A Foundation for Enterprise Imaging: HIMSS-SIIM Collaborative White Paper

A successful Enterprise Imaging (EI) program encompasses a number of key elements, each of which will be briefly summarized in this introductory paper. Many of these topics will be discussed in more depth in upcoming HIMSS-SIIM collaboration white papers dedicated to the subject.

These key elements for a successful Enterprise Imaging program are as follows:

Effective and engaged governance is key to a successful EI strategy implementation. The HIMSS-SIIM collaborative workgroup defines Enterprise Imaging governance as “the decision-making body, framework, and process to oversee and develop strategies for the enterprise imaging program, technology, information, clinical use, and available financial resources.” A dedicated HIMSS-SIIM white paper on EI governance is in press. [11]. Like non-imaging healthcare governance groups, the governance bodies for EI are responsible for bringing together a wide group of clinical, administrative, and information technology stakeholders to make decisions [12‐15].

Governance must be addressed early in embarking on EI. When just starting, typically each department owns, maintains, and controls its own imaging operations, data governance, IT support personnel, and technology infrastructure. There is typically little cross-department knowledge or infrastructure sharing. Personal and professional culture between departments may be non-existent or even negative because of historical turf battles. It is critical to start anew with a core group of constructive stakeholders, executives, and sponsorship in overseeing all imaging activities and service lines.

A thoughtful and on-paper strategic plan vetted through governance should make clear the needed EI infrastructure and services. A beginning strategy and roadmap will at least include plans for the aforementioned key elements (governance, the enterprise image management platform, content, content viewing, EHR integration, image exchange services, and analytics).

Enterprise Imaging requirements, scope, and outcome strategic decisions are particularly critical [16]. These decisions inform the suitability of any already owned technology for the EI platform and thus also make apparent capital or operational funding requirements to implement the program. They also make clear those applications, storage, and viewers that do not fit with long-term enterprise plans. Roadmaps to sunset these components should be considered early, objectively, and transparently to all stakeholders. A restructuring of IT support away from departmental approaches toward enterprise may be necessary, and may involve new support personnel, or reassignment or release of personnel with deep relationships and long track records. All of these contribute to timeline decisions, as EI cannot be implemented without the infrastructure that supports the departmental imaging workflows. The above are delicate and often contentious discussions where strong governance proves invaluable. Engaging the specialties creating images in the strategic decisions will become important when clinical champions become needed during infrastructure and application selection, workflow redesign, and implementation.

The EI platform provides the infrastructure, modalities, devices, and integration points on which strategies can be based (Fig. 1). A standards-based DICOM and non-DICOM clinical image and video storage repository is central to the EI platform. This central repository may be a vendor neutral archive or, if it meets defined enterprise-wide requirements, an existing PACS archive. It includes an index of the image and meta-information content in the archive. The archive should be modality agnostic, modality vendor agnostic, specialty and service line agnostic, and viewer agnostic. Standards-based interfaces and communications, including DICOM, HL7, and standards-based Web Services, connect, enable, and support image acquisition workflows across modalities and departments [17‐19]. Image acquisition devices that support these standards may store their images, with meta-information, into the VNA. Acquisition devices that are supported include departmental DICOM imaging modalities, point-of-care acquisition modalities, handheld device photo or video apps, digital capture systems in procedure rooms, image exchange gateways, and software designed to import content saved on a disk or received by referring or patient portals.

Core services of the platform include the following:

In the past, the primary role of an EHR in the imaging domain was to create and send orders to a Radiology Information System (IS) or PACS, to receive the final imaging results (the signed-off diagnostic report) and a link to view the images in the departmental PACS viewer. EHRs today commonly support “information system” services and workflows for radiology, cardiology, obstetrics, and other imaging departments.

As with other medical procedures, the EHR can use order information to index and track imaging procedures [20]. Some imaging procedure workflows are order based, while some are encounter based [21]. For encounter-based imaging procedures, the EI platform can either use an image availability notification to report “unsolicited” imaging results or it can request the EHR to create an order under which the imaging results can be associated. In this way, imaging procedures that are order based and those that are not order based can be notified, indexed, and made available in the EHR with clinically relevant descriptors.

Images serve many purposes in patient care, with many operational workflows driving their creation [22]. Compartmentalizing the spectrum of content routinely captured today is difficult. Categorizing and supporting content based on the medical specialty that performs the imaging has been the traditional model (Fig. 1). This works reasonably for diagnostic imaging because of consistency in operational workflow within departments. However, now ubiquitous handheld still image and video content does not fit well in this model because any provider with appropriate training and security access may store content in the EHR, sometimes bridging historical specialty lines. With increasing provider integration and graduate medical education programs requiring imaging competencies, even within the same enterprise, often many different specialties may perform nearly identical procedures; extremity radiography and bedside ultrasonography are examples of such procedures [23‐27]. Images coming in from external entities are also hard to classify by specialty, as for patient convenience, they are stored commonly on the enterprise repository without a local specialist imager reviewing them. Finally, the traditional specialty-based model for grouping imaging content also does not reflect the depth, breadth, or purpose of images captured within each specialty or department.

