A comparison of quality and utilization problems in large and small group practices.
Fewer quality and utilization problems were expected after physicians went from small or independent group practices to large group practices because of the influences of the large practice environment. Quality and utilization problems of physicians who practiced within large groups throughout the study were expected to be fewer than those of physicians in small or independent practices and to remain constant over the study period.
Utilization review and claims data maintained by the Iowa Foundation for Medical Care (IFMC) were used to identify the incidence of quality and utilization problems. IFMC conducts utilization review as Medicare's Peer Review Organization in Iowa. EFMC also conducts review on inpatient hospitalizations for the Iowa Medicaid program. The data used in the study were from claims and utilization reviews of Medicare and Medicaid inpatient admissions during calendar years 1989, 1990, and 1991.
The utilization review process begins with identification of admissions to be reviewed. HCFA requires review of a random sample of all admissions, readmissions that occur within 30 days, transfers between acute care facilities, and certain DRGs. Determinations about the necessity and quality of care provided are made through a review of the medical record by a registered nurse. The nurse reviewer applies screening criteria that measure the intensity of care provided and the severity of the case (intensity of service/severity of illness). The criteria used are specific to the body system evaluated. For example, neurologic criteria are used to assess a person admitted for a stroke. In addition to the screening criteria, the quality of care on each admission is evaluated by applying generic quality screens. Any case that does not meet either the intensity of service/severity of illness or the generic quality screens is referred to a physician. Only physicians make determinations of whether care was substandard.
A letter was sent to the medical directors of three large clinic practices in Iowa--McFarland Clinic (Ames), Mercy Clinics (Des Moines), and Iowa Physician Clinics (Des Moines). The letter requested the date each physician in the group joined and each physician's practice experience prior to joining the large group. The prior practice experience was designated as either "small clinic," "large clinic," "independent practice," or "student." Each medical director was called several days after the letters were sent to discuss the project. The operative definitions used in the study were explained, and any questions or concerns with the study answered.
Cases that underwent utilization review were identified for physicians in the three large clinic practices using utilization review data. The cases for all three practice groups were combined and assigned to one of the following three study groups:
* Small group. Cases that occurred prior to joining a large group practice.
* Large group. Cases that occurred after joining a large group practice.
* Continuous large group. Cases of physicians who were in a large group practice throughout the study period.
The prior practice experience groups of "student" and "large group" were not included in the analysis, because only three reviews were conducted on physicians from large practice groups and only four reviews were conducted on physicians who were students prior to joining the large practice.
Study group #3, the "continuous large group," comprised physicians who joined a large group practice prior to the study period. This group served as a control group, because it was assumed to represent typical incidences of utilization and quality problems that occur in large group practices.
* Quality of care. Problems of quality of care were identified by generic quality screens applied in all reviews conducted by IFMC. The generic quality screens (see figure 1, page 31) measure the adequacy of discharge planning, medical stability, deaths, nosocomial infections, trauma suffered in the hospital, and unscheduled return to surgery. A case was counted as having a quality problem if a physician reviewer determined that one or more of the generic quality screens failed. A case was also counted as having a quality problem if a physician reviewer determined there was a premature discharge.
* Inpatient utilization. A high number of admission denials and greater lengths of stay were interpreted as indications of inappropriate use of hospital inpatient services. The length of stay data were obtained from claims data. Admission denials were obtained from utilization review data.
* Inpatient charges. The charges measured were billed charges obtained from claims data. Billed charges were not necessarily the same as the actual amount paid. The amount paid may be subject to such variables as the negotiated leverage of health plans, benefits of the health plans, and the like. Billed charges were used because they are "before negotiations" and are therefore assumed to be relatively constant between providers. The large groups represented in the study did not, however, negotiate hospital rates nor did they own any insurance products or managed care plans.
Cases from all three medical practice settings were combined to obscure the identity of any individual physician or group practice. Individual physicians and medical practices were not compared. The Manager for Research and Development at the Iowa Foundation for Medical Care maintained the coded fist of the physicians and medical practices in a secured area at IFMC. Only the study groups are reported in the research data.
Closed system. A group of 30 or more primary care physicians in Iowa with or without specialty colleagues, with direct reporting relationships to other physician directors.
Adult primary care physicians. Physicians who have declared their practice specialty as family practice, internal medicine, or general practice.
Independent practice physician. All adult primary care physicians in Iowa who practice independently or in groups of fewer than 30 physicians.
