Preventing Unnecessary Hospitalizations in Medi-Cal: Comparing Fee-for-Service with Managed Care February 2004 Preventing Unnecessary Hospitalizations in Medi-Cal: Comparing Fee-for-Service with Managed Care Prepared for: CALIFORNIA HEALTHCARE FOUNDATION Prepared by: Primary Care Research Center University of California, San Francisco Authors: Andrew B. Bindman, M.D., Arpita Chattopadhyay, Ph.D., Dennis Osmond, Ph.D., William Huen, M.S., M.P.H., and Peter Bacchetti, Ph.D. February 2004 About the Authors This study was performed by faculty and staff members within the University of California, San Francisco’s Primary Care Research Center at San Francisco General Hospital. Andrew Bindman is a professor of medicine, epidemiology, and biostatistics and is director of the Primary Care Research Center. Arpita Chattopadhyay is a senior statistician. Dennis Osmond is a professor of epidemiology and biostatistics. William Huen is a student in the University of California, Berkeley and the University of California, San Francisco joint medical program. Peter Bacchetti is a professor of epidemiology and biostatistics. About the Foundation The California HealthCare Foundation, based in Oakland, is an independent philanthropy committed to improving California’s health care delivery and financing systems. Additional copies of this report and other publications can be obtained by visiting www.chcf.org. ISBN 1-932064-63-X Copyright © 2004 California HealthCare Foundation Contents 4 Overview 5 I. Background 7 II. Methodology 8 III. Findings Preventable Hospitalization Rates in Medi-Cal Medi-Cal Managed Care Versus Fee-for-Service Comparisons by Insurance Type 20 IV. Conclusions Summary Discussion 23 Appendix: Methodology 26 Endnotes Overview IN THE 1990S, CALIFORNIA EXPANDED MEDI-CAL managed care enrollment with the stated goal of improving ben- eficiaries’ access to health care. Unlike Medi-Cal fee-for-service, Medi-Cal managed care requires a beneficiary to select a primary care provider to serve as his or her usual source of care. One method of measuring access to care is through preventable hos- pitalization rates. Preventable hospitalizations are admissions for ambulatory conditions such as asthma, diabetes, and hyperten- sion that can often be managed in an outpatient setting. Patients with these conditions who do not have adequate access to ambulatory care can experience a decline in their health, increas- ing the likelihood that they will need to be treated in a hospital. This report summarizes the results of a study that used preventa- ble hospitalization rates to compare ambulatory care delivery in Medi-Cal fee-for-service with Medi-Cal managed care. The results show that from 1994 to 1999 the preventable hospitaliza- tion rate was significantly lower for Medi-Cal beneficiaries enrolled in managed care than those using fee-for-service. Key findings include: s For the largest group of non-elderly beneficiaries who qualify for Medi-Cal—those eligible through the California Work Opportunity and Responsibility to Kids (CalWORKs) pro- gram—the average annual preventable hospitalization rate was more than a third lower in managed care than in fee-for- service. This suggests that managed care was associated with an average of more than 7,000 fewer hospitalizations per year, saving an estimated $66 million in hospital charges. s Among beneficiaries with disabilities who are eligible for Medi-Cal through Supplemental Security Income (SSI), the average annual preventable hospitalization rate was about one- quarter lower with managed care than with fee-for-service. s Of the three types of Medi-Cal managed care models, the Two-Plan Model had the lowest overall rate of preventable hospitalization. These findings suggest that the requirement of a usual source of care for Medi-Cal beneficiaries is associated with improvements in these patients’ access to ambulatory care and their overall health. The large reductions in preventable hospitalizations for Medi-Cal beneficiaries in managed care suggests that there is an enormous opportunity to improve access to ambulatory care among Medi-Cal beneficiaries and reduce Medi-Cal expendi- tures for hospital care. 4 | CALIFORNIA HEALTHCARE FOUNDATION I. Background MEDICAID ORIGINATED IN THE MID-1960S AS A jointly financed federal and state health insurance program for low-income, disabled, and elderly Americans. As of 2002, Medi-Cal, California’s Medicaid program, was providing health insurance to roughly 6.5 million Californians at an estimated annual cost of more than $29 billion, making it the largest state Medicaid program in the country. However, enrollment in the Medi-Cal program does not necessarily ensure access to health care services. Surveys of California physicians have found that a little more than half accept Medi-Cal patients1 and that the supply of primary care physicians available to Medi-Cal beneficiaries is below recommended federal standards.2 Correspondingly, a survey of Medi-Cal beneficiaries conducted in 1999 found that 56 percent of beneficiaries reported diffi- culty in finding doctors who were willing to treat Medi-Cal patients; 94 percent of beneficiaries stated that getting more doctors into the program was important.3 Between 1994 and 1999, California expanded Medi-Cal man- aged care enrollment from 16 percent of all Medi-Cal benefi- ciaries to 50 percent statewide. One of the stated goals of this expansion was to improve beneficiaries’ access to care. 4 Medi- Cal managed care was implemented county by county through a combination of voluntary and mandatory managed care programs. These programs mainly target beneficiaries (predominantly women and children) who are eligible for federal Temporary Assistance for Needy Families, which in California is referred to as the CalWORKs program. Unlike Medi-Cal fee-for-service, Medi-Cal managed care requires a beneficiary to select a primary care provider to serve as his or her usual source of care. Access to a primary care physician as a usual source of care can facilitate timely medical attention in an outpatient (ambulatory care) setting. One measure of Medi-Cal beneficiaries’ access to ambulatory care is preventable hospitalization rates. Preventable hospitaliza- tions are admissions for ambulatory-care-sensitive conditions such as asthma, diabetes, and hypertension, which can often be managed with timely and effective treatment in an outpatient setting, thereby preventing hospitalization. Hospital admissions for these conditions reflect a decline in health status, and high- er rates of admission for these conditions are associated with worse access to care.5 Numerous studies have found that pre- Preventing Unnecessary Hospitalizations in Medi-Cal | 5 ventable hospitalization rates are higher in the The time frame of this project, 1994 through United States among low-income people, 