Business Context > Business Rules

Business Rules

PCR Source Models

Contributing Sources are organizations that contribute information to the PCR and other EHR assets. Differences in an organization’s legal structure and their approach to managing medical records at the patient identifier level impact how the patients are contributed & consumed to and from the PCR.

These concepts are important to understanding the different PCR source models:

Concept Description
Legal Entity The entity legally responsible for the Contributing Source and the information contributed. May consist of one or more facilities. The legal entity typically equates to the Health Information Custodian (HIC) per contribution of data to the PCR.
Facility A facility that provides care, maintains medical records, and contributes to the EHR.
Definitional Source ID Primary ID value that uniquely identifies a source patient [record] in the PCR. In business practice, this translates to one of; CORPID (for Corporate ID Model sources), MRN (for Standard Model sources), or in the single case of the authoritative MoHLTC RPDB feed, Ontario HCN.
Informational Source ID Secondary ID value that uniquely identifies a patient under a given definitional source record. In business practice, this translates to either HCN or MRN.
MRN An identifier that uniquely identifies a patient and their records within a facility. The MRN is typically visible to health care providers and the users of local information systems. From a business perspective, this translates to a Definitional Source ID (for Standard Model sources) and/or an Informational Source ID (for Standard & Corporate ID Model sources)
CORPID A Corporate Identifier that uniquely identifies a patient and their records within the Legal Entity. The same patient that visits multiple facilities in the organization and is assigned different corresponding MRNs is ultimately managed under a single Corporate Identifier (and typically demographics) within in the Legal Entity. The CORPID is typically not visible to health care providers and the users of local information systems.

The following diagram illustrates the PCR source models:

PCRsourcemodels-png

Standard Model

Legal entities that use single pool of Medical Records Numbers (MRNs) to identify all patients and their records at all points of care (often a single facility) use the Standard model.

  • MRN is contributed to PCR as the primary patient identifier for the organization (referred to as the Definitional Source identifier in the PCR)
  • MRN is also contributed as a secondary identifier for consistency (referred to as an Informational Source identifier in the PCR)

Corporate ID Model

Legal entities which use different pools of Medical Record Numbers (MRNs) to identify patients and their records within different facilities use the Corporate ID model.

  • CORPID is contributed to PCR as the primary patient identifier for the organization (referred to as the Definitional Source identifier in the PCR)
  • MRN(s) are contributed as a secondary identifier (referred to as an Informational Source identifier in the PCR)

Matching Algorithms

The PCR uses a deterministic and probabilistic algorithm to link a patient’s distinct hospital records under their given PCR Enterprise Client Identifier (ECID), and to match candidates for PCR queries.

Overview

The PCR linking and matching algorithm:

  1. Optimizes demographic and identifier data for statistical comparison (e.g. data normalization)
  2. Finds all potential matches using threshold analysis (e.g. uses real data to define weights associated with the values)
  3. Scores via deterministic and probabilistic algorithm (refer to the sub-sections below)

Deterministic Matching

Deterministic scoring takes into the account the presence or absence of a value to derive a matching score against the PCR. Holistically, it represents an ‘all or nothing’ static score based on the value in question matching the value in PCR exactly, or not. PCR performs ‘exact match’ deterministic scoring (for linking records on contribution and matching candidates on queries) on the following attributes, which are expected to be found across different source records for the same patient;

  1. Health Card Number (HCN)
  2. Gender (Administrative Sex)

Probabilistic Scoring

PCR performs ‘probabilistic scoring (often referred to as ‘fuzzy’ matching) for linking records on contribution, and matching candidates on queries. The following string value attributes, which are expected to be found across different source records for different patients are used for probabilistic scoring;

  1. Name (Legal, Alias)
  2. Date-of-Birth
  3. Address (Home, Mailing, Temporary)
  4. Phone (Home, Work, Temporary, Other, Mobile)

Probabilistic scoring takes into consideration the frequency of the occurrence of a data value within a particular distribution when matching against the PCR. Holistically, it represents a variable match score based on the value in question matching the value in PCR to a measured degree, as per the following example criteria and thresholds:

