Unique Specimen Identifier Requirements
Anatomic Pathology Working Group
November 28, 2012
Contents
Overview and Business Case
Present State of Identification in the Anatomic Pathology Laboratory
Activity Diagram of Current State
Proposed Future State
Unique ID Technical Schema
Workflow and Physical Requirements
Information Requirements for the Unique ID
Unique ID Use Cases
Actors
Use Cases
Use Case Diagram
Activity Diagram
Glossary
References
Overview and Business Case
This work covers the necessary and sufficient requirements for a unique identifier (UID) standard. It is presented as a Domain Analysis Model (DAM), and the purpose of this work is to ensure interoperability of specimen identifiers both within and between anatomic pathology laboratories. The scope of the DAM is to describe the relevant attributes that are specific to the identification of an anatomic pathology specimen.
Present State of Identification in the Anatomic Pathology Laboratory
As a pathology case moves through the laboratory, it must be correctly identified at all points in time and location to ensure the specimens, correct processing steps, and results are matched to the right patient. Presently, laboratory staff, laboratory information systems (LIS), and specimen tracking systems (STS) create and read identifiers on anatomic pathology specimens.
Starting at the point of accessioning where the case enters the laboratory and receives a case number, LIS and STS systems generate identifiers for the case, its specimens, and any derived specimens produced as the case moves through the laboratory. Unlike other laboratory specimens (such a tube of blood), a derived portion of an anatomic pathology specimen (e.g. a subsection of tissue) is unique and distinct from other derived portions. This uniqueness must be preserved in the identification scheme used in the laboratory.
Identifiers in the laboratory come in two general types: human-readable and machine-readable. A human-readable identifier is made to be read by laboratory staff, such as “Brown, Fred S-12-1234-1A, CD20”. Systems and devices in the laboratory do not interpret human-readable identifiers. Machine-readable identifiers are constructs to be read in an automated fashion by laboratory devices or systems like LIS or STS. Barcodes (1D and 2D) are the current format of machine-readable identifiers in the laboratory, but technology like radio frequency identification (RFID) is making in-roads into the laboratory.
A commonly used identification scheme is an alphanumeric with meaningful sub-identifiers, such as ”S-12-1234-A”. For this identifier, the “S” denotes a surgical pathology case, “12-1234” is the case number, and “A” denotes the block. In a modern pathology laboratory, this ID scheme is reflected in both the human readable label on the item and the barcode (either 1D or 2D). Since LIS and STS systems are designed to produce unique identifiers for a given laboratory site, this method of identification is effective and eliminates identifier collision within the realm of the given LIS or STS for that laboratory. Identifier collision occurs when a specimen is mis-identified by the duplicate use of an identifier. For example, two anatomic pathology specimens from two separate patients arrive in the laboratory at the same time. During accessioning of the specimens, each is given a unique identifier in the LIS and a corresponding printed label is applied to each specimen. If the same label is applied to both specimens, both specimens are then identified as being the same. Given that similarity, the orders for the correctly labeled specimen could be performed on the mislabeled specimen, setting the stage for collision. The collision of identifiers is a serious medical risk, which can have devastating impact to patient care by promoting case or specimen misidentification, leading to incorrect work and/or diagnoses (need definitions here) being applied to a patient.
Where the aforementioned method of identification breaks down is in the usage of identifiers across separate systems. For example, consider the movement of a specimen between a LIS and a STS, specifically a single cassette originally identified and labeled by the LIS and a pending order for microtomy. For a STS system to match the order for microtomy to the correct block, it must either be able to interrogate the machine-readable identifier to correctly match the order to the specimen (the block) or a laboratory technologist must manually match the order with the specimen via interaction with the STS. For scenarios where the format (the structure of the identifier) of the machine-readable identifier can be shared between LIS and STS, the first scenario can occur. This is typically not the norm and usually involves some degree of customization at every installation of Brand A’s LIS with Brand B’s STS to ensure interoperability. For the later scenario, this manual step often involves the relabeling of the specimen in question so an appropriate machine-readable identifier can be used to drive the matching of the order with the specimen. This relabeling step is both time-consuming and potentially error prone. A similar re-labeling (and error-prone) scenario occurs when a pathology case moves from one institution to another. To prevent the collision of identifiers, case material sent to another laboratory must be re-labeled with identifiers from the receiving institution to prevent collision.