1999, corresponds to the period of a “natural African Americans, Medicaid beneficiaries, and experiment” during which the use of managed the uninsured.6 Medicaid patients who have care in the Medi-Cal program increased substan- more continuity of care from a usual source have tially. Prior to 1994, five of California’s 58 coun- been found to have lower rates of hospitalizations ties participated in a demonstration project that for ambulatory-care-sensitive conditions.7 required Medi-Cal beneficiaries to receive services through managed care. During the period of the There has been only limited study of the impact study, most of the remaining large urban counties of Medicaid managed care on preventable hos- in the state (where about 80 percent of Medi-Cal pitalization rates. Some policy analysts have beneficiaries in the state reside), began to require been concerned that the resource limitations that all of their CalWORKs-eligible Medi-Cal within managed care could result in an increase beneficiaries receive services through managed in preventable hospitalizations in Medicaid care. The implementation of Medi-Cal managed managed care programs. On the other hand, care occurred on different dates within the study the requirement that beneficiaries have a regular period and took somewhat different forms in the source of care and the financial arrangements affected counties. One of the main differences within Medi-Cal managed care would appear was whether the managed care was provided by a to create an incentive for reducing unnecessary county-operated health plan (County Organized hospitalizations. Medi-Cal managed care plans Health System, or COHS), competing commer- are paid a capitation rate from the state based cial health plans, or a local initiative health plan on the number of beneficiaries who sign up that in most counties was a public plan in com- with their plans on a monthly basis. The capita- petition with a commercial plan. The small tion payment is used to cover beneficiaries’ inpa- number of counties that implemented Medi-Cal tient and outpatient costs. Medi-Cal managed managed care using the COHS model required care plans are at risk for the cost of their patients’ SSI-eligible beneficiaries with disabilities as well care, and they have a financial incentive to as CalWORKs-eligible Medi-Cal beneficiaries increase the use of less expensive outpatient treat- to use managed care. ment when doing so would reduce the use of expensive hospital-based care. In a small study comparing two California coun- ties, Lo Sasso et al, reported that preventable hospitalization rates increased over time in San Mateo’s Medi-Cal managed care plan, compared with Ventura County’s Medi-Cal fee-for-service program.8 One of the aims of this study was to determine whether the findings from these two counties could be generalized to the entire state of California and whether Medi-Cal managed care is associated with changes in preventable hospitalization rates. 6 | CALIFORNIA HEALTHCARE FOUNDATION II. Methodology THIS STUDY USED PREVENTABLE HOSPITALIZATION rates to compare ambulatory care delivery with Medi-Cal fee-for-service compared to Medi-Cal managed care for CalWORKs-eligible and SSI-eligible beneficiaries. The analytic strategy assumed that if Medi-Cal managed care was having a positive effect on Medi-Cal beneficiaries’ access to ambulatory care, then preventable hospitalization rates would be lower among Medi-Cal beneficiaries in managed care than among those in fee-for-service. It was conducted by linking Medi-Cal eligibility files from the California Department of Health Services (DHS) with hospital discharge data available from the California Office of Statewide Health Planning and Development (OSHPD). Because older Medi-Cal beneficiaries are also likely to have Medicare insurance, the analysis was lim- ited to individuals under the age of 65. Preventable hospitaliza- tion rates for Medi-Cal beneficiaries eligible through the SSI program were analyzed separately from those in CalWORKs, in recognition that the former are eligible on the basis of a dis- ability while a large proportion of the latter are generally healthy children or pregnant women—the group that account- ed for most of the growth in Medi-Cal managed care. More information on the methodology of this study can be found in the Appendix. Preventing Unnecessary Hospitalizations in Medi-Cal | 7 III. Findings Preventable Hospitalization Rates in Medi-Cal This study found clear trends in preventable hospitalization rates among people enrolled in Medi-Cal. Between 1994 and 1999, the annual preventable hospitalization rates for all Medi- Cal beneficiaries under age 65 increased slightly, rising from 18.2 hospitalizations per thousand beneficiaries to 18.6 per thousand. Because the number of preventable hospitalizations grew faster than the overall number of non-pregnancy-related hospitalizations among Medi-Cal beneficiaries, the rate of preventable hospitalizations as a percentage of total hospitaliza- tions for this population increased from 22 percent to 25 per- cent over the same time period. There were dramatically different rates of preventable hospitalizations among two large populations who receive Medi-Cal services, namely CalWORKs-eligible beneficiaries and those who qualify for Medi-Cal through a link with SSI. For example, between 1994 and 1999, the annual rate of pre- ventable hospitalizations was nearly eight times higher among SSI-linked Medi-Cal beneficiaries (including beneficiaries dually eligible for Medi-Cal and Medicare) than among CalWORKs-linked beneficiaries (75.9 and 9.9 per 1,000, respectively). This is not entirely unexpected, given the vast difference in the underlying health status of these two groups, as well as differences in how they were incorporated into managed care. In contrast to the results for the overall Medi-Cal population, the average annual preventable hospitalization rate among CalWORKs-eligible Medi-Cal beneficiaries decreased during the six-year study period, dropping from 9.5 to 8.6 per 1,000 (Figure 1). Among SSI-eligible Medi-Cal beneficiaries under age 65, the average annual admission rate increased from 73.4 to 77.8 per 1,000. Adjusting these rates for changes in the demographics of the beneficiaries during this time period did not have any appreciable effect on the results. 