Medical Informatics

Shmuel Rotenstreich

Friedman

“Medical Informatics is not about usingMicrosoft Word to enter patientinformation…”

Charles Friedman, PhDUniversity of Pittsburghat the UW Symposium, Fall 2000

Shortliffe

“ Medical informatics is the rapidly developing scientific field that deals with resources, devices and formalized methods for optimizing the storage, retrieval and management of biomedical information for problem solving and decision making”

Edward Shortliffe, MD, PhD1995

Computers in Medicine• Information central to biomedical research and

clinical practice• Type

– integrated information-management environments – affect on practice of medicine and biomedical

• Method– medical computing – medical informatics– clinical informatics– bioinformatics

Value• Value of medical-informatics and informatics

applications• Computers and the Internet in biomedical

computing• Relation among

– medical informatics – clinical practice – biomedical engineering– molecular biology– decision support

Difference

• information in clinical medicine and “regular” information

• Changes in computer technology and change in medical care and finance

• Integration of medical computing into clinical practice and “regular” computing integration

Areas• Medical Decision making• Probabilistic medical reasoning• Patient care and monitoring systems• Computer aided surgery• Electronic patient records• Clinical decision support• Standards in medical informatics• Imaging• Image management systems• Telemedicine

Medical Informatics

• Medical Education

• Patient Data Collection and Recording

• Clinical Information Retrieval

• Medical Knowledge Retrieval

• Medical Decision Making

Medical Informatics is Multidisciplinary

• Applies methodologies developed in multiple areas of science to different tasks

• Often gives rise to new, more general methodologies that enrich these scientific disciplines

Example of Scientific Areas Relevant to Medical Informatics

• Medicine/ Biology• Mathematics• Information Systems• Computer Science• Statistics• Decision Analysis• Economics/Health Care Policy• Psychology

The Diagnostic-Therapeutic Cycle

Patient

Data collection:-History-Physical examinations-Laboratory and other tests

Decisionmaking

Planning

Information

Diagnosis/assessmentTherapy plan

Data

Levels of Automated Support(Van Bemmel and Musen, 1997)

Medical Decision-Support Systems• Task:

– Diagnosis/interpretation– Therapy/management

• Scope:– Broad (e.g., Internist-I/QMR: internal medicine Dx;

DxPlain; Iliad; EON for guideline-based therapy)– Narrow (e.g., a system for diagnosis of acute

abdominal pain; MYCIN: infectious diseases Dx; ECG interpretation systems; ONCOCIN: support of application of oncology protocols)

Types of Clinical Decision-Support Systems

• Control level:– Human-initiated consultation (e.g., MYCIN,

QMR)– Data-driven reminder (e.g., MLMs)– Closed loop systems (e.g., ICU ventilator

control)• Interaction style:

– Prescriptive (e.g., ONCOCIN)– Critiquing (e.g., VT Attending)

Diagnostic/Prognostic Methods• Flow charts/clinical algorithms• Statistical and other supervised and

nonsupervised classification methods– Neural networks, ID3, C4.5, CART, clustering

• Bayesian/probabilistic classification– Naïve Bayes, belief networks, influence diagrams

• Rule-based systems (MYCIN)• “Ad hoc” heuristic systems (DxPlain)• Cognitive-studies inspired systems (Internist I)

de Dombal’s System (1972)• Domain: Acute abdominal pain (7 possible diagnoses)• Input: Signs and symptoms of patient• Output: Probability distribution of diagnoses• Method: Naïve Bayesian classification• Evaluation: an eight-center study involving 250 physicians and

16,737 patients• Results:

– Diagnostic accuracy rose from 46 to 65%– The negative laparotomy rate fell by almost half– Perforation rate among patients with appendicitis fell by half– Mortality rate fell by 22%

• Results using survey data consistently better than the clinicians’ opinions and even the results using human probability estimates!

Definitions• Medical Informatics: the science of medical

information collection and management

• Medical Decision Making: quantitative methods for reasoning under uncertainty

• Medical Computing: computer applications for information management

• Medical Decision Support: computer-based information processing to help human decision makers

Case PresentationDescription: 74 female, history of right CVA (cerebrovascular accident*) in 1989

(LLE weakness), one week of productive cough and increased debility.

Exam consistent with bronchitis, oral antibiotic prescribed, but patient had a tonic grand mal seizure in clinic

Became flaccid, unconscious, pulseless, apneic, but upon positioning for CPR, developed pulse and spontaneous respirations and awoke about 2 minutes after start of episode, complaining of lower sternal chest pain.

