The politics of informatics.

The politics of informatics may not appear to be crucial to what would seem to be technical procurements, such as those for transaction systems, communication networks, or data warehouses for retrospective analysis, but they determine the eventual success of those ventures in automation. Nowadays, providers are aggregating into larger organizations to take care of populations of people over time, and they want to share data about their patients electronically, across political lines of group practices, hospitals, and other ambulatory treatment facilities, and between departments within group practices and hospitals. Moving information about patients electronically takes far more standardization of data across those organizational boundaries than does the movement of paper records.

People are good at pattern recognition and can glean the information they need from paper records of all shapes and sizes, from text and images formatted in a wide variety of ways. Computers and communication networks fail at pattern recognition. They move information quickly, at minuscule marginal cost, provided the data are standardized and formatted exactly from one computer to another. Without that excruciatingly exact formatting, data cannot flow electronically from the laboratory system of a hospital to a communication network and into the computers in the offices of physicians. And those data will only flow electronically into the offices of physicians who have installed specific software, formatted just so, on their office computers to allow them to retrieve those data.

Frankly, it is amazing to me that technologies that move electrons and photons from place to place work at all, given the insight and flexibility inherent in people and absent in data processing equipment. So, because our health care communities want to move information about patients from one location of care to another electronically, standardization of those systems is necessary, and standardization introduces politics, across departments within institutions and between institutions.

What has brought politics into the process of selecting systems is that our interest in and need and capability for electronic communication of peoples' administrative and medical records within organizations, and between organizations, have grown substantially. We have discovered networks. Electronic communication networks permit movement of data electronically from computer to computer, regardless of the distance between them, with incredibly low marginal cost, but with high start-up costs of standardization of data dictionaries, data formatting, network hardware and software, computer operating systems, and application programs. Because we tend to associate human capabilities with computers that do some things so much better than we can, such as sorting huge numbers of patients by name in seconds, we tend to underestimate the human effort required to set up computers to perform useful work, especially when more than one computer must share data with others across communication networks.

Politics tend to come into play when substantial resources must be allocated. Politics and economics are inextricably entangled. Within an institution with an influential director, politics may appear quiescent. The director considers rational arguments for this or that project and approves some and does not approve others. Life goes on, without interminable committee meetings in which people try to persuade others to their points of view about proper allocation of resources. A private practice of a single physician, or a small group practice with a single owner, may appear to be free of political wrangling over funds.

Even a large group practice, hospital, or health plan with a dominant leader and a clear mission and goals may appear free of political tension. In those organizations, computer technologies may have been introduced in the past for financial accounting without dissension, and they probably work well. Even departmental clinical systems may work well. They perform their automated information management in specific departments--radiology, pharmacy, and laboratory--and produce paper reports that fill paper medical records.

A medical community including a solo physician's office, a small single-specialty group practice, a larger multispecialty group practice, and a hospital probably move patient data between them with paper records. Each of these provider organizations electronically calculates its bills for services it has performed in its own patient accounting system, and each produces claims forms that its patient accounting system prints on paper to mail to the health plans used by its patients. The hospital and larger group practice may send claims electronically to some payers, from their computers to the computers of those insurers over standard telephone lines. Provider organizations do not share financial risk for the care of those patients. They rely on fax transmissions of small documents, and couriers for large medical recording, to move information about patients quickly;, from location to location.

Into this happy scene, introduce managed care and financial risk sharing among provider organizations. Now the pressure mounts among physicians and hospitals for cost control. Information systems remain departmental within the hospital and group practices, and paper only is exchanged between them. Each organization reduces staff, and each saves money wherever it can by purchasing supplies and durable goods with more attention to their costs than they did before. And those initiatives work for a few years, while fees for services rise annually less than they did in the past. The physicians' practices and the hospital may contemplate sharing financial risk for capitated covered lives, but they retain financial independence, and each maintains its own patient accounting and financial management information systems. There may be tough negotiations between hospital and physicians over how to divvy up the pot of money they receive monthly from the health plan with whom they share financial risk, but they can handle those negotiations using reports from their financial systems to guide them.

At the same time, however, health plans demand more data from the physicians' practices about the screening procedures they perform on their patients, and the health plans and state regulatory agencies demand more clinical data from the hospital about the treatment and outcomes of its patients. Those data collection efforts require expensive manual chart review. Add to this simmering mixture of requirements and financial pressures technological change that permits computer-based patient records for immediate patient care and retrospective health services research to identify opportunities for quality improvement and cost control, and the physicians' practices and the hospital begin to contemplate automating their clinical record systems.. While the physicians' practices and hospital would not contemplate integrating their financial systems until they merge their organizations, they will consider sharing computer-based patient records. Enter politics.

Sharing electronic information between divisions of a single organization is difficult to accomplish, and nearly impossible between organizations, without politics. In fact, politics enables the process of information systems planning and procurement for electronic communication across institutional boundaries. Ignore politics and doom the process to frustration and failure. Recognize politics, prepare for give and take about who pays for what in order to share data electronically across institutional boundaries, and the process of governance that necessarily must precede the process of technical selection and implementation can start.

Take the solo practice, the small single specialty group, the larger multispecialty group, and the hospital to which we referred before. Imagine that someone in each organization simultaneously decides a computer-based patient record (CPR) would make patient care aild quality improvement studies more efficient and effective. Interest is gained, and each organization starts to define its requirements and to select a system.

