F3 Aller Disease Surveillance APIII 2006

Disease Surveillance – Applying Pathology Informatics to Public Health: 

Disease Surveillance – Applying Pathology Informatics to Public Health Raymond D. Aller, M.D. Contributing Editor, CAP Today and Director, Bioterrorism Preparedness and Response Section Acute Communicable Disease Control Program Los Angeles County Department of Public Health 18 August 2006

Pathology as a clinical informatics pioneer: 

Pathology as a clinical informatics pioneer First information systems used in routine patient care First electronic longitudinal patient records First nationwide clinical records Comfort and familiarity with building and maintaining large analytical databases

.. and as pioneer of technologies and standards for PH disease surveillance: 

.. and as pioneer of technologies and standards for PH disease surveillance SNOMED – 1975 (SNOP 1965) In 1970’s first automatic transfer of coded data to a PH system (tumor registry) HL7 – from glimmer in 1979 to publication in 1987 LOINC – cooperative development with extensive pathology/lab medicine involvement - 1994.

Clinical Informatics: 

Clinical Informatics Not about computers or information technology Is about how we acquire, analyze, and apply clinical information We are informaticists (Informaticians work with dead information)

Keys to success : 

Keys to success Use of data produced as a byproduct of clinical care Automatic transfer of that data from clinical care setting, to public health database Translation functions to standardize the incoming vocabularies and codes Mechanisms to detect outbreaks designed for both naturally occurring (epidemics) and malicious (bioterrorism) Informatics is today a key tool in defense against natural and man-made outbreaks.

Objectives: 

Objectives Describe at least 3 types of data from clinical medicine being used to detect disease outbreaks Discuss 4 or more non-technical barriers delaying connection of clinical data sources to the public health database List at least three types of unique contributions, critical to PH disease surveillance, that originated in pathology Eliminate excuses as to why his/her laboratory isn’t pursuing connection with their local PH disease surveillance system.

Types of data from the lab to public health: 

Types of data from the lab to public health First implemented: tumor diagnoses from surgical path to the tumor registry. An entire subject in itself, and I won’t attempt to cover today In recent years, I’ve focused on a few other areas, from clinical lab to PH: Transfer of reportable (mostly infectious) disease information (for case management) Antimicrobial susceptibility (de-identified) All lab orders (de-identified)

Knowing that there is a problem: 

Knowing that there is a problem Natural (epidemic) vs. malicious (bioterrorism) Most agents (biological, nuclear, incendiary, chemical, explosive, radiologic - b-NICER) are overt Biological agents - typically covert A major portion of preparation for bioterrorism is about detection.

Detection – two sides of the coin: 

Detection – two sides of the coin Detecting the organism (or toxin) itself Example – the post office Biohazard Detection System A point of care laboratory assay! Detecting early signs of disease Disease reporting by clinicians, labs Symptoms and behavior Animal diseases

Our most important data sources: 

Our most important data sources The astute clinician and the telephone! Message: If you see something that looks wrong, let public health know.

The Scope of PH Informatics (1): 

The Scope of PH Informatics (1) Vital records: birth and death registries Public Health laboratory: biological chemical/ environmental (potable and recreational water, milk, etc.) newborn screening/ management for genetic and infectious disorders Management of chronic conditions (disabilities, newborn abnormalities, etc.)

PH Informatics (2): 

PH Informatics (2) Population health and perception measurement, monitoring - data acquired from: Population-based clinical care systems (e.g., Indian Health Service) Telephone and other population surveys Environmental health Reportable disease: National electronic disease surveillance system Management of communicable disease: Acute communicable disease, Sexually transmitted diseases, HIV epidemiology, Tuberculosis, Immunization registry

PH Informatics (3): 

PH Informatics (3) Lead poisoning case management Electronic laboratory reporting: Syndromic surveillance Alerting and partner communications Strategic national stockpile inventory and warehouse management Countermeasure response administration (mass prophylaxis) Education and training of medical professionals and the public Tracking/managing that training (learning management) Emergency management information systems: Coordination with other first responders, such as fire, police

Public Health Informatics: 

Public Health Informatics Today’s talk Encourage reporting of disease by clinicians Electronic laboratory reporting Syndromic surveillance Other areas of interest to pathology, will not be discussed today Lab information systems – PH, veterinary Many others

How do we detect disease?: 

How do we detect disease? Reportable disease Clinician – paper (underreporting) Infection control practitioners – web entry Electronic reporting from labs Automatic reporting from clinical systems Symptoms Emergency departments (symptoms, volume - Reddinet) 911 calls, ambulance dispatch, nurse call lines Byproducts of care Ambulatory diagnoses, lab orders (Biosense) Behavior Purchase of over the counter medications School or work absenteeism