8 | CALIFORNIA HEALTHCARE FOUNDATION Figure 1: Unadjusted Average Annual Preventable Hospitalization Rates among Non-Elderly CalWORKs and SSI-Eligible Medi-Cal Beneficiaries 80 77.8 70 73.4 73.7 73.8 74.9 69.7 60 50 SSI 40 CalWORKs 30 20 9.5 10.4 9.2 9.5 9.1 8.6 10 0 1994 1995 1996 1997 1998 1999 Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 For the population receiving CalWORKs, there was an approximately threefold range in the aver- age annual preventable hospitalization rate across counties: from 5.4 per 1,000 in Del Norte to 15.2 per 1,000 in Imperial (Table 1). There was a sim- ilar range of difference across counties for the SSI population: from 32.6 per 1,000 in Marin to 89.9 per 1,000 in Riverside. Preventing Unnecessary Hospitalizations in Medi-Cal | 9 Table 1: Average Annual Preventable Hospitalization Rates among Non-Elderly Medi-Cal Beneficiaries by Eligibility Criteria and County 1994-1999 County CalWORKs SSI All Medi-Cal Alameda 12.6 71.2 25.0 Butte 8.5 58.3 18.3 Calaveras and Amador 8.9 61.8 17.8 Contra Costa 9.5 73.6 23.2 Del Norte 5.4 35.8 13.4 El Dorado 7.3 54.1 18.1 Fresno 9.2 58.2 16.2 Glenn and Colusa 7.4 61.8 15.1 Humboldt 10.5 50.7 20.9 Imperial 15.2 68.8 21.7 Inyo, Alpine, Mono, and Mariposa 9.9 56.5 16.2 Kern 8.5 79.4 18.6 Kings 13.9 87.7 22.6 Lake 8.5 52.9 19.5 Lassen and Modoc 9.0 55.0 16.3 Los Angeles 10.0 84.7 20.2 Madera 9.5 60.5 14.9 Marin 7.5 32.6 17.1 Mendocino 9.8 60.0 20.3 Merced 7.2 64.7 13.4 Monterey 10.1 70.1 17.5 Napa 8.8 47.0 19.1 Nevada, Sierra, and Plumas 11.2 77.0 25.9 Orange 7.1 60.6 15.8 Placer 7.7 55.1 19.0 Riverside 13.2 89.9 25.1 Sacramento 7.1 52.1 15.6 San Benito 8.6 72.8 15.6 San Bernardino 12.4 87.1 22.7 San Diego 9.1 62.2 18.2 San Francisco 11.3 79.9 33.7 San Joaquin 7.5 58.3 17.1 San Luis Obispo 9.0 57.3 20.6 San Mateo 8.3 52.5 17.4 Santa Barbara 7.1 51.4 14.3 Santa Clara 8.9 55.8 16.7 Santa Cruz 5.8 48.2 14.7 Shasta 9.1 56.2 19.9 Siskiyou and Trinity 8.8 46.3 18.9 Solano 7.0 49.0 14.8 Sonoma 7.1 48.6 17.4 Stanislaus 8.0 67.7 18.0 Sutter 11.3 72.0 21.4 Tehama 6.7 53.6 16.1 Tuolumne 7.5 74.6 15.7 Tulare 7.2 48.7 17.0 Ventura 9.3 59.8 17.5 Yolo 7.3 63.9 16.1 Yuba 10.1 75.8 21.9 Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 Note: All figures represent hospitalization rates per 1,000 beneficiaries. 10 | CALIFORNIA HEALTHCARE FOUNDATION Medi-Cal Managed Care Versus Compared with CalWORKs-eligible Medi-Cal Fee-for-Service beneficiaries, a much smaller percentage of SSI- eligible Medi-Cal beneficiaries were in managed Enrollment in Managed Care care in 1994, and the rate of growth through Enrollment in managed care by CalWORKs-eli- 1999 was also much slower. In 1994, 7 percent of gible Medi-Cal beneficiaries grew dramatically SSI-eligible Medi-Cal beneficiaries were in man- from 1994 to 1999. In 1994, 23 percent of these aged care. By 1999 this percentage reached 18 beneficiaries were in managed care. By 1999 the percent (Figure 3). SSI-eligible Medi-Cal benefi- proportion had reached 78 percent (Figure 2). ciaries were required to be in managed care only Figure 2: Percentage of Non-Elderly CalWORKs-Eligible Medi-Cal Beneficiaries Enrolled in Fee-for-Service and Managed Care 100% – 90% – 80% – 70% – 78 67 60% – 50% – 51 40% – 30% – Fee-for-Service 33 27 Managed Care 20% – 23 10% – 0% – 1994 1995 1996 1997 1998 1999 Source: Department of Health Services 1994-1999 Figure 3: Percentage of Non-Elderly SSI-Eligible Medi-Cal Beneficiaries Enrolled in Fee-for-Service and Managed Care 100% – 90% – 80% – 70% – Fee-for-Service Managed Care 60% – 50% – 40% – 30% – 18 20% – 14 16 17 9 10% – 7 0% – 1994 1995 1996 1997 1998 1999 Source: Department of Health Services 1994-1999 Preventing Unnecessary Hospitalizations in Medi-Cal | 11 in the eight COHS counties, and most SSI-eligi- increase, after adjusting for changes over time in ble Medi-Cal beneficiaries who were in managed beneficiaries’ demographics, their county of care lived in those counties. residence and month of admission. (Figure 4). Instead, with the expansion of Medi-Cal man- The Influence of Managed Care aged care to 78 percent of the CALWORKs-eligi- Preventable hospitalization rates for both ble beneficiaries, the observed preventable hospi- CalWORKs and SSI beneficiaries were signifi- talization rate actually decreased over time to 8.6 cantly lower among those enrolled in Medi-Cal per 1,000 beneficiaries in 1999. In other words, managed care compared to those enrolled in there were 22 percent fewer preventable hospital- fee-for-service. From 1994 to 1999, the average izations associated with the growth of Medi-Cal annual preventable hospitalization rate for managed care in 1999. Between 1994 and 1999, CalWORKs-eligible Medi-Cal beneficiaries was the growth of Medi-Cal managed care was associ- more than a third lower in managed care than in ated with a reduction of an average of 7,000 fee-for-service: 7.2 preventable hospitalizations hospitalizations per year. The average charge for per 1,000 managed care enrollees versus 11.4 these preventable hospitalizations was about per 1,000 for fee-for-service enrollees. Had the $9,500. Thus, the annual reduction of preventa- penetration of Medi-Cal managed care been held ble hospitalization charges was more than $66 stable at the 28 percent level observed in 1994, million less in Medi-Cal managed care than it the average adjusted annual preventable hospital- would have been in fee-for-service for ization rate would have been expected to CalWORKs. Figure 4: Observed and Expected Average Adjusted* Annual Preventable Hospitalization Rates among Non-Elderly CalWORKs-Eligible Medi-Cal Beneficiaries 12 11.0 11.0 10.6 10.6 9.5 9.6 10 10.4 9.5 9.5 9.5 9.1 8 8.6 6 Expected rates without increase in managed care 4 Observed rates with managed care increase 2 0 1994 1995 1996 1997 1998 1999 * Age, sex, race/ethnicity, county, and month of admission Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 12 | CALIFORNIA HEALTHCARE FOUNDATION Among SSI-eligible Medi-Cal beneficiaries under After adjusting for changes over time in benefici- age 65, the average adjusted annual rate of pre- aries’ demographics, their county of residence ventable hospitalization was almost a third higher and their month of admission and holding the in fee-for-service than in managed care: 76.4 per penetration of Medi-Cal managed care stable at 1,000 beneficiaries versus 57.5 per 1,000 benefici- the 7 percent level observed in 1994, the annual aries, respectively. The difference in admission preventable hospitalization rate would have been rates for preventable hospitalizations between expected to have increased from 73.4 to 79.8 per Medi-Cal beneficiaries in fee-for-service and 1,000 beneficiaries in 1999 (Figure 5). Instead, managed care was similar for CalWORKs and with the expansion of Medi-Cal managed care to SSI-eligible beneficiaries in percentage terms, 18 percent of the SSI-eligible beneficiaries, the even though the overall admission rate was sub- observed preventable hospitalization rate actually stantially higher for SSI-eligible beneficiaries. rose only to 77.8 per 1,000 beneficiaries. Figure 5: Observed and Expected Adjusted* Average Annual Preventable Hospitalization Rates among Non-Elderly SSI-Eligible Medi-Cal Beneficiaries 82 Expected rates without increase in managed care 79.8 80 Observed rates with managed care increase 78 