Overview
In the twenty-first century, perioperative information systems will not only provide answers to clinical questions in near real-time, they will also anticipate these questions. James M, Berry, MD
This vision of technology may soon be realized by the careful development and deployment of medical informatics systems in the perioperative environment. Informatics systems hold the potential to transform operating room (OR) management and produce significant improvements in both efficiency and patient safety. Real-time access to both current patient data and historic patterns make the implementation of quality assurance, practice improvement, and safety initiatives both simple and transparent.
Compared with other medical or nonmedical environments, the density and complexity found in the OR make technology implementation much more challenging. The need to frequently move equipment and reconfigure the operating room for different procedures makes even simple connectivity an ongoing challenge. Also, the high-noise environment makes alarm sounds or other audio notification problematic. The numerous, often incompatible electronic devices used produce electrical noise, cross-talk, and other interference. Connecting physically separate devices—such as monitors, towers, input devices, and storage—requires tangles of highly specific, and often poorly shielded, data cables ( Fig. 22-1 ).
Data
Data Gathering
The foundation of any information system is the ability of the system to gather accurate and timely data. Data may come from a variety of sources, either primary, data captured directly from the patient, or secondary, preexisting data retrieved from other systems and processed into a more specialized database; perioperative data management systems would be a robust example of such a secondary data system. The importance of integrating this data cannot be underestimated; advanced decision support systems will require not only current data but also past data from the patient’s history, comparative data from other patients, and best-practice rules to formulate recommendations for clinicians.
Data Sources
Patient vital signs and ventilatory parameters are fundamental to successful perioperative management, because they serve as objective and quantitative measures of the patient’s condition and its evolution over time. Because these data serve as the foundation for most perioperative documentation and clinical decision making in all phases of care—preoperative, intraoperative, and postoperative—they must be readily accessible to all clinicians, including surgical, nursing, and anesthesiology providers. Complicating this processing is the sheer volume of data and the multiple, complex networks from which data are derived. For example, a modern operating suite and the associated critical care units may collect 1 gigabyte, the contents of an encyclopedia, daily from vital signs alone.
The problem of data volume exists in many professional arenas, not just in medicine, and new software technologies and concepts are helping to resolve this issue. Beyond basic search engines, modern tools such as The Brain ( www.thebrain.com ) or Wolfram Alpha ( www.wolframalpha.com ) can analyze a dataset, draw inferences, extract data from other sources, and present artificial intelligence–filtered data to the user. In this way, the computer serves not only as a tool to store and retrieve information but also as a tool to intelligently guide users toward the information they are seeking. Taking a similar approach in medicine, in the future, information systems will be able to examine a constellation of findings in an electronic health record and predict, based on past patterns, likely outcomes and potential morbidity. Additionally, similar algorithms will track numerous care plans and ensure that they are followed to maximize efficacy and minimize morbidity.
Monitor Interfaces
Physiologic monitors come in a variety of designs from a handful of manufacturers. Although they may differ cosmetically and provide similar functionality, data output from monitors is not always standardized. Although RS232 serial output, which uses a 9- or 25-pin connector dating from the 1980s ( www.interfacebus.com/RS232_Pinout.html ) is still found on some monitor interfaces, RS422 or universal serial bus (USB) output is capable of much higher bandwidth and thus higher data transmission rates ( Fig. 22-2 ).
Also, monitors tend to be both expensive and long lived (10 years typically) in most inpatient environments. This may result in an institution having a mix of physiologic monitors from multiple vendors, with multiple software revisions or hardware generations, many of which may have different physical data interfaces and different interface “languages.”
Historically, networked monitors have had proprietary physical connectors that required specialized hard wiring from all the locations in which monitoring occurs. This created a complex network of single-use cabling that would have only one purpose and an investment that would only be viable for a short period of time. Fortunately, most newer generation monitors use not only a standard network cabling infrastructure for hardware connections (RJ-45 Ethernet cabling and jacks; Fig. 22-3 ) but also are able to transmit data over a standardized Ethernet network.
