Health care delivery is arguably the most complex industry in existence.

Multiple factors drive this complexity. Health care delivery is a socioeconomic good provided by organizations that have a societal and community mission and traditional business obligations, such as generating a positive operating margin and providing high-quality products and services. As a socioeconomic good, however, strategy and investment decisions for health care organizations have a more multifaceted set of decision criteria than for a traditional business.

Medicine, too, is a complex and rapidly evolving knowledge domain. Last year, according to the National Library of Medicine, approximately 800,000 articles were added to the referred biomedical literature. Ten years ago, that number was slightly less than 400,000. A Learned Publishing article estimates that 50 million articles have been published since formal research began.

If you view patient care as a manufacturing process, you'll find that a typical hospital carries out hundreds of unique processes that are variably performed. These processes "produce" outputs for which the quality of the output can be difficult to measure. And the processes are performed on "inputs" or patients who exhibit significant variability in symptoms and histories. Therefore, hospitals execute an exceptionally complex manufacturing process that has highly variable inputs, processes and outputs.

Finally, the structure of the health care delivery ecosystem requires coordination among a wide range of types of organizations. These organizations provide acute and long-term care services, or serve patients through physician practices, public health agencies, diverse purchasers of care and government entities.

Given all of the above, it should not surprise us that Peter Drucker wrote that the hospital is one of the most complex organizations ever created.

The Primary Value Proposition

Information technology has transformative power. It can accelerate processes and make them less error-prone and more efficient. It can offer new services that overcome distance, time and the need for a physical structure, such as a storefront. It can deliver information instantly and in novel ways to decision-makers. And it can run algorithms to monitor equipment and correct minor problems before they become major problems.

Sophisticated analytics have enabled us to predict the weather, launch satellites that visit remote planets, capitalize on split-second opportunities in the financial markets and model the actions of proteins.

We see this power in all facets of our lives. And, yet, if you step back and ask, "What is the most fundamental value proposition of information technology?" you could very well answer, "It enables us to conquer complexity."

At times, information technology is used to uncover complexity, such as determining the particle composition of the universe. At other times, information technology is used to create or enable a complexity that is valuable and so user-friendly that you can use your smartphone to find the nearest fancy restaurant. And still, at other times, the technology is used to master complexity or make complexity more manageable; for example, searching the literature more efficiently.

In health care, electronic health records enable providers to manage the complexity of a patient's care more efficiently over multiple venues, time and conditions. EHRs do so by allowing clinicians to retrieve and exchange their patients' data easily, thereby helping the clinicians to make optimal clinical decisions, providing workflow technologies to monitor and assess care decisions and processes, and capturing data that can be analyzed to assess care patterns.

On the administrative side, revenue-cycle systems help a provider to manage an increasingly bewildering array of payment schemes, purchaser contracts and medical necessity rules.

Mastering Three Sources of Complexity

The fundamental value of health care information technology is to tame, as well as we can, the complexity that characterizes the delivery system. This "taming" or mastering should enable a reduction in care costs, a more consistent delivery of care according to the medical evidence, a reduction in errors associated with errant clinical processes, and superior service for patients and clinicians.

There are three fundamental sources of complexity in health care: data, processes and medical knowledge. Information technology can reduce or manage complexity in all three sources.

Data. A patient's medical record can contain thousands of different data elements in a wide range of types (such as numbers, text, images and videos). These data may be plagued by different coding schemes, suffer from problems of inaccuracy and incompleteness, and be distributed across dozens of organizations. There is no customer record as complex as the medical record.

Multiple strategies are under way, many of these led by the federal government, to reduce the complexity of patient data.  For example, data and transaction standards are under development, and their use is being enforced.  In addition, health information exchanges, or marketplaces, are being introduced to create more complete sets of patient data.

New methods are available in which the computer assists the caregiver by highlighting the data that are likely to be relevant, given the patient's history, current complaints, and evidence-based treatment guidelines and protocols. These methods help prevent caregivers from becoming overwhelmed by the sheer volume of data.

Also, we can anticipate advanced techniques that use pattern recognition and machine learning to correct deficiencies in the data; for example, noting that the patient's data indicate that he or she is a diabetic even though a diabetes entry isn't in the problem list. Methods that identify patterns in the data, such as a predictive algorithm that indicates a high readmission risk, are already in use. These methods enable clinicians to focus on the pattern and not the underlying data.