Slide17: 

80+ reportable diseases (mostly infectious) must (by law) be reported to Public Health

Reportable diseases: 

Reportable diseases In the past, handwritten paper form via mail or fax Reporting is slow, and rates are often abysmal (5% or less) Improved by use of a web-based system However, findings diagnostic of many of these diseases (e.g., hepatitis B) are stored in hospital, laboratory and/or clinic computers Automatic electronic transfer of these cases to public health (e.g., Electronic Laboratory Reporting) has been shown to greatly improve the reporting rate.

Web-facilitated reporting: 

Web-facilitated reporting Web-visual Confidential Morbidity Report (Web-vCMR) Now called the Community Reporting Module Anywhere with internet access Security: digital certificate on your workstation, plus username/password

From manual to automatic: 

From manual to automatic When a reportable diagnosis is recorded in an EMR/EHR Automatically report - or – Pop-up question to clinician – OK to report? Recognizing patterns in the EMR

Electronic Lab Reporting (ELR): 

Electronic Lab Reporting (ELR) Software on the laboratory information system automatically selects from all laboratory results, those which are reportable to public health In other cases, ALL results transferred to a filtering system, that selects reportables These may be enhanced by other findings of public health importance Antimicrobial susceptibility testing Syndromic indications ELR commonly doubles the reporting rate

Key characteristics of ELR: 

Key characteristics of ELR Unbroken (no manual steps) electronic linkage between the database of the laboratory’s information system and the database of the public health disease reporting system. HL7 format Flat file format For a very small lab, web entry Types of data that may be sent: immunology, microbiology, tumor diagnoses Systems the data is sent to: communicable disease reporting systems, syndromic surveillance systems, tumor registries

Why Electronic Lab Reporting?: 

Why Electronic Lab Reporting? Community/clinician reporting rates abysmal (often less than 5%) Laboratories typically have much better administrative organization Positive laboratory findings more definitive than a clinical impression Even without ELR, labs often achieve 50% or better reporting rates. ELR permits close to 100% reporting rates

ELR Benefits to the Lab: 

ELR Benefits to the Lab Results to PH as soon as available – compliance with <24-hour reporting law Every case that meets criteria is sent automatically Some states now mandate electronic reporting HIPAA disclosure records complete Lab staff time savings.

ELR Implementation: 

ELR Implementation Format – HL7 v2.3.z, v2.3.1, now v2.5 target Security – sFTP, VPN Important to capture patient address and phone (for PH followup) – may require additional data from ADT system, etc. Codes converted from local to standard Result names standardized – LOINC Result values standardized – SNOMED Specimen source – HL7 table, SNOMED Appropriate tests/results to send – “Dwyer/Sable tables”

LOINC: 

LOINC Logical Observation Identifier Names and Codes In ELR, used for test names Published beginning in 1994 Freely available – copyright but royalty free Now mandated by Federal government for all governmental healthcare programs (VA, DOD, IHS). The standard for reporting of public health data. Future standards for physicians office systems www.loinc.org

SNOMED: 

SNOMED Systematized Nomenclature of Medicine In ELR, used for organism names Under development since the late 60’s Encompasses all areas of clinical medicine Mandated for all medical records in the UK Also used by many organizations (Kaiser) and countries Licensed for use throughout the United States. www.snomed.org

Data Transformation : 

Data Transformation Hospital Systems Public Health Information System Public Health Information System Public Health Information System Web page Data entry Format converter De-identification (of “non-reportable” findings) Collation Text handling Routing PH compliant HL7 messages PH compliant HL7 messages PH compliant HL7 messages These functions may be performed in the Public Health Department, Data Producing Facility or an Intermediary Filter

To Do a LOINCing: 

To Do a LOINCing Download lab’s test dictionary Test name, test code, units, method, specimen type Done by Public Health staff or outside services (probably not the lab)

Getting from internal test codes to standard LOINC codes: 

Getting from internal test codes to standard LOINC codes Today – Laborious, mostly manual process to match description, units, method, etc. to appropriate code Some vendors (e.g. 3M) have a suite of automated tools Tomorrow – Vendors (instruments, kits, reference labs) supply the appropriate codes Statistical assessment of a high-volume HL7 data stream (e.g., mean and SD of results)

What Data is Transmitted?: 

What Data is Transmitted? Now: Reportable diseases Supporting lab findings – liver enzymes and bilirubin on cases of positive hepatitis serology Soon: Lab orders that may help define syndromes Future: Antimicrobial susceptibility testing – on ALL organisms

