New governance for a new era: issues and challenges for integrating systems.

HEALTH CARE BYTES

In this second part of a two-part column, Dr. Ruffin providers greater detail on seven key factors that he believes will govern the operations of integrated systems. Successful systems, he says, will be those that attract and retain physicians, have disciplined governance, integrate care through sharing of information among system elements, conserve capital, ensure strategic growth, control costs, and are proficient at processing information. It is important to understand, he says, that, in the movement from a fee-for-service payment mechanism, in which the various elements of the health care field bill for their services independently and according to rules designed for their benefit, to an integrated system, in which such independence can only lead to chaos, very substantial changes will be required in the governance of our health care institutions and organizations.

In August in the first part of this column, I identified and briefly described seven factors that I believe will define an integrated system's ability to successfully compete in the new marketplace health care services that is now unfolding. Because these factors are so critical to success, I want to expand on my earlier discussion.

1 Attract and Retain Physicians

Without physicians, the integrated delivery system (IDS) will not take care of patients, will not win contracts with employers, insurers, or health plans. The IDS cannot grow any faster than it can acquire physicians to treat patients. The IDS can do many things to attract physicians into its organization. Most physicians complete training with considerable debts to colleges and medical schools they attended. Most physicians look for a safe way to start seeing patients, without taking on more debt and risk to start a solo practice. The IDS can reduce a new physician's practice costs by including him or her in an existing group practice, giving him or her a salary with bonus for treating more patients than expected, and making available state-of-the-art information systems that he or she may have used in training but could not afford were he or she to start practice alone. The benefits of group practice - guaranteed minimum income, insulation from administrative hassles, camaraderie with partner physicians, access to sophisticated practice managers, more leverage in managed care contracting, ability to participate in capitated contracts with relatively little personal financial risk, sense of security and future prospects for practice growth, continuing clinical education, access to advanced information systems, including computer-based patient records and telecommunication networks linking physicians' offices and facilities, to name a few benefits - will appear increasingly more persuasive to physicians.

2 Disciplined Governance

A loose consortium of facilities that call themselves collectively an integrated delivery system is practicing deception. In a real integrated system, information processing and distribution are standardized throughout. Information systems and data dictionaries need to be standardized; business processes need to be standardized; and senior managers, physicians, and boards of directors of the various operating units need to cede some of their authority to senior managers, physicians, and board of directors of the parent organization. In practice, most emerging IDSs still receive as much, or more, income from fee-for-service operations as they do from capitation and other contracts that limit revenue to some multiple of covered lives.

During the transition from fee-for-service to capitation, member organizations of IDSs will suffer considerable ambivalence, recognizing the theoretical value of standardized ways of collecting, storing, communicating, retrieving, and analyzing data about the services they render and the outcomes of their patients, but yearning for the independence to establish whatever management and clinical practices that make the most sense to them, with whatever information and communication systems appeal most to them. They may choose to ignore the enormous boon to any economy that happens when standards emerge, and they may rationalize ignoring standards for practice guidelines and automated information systems because their needs are different from, and more urgent than, those of any other operating unit of the IDS.

The emerging role of governance of the system is to force independent operating units to adhere to standards where standards will allow the entire system to operate more effectively and efficiently. Standards are indispensable for integration of the electronic flow of information about health plan members and patients within the system, and they are essential if data about members and patients are to be collected in relational databases for systematic, retrospective analysis that will lead to steady practice improvement.

3 Integrate Care

We cannot integrate care without integrating information about the members and patients we serve and about the health care services we produce. Care is organized, and integrated, by the sharing of information. The home health agency needs to know about the diagnoses and treatments given to patients in hospital. Subspecialists need to know what primary care physicians want them to do in their consultations. Clinicians need to know the results of diagnostic studies performed in radiology centers and laboratories. Management needs to know about the services provided and the resources consumed throughout the organization and about yet unmet needs of patients, members, and providers if they are to allocate capital effectively and efficiently. Integrating care reduces unnecessary duplication of diagnostic tests, repetition of treatments that have been ineffectual in the past, and the time required for consultants to learn about the past medical history of patients on whom they are consulting.

Integrating care also helps to standardize care and the data collected about care, which greatly enhances opportunities for retrospective studies to reveal opportunities for quality improvement. If every physician uses his or her own idiosyncratic style in treating every patient, reliable data for clinical research will be hard to come by and relatively useless to study. But if clinicians in the IDS collect data about their patients and introduce treatments in a standardized way consistent with good clinical research discipline, observations made on patients' outcomes can be enormously informative to the IDS to help both clinicians and managers improve the processes of care for members and patients.

