The Development and Implementation of a Shared National Communicable Disease Reporting System in Ireland 2000 - 2006
John Brazil, Suzanne Cotter, Gillian Cullen and Colm Grogan.
Health Protection Surveillance Centre, Ireland
1
Republic of Ireland
• Population
– 4,234,925 (2006 census)
• Location
– Western Europe, occupying fivesixths of the island of Ireland in the North Atlantic Ocean
•
Area
– 70,280 sq km / 27,135 sq mi
2
CIDR Programme Timetable
• • • • • • • • • HPSC Established Needs Analysis Project Setup Procurements Information Governance System Design System Development System Pilot System Implementation • • • • • • • • • 1998 2000 2001 2000 - 2003 2001 - ongoing 2002 2003 - 2004 2004 2005 - 2006
3
Infectious Disease Surveillance and Control
Patient Sick
Sees Doctor
Lab Test
Information and Advice
Public Health Action
4
Key Themes - 1
• MULTIDISCIPLINARY MULTIORGANISATIONAL PARTNERSHIP • DETAILED REQUIREMENTS GATHERING AND SYSTEM SPECIFICATION. • A FLEXIBLE AND SCALEABLE SYSTEM.
5
Key Themes - 2
• The need to provide APPROPRIATE ROLE-BASED INFORMATION FEEDBACK. • The Importance of SHARING EXPERIENCE with other jurisdictions and public health information system developments, including the Public Health Information Network.
6
CIDR Partners
Food Safety Promotion Board Department of Health and Children Food Safety Authority
Irish Government Virtual Private Network
CIDR Database
Health Protection Surveillance Centre
Health Protection Units (9)
Clinical Laboratory (~53) HSE Areas (4)
Local Health Office (32)
7
Challenges of a multi-agency initiative • Clear vision • Clear control • Communication
8
CIDR System Design
• Detailed Business Process Analysis • Informed by Public Health and Laboratory Domain Experts • Use of a series of prototypes • Business Process Analysis and System Design (and subsequently development) carried out in conjunction with Fujitsu
9
Detailed Business Process Analysis
Initial suspicion
Start PT01-Patient attends GP/ Clinician PT02-Follow course of treatment No End 2 GP system that sentinel GPs may be using
Other start points
Start
Core process is disease based (i.e. select disease then patient) - for initial notification and for Lab result entry. Are any variants not disease based? Are AMR organisms linked to a disease? PT03 Public suspicion Subprocess to be performed each time data is entered to CIDR, i.e.: CL03;RL07;MO01
Data entry 1. Data entry
Enter data to CIDR
Data only visible within reporting area until authorised (where required) or until event created
2. Database update
Including fields for: - Consent Y/N; - Outbreak related Y/N; - Outbreak number: # / not known
Proposed prototypes Variant: Anti-micobial
Yes DE03 Accept association with existing patient
Note re sentinal reporting Common diseases that are usually identified in General Practice. Sentinel is an early warning system. Sentinel reporting is in place for: Flu; Gastro; Mumps (with an everincreasing set of diseases and sentinel sites).
No MDCPositive test? MDDAlready notified on suspicion? MDE- Provide extra data/ confirmation?
DE02 Review possible associations Forward to NDSC (via Irish College of GPs) Appropriate data entry person (Labs or DPH) DE01 Enter data to CIDR Yes DEA Association exists?
DEB Accept patient match?
Variant: Sentinel reporting
MD01-Preliminary diagnosis Based on sentinel rules: e.g. Is this the first case this week? MDA-Test required? No
No
MD03-Follow up with patient
Yes
Yes
Record in GP system
Weekly export from GP system (IT supplier)
Anonymise data (IT supplier)
No
DE04 Create new patient
MDB-CIDR/ notifiable disease suspected?