Actions:

– Transfer to Emergency Room– Examination– Bloodwork– Chest Xray– Cardiogram– Admission and therapy

* Of or relating to the blood vessels that supply the brain

Demo - Part I• Lab Data: ABG and CPK/Isoenzymes• Radiology: CXR, VQ, Doppler• Cardiology: ECG, Cardiac Cath• Medications• Alerts• Discharge Summary

ABG - Arterial blood gas CPK - blood test CXR – Chest X-RayEKG: Electrocardiogram (ECG) Cardiac Cath - Interventional heart catheterization

Case SummaryDescription: bronchitis, bed-bound, venous thrombosis, pulmonary embolism, myocardial infarction, ventricular arrhythmia, hypotension, seizure, adult respiratory distress syndrome, methicillin-resistant Staph aureus

Discharge Plan» Where?» What happened?

Outpatient Follow-up» Medications» Laboratory» Health Maintenance

Demo - Part II

• Demographic Information

• Additional Hospitalizations?

• More Discharge Summaries?

• Recent Lab Results

• Outpatient Notes

How Did We Do It?

• Information Science

• Standards

• Integration

Ambulatory Care• Aka Primary Care, Office Medicine…• Roles (information specific):

– Patient– Scheduling, Registration– Nursing, Triage– Physician– Ancillary Services

• Radiology

Patient

• Able to request an appointment!• Check meds!• Self reported SF-36 functional• Insurance Information!

Clinic Receptionist

• Appointment scheduling• Check-in• Insurance Information• Billing• Follow-up visit

Nurse

• Triage (certain settings)• Chief Complaint• Brief History• Vital signs & Initial Exam• Pulse, BP, Respirations, Pulse Oximeter• Psychosocial Assessment• Discharge Instructions (Pt Education)

Physician

• Review Chart Data, Studies• Document History and Physical Exam• Dx, Tx plan (orders, follow-up)• SOAP note

– Subjective– Objective– Assessment– Plan

Ancillary Studies: Radiology Tech

• Schedule Exam• Review Allergies, Pregnancy• Review Clinical Indication• Enter Exam Data

Conventional data collection for clinical trial

Clinical trial design• Definition of data elements•Definition of eligibility•Process descriptions•Stopping criteria•Other details of the trial

Data sheets

Computer database

Analyses

Results

Medical records

Role of EMR in supporting clinical trials

Clinical trial design• Definition of data elements•Definition of eligibility•Process descriptions•Stopping criteria•Other details of the trial

Clinical trial database

Analyses

Results

Medical records systems

Clinical datarepository

Networking the organization

Enterprise network

Patientworkstation

Clinical workstations

Clerical workstation

Researchdatabeses

Administrative systems(e.g. admissions, discharges and transfers)

Libraryresources

Radiology

Billing andfinancial systems

Costaccounting

Microbiology

Pharmacy

Clinical databasesElectronic medical

records

Personnelsystems

Materialmanagement

Educationalprograms

Clinicallaboratory

Datawarehouse

Moving beyond the organization

Patients

Healthyindividuals

Providersin officesor clinics

Informationresources(Medline..)

Governmentmedical research

agencies

3rd partypayers

The Internet Governmenthealth insurance

programsOther hospitalsand physicians

Pharmaceuticalsregulators

Communicabledisease agencies

Health ScienceSchools

Vendorsof various types

(e.g. pharmaceuticalscompanies

Healthcare institutes Needs

• Healthcare institutes are seeking Integrated clinical work stations that will assist with clinical matters by:– Reporting results of tests– Allowing direct entry of orders– Facilitating access to transcribed reports– Supporting telemedicine applications– Supporting decision-support functions

The Heart of the Evolving Clinical Workstation

• Electronic • Confidential• Secure• Acceptable to clinicians and patients. • Integrated with non-patient-specific

information

Bioinformatics vs. Clinical

• Bioinformatics - The study of how information is represented and transmitted in biological systems, starting at the molecular level.

• Clinical informatics deals with the management of information related to the delivery of health care

• Bioinformatics focuses on the management of information related to the underlying basic biological sciences.

NIH maintains a database and tools of macromolecular 3D structures for visualization and comparative analysis

MMDB - Molecular Modeling Database - contains experimentally determined biopolymer structures obtained from the Protein Data Bank

National Library of Medicine Medline

Medical Informatics Standards

• Medical Information Bus - IEEE 1073– Standard for connecting up to 255 medical devices– Not all devices compatible– Decreases errors in data capture

• HL-7 Health Level 7– Domain: clinical and administrative data. – Mission: "provide standards for the exchange, management and

integration of data that support clinical patient care and the management, delivery and evaluation of healthcare services. Specifically, to create flexible, cost effective approaches, standards, guidelines, methodologies, and related services for interoperability between healthcare information systems."

• DICOM - Digital Imaging and Communications in Medicine

A protocol for the exchange of health care information

1 Physical2 Data Link3 Network4 Transport5 Session6 Presentation7 Application

HL7

Medical Information Bus IEEE 1073

• Standard for medical device communication • A family of standards for providing

interconnection and interoperability of medical devices and computerized healthcare information systems.