Integrating care includes many features, among them instituting standard means of collecting and storing data about members and patients; standardizing contracting so that every operating unit of the IDS works under a standard contract with any given payer; cultivating a standard clinical and managerial culture that includes all providers and managers of facilities; standardizing the data dictionary of electronic transaction systems so that procedure costs, formularies, and laboratory reports have the same consistent terms throughout the organization; and vesting in an information systems council for the IDS the power to dictate to operating units that they adhere to standards set by the organization for information and communication systems. Standardization of data and automation of data and information processing are key characteristics of growing, consolidating, competitive industries such as the health care industry is becoming.

4 Conserve Capital

No business can afford to squander capital for long. Since the 1960s, when generous indemnity health insurance was made available to most Americans, health care providers have enjoyed the ability to pass along their operating costs to consumers of their services (patients and insurers) and thereby guarantee profits year in and year out. Consequently, health care costs have risen faster than general inflation every year, until recently. As delivery systems compete for patients more aggressively and as insurers reduce the fees they pay providers, IDSs will need to husband their capital carefully. They will have major demands for capital in the near future, among them to recruit and retain physicians, to establish health plans and managed care insurance products, and to modernize their information and communication systems.

Most hospitals are not-for-profit and many IDSs are not-for-profit, so the capital they obtain comes from savings and debt. As for-profit managed care organizations grow and capture a larger share of the insurance market (currently 69 percent of health plans, covering 55 percent of enrollees, are for-profit), provider-sponsored IDSs will need all the sources of capital at their disposal. They don't pay taxes, but they also don't have access to equity capital. They need to concentrate on efficient operations and on saving money for that rainy day that may be just around the corner.

5 Growing Strategically

Strategic growth is a euphemism for smart growth, not the growth all-too-often seen among IDSs that have emerged from regional hospital systems that grew horizontally, acquiring community hospitals whose administrators were good buddies of the senior executives of the hospital-based IDS. Hospitals had capital to waste. They don't any longer, as more care is delivered outside the hospital. Hospitals need to save their money for alliances with physicians and for health plans and information systems. Technology is reducing the value of hospitals as warehouses of beds, and the increasing importance of efficient health care processes is raising the value to IDSs of alliances with physicians committed to their financial success and growth. The IDS needs the capital to invest in information and communication systems that will enhance the efficiency of practice for physicians and in health plans that will help physicians market their services more successfully.

Strategic growth means growing into business activities not commonly associated with hospitals - health plans, group practices, and networks of participating physicians. The cultural changes associated with this shift are difficult for many hospital administrators and board members to accomplish. They wish they could go on forever acquiring and managing hospitals and dealing with hospital people. But they can't. They need detailed, long-range strategic plans that say more than, "Do good, bill for all, and keep everybody happy." They need to specify how, and in what ventures, they will invest their capital, even if those statements alienate portions of the medical staff or other community leaders, and reduce the significance of the hospital in the larger IDS.

6 Control Costs

Controlling costs of health care services is the only means of generating revenue in excess of costs (profits) available to providers and IDSs that receive most of their income from capitation. Group- and staff-model HMOs have lived with this reality for decades, but now IDSs emerging from regional hospital systems receive most of their revenue, and most of their profits, from managed care contracts. In fact, the parts of those organizations that are growing most quickly are their home health agencies and their health plans, and the health plans return income to the provider portions of the organizations with capitation. Primary care capitation has become common in most urban areas, and group practices are becoming accustomed to carve-outs and fixed price contracts more quickly than hospitals. Controlling variable costs means controlling the costs and volume of supplies and services rendered to patients. Controlling fixed costs means controlling the costs of capital, insurance, and other costs of doing business associated with the location of services rendered and the overhead of the organization providing those services.

In the long run, all costs are variable and can be changed. In the short ran of one year or less, many costs appear to be fixed. Most of the costs of care are direct clinical costs. Consequently, systematic clinical data analysis for quality improvement studies may reveal many opportunities to reduce clinical costs. For most common ailments, there is a wide variation in the processes of care used by clinicians, with attendant wide variation in the costs of that care, without nearly as dramatic variation in clinical outcomes and functional status of patients. I believe clinical data must be adjusted for patients' prior probability of outcomes of most interest to researchers (and charges, or costs of care, are among the outcomes most frequently studied). Once such an adjustment is made, the efficiency and effectiveness of clinicians' practice habits can be compared.

Risk adjustment is not a perfect science, but useful observations can still be made with less than perfect data. We highly recommend that analytical staff skilled in clinical risk adjustment, and familiar with clinical practice, select the risk-adjustment methods to which patients' data are subjected and obtain approval for the model used by clinicians before the data at hand are adjusted for the prior probability of the outcome in question. After a risk-adjustment methodology is accepted by clinicians, they are more likely to accept the adjusted results as valid and more likely to participate in quality improvement programs based on the results of the studies.