No AMR related diseases are not currently notifiable MD04-Complete relevant form(s) For AMR question becomes: Does AMR-relevant organism exist? If so, is this the first test this quarter? Once yes to both questions (regardless of whether or not it proves to be susceptible) test results should be reported AMR relevant organism? CLB Positive test? First test for this patient this qtr? Yes
Yes
Options: 1. Paper form (mail or fax) 2. No form - telcon recorded on form by DPH 3. On-line 4. Integrate: Key to GP system and transfer to CIDR
Yes Designated GP practices that volunteer for Sentinal. NB: Practice ID required. GPs associated with practice (many-toone-or-more) MD02-Request appropriate test(s)
CIDR system
CD11 Search for patient match
CD12 Compile possible patient matches
Highest matches presented first Lesser matches listed below User allowed restricted detail: excludes access to disease details and history
CD13 Update system
Yes
Clinical laboratory
Four organisms subject to this regime: - Escherichia coli (Blood and CSF) - Streptococcus pneumoniae(Blood and CSF) - Staphylococcus aureus (Blood) - Enterococcus faecium/faecalis(Blood)
CL02- Inform GP/Clinician of results
Based on reporting areas (8 Health Boards) MD05Communicate with / transmit form(s) to medical officer Patient ID (created by originating hospital) is unique identifer. Further data (Name; Address) would require patient consent because diseases are not notifiable. Do we need to capture Specimen ID?
6. Lab authorisation (incl. fast-track authorisation)
Y/N CL03 Enter data to CIDR CLE Authorisation required? Potential for future: Hospital or community acquired? Open question as to how to populate
Based on NATIONAL file of patients: Name: Fullname; Initials; Surname; Surname variants DOB: Exact; Same month; Same year Gender: Exact Address: Full address; 1st line of address; County PPSN: Exact Health Board: Exact Disease: Exact Reconcile with SEARCH diagram (seperate page) Comprising: 1. Own (clinical) lab data 2. Related reference lab data
Follow-up
CL05 Compile extra data Yes
CLC Extra test required? Yes
Some enhanced data will not be received. Judgement required as to whether to continue to pursue extra data or not.
CL06 Tag as unavailable. (Removed from open list)
No
CLF Can enhanced data be rec'd
Participating labs (22) Expected to be National in due course
CLA Can CL do the test?
Yes
CL01- Conduct relevant tests
No
CLD Reference lab? Yes RL03 Conduct relevant tests
May be: - detected from LIMS - Where no LIMS, relies on manual checks In certain instances all blood culture isolets are sent to NDSC. (St James; Tallaght) Relevant data and deduplication distilled by NDSC
Standardised approach to susceptiblity testing CL04 not yet established. NCCLS being promoted but Yes Authorise data is at least a year away. System needs to cater for 2 different test methods and qualitative or quantitative results.
MO09 Investigate complaint
End
Reference laboratory
RL02 Inform originating Lab
No
RL04 Complete relevant forms and transmit No
If quantitative results entered, table used to RL06 assign related qualitative result. Enterquant. enter If no to CIDR qual. direct.No host lab auth. queue
MO10 Communicate with / transmit form(s) to medical officer
Yes
MOF Notifiable/ CIDR disease?
No
End
Under
MO14 Associate two events
Yes
No RL01 Conduct relevant tests RLA Positive test? Yes
RLB Ref lab on CIDR?
Yes
RL07 Enter data to CIDR
RLC Authorisation required?
Yes
RL05 Authorise data No
Start
MO11 Identify association error
MO12 Call up relevant event(s)
MOG Error type?
Over
MO13 Break association MO07 EHO follow up as required Yes MOE Data originates from Lab?
5. Follow-up action: data recording
Negatives only become relevant for: 1. Anti-microbial 2. TB and other specific diseases ... 3. Outbreaks
CIDR preprocessing
MOA Consent received? No
Yes
MO01 Enter data to CIDR
Should be reflected in CD06 'To do' list
MO02 Request consent
MO17 Delete (Logical)
From GP on suspicion From GP on confirmation From DPH after complaint investigation
Event creation 3. Event creation
Drilldown to some extra detail (not full records) - restricted to area: full notification only visible where patient is within user's authorised region. - and disease: patient details only accessible re current disease
MO08 Compile extra data
Yes
MOH Can enhanced data be rec'd No DPH to review list of 'unavailables' regularly for control purposes MO15 Tag as unavailable. (Removed from open list)
No
Completed by Lab for Anti-microbial resistance organisms May be completed by originating clinic in other cases (e.g. HIV/AIDS) ... together with related follow-up
MO04 Review possible associations Yes MOB Association exists? MOJ Wait? No MOD Accept event match? No MO05 Create new event MO06 Accept association with existing event MO16 Accept / overwrite proposed updates to status of events Yes
The allocation of tasks among individuals within the DPH (Medical Officer; Surveillance Scientist; Data Entry Clerk) will vary by department (based on size; local practice; remit of surveillance scientist etc.)
