The importance of data warehouses for physician executives.
A data warehouse for health care populations consists of integrated information about patients, their symptoms and signs of disease, their diagnoses and procedures, their clinical laboratory and radiology results, and their charges and associated costs in a relational database useful for clinical and health services research studies. Data warehouses contain millions of records about hundreds of thousands, in some cases millions, of patients. The data warehouses of insurance companies have claims data filed with procedures, diagnoses, charges, and payments.
Insurers do not have access to clinical data, such as laboratory and radiology reports, so those clinical details will not be in the data warehouses they build. Conversely, data warehouses that providers build will have inpatient and outpatient claims data, laboratory and radiology results, and pharmaceutical data, but they will not include data about patients treated b competitors. If providers have cost accounting systems, they may include data on charges and payments associated with services.
To the best of my knowledge, no large group practice, hospital, or health care system has a data warehouse with standardized measures of the functional status of most of its patients, both before and after treatment. Standards for identifying diagnoses (ICD-9-CM) and procedures (ICD-9-CM and CPT4) are fairly well developed and are widely used by providers, because most providers in the civilian health care system are paid for their services on a piece-rate, fee-for-service basis that requires identification of diagnoses and procedures for payment. But there are no widely used standards for defining patients' symptoms or physicians' findings on physical examination of patients. Why? Because we have not needed to identify these symptoms or findings for reimbursement. Why do we need to do so now? Because buyers of health care services are beginning to ask sellers of health care insurance, either managed care plans or provider organizations directly contracting with employers, to produce evidence of acceptable clinical outcomes and to show continuous improvement in those outcomes.
The American Health Security Act proposed by the Clinton Administration describes a national information infrastructure, including standardized data about diagnoses, procedures, and clinical outcomes that every provider would be obligated to collect after every clinical encounter. The purpose of a national standardized database of patients' functional status before and after treatment, diagnoses given, and procedures performed is to produce a national resource for comparison of clinical practices across the United States.
In the past, group- and staff-model health maintenance organizations shadow-priced indemnity insurance health plans, pricing their insurance products about 10 percent below the high-option indemnity plans and still offering more generous services. Staff- and group-model HMOs, such as Kaiser Permanente Health Plan, never needed to collect detailed HCFA 1500 forms for the office visits of their patients, because they did not produce bills for service. Now, as integrated health care delivery systems begin to emerge in the previously fragmented, private practice community of providers, even Kaiser needs to find the discipline of systematic and standardized data collection for all the services rendered for all its members to create a database its leaders can use to analyze patterns of treatment, institute systematic quality improevemnt, and compete for patients more effectively.
Where should the data go that physicians and hospitals collect electronically for patient care and billing and that they want to be able to study to identify opportunities to reduce costs and improve quality of patient care? The answer is that they should flow into an electronic data warehouse in which all relevant clinical and financial data about patients can be stored, including the results of patient satisfaction and functional status surveys; laboratory results, including anatomical and clinical pathology studies; orders; and coded radiology interpretations.
All too often, data in electronic form are printed on paper to produce bills and laboratory and pharmacy reports and are purged from electronic storage after patients have completed office visits or have been discharged from a hospital. Chief financial officers may be delighted to print data on paper, because they think paper storage of medical and administrative records is cheaper than electronic storage on magnetic disk drives, optical disk drives, or magnetic tape, but CFOs, who have never suffered through chart review, do not understand the hideous inefficiency of abstracting clinical data from paper records in order to perform clinical research, quality improvement, and utilization and risk management studies.
Computer technologies are becoming more robust, less expensive, and more efficient faster than the attitudes of most health care managers are changing about those technologies. The principal asset of a health care organization is its medical records, on which clinicians depend for patient care and continuous quality improvement. Organizations need to realize that continuous quality improvement requires frequent, ad hoc access to data about the processes that may need improvement. Organizations can facilitate quality improvement studies and reduce the expensive hours spent in data collection by investing in one-time creation of a corporate data warehouse that will serve many users countless times in their searches for data.
Intermountain Health Care discovered that an adequate study of a well-defined clinical process still required some $50,000 in time by medical record specialists abstracting several hundred clinical details from 1,000 charts. Why so many charts? There are so many variations in the conditions of patients before a surgical procedure or a medical treatment, so many variations in treatments, and so many ways that patients may respond to treatments that large numbers of patients need to be studied to identify statistically significant trends attributable to treatment. It makes far more sense to leave the data about patients in electronic form and copy clinical and financial data from transaction systems to data warehouses than to grope through paper records containing data that were at one time in electronic form.
I've noticed that physicians readily embrace the concept of a data warehouse for health services and clinical research on their own patient populations. Physicians have not been given clinical data about the outcomes of their patients compared to the outcomes of similar patients treated by other physicians in their communities. Physicians are trained in academic health centers where clinical research studies occur every day. They are familiar with standardized data collection and data analysis. Given reliable data about patient outcomes, they can consider changing practice habits and adopting practice guidelines that will lead to improvement in the processes of care for their patients.
Until a health care organization invests in a data warehouse for integration of clinical, financial, and outcomes data, claims of commitment to quality improvement made by administration ought not to be taken seriously. The organization does not benefit from managers (and clinicians) squandering countless hours in committee meetings about quality improvement trying to imagine root causes of problems, when they could be analyzing large clinical databases for important relationships.
The data warehouse is a gold mine in which to find opportunities for clinical quality improvement. Leaders of an organization who choose not to build a data warehouse are saying they are not interested in putting sources of data in the hands of users at an economical price for quality improvement activities. They are more interested in the appearances of quality improvement, which they control, than in widespread pursuit of CQI studies, which cannot be in their control. This resource is too valuable to be left in the hands of managers who may be threatened by the undermining of organizational hierarchy that widespread access to corporate data about patients will tend to produce. Physician managers, who have an understanding of both organizational and clinical needs, are in the best position to champion the creation of data warehouses in their organizations.
(*) Salem-Schatz, S., and others. "The Case for Case-Mix Adjustment in Practice Profiling: When Good Apples Look Bad." JAMA 272(11):871-4, Sept. 21, 1994.
|Printer friendly Cite/link Email Feedback|
|Date:||Nov 1, 1994|
|Previous Article:||Managing the new technical employee.|
|Next Article:||Physician financial relationships in the new regulatory environment.|