76.6 75.4 77.8 76 74.0 73.4 74 74.9 73.4 73.7 73.8 72 70.8 70 69.7 68 66 64 1994 1995 1996 1997 1998 1999 * Age, sex, race/ethnicity, county, and month of admission Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 Preventing Unnecessary Hospitalizations in Medi-Cal | 13 Comparing Managed Care Models Medi-Cal Managed Care Models Variation among Models The California Department of Health Services Among the three types of Medi-Cal managed (DHS) contracts with 22 health plans to serve Medi-Cal beneficiaries. These health plans care models, the Two-Plan Model had the lowest operate within a managed care system that is overall rate of preventable hospitalization. built primarily around three different organizing Grouping the counties by Medi-Cal managed models. care model type (Two-Plan, COHS, GMC, or s Two-Plan Model Under the Two-Plan none [see sidebar]) revealed that CalWORKs- Model, DHS contracts with one county- eligible Medi-Cal beneficiaries who lived in developed health plan, called a Local counties that implemented any of the three main Initiative, and one commercial plan. models of Medi-Cal managed care had lower Enrollment in the Two-Plan Model is manda- average adjusted annual rates of preventable hos- tory for the CalWORKs-linked population. pitalizations between 1994 and 1999 than did Voluntary enrollment of other Medi-Cal ben- eficiaries is permitted. Eleven health plans CalWORKs-eligible Medi-Cal beneficiaries in participate in the Two-Plan Model, which fee-for-service counties (Figure 6). CalWORKs- operates in 12 counties (Alameda, Contra eligible Medi-Cal beneficiaries in Two-Plan Costa, Fresno, Kern, Los Angeles, Model counties had the lowest average adjusted Riverside, San Bernardino, San Francisco, annual preventable hospitalization rates (7.8 per San Joaquin, Santa Clara, Stanislaus, and 1,000 beneficiaries), followed by those in GMC Tulare). (8.4 per 1,000 beneficiaries) and COHS (8.8 s Geographic Managed Care Model The per 1,000 beneficiaries) counties. The Two-Plan GMC Model allows many different health and GMC county preventable hospitalization plans to operate within a designated county, rates were significantly lower than that for similar to most other states’ managed care fee-for-service counties, but the observed differ- programs. Beneficiary enrollment in a health plan is mandatory for the CalWORKs-linked ence between COHS counties and fee-for-service population. Other categories of Medi-Cal was not statistically significant. beneficiaries may voluntarily join these plans. There are two GMC counties CalWORKs-eligible Medi-Cal beneficiaries in (Sacramento and San Diego) with a total of managed care may receive that care through nine participating health plans. either public or commercial health plans depend- ing upon their county’s Medi-Cal managed care s County Organized Health System Model The COHS Model is one in which counties model type. In COHS counties, all CalWORKs- operate a health plan. Counties negotiate eligible Medi-Cal beneficiaries receive managed their contract with the California Medical care through a public plan, and in Two-Plan Assistance Commission. Enrollment in the counties they may choose between a local initia- COHS is mandatory for virtually the entire tive and a commercial plan. Local initiative plans Medi-Cal population in that county and are required to contract with traditional Medi- occurs concurrently with enrollment in the Cal providers. In most Two-Plan counties, the Medi-Cal. Five COHS plans service eight counties (Orange, Monterey/Santa Cruz, local initiative plan is a public plan. One argu- Santa Barbara, San Mateo, and ment for the Two-Plan Model was to create com- Solano/Napa/Yolo). petition that might improve managed care plan performance. Among CalWORKs-eligible Medi- Cal beneficiaries in public managed care plans, the average adjusted annual preventable hospital- ization rate was lower in Two-Plan counties than 14 | CALIFORNIA HEALTHCARE FOUNDATION Figure 6: Average Adjusted* Annual Preventable Hospitalization Rates among Non-Elderly CalWORKs-Eligible Medi-Cal Beneficiaries by County Managed Care Model 1994-1999 12 9.8 10 8.8 8.4 7.8 7.8 8 6 4 2 0 Two-Plan COHS GMC FFS * Age, sex, race/ethnicity, county, and month of admission Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 in COHS counties (7.8 versus 8.8 per 1,000 ventable hospitalization rate among CalWORKs- beneficiaries), suggesting that perhaps the compe- eligible Medi-Cal beneficiaries was lower in Two- tition from the commercial plan exerted some Plan commercial plans (7.6 per 1,000 beneficiar- positive effect on the performance of the public ies) than in GMC commercial plans (8.4 per plan in Two-Plan counties. 1,000 beneficiaries). An alternative explanation for the better per- Variation within Models formance of the public plan in Two-Plan coun- During the study period, there was a wide ties compared to COHS counties might be that range of variation in hospitalization rates for there was risk selection of healthier CalWORKs- preventable conditions for CalWORKs-eligible eligible Medi-Cal beneficiaries to the public Medi-Cal beneficiaries across counties that man- rather than the commercial plan in the Two-Plan dated Medi-Cal managed care in California, counties. However, one would not expect that even within counties with the same managed care healthier beneficiaries would tend to choose a model. For example, in 1994 the preventable public over a commercial plan, and this hypothe- hospitalization rate in Alameda was almost sis is also not supported by the finding that com- twice that of Tulare (Table 2). In many but mercial plans in Two-Plan counties (which would not all of the mandatory Medi-Cal managed have received the “sicker” Medi-Cal beneficiaries) care counties, the preventable hospitalization also outperformed commercial plans in GMC rate decreased over time, particularly after the counties. Taken as a whole, these findings suggest implementation of Medi-Cal managed care. that competition between public and commercial For example, in Los Angeles the admission rate plans has a mutually beneficial effect that is not for ambulatory-care-sensitive conditions achieved to the same degree by public plans decreased from 9.9 per 1,000 beneficiaries in alone or by having commercial plans in competi- 1994 to 7.9 per 1,000 beneficiaries in 1999. tion with one another. The average adjusted pre- Preventing Unnecessary Hospitalizations in Medi-Cal | 15 Table 2: Average Adjusted* Annual Preventable Hospitalization Rates among Non-Elderly CalWORKs-Eligible Medi-Cal Beneficiaries by County, Managed Care Model, and Year County Managed 1994 1995 1996 1997 1998 1999 Care Model Alameda Two-Plan 12.7 12.9 11.0† 11.9 13.5 12.6 Butte None 8.0 9.5 8.6 8.7 6.6 7.6 Calaveras and Amador None 8.3 11.0 10.5 5.7 5.4 11.6 Contra Costa Two-Plan 9.2 10.5 9.0 