Will logical delete suffice where no consent received? Or should data be kept but anonymised? Possibile outcomes include: A. Denotification; B. Reclassification 1. of disease status(probable; confirmed; etc.) 2. to another disease Denotification will be the responsibility of the DPH (possibly after liaising with or on prompt from clinician). Denotification will apply especially for certain enhanced surveillance diseases. NB: denotification may result in specific suspicion being reclassified as general suspicion.
MO03 Run search (for each) on latest data Duplicate across quarters?
Choice up to each event creator as to whether: a) system defaults to YES or b) review of individual alerts required End No MOI Release prompt?
No
No
Duplicate within quarter?
No
CD03 Search for record match
CD04 Compile possible associations
Outbreaks
Outbreaks
CDB CIDR disease?
No
CDA Notifiable disease? Yes CDC Mandatory data set exists? CD01 Generate list of unauthorised records
Disease: Exact Date of onset: Exact; Same month; Same year Date of diagnosis: Exact; Same month; Same year Patient identifier PLUS Search criteria identifierdynamic link specimen may be used to based on disease frequency results and to exclude duplicates. Records will only be matched with other AMR records reported. Matching with other records while possible (possibly authoriser 1. What is the using DOB) will not be undertaken due to privacy geography based on : source issues. loacation or patient location? 2. Search within authoriser's geography only or nationally
New event rules differ by disease, mainly time-based, e.g. Gastro; TB
CD05 Update system
CD10 Compile proposed updates to status of events
Outbreaks
Yes Note re Outbreaks: Established outbreak criteria in place for: Salmonella; Flu; Other Gastro. Criteria being agreed for an increasing number of diseases. Including status of outbreak Post component B outbreak number must be entered to LIMS and be part of translation Enter data to CIDR CDD Consent confirmed?
Yes (On the record) Varies by disease (and disease category: CIDR; Notifiable)
Yes
Practice is only to consider first test for each patient. However reporting periods are typically quarters. Current process doesn't prevent double counting of cases across two or more reporting periods.(WHONET protocals for EARSS reporting specifically request that data be collected on on Yes first case per quarter basis).
(For newly updated events) System suggest status based on case definitions. Case definitions: Normal surveillance: will use European. Enhanced: 2 sets - National and European. NB: National = Possible;Probable;Confirmed (dependent on other status types, e.g. TB test status). European = Probable;Confirmed EXAMPLE NEEDED How do statuses relate to each other
Based on disease type, e.g. EHO would need to follow-up in cases of gastro-enteritis CDE Does EHO need to followup? No CDF Minimum data set exists? CD08 Generate prompt for EHO
Yes
On event On event
Yes
Medical Officer (nominee within Department of public health or CCA) Preliminary investigation Outbreak? Yes Confirm outbreak status Initiate action Follow-up as required
No Publicise to Outbreak control team
Based on authoriser. Determined with reference to : 1. organism (where AMR); 2. originating clinic (where HIV); 3. DPH for patient in all other cases (derived from patient address or location of GP/Clinician) Does one person in each DPH do all event creation or is authorisation further specialised?
4. Alerts to follow-up
Only applicable for selected diseases NB: Putting specific diseases on enhanced surveillance will only be done at a national level within the new system.
CDG Enhanced data set exists?
