Understanding why and how unintended consequences occur will help you identify and fix your current EHR-related problems and will also help you avoid future unintended consequences.
The management expert, Peter Drucker, called health care workplaces "the most complex human organization[s] ever devised." Interactions between these complex environments and increasingly complex EHRs can spawn subtle unintended consequences of EHR implementation. These consequences do not result from malfunctions within the EHR, but from the interactions between the EHR and the work environment or between the EHR and the technical and physical infrastructure.
In this section we describe Interactive Sociotechnical Analysis (ISTA), a framework to help you understand the types of interactions that can result in unintended consequences. The ISTA framework has four key elements:
- The EHR (as designed), or how the developers envisioned that the EHR would be used
- The work environment: The policies, priorities, hierarchies, and relationships within the organization.
- The technical and physical infrastructure: Other IT, medical devices, building design and layout.
- The EHR (as used): The product of interactions between the EHR and the work environment and the physical and technical infrastructure.
Elements of the ISTA framework
visual flowchart of the elements of the interactive sociotechnical analysis, with four boxes linked directionally or bi-directionally; a full description of the flowchart is provided in the text below
1. The EHR (as designed) interacts with the work environment.
EHRs can alter communication and relationships among clinicians in undesirable ways, even while the EHR helps eliminate other problematic and dangerous forms of communication (such as illegible prescriptions).
Post-EHR Changes in Communication: A survey of commercialEHR physician users found that communication among clinicians, and between clinicians and their patients, benefitted from EHRfeatures such as e-mail, instant messaging, improved access to patient information, and improved access to clinical guidelines. The survey also revealed that the introduction of an EHR resulted in some new communication barriers.
2. The EHR (as designed) interacts with the physical or technical infrastructure.
A poor fit between an EHR and other IT or the physical infrastructure is a common source of unintended consequences. Problems involving the interface with other health IT systems can lead to poor decisions, delays, data loss, errors, unnecessary testing, and system downtimes. "Dueling systems" can result if paper-based or legacy systems continue to be used after the implementation of an EHR. Features of the physical layout such as the ease with which computers can be accessed, noise, overcrowding, and illumination affect work performance and safety and may have unanticipated, negative effects on use of the EHR.
Conflicts Between Technology and Physical Layout: As a cost-saving alternative to installing computers in each patient room, a large academic hospital chose to invest in computers on wheels (COWs). The COWs were designed to roll anywhere they were needed and were intended to be especially useful at a patient's bedside. However, problems began to emerge after the COWs were put to use in the hospital.
3. The work environment interacts with the EHR (as used).
Policies, priorities, hierarchies, and workplace relationships shape how the EHR is used and for which tasks. At the same time it is also possible that the way the EHR is used can lead to changes in organizational policies, procedures, and hierarchies.
EHR Safety Check Results in a Potentially Dangerous Workaround: To reduce the risk of medication overdoses, a nursing home implemented "intentional blocks" in the order entry system. These blocks would not allow providers (most often nurses) to order medication doses that exceeded typical thresholds. To circumvent the blocks, nurses simply ordered multiple doses of the same medication in order to obtain the full dose that they desired.
EHR-in-Use Alters Clinical Authority and Oversight: A hospital policy at an academic medical center required infectious disease (ID) fellows to review residents' CPOE orders of broad-spectrum antibiotics. However, no restrictions on the resident's ordering privileges were implemented in the system. In order to avoid the hassle of dealing with the ID fellows, some residents would resort to "stealth dosing," that is, waiting until the ID fellows went off duty to prescribe the restricted medications.
4. The physical or technical infrastructure interacts with the EHR (as used).
There is also a two-way interaction between the EHR (as used) and the IT and physical infrastructure. The example below highlights some of the problems that can emerge when attempting to integrate an EHR with other IT systems.
System Integration Problems: After implementing an EHR, a small hospital discovered that test results from an outside lab were not being loaded properly into the EHR. Lab results were being attached to the wrong patient records.
5. User reactions to EHR features may require redesign.
Finally, sometimes actual use of the EHR diverges so dramatically from the original design that it becomes necessary to reconfigure some EHR features. The next example illustrates some options for reconfiguring the system when "alert fatigue" sets in.
Responding to Alert Fatigue: EHRs often include decision support functionalities such as drug-drug interaction, drug-dose, drug-lab, and contraindication alerting. Several studies have identified "alert fatigue", that is, choosing to ignore alerts, as a common condition amongst clinicians using EHRs with decision support.
In the next section of Module III, we provide information and tools that can help you detect unintended consequences as they emerge. We outline how to develop and maintain an "Issues Log," a central repository of EHR-related problems.
Other Resources for Understanding Unintended Consequences: Several other researchers have proposed frameworks for understanding EHR-related unintended consequences, Including Sittig's sociotechnical model for studying health information technology in complex adaptive health care systems, Henriksen's Human Factors Framework, Vincent's Framework for analyzing risk and safety in clinical medicine, and Carayon's SEIPS model