatically performing treatments.
3
Motivating Example: Cardiac Arrest Resuscitation
§ Cardiac arrest is the abrupt loss of heart function and can lead to death within minutes.
§ American Heart Association (AHA) provided resuscitation guidelines for the urgent treatment of cardiac arrest
§ An epinephrine example for cardiac arrest resuscitation • Epinephrine is a commonly used drug for improving patient’s cardiac output. • Epinephrine may not be effective if patient’s blood pH value < 7.4 • Epinephrine may adversely raise patient’s blood pressure.
§ Preventable medical errors: • Shock a patient when the EKG shows an un-shockable rhythm. • Sodium bicarbonate is injected with calcium chloride • etc
4
Technical Challenges § The preconditions and corrective treatments may cause a cascade effect § Potential side effects are dynamic and may interfere with other treatments
or invalidate the preconditions. § The effectiveness of the treatments are non-deterministic,
6
Ac#vateDefibrillator
Rhythm == Shockable
InjectEPI
BloodPH > 7 . 4
UrineFlow > 12 mL / s
Inject -‐ SodiumBicarbonate
Airway & Breathing
Treatment
Preconditions We propose a Treatment Precondition and Correction (TPC) tree to structure treatments and preconditions.
Technical Challenges § The preconditions and corrective treatments may cause a cascade effect § Potential side effects are dynamic and may interfere with other treatments
or invalidate the preconditions. § The effectiveness of the treatments are non-deterministic,
7
Ac#vateDefibrillator
Rhythm == Shockable
InjectEPI
BloodPH > 7 . 4
UrineFlow > 12 mL / s
Inject -‐ SodiumBicarbonate
Airway & Breathing
AssistedVen#la#on
The side effect of bicarbonate may adversely affect patient’s breathing.
Urine flow rate may become lower than 12 due to kidney function degradation.
IncreaseIVFluid
Technical Challenges § The preconditions and corrective treatments may cause a cascade effect § Potential side effects are dynamic and may interfere with other treatments
or invalidate the preconditions. § The effectiveness of the treatments are non-deterministic,
8
Ac#vateDefibrillator
Rhythm == Shockable
InjectEPI
BloodPH > 7 . 4
UrineFlow > 12 mL / s
Inject -‐ SodiumBicarbonate
Airway & Breathing
AssistedVen#la#on
The side effect of bicarbonate may adversely affect patient’s breathing.
Urine flow rate may become lower than 12 due to kidney function degradation.
IncreaseIVFluid Unlike traditional cyber validation mechanisms, the system cannot fully control, lock or rollback, the physical components, such as patient conditions.
System Architecture § Medical Device Plug and Play
(MDPnP) is a centralized supervisory framework for integrating networked medical devices.
§ The proposed Treatment validation protocol gathers physiological information and command medical devices through MDPnP controller.
Treatment Valida,on Protocol
MDPnP Controller
User Interface Medical Staff
Pa#ents
Treatment Request
Pa#ent condi#ons and system states
Treatment Ac#ons
Pa#ent condi#ons and devices states
EKG Monitor Adapter
Defibrillator Adapter
Lab Value Monitor Adapter
Infusion Pump
Adapter HR / BP Sensor Adapter
9
Physical Models and Definitions § PhysiologicalCondition (PC) is defined as a tuple < Checker, PM, Operator, RV
>, where • Checker is the entity capable of checking the, Checker ∈{System, MedicalStaff}, • PM (physiological measurement) ∈ {EKGRhyhtm, HeartRate, BloodPressure...}, • Operator ∈ {>, <, =, ≤, ≥}, • RV is the reference value, which can be threshold, trend, or pattern of the
physiological measurements. § A treatment is defined as a tuple < Agent, Action, PS, SS, ES >, where
• Agent is the entity that performs the treatment, Agent ∈ {MedicalDevices, MedicalStaff},
• Action is the set of executable instructions, • PS is a set of preconditions that must be satisfied before preforming the Action • SS is a set of potential side effects • ES is a set of expected responses after performing the treatment
10
A Treatment Validation Protocol § The proposed treatment validation protocol consists of three phases
• Treatment precondition and collection tree construction phase » Check preconditions and request medical staff to specify a corrective
treatment if any precondition is not satisfied • Executing and monitoring phase
» Send requests to medical devices to perform the treatment » Monitor the potential side effects
• Checking expected responses phase » Check if the patient’s response is as expected and adjust the treatments
accordingly.