The outcomes of all Medicare and Medicaid utilization reviews were identified for each of the physicians practicing in the three group practices. No differentiation between the Medicare and Medicaid pay sources was made. The incidence of quality and utilization problems for physicians before and after joining a large group and for physicians in a large group throughout the study period are listed in the table on page 32.
[TABULAR DATA OMITTED]
Average hospital charges. The impact of a large group practice was greatest on average hospital charges, which decreased $3,065 (48 percent) after physicians joined a large group practice. This amount was nearly the same (within 2 percent) of the average hospital charge of physicians who practiced exclusively in large group practices throughout the study period. The change in average hospital charges before and after exposure to a large group practice is illustrated in figure 2, page 32.
The average hospital charges ($6,548) for physicians in the small or independent practice study group are similar to those for large hospitals ($6,532) in the Average Hospital Charges report published by the Iowa Health Data Commission. This report, published March 1992, lists the average charge per patient for small (1-100 beds), medium (101-299 beds), and large hospitals (300 or more beds) in Iowa during calendar year 1990. Average hospital charges for physicians after joining a large group ($3,483) or continuously practicing in a large group ($3,397) approximates the average hospital charges of $3,089 for small hospitals.
Indexing the severity of all sample cases could help to explain the similarity in average hospital charges between the study groups and hospitals of certain sizes. Differences in average hospital charges between large and small hospitals are often due to severity or seriousness of illness. Large hospitals often treat more severe and costly cases because of their greater resources. Severity indexing in future analyses could help to explain whether the decrease in average hospital charges was due to a change in the severity of cases or to benefits from a large practice environment.
Average length of stay. The average length of stay decreased 23 percent, from 18.1 to 13.9 days, after physicians joined a large group practice, but it remained significantly longer (48 percent) than those of physicians practicing exclusively in a large group practice (figure 3, below).
Adverse utilization review decisions. The proportion of total reviews that resulted in DRG changes, coding changes, and generic quality screen failures for physicians before and after they joined a large group practice are illustrated in figure 4, page 33.
DRG changes. The proportion of DRG changes made on all cases of physicians after joining a large group practice decreased 27 percent, but was still 60 percent higher than that for physicians who practiced exclusively in large group practices. Results of DRG changes could be skewed because of the small sample sizes (n=13 in study group 1, n=14 in study group 2, and n=6 in study group 3).
Coding changes. Coding changes decreased 22 percent after physicians joined a large group practice, but remained higher (42 percent) than for physicians who practiced in a large group throughout the study period.
Generic Quality Screen Failures. The incidence of generic quality screen (GQS) failures remained constant for physicians before and after joining large group practices. The incidence of GQS failures for physicians entering a large group practice was also about the same as for physicians who practiced in a large group throughout the study period. Results of generic quality screen failures could be skewed because of the small sample sizes (n=6 in study group 1, n=10 in study group 2, and n=6 in study group 3).
Admission Denials. The incidence of admission denials in the study group (n=2 for before, n=2 for during, and n=1 for control) and the control group (n=2 in 1989, n=2 in 1990, and n=1 in 1991) were too small to make any inferences. Premature Discharges. The incidence of premature discharges in the study group (n=0 for before, during, and control) and control group (n=0 in 1989, n=1 in 1990, and n=0 in 1991) were too small to make any inferences.
This pilot study illustrates associations between increased efficiencies of hospital inpatient utilization when physicians begin to work in large medical practices. The association appears to be strongest for average hospital charges. A weaker association was also observed for DRG changes. The data also suggest that length of time in a large group practice may generate greater efficiencies in utilization. The decreases demonstrated for DRG changes, coding changes, average length of stay, and average charges after physicians joined large clinic practices were never as low as those amounts for physicians in a large group practice throughout the study period.
Inferences from data. The inpatient data used in this study constitute only a small portion of the total encounters the patient may have with the health care system. The study omits data from settings (offices/clinics) where differences in cost savings and quality of care may be likely to occur. There is also a danger that the data do not actually reflect physician practice style, but rather the treatment provided by the hospital.
Comparison of pre-, poststudy periods. The ideal methodology would be to compare only those patients seen in both the pre- and poststudy periods, rather than just physicians.
Application of generic quality screens. Potential generic quality screen "failures" identified by nurse reviewers were more frequent prior to the current PRO contract because of stricter application of GQS review guidelines. Physician reviewers and their ultimate decisions on whether a screen "fails" are not bound to nurse review guidelines. However, the frequency of exposure to potential GQS failures may inflate the incidence of adverse review decisions prior to the current PRO contract.