9.0† 9.2 9.2 Del Norte None 3.7 5.6 5.0 4.3 6.2 8.3 El Dorado None 6.2 9.6 6.9 7.6 5.0 7.0 Fresno Two-Plan 8.8 9.3 9.7 8.6† 9.2 8.1 Glenn and Colusa None 7.3 6.8 5.1 7.3 7.5 10.1 Humboldt None 11.1 11.3 8.9 11.8 10.3 9.3 Imperial None 16.9 16.8 13.0 15.7 13.1 14.0 Inyo, Alpine, Mono, and Mariposa None 7.7 7.8 8.6 7.5 12.7 16.5 Kern Two-Plan 7.9 8.8 7.8† 8.3 8.1 8.9 Kings None 11.5 14.2 13.0 15.4 12.7 14.5 Lake None 6.8 7.2 7.6 8.6 11.7 10.3 Lassen and Modoc None 12.7 7.5 8.3 11.3 5.9 7.3 Los Angeles Two-Plan 9.9 10.9 10.3 10.4† 9.2 7.9 Madera None 10.7 10.8 8.1 8.3 9.6 9.1 Marin None 7.6 6.9 8.5 6.3 9.4 6.3 Mendocino None 10.1 10.7 10.9 7.4 9.0 9.0 Merced None 7.7 9.2 7.0 7.5 6.3 5.4 Monterey COHS 8.1 9.7 10.0 11.4 9.4 11.3† Napa COHS 14.5 6.4 6.9 6.6 7.9† 11.3 Nevada, Sierra, and Plumas None 13.4 9.5 9.8 11.9 10.2 11.8 Orange COHS 7.7 8.0† 5.3 5.8 6.4 8.2 Placer FFS/MC 7.3 8.5 6.3 7.6† 6.9 9.0 Riverside Two-Plan 13.5 14.8 11.2† 12.5 12.4 12.9 Sacramento GMC 6.5† 7.4 7.7 6.8 6.7 6.4 San Benito None 11.3 7.6 7.7 8.5 7.4 3.4 San Bernardino Two-Plan 11.5 14.3 11.8† 11.7 11.2 11.0 San Diego GMC 9.2 10.0 8.5 8.7 8.5† 8.1 San Francisco Two-Plan 12.6 13.7 10.4† 8.5 9.4 10.8 San Joaquin Two-Plan 7.3 7.2 6.7† 7.2 7.7 7.6 San Luis Obispo None 7.2 12.2 9.7 7.1 7.9 8.4 San Mateo COHS 8.1 8.1 7.3 8.4 8.2 9.2 Santa Barbara COHS 7.4 7.6 6.1 6.4 7.2 5.9 Santa Clara Two-Plan 9.1 9.3 8.3 8.3† 8.8 7.9 Santa Cruz COHS 6.8 5.4 5.2† 4.6 5.9 6.3 Shasta None 9.0 10.1 7.8 8.3 9.6 9.0 Siskiyou and Trinity None 8.8 10.8 9.8 6.2 8.1 8.3 Solano COHS 6.8† 7.7 5.2 6.2 7.7 8.1 Sonoma FFS/MC 6.3 8.1 6.7 5.9† 7.7 8.0 Stanislaus Two-Plan 7.3 8.5 7.2 7.8† 8.9 7.0 Sutter None 8.9 13.2 11.8 10.1 9.6 12.3 Tehama None 7.6 5.0 6.2 5.8 8.6 6.2 Tuolumne None 7.5 5.4 6.9 6.8 8.0 8.8 Tulare Two-Plan 6.6 6.9 7.1 7.8 7.7 8.0† Ventura None 7.6 8.8 9.3 9.9 10.2 9.1 Yolo COHS 6.8 7.2 5.9 8.8 7.1 7.1 Yuba None 8.1 11.4 9.6 9.1 9.7 11.7 Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 Note: All figures represent hospitalization rates per 1,000 beneficiaries. * Adjusted for age, sex, race/ethnicity, and month of admission † Year county transitioned to mandatory managed care. Counties with no † transitioned to mandatory managed care before 1994 16 | CALIFORNIA HEALTHCARE FOUNDATION Table 3: Average Adjusted* Annual Preventable Hospitalization Rates among Non-Elderly SSI-Eligible Medi-Cal Beneficiaries by County, Managed Care Model, and Year County Managed 1994 1995 1996 1997 1998 1999 Care Model Alameda Two-Plan 76.2 71.8 72.1 76.7 79.7 78.4 Butte None 53.9 56.9 68.5 55.4 66.0 71.5 Calaveras and Amador None 62.2 69.4 64.3 58.0 52.8 69.2 Contra Costa Two-Plan 76.4 74.7 79.4 76.9 82.5 74.5 Del Norte None 46.3 34.9 44.4 22.1 36.0 40.6 El Dorado None 54.5 55.8 67.2 54.1 52.0 42.0 Fresno Two-Plan 58.6 59.9 71.5 60.9 67.8 67.5 Glenn and Colusa None 66.9 67.0 63.9 48.5 72.3 67.0 Humboldt None 62.9 48.7 51.7 56.0 63.9 57.5 Imperial None 65.1 67.5 64.1 80.6 76.3 66.7 Inyo, Alpine, Mono, and Mariposa None 47.9 59.7 66.6 72.4 63.8 57.2 Kern Two-Plan 78.5 98.2 84.3 86.1 82.1 78.1 Kings None 77.0 84.4 98.5 100.7 98.2 98.9 Lake None 52.8 56.4 41.8 55.7 64.6 46.7 Lassen and Modoc None 53.4 60.4 49.3 66.7 63.0 48.6 Los Angeles Two-Plan 96.1 93.8 102.4 100.8 102.2 118.2 Madera None 52.4 63.6 60.5 57.5 65.3 72.4 Marin None 37.3 25.8 29.9 35.2 34.8 37.7 Mendocino None 43.9 63.4 78.7 66.2 66.0 63.4 Merced None 64.9 80.7 82.6 63.9 61.8 62.9 Monterey COHS 81.7 72.3 82.4 68.8 75.2 68.52† Napa COHS 52.0 48.5 40.6 51.7 40.8† 51.0 Nevada, Sierra, and Plumas None 67.1 58.4 69.1 91.3 89.4 93.0 Orange COHS 74.3 66.1† 64.2 64.8 66.6 66.7 Placer FFS/MC 65.8 58.3 55.9 55.9 51.1 58.2 Riverside Two-Plan 111.1 97.7 92.4 94.5 89.5 105.9 Sacramento GMC 55.5 54.8 61.1 53.0 52.5 63.0 San Benito None 81.7 86.9 68.0 70.5 76.7 54.1 San Bernardino Two-Plan 101.8 98.2 105.9 105.1 105.7 111.6 San Diego GMC 73.4 72.9 65.1 62.5 69.3 80.0 San Francisco Two-Plan 98.2 87.6 81.5 82.2 80.9 90.6 San Joaquin Two-Plan 63.4 60.9 56.5 57.9 66.8 64.2 San Luis Obispo None 63.0 59.2 59.3 60.3 56.8 61.8 San Mateo COHS 61.5 58.0 64.7 56.3 53.9 48.4 Santa Barbara COHS 56.1 60.1 54.8 56.0 50.8 48.9 Santa Clara Two-Plan 63.6 70.1 65.2 62.7 62.1 57.9 Santa Cruz COHS 48.5 54.0 55.3† 43.8 47.6 49.7 Shasta None 66.8 63.8 63.9 49.7 58.1 61.5 Siskiyou and Trinity None 48.6 54.6 45.1 55.4 39.1 43.2 Solano COHS 43.9† 49.9 48.6 45.5 51.7 56.3 Sonoma FFS/MC 53.6 56.3 50.4 49.1 47.0 54.3 Stanislaus Two-Plan 70.5 70.5 67.4 68.3 69.9 67.5 Sutter None 51.8 58.3 88.6 78.4 85.8 81.2 Tehama None 40.3 49.4 50.1 56.1 61.3 58.4 Tulare None 73.7 71.4 75.7 76.9 84.8 86.0 Tuolumne Two-Plan 45.5 44.8 56.2 51.3 47.9 58.3 Ventura None 62.7 59.4 64.8 75.6 55.5 62.5 Yolo COHS 75.2 77.4 61.3 62.4 62.0 57.5 Yuba None 59.7 72.5 81.1 92.1 83.6 88.9 Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 Note: All figures represent hospitalization rates per 1,000 beneficiaries. * Adjusted for age, sex, race/ethnicity, and month of admission † Year county transitioned to mandatory managed care. Managed care counties with no † transitioned to mandatory managed care before 1994 Preventing Unnecessary Hospitalizations in Medi-Cal | 17 A similarly wide variation occurred among coun- Nonetheless, it is possible to get a general sense ties in preventable hospitalization rates for SSI- of the magnitude of preventable hospitalization eligible Medi-Cal beneficiaries. For example, in rates across insurance groups by combining 1994 the average adjusted annual rate in counties counts of preventable hospitalizations by the ranged from 37.3 per 1,000 beneficiaries in expected payer source coded in the OSHPD Marin to 111.1 per 1,000 in Riverside (Table 3). hospital discharge file with the annual estimates from the Current Population Survey of the num- ber of people who are covered by different types Comparisons by Insurance Type of health insurance. Results linking lower rates of preventable hospi- talizations and managed care in the Medi-Cal One of the main limitations of this approach is program raise the question of how those findings that many uninsured people gain Medi-Cal bene- might compare with the experience of people fits as a result of a hospitalization. Thus, prevent- with private insurance, as well as those who have able hospitalization rates for Medi-Cal beneficiar- no coverage at all. The availability of a special ies calculated from routine hospital discharge data set that linked hospital discharge data with data—which record the insurance status of the Medi-Cal eligibility files made it possible to patient only at the time of hospitalization—may accurately describe preventable hospitalization be biased upward while those for the uninsured rates in the Medi-Cal population over time. are biased downward. In reality, preventable hos- Unfortunately, no comparable eligibility files are pitalizations attributable to