Another aspect of identification that is worth mentioning but out of scope for this DAM is the movement of specimens between the clinical laboratory and the research laboratory. Correct identification is essential for both the accurate movement of specimens from the clinical realm to the research realm and the matching of clinical data associated with the specimen. In addition, to ensure the necessary “blinding” aspects of research and experimental design which the clinical data follows, the identifier must also be “blind”. This “blind” identifier must not have any discernable information in its structure or format that would allow the determination of the origin of the specimen (such as the laboratory of origin). A proposed standard is being assembled by caBIG for this need, and this DAM is very similar in nature to what is presented here, minus the “blind” identifier requirement. A globally unique identifier can definitely improve the workflow and interaction between the clinical and research spaces.
Activity Diagram of Current State
Below is asample activity diagram to describe the current state of laboratory workflow (both intra-lab and inter-lab) with relabeling of specimens as they move between an actor in the lab that assigns an identifier and another actor that interrogates an identifier.
Proposed Future State
A unique identifier standard will allow the interoperability of machine-readable identifiers between systems and devices in the laboratory, eliminating costly relabeling steps while preventing identifier collision. To accomplish this and have rapid, efficient adoption, this standard must have the following attributes:
- The unique identifier must be considered a separate entity from what is currently used in laboratories today with respect to the human readable identifier in place now. Currently, the information contained in the human readable identifier is typically the same which is contained within machine-readable identifier. Two reasons drive the separation of the human readable from the machine-readable ID. First, the physical space requirements present with some of the derived specimen types in anatomic pathology restrict the amount of information that can be present in the human readable part of the label. This is clearly seen in the labeling done for cassettes and slides, where the available area for a label is small given the size of the item. Another reason for the clear separation is the potential length of the unique ID so that the laboratory sub-identifier can be long enough to handle the current and future number of laboratories globally.
- An assigning authority which can assign globally unique identifiers will be leveraged for identification of the laboratory location with a unique identifier. Coupling this globally unique identifier with a locally unique specimen identifier (present state in the laboratory today).
- Be able to handle the concept of multiple specimens in a single container. Typically, the relationship between specimens and containers is one-to-one. For example, one piece of tissue (the specimen) is placed in a cassette. However, with emerging molecular techniques and other testing modalities that have multiple specimen types co-located, this additional scenario is becoming more common. Having healthy tissue and diseased tissue from a patient on the same slide is an example.
- Have the ability to handle the lack of a human readable or a machine readable for a specimen. These could be scenarios where an identifier exists in the system but the specimen is not labeled with both types of identifiers.
Unique ID Technical Schema
Workflow and Physical Requirements
- The standard must enable the creation of a globally unique identifier for both specimens and containers.
- The standard must be able to handle the relationship between specimens and containers. As stated in the glossary, every specimen is held by a container. From the perspective of the specimen, this is a one-to-one relationship. While the current majority of containers contain a single specimen, the relationship from the perspective of the container is a one-to-many. There are a growing number of scenarios in which a single container houses a number of separate, unique specimens.
- Given the one-to-many nature of the container-specimen relationship, the standard will place the unique container ID in the machine-readable component of the label. This design will allow for efficient and scalable lookup of specimens held within the container.
- This standard will allow for a distinct detection of a machine-readable ID in keeping with the standard to prevent mis-interpretation with a currently implemented schema in use.
- The length of these IDs will be amenable to current LIS storage and implementation abilities and the available technology for identification
- All systems that interact with Unique IDs must be able to store two identifiers related to the specimen, the container and specimen identifiers.
Information Requirements for the Unique ID
- Location ID – This identifier represents the physical location where the specimen was accessioned and will be globally unique.
- Specimen ID– This is the unique identifier for the specimen. This ID is unique for a given location.
- Container ID – This is the unique identifier for the container. This ID is unique for a given location, and for the typical scenario of a single specimen in a container, it will be the same as the specimen ID.
Unique ID Use Cases
Actors
Laboratory Staff
These are persons in the laboratory that interact with specimens. Examples include pathologists and medical technologists.