Cost measurement. Claims data may not accurately reflect the actual costs of care. Urban hospitals will receive DRG payments that are larger than those received by rural hospitals. Therefore, rural hospitals will appear to provide less costly care. The influence of closed physician practice systems on cost of care can be demonstrated if the marginal differences are assumed to be constant over the experimental period, and this relationship is the same across DRGs and all hospitals.
Physician specialty. The physician's practice specialty is obtained from physician-reported data at the time of licensure. It does not reflect changes in physician practice that may occur since licensure.
Management in a closed physician practice system. No attempt was made in this study to define or analyze the management variables that make a closed physician practice system different from the environment of the independent physician practice. In addition, the commonality of these variabilities among closed systems and their relative influence on dependent variables were not examined.
Definition of independent practice. The distinction between independently practicing physicians and physicians practicing in a closed system is difficult to measure. The ranges of sizes of physician practice groups is on a continuum.
Dependent variables. The dependent variables measured from utilization review data represent statistically rare events. Not all admissions that occur during the study period are subject to review. Of those reviewed, few result in adverse review decisions that can be measured as a dependent variable in the study. The measures of these rare events may not be sufficiently sensitive to pick up differences in quality and appropriateness within and among various study groups. Only the average cost per stay can be measured on every hospital admission during the study period, because it is derived from claims data. More in-depth analysis using larger databases over longer periods is necessary to study associations identified in this pilot study.
Appropriate use of inpatient setting. Admission denials only measure whether inpatient care was unnecessary. This study does not identify care that was necessary but not received.
RELATED ARTICLE: Figure 1. Generic Quality Screens--Hospital Inpatients
1. Adequacy of Discharge Planning(*) No documentation of discharge or appropriate follow-up care with consideration of physical, emotional, and mental status needs at time of discharge.
2. Medical Stability of Patient
a. BP within 24 hours of discharge (systolic
less than 50 or greater than 110). b. Temperature within 24 hours of discharge greater than 101[degrees]F (38.3[degrees]C) oral,
greater than 102[degrees]F (38.9[degrees]C) rectal). c. Pulse less than 50 (45 if the patient is on a beta blocker) or greater than 120 within
24 hours of discharge. d. Abnormal diagnostic findings that are addressed and resolved or where the
record does not explain why they are not resolved. e. Purulent or bloody drainage of wound or open area within 24 hours prior to discharge.
a. During or following any surgery performed during the current admission. b. Following return to intensive care, coronary care, or other special care unit within
24 hours of being transferred. c. Other unexpected death.
4. Nosocomial Infection(*)
5. Unscheduled Return to Surgery Within same admission for same condition as previous surgery or to correct operative problem.
6. Trauma Suffered in the Hospital a. Unplanned surgery, which includes, but is not limited to, removal or repair of a
normal organ or body part (i.e., surgery not addressed specifically in the operative
consent). b. Fall(*) c. Serious complications of anesthesia. d. Any transfusion error or serious transfusion reaction. e. Hospital-acquired decubitus ulcer and/or deterioration of an existing decubitus.(*) f. Medication error or adverse drug reaction with serious potential for harm or
resulting in measures to correct. g. Care or lack of care resulting in serious or potentially serious complications.
(*) PRO reviewer is to record the failure of the screen but need not refer potential severity Level I quality problems to physician reviewer until a pattern emerges.
Figures 2 to 4 [ILLUSTRATION OMITTED]
[1.] Madison, D., and Konrad, T. "Large Medical Group Practice Organizations and Employed Physicians: A Relationship in Transition." The Milbank Quarterly 66(2):240-82, 1988. [2.] Goodman, L., and Swartwout, J. "Comparative Aspects of Medical Practice." Medical Care 22(3):255-67, March 1984. [3.] Munger, C."MMGMA: Providing Better Business for Doctors." Michigan Medicine 88(3):38-9, March 1989.
Stephen C. Gleason, DO, is Chairman and Chief Medical Officer, Mercy Clinics, Inc., Des Moines, Iowa. Michael J. Richards, MD, is President, Iowa Physicians Clinic Medical Foundation, Des Moines. John E. Quinnell, MBA, is a researcher with Sunderbruch Corp., West Des Moines. Iowa. Technical support for the article was provided by Iowa Foundation for Medical Care, West Des Moines.
|Printer friendly Cite/link Email Feedback|
|Author:||Quinnell, John E.|
|Date:||Dec 1, 1995|
|Previous Article:||How to keep the Joint Commission happy and plaintiff's attorneys frustrated.|
|Next Article:||Approaches to patient management problems.|