Clinical Data Exchange Models

As we look to understand the many interoperability challenges that exist in the healthcare industry, it can be helpful to take a step back and focus on one piece at a time. In today's post, we'll take a closer look at three of the most common architectural models for a Health Information Exchange (HIE). Those models are: Centralized, Federated/Decentralized, and Hybrid.

Centralized Exchange Model

With a Centralized exchange model, data is collected and stored in a central repository such as a data warehouse. The organization that maintains the exchange has full control over the sharing of data in this model. This includes control over the authorization and authentication process to access the data. As well as the process for authorizing and recording data transactions among exchange participants.

One of the key benefits in this model is the response time for data returns. This is due to the simple fact that it's quicker to obtain data from a central, consolidated source that it is from multiple sources. With a strong governance model in place, the ability to analyze and audit transactions can also be easier with a single centralized model.

Unfortunately, this model also comes with some challenges. It is usually quite expensive and resource intensive to implement. It also relies on consistent, timely data transfers from source systems ... which may occasionally present issues. With a centralized model, it's important to make sure that patient matching algorithms are in place and effective. Unless there is a shared unique identifier present, data matching can be difficult. Increasing the risk of data mismatches, data omission, or duplicates.

Federated / Decentralized Exchange Model

In a Federated exchange model, data remains at the source or point-of-service. Participating entities agree to share their data with other members of the exchange. This model is sometimes referred to as Decentralized exchange.
Participating entities maintain ownership and control of the data within their source databases. Data gathering and authorization occur on an as-needed basis. Ensuring that responses always contain the most up-to-date information. A single governing body or administrative group maintains a Record Locator Service (RLS) to facilitate the data transfers using standard integration methods. This RLS maintains a patient list along with details on the participating organizations who have provided services. Which many see as a strong benefit.
Patient matching concerns do not disappear with a Federated model. They become the responsibility of the source organization to ensure their data can be accessed and shared. With numerous connection points needing to always be available, maintenance and security can also be challenging. Individual consent for data sharing also plays a significant role in ensuring that this type of exchange model is successful. (We'll dive into that more next week. 😊) As with any technical architecture that relies on having numerous systems constantly connected, maintenance and security can be challenging.

Hybrid Exchange Model

The Hybrid exchange model is a cross between Centralized and Decentralized. This model is meant to leverage the 'best of both worlds'. Participating entities are able to maintain ownership and control of their data by having key elements stored in logically separated 'vaults' managed in a central location.

The HIE database in this model manages the requests for information, which are then distributed across the network. Algorithms are put in to assist with increasing the probability of successful data gathering for the response. Often leveraging these key elements and distinct identifiers for the patient and provider(s). Source database connections in this model are often dependent on federated EHR adaptors. Source database connections in this model are often dependent on federated EHR adaptors.

Now, before we wrap up ... let's take a quick look at the different forms of data exchange we often see in healthcare.

Directed Exchange

A directed exchange occurs when the source party sends information to a known recipient. Data is sent in an encrypted, reliable, and secure manner between two trusted entities. Some examples of directed exchange are:

• A hospital sending a Summary of Care via direct message to a follow up provider.
• A provider/hospital sending immunization data to public health organizations or state registries.
• Quality measure reports being electronically submitted to government agencies, such as CMS.

Query-Based Exchange

Query-based exchange is often used in situations where unplanned care is being delivered. For example, an emergency department or urgent care visit. With this form of exchange, providers 'query' to search for available and accessible clinical data on the patient. Potential sources will depend on the information exchanges the provider is connected to. Epic Care Everywhere functionality is an example of this form of data exchange.

Consumer-Mediated Exchange

Consumer-mediated exchange is becoming more and more prevalent in global healthcare. With this form of exchange, patients have access to their own health information along with the ability to collate and control the distribution of information among providers. In a broad sense, this is sometimes compared to an individual's access to their financial information. (Think of transferring funds between accounts, or using a service like paypal to send money to friends.)