Laboratory Devices
These are devices or systems that perform work on pathology specimens and containers and use identifiers to match work orders with the correct specimens and containers. Examples include the aforementioned laboratory information systems (LIS), specimen tracking systems (STS), and laboratory devices like a slide stainer.
Use Cases
Assign Identifier
When a specimen either enters the laboratory or it created, it must be assigned an identifier. This is done either by the laboratory staff or a laboratory device (typically the LIS or STS) to generate the identifier and apply it to the specimen.
Interrogate Identifier
To enable the correct identification and matching of orders to specimens in the laboratory, the identifier must be “read” or interrogated. This may be a standard reading of the identifier by laboratory staff or a machine reading by a laboratory device, such as reading a bar code.
Use Case Diagram
The diagram below highlights the processes at work in the laboratory.
Activity Diagram
The following diagram captures the flow of specimens though the laboratory to highlight the role the unique identifier in the matching of work orders to the correct specimens without the need for relabeling of the specimen.
Glossary
Accessioning
This is the act of registering a specimen in the laboratory. Typically, this involves the creation of a top-level case identifier and the detailing of the various components of the case along with data associated with patient’s clinical condition and the collection and transport of the specimen.
Case –From DICOM Supplement 122 with minor edits
A pathology case is a unit of work resulting in a report with associated codified, billable acts.
Container– Taken from DICOM Supplement 122 with minor edits
Specimen containers (or just “containers”) play an important role in laboratory (diagnostic) processes. In most, but not all, process steps, specimens are held in containers, and a container often carries its specimen’s ID. Sometimes the container becomes intimately involved with the specimen (e.g., a paraffin block), while other times the container is simply that, a container.
Containers have identifiers that are important in laboratory operations and in some imaging processes (such as whole slide imaging). The Container ID and the Specimen ID are distinct entities, making them different data elements. In many laboratories where there is one specimen per container, the value of the specimen ID and container ID will be same. However, there are use cases in which there are more than one specimen in a container. In those situations, the value of the container ID and the specimen IDs will be different.Once specimen ID will match the container ID and the others be uniquely distinct but not match the container ID.
Derivative Specimen
Specimens are sampled and processed during a laboratory’s (diagnostic) workflow. Sampling can create new derivative specimens. These derivative specimens are full specimens in their own right (they have unique identifiers and are direct subjects in one or more steps in the laboratory’s (diagnostic) workflow. This property of specimens (that can be created from existing specimens by sampling) extends a common definition of specimen which limits the word to the original object received for examination (e.g., from surgery).
Identifier Collision
This is the misidentification of an entity due to identical identifiers from different sources being used in the same environment. For example, a LIS in Laboratory A labels a specimen from Mr. Jones as “S-12-1234-1”, while Laboratory B labels a specimen from Mr. Smith as “S-12-1234-1”. If one were to send Mr. Jones’s case from Laboratory A to Laboratory B, the identifiers on Mr. Jones’s case would collide with those of Mr. Smith’s and allow for the matching of orders and results with the wrong patient.
Machine-readable identifier
An identifier that is read by a device, such as a barcode or RFID tag
Matching
This is the act of comparing the orders for work in the laboratory with specimens to make sure that the correct work is done on the correct specimen. The specimen identifier is the linking item between the order and the specimen to enable this matching to happen.
Orders
An order is requested work on a specimen. The order contains a description of the work requested along with the identifier of the specimen to enable matching of the order with the specimen.
Processing
This is the act of manipulating a specimen via physical, chemical, or other means. Processing may result in the creation of multiple derivative specimens (such as the sectioning of a piece of tissue from surgery into smaller portions for submission in cassettes), or the act of processing may simply change the characteristics of a single specimen (such as the application of a stain to a slide).
Specimen– From DICOM Supplement 122 with minor edits
A physical object (or a collection of objects) is a specimen when the laboratory considers it a single discrete, uniquely identified unit that is the subject of one or more steps in the laboratory (diagnostic) workflow.
References
DICOM. Supplement 122: Specimen Module and Revised Pathology SOPClasses.June, 2008
caBIG® Platform Independent Model and Service Specification. CS13: Specimen Identifier Management Service. February, 2010
1