Medi-Cal may not be available for the uninsured and privately insured as high as routine hospital discharge data suggest, populations to link with hospital discharge data and those for the uninsured may not be so low. to permit equally accurate estimates of preventa- Another source of error is that some Medi-Cal ble hospitalizations in these insurance groups. beneficiaries in private managed care plans may be mistakenly categorized as privately insured. Figure 7: Unadjusted Annual Preventable Hospitalization Rates among Medi-Cal Beneficiaries in 1994 and 1999 25 23.1 Patient Discharge File 19.8 Medi-Cal Linked Patient 20 18.2 18.6 Discharge File 15 10 5 0 1994 1999 Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 18 | CALIFORNIA HEALTHCARE FOUNDATION This would result in an underestimate among (Figure 8). For example, the data indicate that in Medi-Cal beneficiaries and an overestimate 1994 the preventable hospitalization rates were among privately insured in their preventable hos- about four times greater among Medi-Cal bene- pitalization rates. Assuming that the linked ficiaries than among privately insured people OSHPD-Medi-Cal data provides the most accu- (19.8 versus 5.5 per 1,000) and in 1999 they rate estimates, the error in Medi-Cal preventable were more than five times greater (23.1 versus hospitalization rates from routine OSHPD 4.5 per 1,000). patient discharge data was greater in 1999 than Given the misclassification error for some Medi- in 1994 (Figure 7). This most likely reflects the Cal managed care beneficiaries in private man- growth over time in Medi-Cal managed care and aged care plans, the true difference between the associated misattribution of Medi-Cal benefi- Medi-Cal and privately insured preventable hos- ciaries who were in private managed care plans to pitalization rates may be even greater. The cal- private insurance. culated differences between privately insured However, even if one assumes that the Medi-Cal and uninsured preventable hospitalization rates and privately insured preventable hospitalization appears to be small, but in reality the expected rates are lower than calculated from routine hos- errors in calculating these rates for privately pital discharge data, and the uninsured rates are insured and uninsured people run in opposite higher, Medi-Cal beneficiaries appear to have a directions, making it likely that the uninsured markedly higher rate of preventable hospitaliza- rate is truly higher than it is for those who are tion than individuals in other insurance groups privately insured. Figure 8: Unadjusted Annual Preventable Hospitalization Rates in California for the Non-Elderly, by Insurance Type, in 1994 and 1999 25 23.1 19.8 20 1994 15 1999 10 5.5 4.5 4.1 4.1 5 0 Medi-Cal Private Uninsured Source: Office of Statewide Health Planning and Development/Department of Health Services 1994-1999 Preventing Unnecessary Hospitalizations in Medi-Cal | 19 IV. Conclusions Summary The rate of preventable hospitalizations among non-elderly beneficiaries was significantly lower in managed care than in fee-for-service. This difference in hospitalization rates between fee-for-service and managed care persists even after controlling for differences in the characteristics of patients and seasonal and secular trends, as well as county effects. Among the three types of Medi-Cal managed care plans, the Two-Plan Model had the lowest overall rate of preventable hos- pitalization. Furthermore, the Two-Plan Model achieved this benefit by having lower admission rates for ambulatory-care- sensitive conditions in both the public and commercial plans, compared with the corresponding plans in the other county model types. Some caution should be used in interpreting the performance of the county model types and public and com- mercial plans because the type of model a county used to implement Medi-Cal managed care was not chosen at random, and unmeasured differences in the counties and their Medi-Cal beneficiaries may have confounded the results. During the period of the study, there was much greater growth in the use of managed care among Medi-Cal beneficiaries who were CalWORKs-eligible than the SSI-eligible. Although there are fewer observations upon which to draw conclusions, the magnitude of the effect of managed care on lowering hospital- izations for ambulatory-care-sensitive conditions is similar among SSI-eligible beneficiaries as among CalWORKs-eligible beneficiaries in percentage terms. Because SSI-eligible benefici- aries are hospitalized at much higher rates than CalWORKs- eligible beneficiaries, many more Medi-Cal patients would be eligible for the potential health and utilization benefits of man- aged care if this program were further expanded among SSI- eligible beneficiaries. However, because SSI-eligible beneficiar- ies are a particularly vulnerable segment of the Medi-Cal popu- lation, appropriate standards and effective oversight would be necessary to ensure that these potential improvements were achieved and that unintended consequences were avoided if managed care were expanded for this population. 20 | CALIFORNIA HEALTHCARE FOUNDATION Medi-Cal beneficiaries have substantially higher Discussion rates of preventable hospitalization than do pri- The requirement of a primary care physician in vately insured Californians and those without Medi-Cal managed care may have contributed insurance. In 1999, one in nine Californians to the lower rates of preventable hospitaliza- under the age of 65 was insured by Medi-Cal, yet tions. Medicaid beneficiaries in other states have more than one in three admissions for preventa- reported an increase in having a regular source ble conditions that year were for Medi-Cal bene- of care after the implementation of Medicaid ficiaries. The higher rate of preventable hospital- managed care.9-11 There is no guarantee of access izations for ambulatory-care-sensitive conditions to primary care for Medi-Cal beneficiaries who among Medi-Cal beneficiaries is consistent with receive their care through the fee-for-service sys- the high rate of California physicians who report tem. In addition to requiring a regular source of they are unwilling to care for Medi-Cal patients, care, Medi-Cal managed care plans have also and Medi-Cal beneficiaries who report difficul- sought to improve access to and quality of care ties finding a doctor who will care for them. by paying physicians above fee-for-service rates However, direct comparisons of the preventable and by incorporating disease- and case-manage- hospitalization rates between Medi-Cal benefici- ment strategies into their approach to health aries and either those with private insurance or care delivery.12 the uninsured must be interpreted with caution. One reason is that many uninsured patients One alternative explanation for the study’s find- become covered by