Patient Care Special Interest Group Minutes May 19, 1993 Morning Session Kathy began the meeting by asking for a new co-chair volunteer. Karen Keeter later agreed to take on this responsibility. Mead Walker gave a presentation on Object Oriented Analysis and briefly explained the goals of the Quality Assurance & Data Modeling committee. Their purpose is to define a data modeling process that the individual technical committee's can apply to their problem domain. In addition, this committee has assumed responsibility for coordinating the definition of a fundamental data model that defines objects common to all technical committees. HL-7 is making the transition from an Entity-Relationship data model to and Object Oriented data model. Mead explained the differences between the two methodologies. The Quality Assurance & Data Modeling committee has published a draft version of the "Message Development Framework for HL-7" which has been rewritten to incorporate the Object Oriented methodology. Previously, this document assumed an Entity-Relationship approach to modeling the data. [Parts of the discussion that are covered in the draft document have been omitted from the minutes. Please see "Message Development Framework for HL-7" for further information.] Members of PCSIG expressed some concern about making the transition from an E-R approach to the O-O approach. Karen Keeter asked whether it made since to begin using O-O since the Framework document is still evolving. Mead suggested that we make the transition as soon as possible. In order to avoid duplication of efforts, Mead suggested that there must be some high level data model that can be used for partitioning the problem domain and identifying common objects. He referred to some work that had started on this over the last two days by the HL7-MEDIX IEEE-P1157 Joint Working Group (hereafter referred to as JWG). He said that this group, under the direction of Medix, will coordinate the definition of a global data model. In addition to developing this model, they will define a process for reconciling differences that arise between committees. Encounter Model Discussion After discussing the transition to Object Oriented analysis, Mead briefly discussed the encounter model that had been formulated by the JWG. The encounter model defines the relationship between 12 different entities that form a common model that could be referenced by all the technical committees and SIGs. The 12 entities are: Episode, Payor, Problem, Healthcare Provider, Healthcare Encounter, Patient Account, Benefit Plan, Patient Financial Transaction, Observation, Service, Service Category, Order. Descriptions were provided for some of the entities: Healthcare Encounter: A span of time in which a patient is in contact with a Healthcare Provider or Practitioner. For example: a doctor visit or inpatient stay. It is possible to have multiple encounters going at the same time. Episode: Used to group multiple encounters for some purpose. Practitioner: A person that provides healthcare related services -- doctor, nurse, lab technician, social worker. Provider: An organization that supervises the provision of care to patients. In a solo practice, the practitioner is also the provider. It was noted that there are differences between the JWG entity definitions and those of PCSIG. Mead suggested that a meeting be scheduled at the next HL-7 conference to discuss these differences and reconcile conflicting data models from all the committees. Mead plans to distribute the JWG model in about 1 month. Kathy will distribute that model and request an analysis of discrepancies to be returned to her in August. In addition, she asked for any feedback regarding the orientation packet and issues log. Afternoon Meeting The group reviewed Draft 6 (April 6, 1993) of the "HL7 Patient Care, Data View and Event View" document and focused primarily on answering questions posed by Hans that were left in the draft version of the document. Karen Keeter volunteered to maintain a log of changes to the document. Jay Margolis asked where in the model the patients physical state was described. Karen suggested that this may be in the Action model under Observations. 2.2 Person Model (page 5) In response to the question regarding when a person is considered a patient, it was agreed that this could be defined outside of PCSIG without affecting the PCSIG model. In response to the question about how a mother and baby should be represented, it was agreed that it should be two separate persons. 2.3 Encounter Model (page 6) A question was raised about Group encounters. The group concept may be covered by the JWG model which has defined organization. Need to follow up on this to ensure that it satisfies PCSIG requirements. It was agreed that we should try to eliminate the Encounter sub-types if possible. It appears that the JWG model has done this by saying that an encounter can be related to both a practitioner and a provider. It was suggested that the JWG model should support practitioner-less encounters. An question was raised about whether we care about encounters for which no data is collected. It was agreed that the Encounter entity needs to have a (1,M) relationship with Practitioner and that Practitioner should have a (0,M) relationship with Encounter. We briefly discussed JWG's Episode entity. Although the definition is not complete, it is a grouping of Encounters. The Episode may be defined by the Payor. A question was raised about the difference between clinical and financial episodes. The clinical view is dependent on the problem that is being analyzed. PCSIG is interested in more detail than the Encounter entity is intended for. Do we need to model a "contact" or this covered in our definition of observations, orders, etc? Should there be a recursive encounter relationship? These questions were not resolved and should be logged as issues. 2.4 Planning Model (page 7) It was agreed that the "satisfies" relationship between Patient/Treatment and Patient/Observation should be removed and that the Goal/Outcome entity should be renamed to Goal/Expected Outcome. It was suggested that the following relations be added to the model: 1. Observation "measures" Goal/Expected Outcome. 2. Treatment "?????" Goal/Expected Outcome. 3. Observation "validates" Diagnosis/Problem. 4. Treatment "is motivated by" Diagnosis/Problem. A number of issues were raised: 1. Do Problems & Diagnoses belong combined or should they be separate entities? 2. Should there be a Protocol entity? Or a Protocol model? 3. How does the model support conditional orders? 2.4.1 Order (page 8) We agreed to modify the Order relationship, item #2. An order is placed for one and only one patient. We all agreed that the PCSIG definition of an Order needs to be compared against the Order Entry/Clinical Results Technical Committee's definition. Specifically, how do they handle multiple occurrences of an order? Linda Joseph and Catherine Rogers agreed to follow up on this. Sections 2.4.3 and 2.4.2 should be switched to improve the consistency of the document. 2.5 Action Model We agreed that the Goal/Expected Outcome entity needs to be added to the Action Model. Item #16 under Diagnosis/Problem relationships should read "A goal/expected outcome is referenced by 1,M diagnoses/problems. The issue of variations was raised. Is it a type of observation? We need to determine how this fits into the protocol model. Agenda For September Meeting: 1. Reconciliation of JWG data model with PCSIG data model (and other committee's models). 2. Protocol model development. 3. Comparison/acceptance of Order Entry model.