7 Be Proficient at Processing Information

Proficiency in processing information is a goal to which all organizations aspire. They all strive to move information from place to place in the organization as quickly and accurately as possible, to reduce their own operating costs and to improve services to customers. In integrated delivery systems, information proficiency implies successful movement of administrative and clinical information about patients. Most health care organizations invest resources in automation of administrative information first, because processing claims accurately and quickly reduces the age and total value of funds in their accounts receivable. Automation helps to capture more of the charges than manual processing of services usually identify and helps to reduce the labor costs of collecting, storing, retrieving, communicating, and managing data about charges.

Now, many insurers demand that claims from providers, especially hospitals, be submitted electronically to expedite payment and reduce the costs to insurers of manual entry of claims into their claims processing systems. Automation of clinical information is common in clinical laboratories and is becoming common in radiology departments. Results are distributed to clinicians sooner and automation allows for more complete accounting of services rendered to patients, so providers can produce more accurate, and larger, bills for services rendered. Automation of medical records is not common except where transcription is performed by word processing systems. In such systems, text transcribed from clinicians' dictations is moved electronically from the word processing system into a patient care information system that clinicians can use to retrieve clinical results from workstations on the nursing units of hospitals and in the offices of group practices.

There are not yet strong economic incentives for providers to invest in automated medical record systems for clinicians to use in their daily care of patients. Technologies for data collection are not efficient enough for physicians to abandon pen and paper, or dictation, to capture their clinical observations about patients. Nevertheless, the day is dawning when clinicians will prefer workstations with voice recognition for entering their clinical observations about patients and hand-held, back-lit, portable computers for retrieval of results and entry of orders and physical findings on rounds. That day must be preceded by considerable investment in capital and new governance initiatives among the operating units of IDSs.

For instance, before a computer-based patient record can be shared electronically among the many locations of care in an IDS, those locations must have adopted a master patient indexing scheme (standardized patient identification process) and have given patients identification and medical record numbers with a common methodology. The operating units must share a common data dictionary, so that data elements in the electronic records mean the same thing in every facility. The transaction systems used in the various operating units need to have the same look and feel, so clinicians moving from one facility to another can use the computer-based patient records wherever they go. The data collected by these standardized transaction systems should be copied to a relational data repository (data warehouse) where clinicians and managers can access them easily to compare the processes and outcomes of care of the various operating units, looking for variation in outcomes that would suggest opportunities for quality improvement. The leaders of the organization need instruction in medical informatics, so they know the implications of these automated information systems and how to analyze data to improve the organization's processes and outcomes of care.

Most health care organizations, including IDSs, will acquire two computer-based patient record technologies - one for on-line transaction processing (OLTP) of physicians' orders and laboratory results, scheduling, and medical records maintenance, and the other for on-line analytical processing (OLAP), with a data model and software tools meant to make retrospective health services and clinical research as efficient and comprehensive as possible. The OLTP system uses a normalized relational database designed for rapid (1-2 second) response times for every change of screen for hundreds, if not thousands, of simultaneous users. The OLAP system uses a denormalized relational database of relatively few relational tables and software designed to make iterative, ad hoc data analysis efficient from users' point of view. The OLAP system is designed to support epidemiology and public health studies, which, in turn, form the intellectual buttress of continuous clinical quality improvement studies.

As information flows, so goes the organization. If information does not flow smoothly from the central part of the organization to a peripheral operating unit, say a group practice or hospital, the peripheral unit really is not an integral part of the organization. If each operating unit (hospital, group practice, HMO, home health agency, urgent care center) maintains its own financial and clinical information systems, and information is conveyed between operating units on paper, the organization probably does not behave like an integrated system, but rather behaves like a loose confederation of otherwise independent entities. On the other hand, if the operating units have invested the time and effort and capital to standardize their operations and their information systems, information flows from one operating unit to another seamlessly, clinicians in one location can obtain computer-based patient records of patients treated in any other location, and financial transactions are integrated so that patients paying fee-for-service receive one integrated bill for all services (facility fees and professional fees) from all providers, the governing body of the organization has created a true integrated delivery system.

Unfortunately, most organizations that aspire, or claim, to be integrated delivery systems are not integrated at all when it comes to the flow of information between their various locations and operating units. Until clinical care of patients is perceived to be a continuum over all the primary care physicians, subspecialists, ancillary services, and facilities of the organization, the organization will appear to be a collection of independent practitioners organized to produce procedures at single points in time and not to manage the welfare of a population of patients over time in the most efficient manner possible. The latter organization is our stated goal, but we persist in protecting our older models of organization. We are changing, slowly. Information technologies we could use to expedite the transformation of our organizations to integrated delivery systems are improving much faster than are our habits of governance and organization.

Marshall Ruffin, MD, MBA, MHA, FACPE, is President and CEO of The Informatics Institute. Until he started the Institute, he was Clinical Information Officer, INOVA Health System, Falls Church, Va. He continues to consult to the Informatics Department of that organization.