Concepts Description
Standardization Normalizes the data (e.g. dashes, area codes in phone numbers, and special characters in names are stripped for comparative purposes)
Frequency Scores based on the frequency of an attribute value and not just whether there is a match (e.g. “John” would score less than “Heinrich”). When a given value is found to be too common in the PCR (e.g. > 1000 occurrences), additional logic “tightens” the search (e.g. Common First + Common Last + DOB requires exact vs. edit distance match on the DOB )
Phonetics Indexes words by sound/pronunciation to identify those that are similar phonetically (e.g. Purdey, Purdie, Purdy)
Equivalence (Nicknames) Evaluates based on equivalency (e.g. Liz, Libby, Beth = Elizabeth)
Field Transposition Accounts for errors where name (First/Middle/Last) or address (StLine1/Stline2) sub-elements are transposed into the incorrect fields (e.g. Last + First vs. First + Last)
Edit Distance Accounts for typos due to character transposition at the attribute value level (e.g. “Micheal”)
Anonymous Values Prevents scoring on values deemed to be “anonymous” or “dummy” data (e.g. Name = John Doe, Baby Boy / Girl, DOB = 0001-01-01, HCN = 2222222222). When the PCR minimum search criteria is met (e.g. First Name + Last Name + DOB), but one or more anonymous tokens are used, PCR will return an error indicating insufficient query parameters provided to perform the search.
Historical Values Performs queries against now inactive source data (e.g. Name, Address, Phone) to match on potential candidates based on historical data (e.g. name change as a result of marriage)
False Positive Filters Prevents improper linkages from occurring when certain conditions are met (e.g. father & son)
Trusted Source Identifies the MOH RPDB as a trusted source for Ontarians insured through OHIP with an Ontario HCN, in order to prevent RPDB records from linking together
OR vs. AND Conditions Generally speaking, PCR treats additional criterion as “OR” conditions, as opposed to “AND” conditions. Additional attributes in a query casts a wider search net, with the highest scoring candidates returned at the top of the candidate list (i.e. in descending order). Refer to the Minimum Score and Maximum Results sections for additional context.

For candidate search queries, PCR totals the deterministic and probabilistic search scores together to generate the candidate confidence score. Note that negative scores can be incurred based on the degree of mismatch under both scoring types, reducing the overall match score. Refer to the Record Linking section for scoring scenarios.

Record Linking

PCR uses its deterministic and probabilistic scoring algorithm and record linking thresholds to determine whether a contributed record should be linked to an existing entity (and assigned that person’s ECID) or not. One of three scenarios are dictated by a linkage threshold scoring range for the member record’s total match score (deterministic + probabilistic) when compared against all the other member records in the PCR. For a given source record added or updated in the PCR, one or more of the following three linkage scenarios can be generated:

  1. No Linkage: Leaves a record in its existing singleton entity based on a ‘low’ matching score below the “Clerical Review” threshold. The demographics on the record are ‘thin’ or ‘sufficiently different’ enough that PCR cannot auto-link or potentially link it to an existing person entity.
  2. Potential Linkage: Flags a record by creating a “Potential Linkage” task based on a ‘medium’ matching score between the “Clerical Review” and “Auto-Link” thresholds. The record in question is left in its current entity (with other linked records, or on its own as a singleton) until the Potential Linkage task is manually resolved by a Data Integrity analyst (linking the record to another existing person entity, or leaving it as-is).
  3. Automatic Linkage: Auto-links a record to an existing person entity based on a ‘high’ matching score above the “Clerical Review” threshold.

Patient Queries

Overview

Consuming systems can query patients in the PCR via:

  1. Get Queries: PCR ECID or Definitional Patient Identifier (e.g. Ontario HCN, CORPID) to ‘get’ either the:

    • Current demographic information and active IDs associated with the patient, or the;

    • Complete list of all identifiers (active, merged, and deleted) for the patient.

  2. Search Queries: Demographics (e.g. First + Last + DOB) or an Informational ID (e.g. MRN) to ‘search’ for candidate patients, returning a subset of their current demographics and active ID list for each matched candidate (to support selection of the appropriate candidate from a ‘pick list’).

The following sections detail PCR get vs. search queries based on input criteria and output results. Note that output data in the PCR is dependent on the data density for a given patient as contributed across sources at the attribute level.