Medi-Cal when they are hos- ings is that Medi-Cal managed care beneficiaries pitalized, including some who were eligible for are healthier and therefore less in need of hospi- Medi-Cal before their hospitalization but were talization than Medi-Cal fee-for-service benefi- not enrolled. A second reason is that, particularly ciaries. This study design makes that explana- for adults, eligibility for Medi-Cal is linked to tion unlikely. First, unlike most reported evalua- having a health condition. Thus, the lower pre- tions of Medicaid managed care, this study sep- ventable hospitalization rate for the uninsured arated Medi-Cal beneficiaries by eligibility cate- compared to those in Medi-Cal is probably not gory, which provides somewhat of a proxy for the result of better health care service delivery to patient health status. Second, most growth in the uninsured but rather of Medi-Cal beneficiar- Medi-Cal managed care was in mandatory man- ies having poorer health. The annual Current aged care programs that would not be subject to Population Survey used to determine the size of selection bias. the population in each insurance group at-risk for a preventable hospitalization does not provide information about patients’ health status. Preventing Unnecessary Hospitalizations in Medi-Cal | 21 Another interpretation of this study’s findings is that the reduced rate of preventable hospitaliza- tions in Medi-Cal managed care represents a decline in beneficiaries’ access to hospital care. The hospital discharge records used in this study do not permit us to determine whether a higher admission threshold was applied to Medi-Cal managed care patients than to fee-for-service patients in California emergency rooms (through which the overwhelming majority of admissions for ambulatory-care-sensitive conditions occur). However, a national study of emergency depart- ments did not find differences in admitting prac- tices by patients’ insurance status or race.13 The lower rate of preventable hospitalizations for Medi-Cal beneficiaries in managed care com- pared with fee-for-service suggests that the financing and organization of Medi-Cal is associ- ated with beneficiaries’ use of services. Judging by the reduction in preventable hospitalizations, the requirement of a regular source of care for Medi- Cal beneficiaries in managed care is associated with improvements in these patients’ access to ambulatory care and health status. Nonetheless, even with the recent growth of Medi-Cal man- aged care, hospitalization rates for ambulatory- care-sensitive conditions remain much higher for the Medi-Cal population than the privately insured population. Although the difference in the health status of patients in different insurance groups most likely explains much of the differ- ence, it would appear that there is an enormous opportunity for Medi-Cal to reduce hospital use and expenditures by expanding the access to and quality of ambulatory care. 22 | CALIFORNIA HEALTHCARE FOUNDATION Appendix: Methodology TO CONDUCT DETAILED ANALYSES OF PREVENTABLE hospitalization rates for different groups of Medi-Cal beneficiar- ies, the annual California hospital discharge data available from the California Office of Statewide Health Planning and Development (OSHPD) was linked with the Medi-Cal eligibility files from the California Department of Health Services (DHS). The annual California hospital discharge record includes informa- tion about admission month and year, patient demographics, and diagnosis and procedure codes. This file also contains a field indi- cating the expected source of payment. By linking the informa- tion available in the annual California hospital discharge file with that available from DHS, it was possible to enhance the accuracy of whether a hospitalized individual was in fact a Medi-Cal bene- ficiary and to capture additional information for the entire year on patients’ month-by-month Medi-Cal enrollment status, aid category, and health plan (where applicable). These data elements combined with DHS-supplied information on the date in which a California county implemented mandatory Medi-Cal managed care enabled us to classify each hospitalization as occurring for a Medi-Cal beneficiary in fee-for-service or man- aged care as well as Medi-Cal beneficiaries under the COHS, GMC, and Two-Plan models of managed care. To correct for out- of-state hospitalizations of California residents, hospitalizations in states that border California were searched for patients with California ZIP codes in Oregon, Arizona, and Nevada hospital discharge abstracts records for the same time period. Preventable hospitalizations of California residents in these three states totaled to less than 0.2 percent of such hospitalizations within California. Because this analysis used hospitalizations as an indicator of ambulatory care prior to the hospitalization, only those Medi-Cal hospitalizations in which an individual had Medi-Cal coverage in the month before hospitalization were counted. In this way, mis- classification of an uninsured individual who gained Medi-Cal as a result of the hospitalization was avoided. However, this approach required that January admissions be excluded from the analysis because information about an individual’s Medi-Cal eligi- Preventing Unnecessary Hospitalizations in Medi-Cal | 23 bility was available only for the calendar year, and who receive benefits through CalWORKs, pre- it was not possible to determine whether some- ventable hospitalization rates for Medi-Cal bene- one with a January admission was a Medi-Cal ficiaries who were eligible through CalWORKs beneficiary in the previous December. Also versus the SSI program were analyzed separately. because the hospitalization discharge and enroll- The numerator of the rate was the count of hos- ment files were linked to a calendar year, it was pitalizations for the specified conditions in a not possible to accurately calculate admission given month. The denominator population for rates for hospital admissions that resulted in dis- calculating the admission rate for each Medi-Cal charges in a different calendar year. Less than 1 delivery model was obtained from the Medi-Cal percent of admissions had discharges in a subse- Monthly Eligibility File. quent year, and these were excluded from the Also, recognizing that non-randomly distributed analysis. patient and county characteristics could con- Data about number, demographics, eligibility found our results, multivariate Poisson regression category, and health plan type of the entire Medi- analysis was used to model the monthly prevent- Cal population (not just those hospitalized) were able hospitalization rate as a function of the obtained from the DHS Medi-Cal Monthly Medi-Cal delivery model (fee-for-service versus Eligibility File. The enrollment files for years managed care or