For the purposes of classifying the various forms of EI content, using modalities alone as guides is similarly inadequate. Ultrasound images for example may take the form of diagnostic echocardiography, may be captured as part of a procedure such as a line placement, may be saved only as documentation of sonographic findings such as during a GI endoscopy, or may be integrated alongside structured and unstructured text as often is done in an ankle-brachial index vascular image-enriched clinical report. Differentiating imaging types using file structure, such as DICOM or JPEG, is also insufficient, as many forms of medical images not initially captured as DICOM are archived and viewed through DICOM communications. Categorizing by image consumer also will not work, as the same performed imaging exam may be used in many different ways; a single extremity radiograph may be consumed by radiologists for diagnosis, by primary care for referral, by orthopedic surgeons for operative planning, by revenue management during revenue justification, and by patients for social media.

For educational and illustrative purposes, this HIMSS-SIIM Definition of Enterprise Imaging workgroup evaluated the above approaches and chose to describe and organize clinical imaging and multimedia content by the intent of its use by the performing providers (Fig. 2). This is to be distinguished from the imaging source (ultrasound modality, digital camera, etc.), the type of images created (DICOM Ultrasound, JPEG Photo, etc.), or the operational workflow that leads to the image(s). This approach defines content across a spectrum or continuum of four broad categories: diagnostic imaging, procedural imaging, evidence imaging, and image-based clinical reports. The framework is intended to be high level and educational, rather than provide hard and fast rules, as even within an individual encounter’s imaging examination, images often serve more than one intent and thus fall into more than one category. For example, during a cardiac catheterization, images may initially be captured and depict slow flow through a coronary vessel (diagnostic), as well as catheter and potentially stent manipulations and timestamps (procedural).

Nevertheless, categorizing the images into these four groups illustrates the complexity of solving EI for an organization. Such categorizing may help guide strategic decisions, as the strategies for image capture, storage, distribution, review, image intent, and value of the images tend to be more consistent within a category.

If an EHR today did not offer providers the ability to input their textual physical exam findings, or did not offer providers the ability to review the textual physical exam findings in another clinic, that EHR would be deemed incomplete and insufficient to meet the needs of the enterprise. Ironically, however, the images depicting physical exam states are commonly not available to providers when they are needed, leaving those EHRs similarly incomplete. At most sites, these images are not available in the EHR because an easy-to-use method of storage and viewing has not been deployed to providers.

One primary goal of an EI initiative is to deliver all forms of imaging to the electronic health record. An enterprise image viewer is necessary to achieve this. Widely distributing images in the EHR improves communication between providers because relevant images toward making a diagnosis are not siloed away from providers needing them [38]. It also may prove to decrease imaging utilization, as a redundant exam may not need to be performed if the original images are easily accessible [39]. Widely available and easily accessed imaging facilitates multidisciplinary learning by providers [40, 41]. With the increased responsibility patients are taking in their health care decisions, and the frequently high cost of imaging, patients are taking more interest in imaging performed; having those images available for consumption or download through patient portals will be expected [42].

Understanding image viewing needs (which vary just like image capture workflow needs) is important for the governance and leadership bodies to recognize in developing the EI strategy. A dedicated HIMSS-SIIM white paper on enterprise viewing is in press [43]. Most enterprise image viewing clinical users can be categorized into one of several groups:

Users in all four groups expect fast and efficient review and manipulation of image datasets on any desktop, laptop, or mobile device. The enterprise viewer will meet the needs of most image reviewers, though pre-surgical and diagnostic image review still typically requires a dedicated viewer with specialty-specific data integrations, functionality, calculations, and user interface for the most advanced patient care and provider efficiency.

Like clinical free text and structured data, images must be shared between providers and patients outside of the walls of the creating hospital or clinic. The EI platform should be well suited to provide both inbound and outbound image exchange services [44‐48]. There are many sources of outside images brought into an enterprise, such as CD/DVD, health information exchanges (HIE), as well as patient portals, referring physician portals, and telehealth portals. These images are stored in the VNA to be indexed and viewed in the EHR, and are available to be pre-fetched onto PACS for comparison or secondary interpretation. For outbound image exchange, the EI platform can leverage the image repository to assemble the studies and reports for export either to a CD/DVD burner or to an electronic exchange gateway. Benefits to an enterprise-wide-level image exchange service include not only a single centralized operational model internally for export but also simplifying the training and operationalizing of outside entities to send inbound by providing a single receiving platform across specialties.

Business and clinical reports associated with the acquisition and management of imaging content outside of radiology and cardiology are felt by the HIMSS-SIIM workgroups to be generally immature. Detailed analytical tools for EI are in their infancy [49, 50]. The EI platform provides an infrastructure that supplies both the image and associated metadata to enterprise data warehouses. Defining and standardizing this metadata across the enterprise provides a repository of data that can provide departments with detailed study statistics, utilization reports, and image acquisition patterns [51‐54]. True deep learning and complex neural networks of image data contents are beginning to be developed that will play large roles in the future of EI [55‐58].