Syndromic Surveillance using lab data: 

Syndromic Surveillance using lab data Real-time public health surveillance using data that is routinely collected for other purposes Not to identify individuals, but to detect atypical patterns of symptoms, orders, findings Therefore, data can be de-identified Real time transmission, analysis, and alerts

Lab Order Defined Syndromes: 

Lab Order Defined Syndromes Blood cultures: fever Stool cultures: GI Sputum cultures: respiratory CSF cell counts: meningeal (e.g., West Nile) This is a nascent area – may be better to get ALL orders, as we learn what constitutes a useful pattern The BioSense – LabCorp experience

Antimicrobial Susceptibility Testing: 

Antimicrobial Susceptibility Testing Antimicrobial resistance is an increasing problem in all communities Traditional – collect antibiograms from hospitals Late 90’s – commercial services such as TSN collected susceptibility data from labs Alternative – collect raw susceptibility results from labs, perform calculations at public health.

Cost of disease reporting from labs: 

Cost of disease reporting from labs Cost of manual reporting – 0.50 to $5 per case Interface Initial – license fees, implementation, personnel time Ongoing – Direct link - maintenance fees Intermediary - monthly use fee. LA County has arranged to reimburse hospitals for costs (see next slide)

For once, a funded mandate: 

For once, a funded mandate LA County has HRSA/NHBPP grant funding to reimburse hospitals for their out-of-pocket expenses (e.g., vendor fees) in establishing data feeds We have already purchased lab interface modules for several hospitals We also provide data conversion (LOINC)

ELR – current status: 

ELR – current status Nationwide, many states (and a few local jurisdictions) are now receiving ELR Most commonly connected – national labs (LabCorp, Quest)

Mandatory use of electronic lab reporting: 

Mandatory use of electronic lab reporting Still voluntary in most jurisdictions Some now require by law/regulation New York State Or for certain tests – Blood lead - California

Potholes in the information superhighway: 

Potholes in the information superhighway When you are trying to travel at 186,000 mps, a pothole is a big deal!

The official salute of the governmental I/T insecurity officer: 

The official salute of the governmental I/T insecurity officer

The general I/T industry is beginning to recognize this: 

The general I/T industry is beginning to recognize this

What is the biggest threat to our informatics preparedness for biosecurity?: 

What is the biggest threat to our informatics preparedness for biosecurity? The information insecurity officer Fingerpointing “Finding a way to get to NO” The security Luddite Forces expensive, error prone and unsafe workarounds

5 I/T syndromes hazardous to the public’s health : 

5 I/T syndromes hazardous to the public’s health The security Luddite The perfect: enemy of the good Governmental bureaucracy Counterproductive hiring policies Inhibitory purchasing procedures Impossible contracting procedures The vaporware merchant Judgments based on Insufficient data

Back to biological disease: 

Back to biological disease Gathering diverse data Looking for patterns Letting key people know about it

Current Data Sources : 

Current Data Sources Disease reports – diagnosed or suspected Telephone reports Paper reports VCMR Labs Syndromic Sixteen hospital emergency departments Over-the-counter pharmacy sales Reddinet© Biosense Coroners service

Conceptual Architectural Overview: 

Conceptual Architectural Overview Public Health Information System Web page Data entry Pharmacy Billing Emergency Department Chief Complaint Clinical Documentation LIS Reference Labs Veterinary / Zoo Labs HIS & other systems Collation, transformation and routing processes PH Compliant HL7 Messages All are Data Producing Facilities (DPFs) Web page Data entry

Assumptions – Bioterrorism detection: 

Assumptions – Bioterrorism detection

Syndromic Surveillance: 

Syndromic Surveillance Presenting complaints are automatically categorized into a particular syndrome Syndrome counts are tracked over time Statistical increase in syndrome counts triggers a signal To detect major trends from baseline patterns, not individual cases

Key steps: 

Key steps Getting the data Analyzing the data Disseminating the findings

Getting ED chief complaint data: 

Getting ED chief complaint data Data use agreement – hospital – PH May take months to get signoff Automatic electronic connection from hospital admitting/hub system to PH information system 12 hours of technical work, but months to get there. sFTP, VPN, or even (+/-) encrypted eMail De-identified – send only age, sex, date/time, chief complaint, zip code, disposition (+diagnosis if rapidly available)

Where do we pull data from?: 

Where do we pull data from? Now - ADT transactions (HL7 A04/8) to extract free text chief complaint Future – Emergency department systems structured observations, impressions and orders from the EMR/EHR