These data efforts fall into two overall approaches:

  • reducing data complexity through standardizing data, correcting problems of data inconsistency, and enhancing data completeness through health information exchange;
  • hiding complexity by highlighting relevant patient data and offering algorithms that guide clinical decisions, and by using methods to identify data patterns.

Processes. At times, there isn't adequate clinical evidence to guide the care that a patient receives. At other times, there is evidence, but it is not followed. Either way, the result is high variability in care processes.

Care processes invariably traverse departments within a hospital or among specialists in the treatment of a patient with multiple chronic diseases or multiple care venues; this occurs, for example, with surgical procedures that require rehabilitation. Since these nodes do not always operate in the same way, and processes that traverse are often ad hoc and idiosyncratic, complexity is inevitable.

There are several IT-based strategies for addressing process complexity.

Improving data liquidity through health information exchange enables each node to be aware of the care given and the results obtained by all parties involved in a patient's care. A fundamental aspect of good care coordination is that everyone involved in a patient's care is aware of the others' actions.

In addition to data accessibility, someone — often the nurse on an inpatient unit — must choreograph care processes. Information technology can provide orchestration support.

Workflow engines can monitor care processes and alert caregivers if the steps in a process are occurring in a suboptimal sequence, or if the interval between steps is too long or steps have been skipped. Joint care plan development and documentation also can assist team-based orchestration of care. Personal health records enable the patient to be more effective as a team member and orchestrate aspects of his or her own care.

These efforts help to address complexity by informing caregivers when a process has crossed a boundary that is deemed undesirable or unacceptable. For example, a workflow engine can alert caregivers that no one has followed up on an abnormal result. It's like the indicators on your car dashboard that inform you of impending trouble.

In a similar way, IT-based support of care teams does not remove or hide complexity, but it does enable caregivers to identify complex processes that are no longer on track.

Medical knowledge. For the caregiver, there is too much to know, and the body of knowledge changes and expands too frequently. Previous efforts to address this problem, such as creating subspecialties, have proven unsustainable and often simply move the complexity problem rather than solve it. For example, while subspecialization may enable a person to master an increasingly narrow body of medical knowledge, the result is increased care fragmentation. The knowledge problem has become a process problem.

There are several strategies that can be applied to address the medical knowledge complexity problem.

Clinical decision support can guide the caregiver's diagnostic and treatment decisions. This support can use computerized provider order entry by, for example, reminding the clinician to perform a genetic test before ordering a particular chemotherapeutic regime. Clinical decision support can guide documentation by highlighting the data that should be gathered to conduct a thorough health maintenance assessment.

Several multifaceted "big data" efforts are under way. Some of these mine EHR data to identify patient characteristics that call for different treatment approaches. Other efforts bring together diverse types of data — EHR, molecular medicine and radiology images — to determine if novel combinations can lead to highly sensitive and specific diagnostic and treatment decision aids.

These efforts generally seek to shift the complex problem of knowledge acquisition, delivery and synthesis from the caregiver to the machine.

Architecture Is Fundamental

Improving the quality, safety and efficiency of health care requires that we address the complexity of care delivery. This complexity has many sources. However, patient data, care processes and medical knowledge are fundamental sources.

Information technology has core capabilities that enable us to uncover, manage and create usable complexity. However, in health care today, our primary interest is in managing existing complexity. This management can take the form of hiding complexity, removing complexity, alerting caregivers if complex processes deviate in unacceptable ways and shifting complexity management from the person to the machine.

The necessary core capabilities of IT are the exchange of health information; rules and workflow engines; algorithms and decision aids; and the support of team-based care. These capabilities are not specific EHR features and functions. They are capabilities that are present (or not) in EHRs and that form their architectural foundation. Hence, a broad range of application features and functions can make use of these capabilities.

A Caveat for Leaders

While we all recognize the importance of the technology, we cannot achieve the goal of complexity management by implementing health care information technology alone. Multiple other clinical and managerial interventions must accompany the deployment of the technology.

Clinical and operational processes must be re-engineered and standardized so that activities that only make things more complex and have no value are removed. Reward systems (such as reimbursement reform) are needed that provide the incentives and revenue to offset the costs of technology and re-engineering.

Finally, leaders must motivate and guide efforts to change how the organization functions. Complexity management is critical to effecting material improvement in health care. And information technology is an essential contributor to this management.

John Glaser, Ph.D., is the CEO of Siemens Healthcare's Health Services in Malvern, Pa. He is also a regular contributor to H&HN Daily.