The solo practice selects a system based on the MS-DOS operating system and creates its own database of clinical findings and treatments it will automate electronically. The small single specialty group selects a system based on the UNIX operating system and creates its own database of clinical findings and treatments to automate, formatting data elements any clinician would collect in a different way from the solo physician and collecting many data elements not collected in the solo physician's system. The larger multispecialty group practice selects another vendor's system, using the VAX operating system, and derives a third formatting scheme for common data elements any clinicians would collect and many data elements unique to its practice. The hospital selects a fourth system, based on an IBM mainframe computer with a fourth operating system, and formats data in a fourth way.

Now each organization has a computer-based patient record, but they cannot share data electronically between them. Paper must continue to flow between their organizations to convey clinical, and financial data, and each organization must pay part of the cost of duplicate data entry when data elements collected electronically in one organization, and then printed to paper, must be re-entered electronically in a second organization to have those data in its CPR.

The best example involves laboratory results. For a physician to keep laboratory results in his office CPR that were obtained and stored electronically in the hospital, he or she must pay to have those data entered manually. For the hospital to include laboratory results obtained in the physician's office prior to admission in its CPR, they must be entered by hand in the hospital. Electronic interfaces between incompatible CPR systems are possible, but the wide variety of CPR systems available for office practices makes it highly impractical for a hospital to plan to write interfaces to all of those selected by physicians' practices without trying to standardize which systems they select. One of the principal reasons for medical service organizations to exist is to help standardize office practice CPR systems for physicians' offices, to permit them to share clinical data about patients electronically.

When organizations that want to share data electronically recognize that they cannot tell each other which information systems to select, or which data elements to put into those systems, but must plan the standards for those systems together to avoid hideous costs of retrofitting and interfacing incompatible systems, early recognition and introduction of a political process to resolve issues about who pays for standardization become paramount. If two group practices share one vendor's CPR system and want to share electronic data with a third practice, who pays for the retrofitting of data dictionaries and electronic interfaces between systems? Data elements must be standardized and electronic bridges must be written to convey data electronically between them. The trauma service of the hospital wants to acquire an electronic trauma registry to win accreditation as a regional trauma center. Who manages and funds standardization of the trauma registry so the data can flow electronically into it from the hospitals' main CPR system and from the trauma registry into the hospital CPR? You might say the hospital does, since the trauma center is part of the hospital, but the trauma center may have funding from external sources and may resist having to standardize its system with the hospital if a less expensive solution would satisfy the accrediting organization.

What process would you use to get multiple hospitals, or multiple group practices, or both that just recently decided to manage populations of patients under a risk contract with a major insurer to standardize their CPR systems so that records from one organization are available electronically at workstations at another organization? If the CPR system of a hospital is perfectly adequate for that hospital, but needs substantial modification to accept data electronically from a partner hospital or from group practices that have recently joined the integrating delivery system those organizations are forming, who pays for the modifications?

As a general rule, leaving funding for standardization up to the various operating units that express interest in sharing computer-based patient records dooms the planning project to failure or to accomplishments far short of expectations. Operating units, be they group practices or hospitals, will find countless excuses not to fund the standardization required to make shared CPR systems function successfully. They will frustrate adoption of standards in the first place by posturing to shift the burden of costs to other operating units. Each organization will argue for the adoption of standards for data elements, operating systems, application programs, and hardware that are closest to its own and that will require the least investment by it to accommodate.

So, are provider organizations that are not economically integrated doomed to failure when they consider planning for CPR systems in common? No, but they do have to address the politics of funding standardization first. They need to meet and agree on the general scope of their efforts. Are they agreed that they want to implement shared CPR systems that permit them to move data from one location of care to another electronically? Do they want to have most, if not all, clinical information about patients available electronically from all locations of care in their integrating delivery system? If they do, they need to create a planning organization of clinicians and managers from all the organizations to establish goals and standards for a shared system. Their work may take three to six months of periodic retreats, with staff work in between, to define their requirements.

The staff also needs to inventory existing computer systems, including communication networks, in all organizations to predict the costs of modifying or replacing them in order to collect and move standardized data from location to location. Senior executives of participating organizations need to understand the likely costs involved and agree to share the costs of standardization to minimize the otherwise interminable haggling over standards.

One standard that has exploded onto the world scene will make planning for shared CPR systems less painful. The standards for creation, display, and communication of documents on the World Wide Web (WWW) of the Internet, using hypertext markup language (HTML) and its successors, will enable health care organizations to share clinical information about patients much more easily than they have until now, as they have tried to standardize on proprietary vendors' software. The WWW has established a set of standards for movement of data over communication networks, and many corporations are using them to define the ways in which proprietary documents will move electronically between locations connected by their corporate networks. Functions to support data entry forms and multimedia database searching from workstations on the WWW appear and improve daily.

In the near future, international standards for community-based medical records, with acceptable security protections incorporated in the networks that convey the data, will be based on Internet standards. In the meanwhile, planning for CPR systems needs to involve multiple organizations and painful standardization of data elements and technology that require political will and foresight from very beginning of planning. Most of the difficult decisions about implementing CPR systems are really capital allocation decisions that involve multiple operating units, and so they are inherently political. Recognize politics early in your planning efforts, so you do not meet with frustration and failure after much time wasted fantasizing on technical possibilities.

Marshall Ruffin, MD, MPH, MBA, FACPE, is President of The Informatics Institute, Falls Church, Va. He may be reached at 810 Gatehouse East, #401-East, Box 11, Falls Church, Va. 22042, 703/205-3901, FAX [phone]. Copyright 1996 by The Informatics Institute.