Interoperability

No one IT system can fully manage a supply chain, or a health system, in its entirety. This fact is a core design intention behind OpenLMIS’ eager approach to working with other IT systems through established open standards. Since OpenLMIS is web-based, has a supply chain focus, and originated in the health space the open standards that OpenLMIS uses also come from those spaces.

This document will cover specific standards and profiles that are in use while the General Interoperability Approach document covers more on the reasoning behind some of these choices in regards to enterprise integration.

Web

  • REST APIs over HTTP w/ JSON preferred
  • Authentication is delegated to OAuth2 with bearer tokens
  • Unique identifiers are UUID
  • Datetimes in RFC 3339
  • Language tags in ISO 639
  • UTF-8 encoding as default

Supply Chain

  • Product Model leverages lessons learned from GS1 and the BI&A Logical Model
  • Configurable file exchange (FTP/AS3) for exchanging Orders and Shipments with Enterprise Resource Planning (ERP) / Warehouse Management Systems (WMS)
  • GS1 identifiers supported on products with preference given to GS1 identifiers, message formats and event transactions where GS1 items are in use. (expanding)
  • Support for EDI (ANSI x12 / EDIFACT) in exchanging inventory reports (planned)

Health

HL7’s FHIR for:

Profiles & Use Cases

Open standards give us most of what we need to start integrating systems, however it can still be useful to refer to standard profiles and/or use cases that use the standards to achieve a specific purpose.

Note on sources of truth and derived definitions

When interoperating or integrating with another system it’s important to take a moment and determine which systems own a particular set of data, and which systems need to know about that data, and potentially add to it.

For this reason we’ll be using two terms in the following sections:

  • Source of Truth (aka Master Data): Is a source, often an IT system though it could be something such as a shared spreadsheet, that defines the one canonical definition of some thing. For some entities (such as a Product), there may be many sources of truth for specific aspects of that entity. It’s important to note however that no two or more sources of truth may try to define the same aspect of the entity.
  • Derived data: Is data which comes from a source of truth. It may enhance/add to the definition that comes from the source of truth. e.g. a Master Facility List (MFR) may be a source of truth that defines the names and locations of all the facilities. In OpenLMIS the Facilities we have would be derived from the MFR, and additionally we might enrich our derived definition with information such as how it fits into the supply chain or whom in OpenLMIS is assigned to it.

A few examples of where OpenLMIS expects to be a source of truth:

  • Re-supply requests and fulfillments that occur inside OpenLMIS
  • A mapping of supply chain workflows and approval processes
  • Cold chain equipment, where it’s installed and functional status
  • Stock cards and associated movements that occur in OpenLMIS

A few examples where OpenLMIS we hope to derive data from:

  • Facilities, geographic and administrative areas
  • Commodities, Items, Lots and Categorizations
  • Cold chain equipment temperature and alerts thereof

Defining Locations (geographic areas, facilities, store rooms, etc)

OpenLMIS needs to know about the Facilities and Geopgraphic Zones that are a part of the supply chain to enable various re-supply workflows: hospitals, clinics, etc. While OpenLMIS needs this information, we believe that the process of uniquely identifying and assigning core attributes (e.g. name, address, etc) of these places works best when the information is curated outside of OpenLMIS, and then shared with OpenLMIS.

The core profile that describes the basic functioning of this is IHE’s mCSD profile of which OpenLMIS leverages Location as the source of truth for:

  • Name
  • If it’s a Geographic Zone or Facility
  • Unique identifier
  • Code(s)
  • Hierarchy (e.g. Acme Clinic is in Maputo district which is in the Country Mozambique)
  • Position (lat & long)

In OpenLMIS we support the ability for an implementation to “follow” a Registry that provides a complete list of Locations. By following such a registry OpenLMIS will allow the administrator to create Facilities and Geographic Zones that are based from the registry as a source of truth, and any updates in the Registry will be reflected in OpenLMIS’ derived definitions.

It’s also possible for OpenLMIS to play the role of this registry which other systems may subscribe to and follow when a more appropriate registry isn’t available, however we’d encourage implementations to take on the extra work of implementing a more appropriate registry for this critical task.

Supply Chain Metrics & Indicators

OpenLMIS has a number of re-supply workflows that produce metrics and indicators relating to the functioning of the supply chain. These are made available through the use of FHIR’s Measure and MeassureReport:

  • Measure: Defines a metric/indicator (e.g. #days stocked out or supply status)
  • MeasureReport: Contains the values for a Measure by Location and a period of time.

This usage is intended to comply with the (upcoming) IHE mADX profile.

By publishing indicators in this way, a connector (planned) can discover new MeasureReport’s as they are published/updated and move them to analytical systems such as DHIS2. More about this approach can be read in OpenLMIS’ Interoperability w/ DHIS2

Cold-chain Equipment & Remote Temperature Monitoring (RTM)

OpenLMIS defines a catalog of cold-chain equipment (e.g. refrigerator), which may be imported from WHO’s PQS, and where that equipment is located. This registry of what equipment has been installed and where it is located is available as a list of FHIR Device.


NOTE

OpenLMIS’ acting as a registry in this case is done in-lieu of a more appropriate source of truth for cold chain equipment installation and location. We’d be happy to learn if there’s a more appropriate open, source of truth, system.


OpenLMIS may also receive status alerts about equipment functionality, which is normally sourced from a Remote Temperature Monitoring (RTM) device, such as Nexleaf’s ColdTrace.