Get Query # 1

Retrieve all provincial demographics and active identifiers for a given patient by definitional identifier

Table: Rules: Get Query #1

Summary Get by Patient ID (ECID or Definitional ID) to retrieve all provincial demographics and active IDs for a given patient.
Operation HL7 FHIR - Patient Read (by ECID), or;
HL7 FHIR - Patient Search (by ECID, Ontario HCN, CORPID, or Standard Model MRN)
Input (Criteria) A definitional ID Issuer & corresponding ID value must be provided in the request.
For Patient Read:
        1. PCR Enterprise Client ID (ECID)
For Patient Search
        1. PCR Enterprise Client ID (ECID), or
        2. Ontario Health Card Number (HCN) as defined by the patient’s record in the MOH RPDB, or
        3. Corporate ID (CORPID) for corporate model hospitals, or
        4. Medical Record Number (MRN) for standard model hospitals.

Note: If a given adopter only has an MRN available for the patient without the knowledge of the source model (Standard or Corporate ID) and corresponding CORPID value, they must perform a Patient EMPI Match.
Output (Results) Summary:

        - Based on exact matching the PCR ECID or Definitional ID for a given active or merged member record, pull the entity view for a given patient.
        - Response includes the entity view for that patient when they have at least one active member record in the entity.
        - Returns the complete set of provincial demographics and list of active identifiers as the ‘golden record’ for a given patient (excludes IDs for merged or deleted members).

Details:
Returns the following list of patient IDs (active only):
- PCR ECID
- Ontario Health Card Number(s)
- Non-Ontario (Out-of-Province) Health Card Number(s)
- Corporate ID(s)
- Medical Record Number(s)
Alternate ID(s)

Returns the following ‘complete’ set of demographics (active only):
- Legal Name - Alias Name
- Date of Birth (DOB)
- Gender (Administrative Sex)
- Address (Home, Mailing, Temporary)
- Phone (Home, Work, Emergency, Other)
- Primary (Communication) Language
- Death Indicator
- Death Date
- Confidentiality Indicator

Get Query # 2

Retrieve all provincial identifiers for a given patient by definitional ID.

Rules: Get Query # 2

Summary Get by Patient ID (ECID or Definitional ID) to retrieve all provincial identifiers for a given patient.
Operation HL7 FHIR - ITI-83: Mobile Patient Identifier Cross-Reference Query (PIXm)
Input (Criteria) A definitional ID Issuer & corresponding ID value must be provided in the request:
        1. PCR Enterprise Client ID (ECID), or
        2. Ontario Health Card Number (HCN) as defined by the patient’s record in the MOH RPDB, or
        3. Corporate ID (CORPID) for corporate model hospitals, or
        4. Medical Record Number (MRN) for standard model hospitals

Note: If a given adopter only has an MRN available for the patient without the knowledge of the source model (Standard or Corporate ID) and corresponding CORPID value, they can perform a Search by Informational ID to retrieve the PCR ECID, and then perform a subsequent ‘get’ by ECID as defined here to return all the patient identifiers.
Output (Results) Summary:
- Based on exact matching the PCR ECID or Definitional ID for a given active, merged, or deleted member, pull the entity view for a given patient.
- Returns all definitional and informational identifiers for a given patient, regardless of associated member status.
- Returns no demographics.

Details:
Returns the following list of patient IDs (i.e. active, merged, or deleted):
- PCR ECID
- Ontario Health Card Number(s)
- Non-Ontario (Out-of-Province) Health Card Number(s)
- Corporate ID(s)
- Medical Record Number(s)
- Alternate ID (s)

Search Query # 1

Retrieves a summary of provincial demographics and active identifiers for a list of candidates that match the provided demographic data.

Rules: Search Query # 1

Summary Search by demographics to retrieve summary provincial demographics and the active ID list for up to 5 matched candidates to present to a user to select a match.
Operation HL7 FHIR - Patient EMPI Match
Input (Criteria) Search by Demographics (minimum criteria) :
1. First Name + Last Name + Date-of-Birth, or
2. First Name + Last Name + Address (minimum of Postal Code)

Note: The MOH & IPC have mandated minimum search criteria and maximum candidates returned when querying based on demographics. Refer to the ‘Minimum Search Criteria’ section for further details.
Output (Results) Summary:
- Based on probabilistic matching on the demographics, pull the entity view for up to 5 matched candidates.
- Response includes the entity view for that matched patient when they have at least one active member record in the entity.
- Candidates are returned in order of match score, highest scoring (most probable) candidate is listed first, to the lowest scoring (least probable candidate) listed last.
- Note that multiple candidates may receive the same (equal) match score.