• Medical devices include a broad range of clinical monitoring, diagnostic, therapeutic equipment

• Computerized healthcare information systems include broad range of clinical data management systems, patient care systems and hospital information systems

THE DICOM STANDARD

• applicable to a networked environment.• applicable to an off-line media

environment. • specifies how devices claiming

conformance to the Standard react to commands and data being exchanged.

• specifies levels of conformance

DICOM Application Domain

MAGN

ETOM

Information Management System

Storage, Query/Retrieve, Storage, Query/Retrieve, Study ComponentStudy Component

Query/Retrieve, Query/Retrieve, Patient & Study ManagementPatient & Study Management

Query/RetrieveQuery/RetrieveResults ManagementResults Management

Print ManagementPrint Management

Media ExchangeMedia Exchange

LiteBox

Standards for Vocabulary• International Classification of Diseases, 9th Edition,

with Clinical Modifications (ICD9-CM)

• Diagnosis-Related Groups (DRGs)

• Medical Subject Headings (MeSH)

• Unified Medical language System (UMLS)

• Systematized Nomenclature of Medicine (SNOMED)

• Read Codes

• Knowledge-Based Vocabularies

ICD9- CM Example003 Other Salmonella Infections

003.0 Salmonella Gastroenteritis003.1 Salmonella Septicemia003.2 Localized Salmonella Infections003.20 Localized Salmonella Infection, Unspecified003.21 Salmonella Meningitis003.22 Salmonella Pneumonia003.23 Salmonella Arthritis003.24 Salmonella Osteomyelitis003.29 Other Localized Salmonella Infection003.8 Other specified salmonella infections003.9 Salmonella infection, unspecified

DRG Example75 - Respiratory disease with major chest operating room procedure, no major complication or

comorbidity

76 - Respiratory disease with major chest operating room procedure, minor complication or comorbidity

77 - Respiratory disease with other respiratory system operating procedure, no complication or comorbidity

79 - Respiratory infection with minor complication, age greater than 17

80 - Respiratory infection with no minor complication, age greater than 17

89 - Simple Pneumonia with minor complication, age greater than 17

90 - Simple Pneumonia with no minor complication, age greater than 17

475- Respiratory disease with ventilator support

538 - Respiratory disease with major chest operating room procedure and major complication or comorbidity

MeSH ExampleRespiratory Tract Diseases

Lung DiseasesPneumoniaBronchopneumoniaPneumonia, AspirationPneumonia, LipidPneumonia, LobarPneumonia, MycoplasmaPneumonia, Pneumocystis CariniiPneumonia, RickettsialPneumonia, StaphylococcalPneumonia, ViralLung Diseases, FungalPneumonia, Pneumocystis Carinii

SNOMED ExampleD2-50000 SECTIONS 2-5-6 DISEASES OF THE LUNG

D2-50100 2-501 NON-INFECTIOUS PNEUMONIASD2-50100 Bronchopneumonia, NOS (T-26000) (M-40000)D2-50100 Lobular pneumonia (T-28040) (M-40000)D2-50100 Segmental pneumonia (T-280D0) (M-40000)D2-50100 Bronchial pneumonia (T-280D0) (M-40000)D2-50104 Peribronchial pneumonia (T-26090) (M-40000)D2-50110 Hemorrhagic bronchopneumonia (T-26000) (M-40790)D2-50120 Terminal bronchopneumonia (T-26000) (M-40000)D2-50130 Pleurobronchopneumonia (T-26000) (M-40000)D2-50130 Pleuropneumonia (T-26000) (M-40000)D2-50140 Pneumonia, NOS (T-28000) (M-40000)D2-50140 Pneumonitis, NOS (T-28000) (M-40000)D2-50142 Catarrhal pneumonia (T-28000) (M-40000)D2-50150 Unresolved pneumonia (T-28000) (M-40000)D2-50152 Unresolved lobar pneumonia (T-28770) (M-40000)D2-50160 Granulomatous pneumonia, NOS (T-28000) (M-44000)D2-50170 Airsacculitis, NOS (T-28850) (M-40000)

Temporal Reasoning and Planning in Medicine

• Almost all medical data are time stamped or time oriented (e.g., patient measurements, therapy interventions)

• It is virtually impossible to plan therapy, apply the therapy plan, monitor its execution, and assess the quality of the application or its results without the concept of time

Time in Natural Language

From—

“Mr. Jones was alive after Dr. Smith operated on him”

Does it follow that—

“Dr. Smith operated on Mr. Jones before Mr. Jones was alive?”

Is Before the inverse of After?