How would data flow from clinics?: 

How would data flow from clinics? Completely automatic – no ongoing manual intervention One possible source – walk-in and same-day patients’ reason for visit A future possible source – electronic medical records systems

Line Listing: 

Line Listing

Key steps: 

Key steps Getting the data Analyzing the data Disseminating the findings

Analyzing the data: Syndrome Classification: 

Analyzing the data: Syndrome Classification Standard Gastrointestinal Respiratory Rash Neurological Special categories Influenza like illness Heat (Trauma) Others as needed

Sample: Syndrome Trend and CUSUM Analysis : 

Sample: Syndrome Trend and CUSUM Analysis

Investigation and Daily Report I: 

Investigation and Daily Report I Investigation Review counts and proportions one day increase or continued increase? coincidence with high profile public event? Review line lists similar chief complaints within and across the hospital, age and gender clustering? Unusually severe and high volume? coincident with traditional disease reports? Review complementary systems Call ED, review charts, follow-up patients

Sample: SaTScan Syndrome Cluster Map: 

Sample: SaTScan Syndrome Cluster Map

Slide60: 

Over-the-Counter Cough & Cold Medication Sales Multiple Signals For 11/28/03

Key steps: 

Key steps Getting the data Analyzing the data Disseminating the findings

Reporting the data: 

Reporting the data Generate summary report, hospitals coded 7 days per week Send to key public health staff, to the LAC Terrorism Early Warning group, as well as to each participating hospital.

Example of daily syndromic report: 

Example of daily syndromic report

What have we found in LA County?: 

What have we found in LA County? Onset of 03-04 flu season Diarrhea outbreak Feb 21 ‘04 – rotavirus Cluster of respiratory distress Mar 3, ‘04 Cluster of rash – Oct 15, 04 “rule out smallpox” – varicella contacts Summer surge in viral meningitis – ‘03 Retrospective analysis for West Nile meningitis

Syndromic surveillance: others findings: 

Syndromic surveillance: others findings Early detection of flu in NYC, Utah, others Case finding for measles, varicella outbreaks Cryptosporidium, Milwaukee, 2001 (OTC) Diarrhea following blackout – NYC, Aug 2003 Asthma, respiratory distress, SD County, Nov. 2003/4 Heat-related illness // Cipro sales after anthrax Fireworks // Dog bites/rat bites // Overdoses West Nile virus spraying // Suicide attempts Carbon monoxide poisoning – Pennsylvania, 2003 Diarrheal outbreaks: norovirus, rotavirus – NYC, 2002

How well do we cover the population of LA County?: 

How well do we cover the population of LA County? Green = well covered Red = poorly covered

Improving the precision of syndromic categorization: 

Improving the precision of syndromic categorization Today – ADT transactions, extract free text chief complaint Tomorrow - structured observations and impressions from the ED record, and EMR/EHR

Public Health and RHIOs: 

Public Health and RHIOs Regional Health Information Organizations Focused on clinical data sharing, but serves many needs Often, PH may be the first to connect multiple hospitals in a community – albeit for a limited data set Public Health is very interested in several of the data types flowing in an established RHIO. One of the best examples is the Regenstrief/Indiana RHIO

Public Health Davies Award: 

Public Health Davies Award To recognize outstanding achievement in using informatics to improve the public’s health Pennsylvania disease reporting system S. Dakota Vital Records Utah – Immunization registry Indian Health Service North Carolina ED surveillance 2006 Awardees to be announced this fall

Key Points about surveillance: 

Key Points about surveillance Avoid manual work by labs and hospitals Rapid detection of nasty disease, tracking slower public health menaces Implementation requires expert and experienced technical support Not a panacea, but gives some reassurance. This is a process that can easily take months – or years HIPAA compliant Need to add more data sources

How can your lab contribute?: 

How can your lab contribute? Talk with your local and state public health agencies Are they ready to receive ELR and ED data? If not, encourage them! Monthly national ELR call, first Tuesday 10a PDT Monthly national ED/SS call, 4th Thursday, q other month Talk with your LIS vendor What software module(s) are available for this activity? We in LA County have had extensive discussions with Misys, Meditech, Cerner, McKesson, some discussions with others – all have useful tools Please let us know if we can be of assistance.

Resources: 

Resources CAP Today – articles, Newsbytes – www.cap.org www.cdc.gov/phin www.loinc.org www.snomed.org

Contact us: 

Contact us Raymond Aller, M.D. raller@ladhs.org 213-989-7208 Thank you !!