11
Supervisory System
User Interface
Defibrillator
Airway Breathing
Rhythm Shockable
1. Ac#vate defibrillator
2. Check Airway & EKG Rhythm
12
Supervisory System
User Interface
Defibrillator
Airway Breathing
Rhythm Shockable
1. Ac#vate defibrillator
2. Check Airway & EKG Rhythm
Epinephrine
BloodPH > 7.4
Urine > 12 mL/s
3. Inject epinephrine
4-‐a. Check precondi#ons 4-‐b. Request correc#ve treatments
13
Supervisory System
User Interface
Defibrillator
Airway Breathing
Rhythm Shockable
1. Ac#vate defibrillator
2. Check Airway & EKG Rhythm
Epinephrine
BloodPH > 7.4
Urine > 12 mL/s
3. Inject epinephrine
4-‐a. Check precondi#ons 4-‐b. Request correc#ve treatments
Sodium Bicarbonate
No Calcium Chloride
Increase IV Fluid
5. Sodium bicarbonate & Increase IV
14
Supervisory System
User Interface
Defibrillator
Airway Breathing
Rhythm Shockable
1. Ac#vate defibrillator
2. Check Airway & EKG Rhythm
Epinephrine
BloodPH > 7.4
Urine > 12 mL/s
3. Inject epinephrine
4-‐a. Check precondi#ons 4-‐b. Request correc#ve treatments
Performing Treatments Sodium Bicarbonate
6. Post order execu#on
Sodium Bicarbonate
No Calcium Chloride
Increase IV Fluid
5. Sodium bicarbonate & Increase IV
15
Supervisory System
User Interface
Defibrillator
Airway Breathing
Rhythm Shockable
1. Ac#vate defibrillator
2. Check Airway & EKG Rhythm
Epinephrine
BloodPH > 7.4
Urine > 12 mL/s
3. Inject epinephrine
4-‐a. Check precondi#ons 4-‐b. Request correc#ve treatments
Performing Treatments Sodium Bicarbonate
6. Post order execu#on
Sodium Bicarbonate
No Calcium Chloride
Increase IV Fluid
5. Sodium bicarbonate & Increase IV
Airway Breathing
BloodPH > 7.4
Sodium Bicarbonate
No Calcium Chloride
7. Side effects adversely affect precondi#ons
8. Pa#ent responses not as expected
9. Excep#ons 10. Alterna#ve Treatment
Assisted Ven#la#on
16
Correctness § Theorem 1. (Preconditions satisfaction) Under the proposed
protocol, a treatment is performed only if all preconditions of the treatment are satisfied.
§ Definition: A tree is well-formed if each unsatisfied precondition have a tree node for correcting it.
§ Theorem 2. (Tree Traversal) The root node of a well-formed tree is reachable if the corrective treatments are effective and the preconditions are not invalidated by the side effects.
§ Theorem 3. (Dynamic adaptability) Suppose side effects of a treatment invalidate any precondition and make the tree become non-well-formed. The protocol updates the tree to be well-formed if the medical staff correctly specifies the corrective treatments.
17
Verification § We model the proposed protocol in UPPAAL to verify both safety
and correctness properties. § The reason we choose UPPAAL is that it provides nice user
interface for the physicians to review the developed models. § The medical devices send the patient’s physiological
measurements, which are modeled as non-deterministic transitions, to the validation protocol.
§ In addition, the medical devices also receive the treatment requests from the protocol and change the states accordingly.
18
UPPAAL Model: Treatment Validation Protocol
19
Check preconditions and construct a tree
Check expected responses
Verification
Medical safety properties
P1: Defibrillator is activated only if the rhythm is shockable and airway and breathing is under control. P2: Epinephrine is injected only if the blood pH value is larger than 7.4 and urine flow rate is higher than 12 mL/s. P3: If the side effect of sodium bicarbonate adversely affects the breathing, the tree is updated with a new treatment node for assisted ventilation.
Protocol correct ness properties
P4: There is no deadlock in the system.
P5: A treatment is performed only if all its preconditions are satisfied.
P6: If side effect does not occur, the root node of the tree is added to the executing list P7: If side effects invalidate a precondition, the tree is updated to be well-formed.
20
Conclusion and Future Work § In order to reduce medical errors, a validation protocol is proposed to
• Check preconditions • Monitor potential side effects • Check expected responses
§ We use a model checking tool to verify both safety and correctness properties.
§ Human computer interaction (HCI) and situation awareness is an important aspect for developing supervisory medical systems with human in the loop.
§ We are implementing a resuscitation assistant system and evaluate medical staff’s mental workload using a trace-driven simulation.
§ In the future, we will check the applicability of the proposed architecture to various medical scenario.
21
Physical Models and Definitions § PhysiologicalCondition (PC) is defined as a tuple < Checker, PM, Operator, RV
>, where • Checker is the entity capable of checking the, Checker ∈{System, MedicalStaff}, • PM (physiological measurement) ∈ {EKGRhyhtm, HeartRate, BloodPressure...}, • Operator ∈ {>, <, =, ≤, ≥}, • RV is the reference value, which can be threshold, trend, or pattern of the
physiological measurements. § A treatment is defined as a tuple < Agent, Action, PS, SS, ES, L >, where
• Agent is the entity that performs the treatment, Agent ∈ {MedicalDevices, MedicalStaff},
• Action is the set of executable instructions, • PS is a set of preconditions that must be satisfied before preforming the Action • SS is a set of potential side effects • ES is a set of expected responses after performing the treatment • L is the life cycle, which specifies the time interval between the treatment being
performed and the treatment has no further effect on the patient
24