Details:
Returns the following list of patient IDs (active only) for each matched candidate:
- PCR ECID
- Ontario Health Card Number
- Non-Ontario (Out-of-Province) Health Card Number(s)
- Corporate ID(s)
- Medical Record Number(s)
- Alternate ID(s)
Returns the following ‘summary’ set of demographics (active only) for each matched candidate, to select the appropriate candidate from the returned list, and perform a subsequent get (e.g. by ECID) if additional demographics are required for the single selected patient:
- Legal Name
- Date of Birth (DOB)
- Gender (Administrative Sex)
- Address(es): Home, Mailing, Home Temporary, Mailing Temporary
- Phone(s): Home, Work, Emergency, Mobile, Other
- Date of Death (if applicable)
- Language(s)
- Confidentiality Indicator
Returns a match score for each matched candidate:
- QRI Value

Search Query # 2

Retrieves a summary of provincial demographics and active identifiers for a list of candidates that match the provided informational identifier.

Rules: Search Query # 2

Summary Search by Patient ID (Informational) to retrieve summary provincial demographics and the active ID list for a given patient.
Operation HL7 FHIR - Patient Match
Input (Criteria) An informational ID Issuer & corresponding ID value must be provided in the request:
1. Medical Record Number (MRN) regardless if under a corporate or standard model hospital record, or
2. Ontario or Non-Ontario Health Card Number (HCN) as defined by the patient’s record in the MOH RPDB and across hospitals, or
3. Alternate ID (ALTID) such as Health Shared Services Ontario (HSSO) CHRIS system GUID.

Note 1: If an Informational ID is defined in a single request in addition to demographics, both the informational ID and demographics provided will be used to search for candidates, even if the ID and demographics correspond to different patients (i.e. overloading the request).

Note 2: If a given adopter only has an MRN available for the patient without the knowledge of the source model (Standard or Corporate ID) and corresponding CORPID value, they can perform a Patient Match by Informational ID to retrieve the PCR ECID, and then perform a subsequent ‘get’ by ECID as defined in the ‘get’ queries above to return all patient demographics or identifiers.

Note 3: Patient Match by Informational Ontario HCN supports retrieving newborn Pre-Assigned Health Numbers (PAHNs) on hospital records that do not yet have a corresponding RPDB (RCB) definitional record in the PCR (i.e. PAHN not yet registered with MOH as non-PAHN) when the corresponding newborn hospital CORPID or MRN is not known by the adopter. It also supports identifying all persons associated with a given HCN across PCR sources, regardless of the authoritative MOH RPDB view (e.g. error or fraud scenarios where the same HCN is used to identify different patients in a given source, but MOH identifies that same HCN as being only associated with one authoritative patient).
Output (Results) Summary:
- Based on exact matching on the HCN or MRN informational ID value, pull the entity view for up to 5 matched candidates (expect only 1 matched candidate in practice).
- Response includes the entity view for that matched patient when they have at least one active member record in the entity.
- Returns a match score, summary subset of provincial demographics (active), and the provincial list of identifiers (active only) as the ‘golden record’ for each matched candidate.
- Candidates are returned in order of match score, highest scoring (most probable) candidate is listed first, to the lowest scoring (least probable candidate) listed last.
- Note that multiple candidates may receive the same (equal) match score.

Details:
Returns the following list of patient IDs (active only) for each matched candidate:
- PCR ECID
- Ontario Health Card Number
- Non-Ontario (Out-of-Province) Health Card Number(s)
- Corporate ID(s)
- Medical Record Number(s)
- Alternate ID(s)
Returns the following ‘summary’ set of demographics (active only) for each matched candidate, to select the appropriate candidate from the returned list, and perform a subsequent get (e.g. by ECID) if additional demographics are required for the single selected patient:
- Legal Name
- Date of Birth (DOB)
- Gender (Administrative Sex)
- Address(es): Home, Mailing, Home Temporary, Mailing Temporary
- Phone(s): Home, Work, Emergency, Mobile, Other
- Date of Death (if applicable)
- Language(s)
- Confidentiality Indicator
Returns a match score for each matched candidate:
- QRI Value