Understanding a Narrative• List all, find at least one, or prove the impossibility of

a legal scenario for the following statements:– John had a headache after the treatment– While receiving treatment, John read a paper– before the headache, John experienced a visual aura

• One legitimate scenario (among many) is:– “John read the paper from the very beginning of the

treatment until some point before its end; after reading the paper, he experienced a visual aura that started during treatment and ended after it; then he had a headache.”

Paper

Aura

Treatment Headache

Monitoring

Determine if an oncology patient’s record indicates a second episode that has been lasting for more than 3 weeks, of Grade II bone-marrow toxicity (as derived from the results of several different types of blood tests), due to a specific chemotherapy drug.

Planning and Execution

If the patient develops sever anemia for more than 2 weeks, reduce the chemotherapy dose by 25% for the next 3 weeks and in parallel monitor the hemoglobin level every day.

Display and Exploration of Time-Oriented Data

Temporal Abstraction

• Many clinical tasks require a great deal of [time-oriented] patient data of multiple types to be measured and captured for interpretation, often using electronic media.

• This is particularly true in the management of patients with chronic conditions. • Diagnostic or therapeutic decisions depend on context sensitive interpretation of

these data. • Most stored data include a time stamp at which a particular datum is valid. • Temporal trends and patterns in clinical data add significant insights to static analysis. • Thus it is desirable automatically to create abstractions (short, informative, and

context-sensitive interpretations*) of time-oriented clinical data, and to be able to answer queries about these abstractions.

• The provision of this capability would benefit both a physician and a decision support tool (e.g., for patient management, quality assessment and clinical research).

• To be of optimum use, a summary should not only use time points such as dates when data were collected; it should also be capable of aggregating significant features over intervals of time.

Temporal Abstraction• Clinical tasks require time-oriented patient data of multiple types to

be measured and captured for interpretation. – Particularly true in the management of patients with chronic conditions.

• Diagnostic or therapeutic decisions depend on context sensitive interpretation of these data.

• Most stored data include a time stamp at which a particular datum is valid.

• Temporal trends and patterns in clinical data add significant insights to static analysis.

• Desirable automatically create abstractions (short, informative, and context-sensitive interpretations*) of time-oriented clinical data, and to be able to answer queries about these abstractions.

• The provision of this capability would benefit both a physician and a decision support tool (e.g., for patient management, quality assessment and clinical research).

• Of optimum use, a summary should not only use time points such as dates when data were collected; it should also be capable of aggregating significant features over intervals of time.

Three Basic Temporal Abstraction

• A model of three basic temporal-abstraction mechanisms: – Point temporal abstraction - a mechanism

for abstracting the values of several parameters into a value of another parameter;

– Temporal inference, a mechanism for inferring sound logical conclusions over a single interval or two meeting intervals; and

– Temporal interpolation, a mechanism for bridging non-meeting temporal intervals.

A Temporal-Reasoning Task:

Temporal Abstraction• Input: time-stamped clinical data and relevant events• Output: interval-based abstractions• Identifies past and present trends and statesSupports decisions based on temporal patterns “modify therapy if the patient has a second episode of Grade II bone-marrow toxicity lasting more than 3 weeks• Focuses on interpretation, rather than on forecasting

Temporal Abstraction:A Bone-Marrow Transplantation

Example.

•

0 40020010050

•

1000

2000

( )

100K150K( )

•••

• • • •

•• •

•••

Granu-locytecounts

• • •

•

Time (days)

Plateletcounts

PAZ protocol

M[0] M[1]M[2]M[3] M[1] M[0]

BMT

Expected CGVHD

Uses of Temporal Abstractions

In Medical Domains• Planning therapy and monitoring patients over time

• Creating high-level summaries of time-oriented patient records

• Supporting explanation in medical decision-support systems

• Representing the intentions of therapy guidelines

• Visualization and exploration of time-oriented medical data

Temporal Reasoning Versus Temporal Maintenance

• Temporal reasoning supports inference tasks involving time-oriented data; often connected with artificial-intelligence methods

• Temporal data maintenance deals with storage and retrieval of data that has multiple temporal dimensions; often connected with database systems

• Both require temporal data modeling

Clinicaldecision-supportapplication

TM TR DB

Medical Image Processing

• Input: X-Ray, CT-scan, MRI, PET, etc.• Tasks:

– Correction of multiple artifacts– Registration:Superimposition to enhance

visualization– Segmentation: Decomposition into

semantically meaningful regions

Conclusion• Multidisciplinary research, development, and

application– inspired by and benefits underlying core

scientific/engineering areas

• Medical Decision support systems: – Tasks: Diagnosis, therapy– Mode: Human initiated, data driven, closed loop– Interaction style: Prescriptive, critiquing

• Multiple diagnostic/therapeutic methodologies