Yes or N/A
Note re test follow ups. What happens if an event has been established based on clinician suspicion but no test results has been received in confirmation? Has test been done? Was test negative? Should case staus be changed? Should event be denotified? Options: 1. DPH treat silence as negative. DPH judgement then applied as to whether to leave case as probable, or to denotify. "Silence" meaning: - Suspected case No - Test overdue - No outstanding queues for Lab (which Lab?) Drawbacks: can't distinguish 'negative' from 'not done'; lax control 2. Labs enter all negative (CIDR) tests Drawbacks: onerous data entry burden on Labs; data sharing issues/justification not clear 3. Labs enter lesser data for negative CIDR tests, e.g. Name; Clinician; Test 4. On silence DPH prompt lab (which lab?) to enter negative results 5. On silence DPH pursue with clinician (not practical) Option 1 preferred Open questions: - Which option to pursue? - Scope? All surveillance or only enhanced surveillance ... normal => silence = negative ... enhanced: => pursue with clinician to identify the lab
Reject : Record reasons
CD09 Update CD06 to reflect new data received
Relevant factors: Disease Relationship between patients Events in time period: Day; Week to date; Week; Month; Year (NB: Seasonality is a factor) Location (down to CC level or even Electoral Division) Note: 20 years history in place (pre July 2000 data in aggregate form) Review events created to date
No No CD06 Generate list of extra data req'd CD07 Transmit to inputer
Alert DPH of potential outbreak
Assign outbreak number
Yes
Aggregate: related to outbreak (rather than new data required on individual events)
Updated daily with latest information receved. e.g. Lab result initially missing, subsequently received. Calendar function? Time driven element required to prompt recording of follow-up actions:(based on disease type and status) Test frequency, e.g. TB every X months dependent on initial test result (or less if and when tests turn negative) Test lead time, e.g. TB X months after initial test dependent on nature and result of initial test (see TB form for examples of logic) List should be prioritised. 1. Based on urgency of public health action. Disease type; number of events (outbreak?) 2. Reporting deadlines: Minimum gaps before enhanced gaps. Minimum gaps are more serious and time sensitive (should be remedied for weekly reports). Enhanced gaps should be remedied for quarterly + reports.
CIDR system
Observed > expected
No
End
Associate events with outbreak
Generate list of extra data req'd
Alert post Data Entry
1. Is data held at the event level? How is it compiled - based on aggregating associated records? 2. If so, are there circumstances in which records may conflict. Once associated in an event any conflicts would have to be addressed. e.g. 2 test results with different dates of onset. - Which is correct? - Which is used in producing related statistical analyses?
10
Key Concepts in CIDR
• Multiple REPORTs relate to a given event in a given individual • An EVENT is the same episode of disease in an individual i.e. within a given timeframe CIDR is a PERSON-based system An OUTBREAK relates to multiple instances of disease in individuals within a given location/timeframe
• •
11
System Dynamics
Steps Notification DataEntry Patient Match Event Creation Outbreak Admin. Follow-up
G P Prototype 3 Labs DataEntry
Prototype 5
DPH
DataEntry Prototype 1
Patient Match
Event Creation Prototype 2
Outbreak Admin.
Follow-up Prototype 4
12
Dataflows in CIDR
Family Physician
C I D R
Clinical Laboratory
Manual Entry via web forms or file upload Manual Entry via web forms or file upload
Public Health
Reports Manual Entry via web forms
Reference Laboratory
Reports Reports
CIDR
Reports Reports Reports Reports
FSAI
FSPB
DoHC
HPSC EU WHO
13
CIDR Software Architecture
• n – tier Architecture • Service Oriented Architecture • . Net development • SQL 2000 database • Web forms for manual data entry • XML-based parser for translation / transformation of data exported from laboratory information systems • Business Objects reporting software for business intelligence
14
Information Governance
• Legislation
– Infectious Disease Regulations – Data Protection Act
• Nationally agreed Business Rules specifying data access rights and levels of service • Information Security Management externally accredited to ISO17799
15
National Implementation of CIDR (July 2006)
North West*
North East West
Salmonella Reference Laboratory
Midland Mid West
East
Meningococcal Reference Laboratory MRSA Reference Laboratory Health Protection Surveillance Centre Virus Reference Laboratory
South East South*
*Public Health only, labs to follow Implemented Implementation in Discussion Pending Public Health Reorganisation
16
CIDR Rollout by Population Coverage
CIDR Rollout
100 90 Population Coverage (%) 80 70 60 50 40 30 20 10 0 May-04 Apr-05 Jul-05 Oct-05 Feb-06 Jun-06 ??? ???
17
Result
• Before CIDR there were numerous diseasespecific databases held at local, regional and national levels. • CIDR has consolidated this data, including both routine and enhanced surveillance information, from clinical, epidemiological, and laboratory sources, into a single database.
18
Ongoing and Upcoming Challenges
• Completing national implementation • Including AMR, TB, Sentinel Surveillance and STIs • Managing user expectations • Developing new or enhanced functionality such as contact tracing • Extension to primary care
19
Built To Last?
Poulnabrone Dolmen, The Burren, Co. Clare, Ireland Dating from the Neolithic period of between 3800 BC to 3200 BC.
20