In this age of health information privacy, a hospital network is required to have significant security measures in place. Virtual local area network (VLAN) or virtual private network (VPN) structures, as well as encryption of wireless traffic, help ensure that network traffic for patient data is segregated and secure. Also, a guaranteed minimum bandwidth available for real-time transmission of critical data and alerts promotes confidence and trust among clinicians that the system is robust, even under stress.
Modern patient monitoring networks typically have intermediate data aggregation points. Data may be stored, transformed, or filtered at these locations. For example, data transmitted from the Philips MP90 bedside physiologic monitors pass through two intermediaries prior to being transmitted to external systems. First, data enter a central station that stores, for 24 hours, all individual data points and waveforms for analysis. Clinicians are able to review, compare, or print data at this central station. The central station may also generate alerts for arrhythmias, hypotension, apnea, and so on to monitoring technicians or nurses. From this hub, data are then transferred to an information server, where the information is both archived and transformed into an externally recognized format for use in other systems ( Fig. 22-4 ).
Physiologic monitoring is but one of a variety of data sources in the perioperative environment. Additional data are required for proper documentation of care and to feed clinical decision-making and decision-support algorithms. Some of these data, such as demographic and laboratory results, are structured in a standard format, but other data are stored as images or in other, nonstructured formats. Demographic databases include patients’ nonmedical data: name, date of birth, address, medical record number or other identifiers, diagnoses, and billing information. Wide access to these data is critical to ensure that documentation is timely and accurate, with minimal need for reentry. Rekeying of demographic and identifying data is both time consuming and subject to operator error, and it should be avoided whenever possible. Careful interfacing can bring the previously entered and validated identifying data to the clinician at the point of care.
Networking Concepts and Open System Interconnection
Communicating data across distances requires certain agreements on protocol and format. From the early evolutionary steps of what we now call the Internet, a system of layers of information and protocol developed that may be visualized as parallel systems of information transport, each serving a particular function. In the open system interconnection (OSI) reference model, the most fundamental layer is physical—literally, the wires connecting the computers, servers, and switches. The highest layer is the application, or the way data are presented on a screen ( Table 22-1 ).
| OSI Model | |||
|---|---|---|---|
| Data Unit | Layer | Function | |
| Host layers | Data | 7. Application | Network process to application |
| 6. Presentation | Data representation, encryption, and decryption | ||
| 5. Session | Interhost communication | ||
| Segments | 4. Transport | End-to-end connections and reliability; flow control | |
| Media layers | Packet | 3. Network | Path determination and logical addressing |
| Frame | 2. Data link | Physical addressing | |
| Bit | 1. Physical | Media, signal, and binary transmission | |
∗ The OSI model was designed to standardize communications functions. Each layer provides services and data to the layer above while receiving information from the layer below.
Interface Languages
Standardized systems’ interface languages ensure the clinician will have easy access to information and save time and effort in formulating treatment plans. For example, the digital imaging and communications in medicine (DICOM) format was developed to ensure interoperability of systems that produce, store, display, process, transmit, or print medical images and structured documents. Integration of common data pipelines, using a common language and interface and shared data validation and security, enables images from various modalities—X-ray, ultrasound, medical photography, and electronic documents—to be freely transferred intact among systems and presented to clinicians in widely accepted visual formats. The Health Level 7 (HL7) standard was developed to support the exchange, integration, and retrieval of health information. HL7 provides a communications messaging model at the application level such that systems can communicate and interoperate without manual intervention at each step. Standards such as these enable different manufacturers’ products to work together in a clinical environment. From the point of view of the clinician, the underlying standard and processing may be transparent, but the displayed results will not be ( Table 22-2 ).