Data Consumption


Agreements

As mandated based on data-sharing agreements with the Ministry of Health (MoHLTC), and reflected in the EHR Services Schedule signed by all PCR adopters, PCR data consumption is bound to the following: Only the following attributes may be downloaded into patient/client registration records:

  1. Health Number and Issuing Province, and
  2. Medical Record Number(s) and Issuing Facility ID(s), and
  3. First, Middle and Last Names, and
  4. Date of Birth, and
  5. Gender, and
  6. Address, and
  7. Phone Number (Work, Home, Mobile, Emergency, and Other), and
  8. Primary Language, and
  9. Death Indicator, and
  10. Death Date.

Enterprise Client ID (ECID)

In addition to the business attributes scoped per the agreement language above, PCR ECID is considered as a non-business attribute used to support patient identification, record assembly, and consent management within the EHR. While the PCR ECID can be consumed accordingly, it must not be displayed on user/clinician/provider facing user interface screens or reports.

PCR adopters SHALL also consider ECID as a point-in-time value. To avoid unexpected results due to dynamic PCR record linking, an ECID must only be used to resolve a patient as part of a ‘real-time’ integration (e.g. a get query immediately following selection of a search result). Historical (static) ECID values must not be used for future integration, given they may change.


Source Patient Identifiers

The PCR is configured to store and return all patient identifiers received from each contributing source - including Ontario Health Card Numbers (HCN), Corporate Identifiers (CORPID) and Medical Record Numbers (MRN). As a result, it is possible for for query responses to return multiple instances for HCN or MRN.

As a result, PCR adopters are expected to:

  • Accept messages which include multiple instances of a given patient identifier (source identifier)
  • Store and use the first instance of the HCN identifier received in the message

Common examples includes:

  • A hospital will send an HCN correction, effectively inactivating the prior invalid HCN, and activating a new valid HCN.
  • In rare cases, the MOH will merge HCNs together under one patient.

Minimum Search Criteria

For Patient EMPI Match operation, PCR requires that "minimum search criteria" to be met in order to process the query. Refer to the Patient EMPI Match page for details.


Maximum Results

As part of privacy governance in managing queries against the provincial level client registry, the Ministry of Health (MoHLTC) and Information and Privacy Commissioner of Ontario (IPC) have mandated a maximum result set of 5 matched candidates on a given PCR response. For Patient EMPI Match Operation, refer to Patient EMPI Match page for details.


Minimum Score

As part of privacy governance in managing queries against the provincial level client registry, PCR EMPI Match Operation returns patient records with matching score that exceeds the minimum score. For details on minimum score, Refer to the Patient EMPI Match page for details.


Merged Records

Performing a get query on a definitional identifier (e.g. get by Ontario HCN) for a record merged in the PCR will return the surviving patient. Note that the surviving patient details may differ from the query criteria used to match on the merged record. Performing a search query on an informational identifier (e.g. MRN) or demographics for a merged record will return no result (neither the surviving nor the non-surviving patient is returned).


Phone Number Convention

Although FHIR does not enforce a phone number convention, it is up to the implementers how to reconstruct and/or parse phone numbers based on the convention used in this implementation guide.

It should be noted that PCR linking and matching algorithm does not score on special characters (e.g. dashes) or spaces, and only scores on the digits in the local number. It is up to the implementers how to reconstruct and/or parse phone number based on the convention used in this implementation guide.


Address Convention

Although FHIR does not enforce a convention for address lines, it is up to the implementers how to reconstruct and/or parse address based on the convention defined in the resource profiles (e.g. pcr-patient-response) for PCR FHIR . It should be noted that PCR linking and matching algorithm does not score on special characters (e.g. dashes) or spaces, and only scores on address lines and postal code.

It is up to the implementers to decide how to reconstruct and/or reformat addresses across all the PCR returned address elements. The logic to reassemble and/or format addresses to assign in the local system will differ across adopters and their corresponding data standards. For example, a consumer requiring no space between Canadian postal code forward sortation area (e.g. M5W) and local delivery unit (e.g. 7E6) must strip out any spaces returned by PCR in the postal code element. Note that PCR currently only maps standardized state/province codes for Canada and the USA. International addresses will return null values for State/Province.