Two-Plan, COHS, GMC versus prior to 1996 contained information only as of fee-for-service) controlling for admission month, the first month of a quarter (January, April, July, admission year, patient age (0–17 versus 18–64 and October). A linear interpolation method was years), sex, race/ethnicity (African American, used to obtain the estimates for the other eight Asian and Pacific Islander, Hispanic, Non- months of those years. Hispanic White, and Other), and county of residence. The use of an appropriate scale factor A commonly accepted list of conditions defined corrected for any remaining over-dispersion in with diagnostic codes for children and adults was our model.16 The independent variables were used to calculate the number of preventable hos- captured for each discharge and then aggregated pitalizations for Medi-Cal beneficiaries (Table to obtain the number of preventable hospitaliza- 4).14,15 These codes generally rely on the primary tions for groups with each combination of char- diagnosis. acteristics. Such an approach can accommodate The analysis was limited to individuals who were changes in individual characteristics over time, younger than 65 because older individuals were such as type of health plan held by a beneficiary likely to also have Medicare insurance. The as managed care or fee-for-service. Because the analysis of preventable hospitalization rates data from out of state admissions did not among Medi-Cal beneficiaries included those include information about Medi-Cal eligibility who are eligible for both Medi-Cal and category, these were not included in the multi- Medicare. These dually eligible patients variate model. The denominator population for accounted for 45 percent of non-elderly SSI- calculating the admission rate was obtained from linked Medi-Cal beneficiaries. For these Medi- the Medi-Cal Monthly Eligibility File, which had Cal beneficiaries, Medicare was the primary detailed information about each of the independ- payer for hospital and ambulatory care services. ent variables. The coefficient estimates from the Poisson regression model were used to obtain Recognizing that Medi-Cal eligibility categories predicted rates standardized for differences in reflect differences in beneficiaries’ health status group composition. and that most growth in Medi-Cal managed care was among low-income women and children 24 | CALIFORNIA HEALTHCARE FOUNDATION To facilitate comparison of preventable hospital- ization rates from different data sources, monthly admission rates were converted to annual rates. All comparisons highlighted in the text of the report are significant at least to the p<.05 level. Table 4: Ambulatory-Care-Sensitive Conditions and ICD-9 Codes Description ICD-9 Code Angina 4111, 4118, 413 Excludes cases with procedure codes [01-86.99] Asthma 493 Bacterial pneumonia 481 4821, 4823, 4829, 483, 485-486 Excludes patients younger than 2 months and cases with secondary diagnosis of sickle cell [2826] Bronchitis 4660 only if secondary diagnosis is 491, 492, 494, or 496 Cellulites 681, 682, 683, 686 Excludes cases with any procedure codes except 860 where it is the only procedure Congenital syphilis 090 (secondary diagnosis for newborn only) Congestive heart failure 428, 40201, 40211,40291, 5184 COPD 491, 492, 494, 496, 492,494, 496 Dehydration 2765 Dental condition 521-523, 525, 528 Diabetes 2500, 2501, 2502, 2503, 2508, 2509 Failure to thrive 7834 Gastroenteritis 5589 Grand mal seizure disorders 7803, 345 Hypertension 4010, 4019, 40200, 40210, 40290 Excludes cases with procedures 36.01, 36.02, 36.05, 36.1, 37.5, 37.7 Hypoglycemia 2512 Immunization preventable conditions (ages 1 to 5) 033, 390, 391, 037, 045, 3200 3202 Iron deficiency anemia 2801, 2808, 2809 (age less than 5 years) Kidney and urinary tract infection 590, 5990, 5999 Nutritional deficiency 260-262, 2680-2681 Pelvic inflammatory disease 614 (cases with surgical procedure of hysterectomy 683-688 (women only) excluded) Ruptured appendix 5400-5401 Severe ENT infection 382 (excludes cases with procedure code 2001), 462, 463, 465, 4721 Skin graft with cellulites DRG 263, 264 (excludes admission from SNF) Tuberculosis 011-018 Preventing Unnecessary Hospitalizations in Medi-Cal | 25 Endnotes 1. Bindman, A., Yoon, J., and Grumbach, K. 9. Cunningham, P., Trude, S. 2001. “Does 2003. “Trends in Physician Participation in Managed Care Enable More Low Income Medicaid: The California Experience.” The Persons to Identify a Usual Source of Care? Journal of Ambulatory Care Management 26 Implications for Access to Care.” Medical (4); 334–343. Care 39 (7); 716–726. 2. Council on Graduate Medical Education. 10. Coughlin, T., Long, S. 2000. “Effects of Patient Care Physician Supply and Medicaid Managed Care on Adults.” Medical Requirements: Testing COGME Care 38 (4); 433–446. Recommendations (Eighth Report). Rockville, 11. Sisk, J., Gorman, S., Reisinger, A., Glied, S., MD: 1996. DuMouchel, W., Hynes, M. 1996. 3. Medi-Cal Policy Institute. Speaking Out: “Evaluation of Medicaid Managed Care. What Beneficiaries Say About the Medi-Cal Satisfaction, Access, and Use.” The Journal of Program. Oakland, CA: 2000. the American Medical Association 276 (1); 4. Street, L. The Medi-Cal Budget: Cost Drivers 50–55. and Policy Considerations. Oakland, CA: 12. Mittler, J., Gold, M. Building and Sustaining Medi-Cal Policy Institute; March 2002. Physician Networks in Medi-Cal Managed Care 5. Bindman, A., Grumbach, K., Osmond, D., and Healthy Families. Oakland, CA: Medi-Cal et al. 1995. “Preventable Hospitalizations and Policy Institute; May 2003. Access to Health Care.” The Journal of the 13. Oster, A., Bindman, A. 2003. “Emergency American Medical Association 274 (4); Department Visits for Ambulatory Care 305–311. Sensitive Conditions: Insights into 6. Pappas, G., Hadden, W.C., Kozak, L.J., Preventable Hospitalizations.” Medical Care Fisher, G.F. 1997. “Potentially Avoidable 41 (2); 198–207. Hospitalizations: Inequalities in Rates 14. Millman, M. Access to Health Care in America. Between U.S. Socioeconomic Groups.” Washington, D.C.: Institute of Medicine; American Journal of Public Health 87 (5); 1993. 811–816. 15. Billings, J., Zeitel, L., Lukomnik, J., Carey, 7. Gill, J., Mainous, A. 1998. “The Role of T.S., Blank, A.E., Newman, L. 1993. “Impact Provider Continuity in Preventing of Socioeconomic Status on Hospital Use in Hospitalizations.” Archives of Family Medicine New York City.” Health Affairs (Millwood) 12 7 (4); 352–357. (1); 162–173. 8. Lo Sasso, A., Freund, D. 2000. “A 16. McCullagh, P. and Nelder, J. Generalized Longitudinal Evaluation of the Effect of Linear Models. 2d ed. London: Chapman and Medi-Cal Managed Care on Supplemental Hall, 1989. Security Income and Aid to Families with Dependent Children Enrollees in Two California Counties.” Medical Care 38 (9); 937–947. 26 | CALIFORNIA HEALTHCARE FOUNDATION