| Code | Description |
|---|---|
| MSH | Message header |
| EVN | Event type |
| PID | Patient identification |
| ZRB | HIS base registration |
| ZP2 | HIS patient information |
| PV1 | Patient visit |
| ZV1 | HIS patient visit |
| ZEN | HIS encounter |
| {ZVP} | HIS patient provider |
| {DG1} | Diagnosis |
| ZDX | HIS diagnosis |
| {PR1} | Procedure |
| ZPR | HIS procedure |
| [ZPN] | HIS PHN |
| {[ZDN]} | HIS dental |
| [ZDP] | HIS dental op |
| {[ZIM]} | HIS immunization |
| {[ZMD]} | HIS medication |
| {[OBX]} | Health factors |
| {[OBX]} | Measurements |
| {[OBX]} | Exams |
| {[OBX]} | CPT |
| {[OBX]} | Labs |
| {[OBX]} | Patient education |
| {[OBX]} | Skin tests |
∗ Using a standard message structure improves interoperability among devices and decreases provider workload by eliminating the need to reenter data into multiple applications.
In addition to traditional sources of patient data, such as monitors and lab systems, modern informatics includes several nontraditional modalities of data collection. Medical imaging and video has become pervasive throughout the practice of medicine, especially in procedural or operative areas. Imaging can be acquired from a variety of sources that include handheld digital cameras, still cameras incorporated into medical devices such as endoscopes, moving video images from surgical instruments (endoscopy, overhead cameras built into lighting fixtures), and other sources. Informatics systems need to be able to capture these images, in real time and only when appropriate, for incorporation into the medical record. Subsequent review of the records for further diagnosis and treatment can then be enhanced with the knowledge provided by having the actual photograph or video loop presented as part of the documentation of the procedure.
Similar to existing telemedicine applications, such as remote consultations between trauma centers and community hospitals, video feeds have also found utility in management of the perioperative suite. Management can take place at any of several levels. Individual anesthesiologists may provide medical services to patients in a group of two to four ORs in a supervisory role, and charge nurses and anesthesiology coordinators may oversee the flow of patients through entire suites or multiple suites in a large hospital. Video technology, such as that provided through the Vigilance system (Acuitec, Birmingham, AL), provides improved situational awareness of ORs and other perioperative locations from non–point-of-care locations such as OR front desks, offices, preoperative holding rooms, and other OR areas. This system integrates real-time video from ceiling-mounted cameras and presents that video in conjunction with other data derived from information systems to produce a single-screen view of one or more ORs. This view is accessible from a fixed workstation ( Fig. 22-5 ), handheld computer, or even an Internet-connected cellular phone, such as an iPhone ( Fig. 22-6 ).
Additionally, critical voice communications can be facilitated by an informatics system. Voice can either be built in using voice over Internet protocol (VoIP) or information systems that facilitate traditional voice communications by providing phone numbers on the screen or by pages to digital paging systems. Together, these technologies are in the process of revolutionizing real-time voice communication.
Central Storage and Servers
Complex information systems, such as those used in a perioperative setting, generate large datasets. After several years of operation, a large hospital may find itself with over a billion data points comprising a database that spans more than one terabyte of data. Compounding the data storage situation is the potential need to keep multiple copies of the database either for back-up or research purposes. In an effort to protect patient privacy, research datasets are often required to be deidentified and stored in such a manner that there can be no linkage from the research dataset to the original data. The U.S. Health Insurance Portability and Accountability Act of 1996 (HIPAA) provides a good standard rule set for data privacy. In addition to removing key patient identifying information, the informatics system may be required to keep a research dataset on a physically separate server to minimize the risk of unintended release of protected patient information. These factors make data storage and security a crucial issue.
Data storage and security are often unseen components of an information system. Data may be stored within an institution or at an external facility that specializes in data management and security. Through the technology of networking, data can flow seamlessly across miles or hundreds of miles of physical distance. The three hallmarks of data storage are 1) capacity, 2) security, and 3) availability. Each of these concepts is important to address, and all are interrelated.
Central storage capacity is fundamentally the available space on the physical devices, typically fixed hard-drive storage, and the configuration of those devices. Storage can either be on a server dedicated to storage or on a mixed-use server that stores data and also runs applications, called processes, which manipulate data or provide the data to applications that use it, called clients . This client/server architecture forms the basis of many medical systems ( Fig. 22-7 ).
