A Comparison of Health Care Experiences for Medicaid and Commercially Enrolled Children in a Large, Nonprofit Health Maintenance Organization

A Comparison of Health Care Experiences for Medicaid and Commercially Enrolled Children in a Large, Nonprofit Health Maintenance Organization

A Comparison of Health Care Experiences for Medicaid and Commercially Enrolled Children in a Large, Nonprofit Health Maintenance Organization Paul W. ...

72KB Sizes 2 Downloads 16 Views

A Comparison of Health Care Experiences for Medicaid and Commercially Enrolled Children in a Large, Nonprofit Health Maintenance Organization Paul W. Newacheck, DrPH; Tracy Lieu, MD, MPH; Amy E. Kalkbrenner, MPH; Felicia W. Chi, MPH; G. Thomas Ray, MBA; Joel W. Cohen, PhD; Robin M. Weinick, PhD Background.—Proponents of Medicaid managed care have argued that this type of care offers the potential to provide mainstream health care for poor children and the elimination of the 2-tier system of care that has long existed for poor and nonpoor children. However, few studies have attempted to assess whether differences in access, utilization, and satisfaction exist between Medicaid and commercially sponsored children who are enrolled in the same managed care plan. Objective.—To systematically answer the following research question: Within the same large, nonprofit, group-model health maintenance organization (HMO), how do children enrolled in Medicaid compare with children enrolled commercially across the domains of access, utilization, and satisfaction with care? Methods.—We compared access, satisfaction, and utilization of services between Medicaid and commercially sponsored children enrolled in Kaiser Permanente of Northern California during 1998 through use of a telephone survey and administrative data. Kaiser Permanente is a nonprofit, integrated, group HMO that serves 2.8 million members in more than 15 counties in northern California. The sample for this survey included 510 Medicaid-enrolled children and 512 commercially enrolled children. An overall response rate of 82% was achieved. Bivariate and multivariate analyses were used to compare Medicaid and commercially enrolled children. Results.—We found few differences between commercial and Medicaid enrollees across the domains of access, utilization, and satisfaction. Where access differences were present (problems in finding a personal care provider, problems getting care overall, and experiencing 1 or more barriers to care), the differences favored Medicaid-enrolled children. That is, Medicaid enrollees were reported to experience significantly fewer access problems and barriers than commercial enrollees, even after adjustment for confounding factors. Only one difference was found between Medicaid and commercial enrollees across the 6 utilization variables examined (volume of emergency department visits), and no differences were found among the 4 satisfaction and 2 global assessments of care received. Taken together, our results suggest that Medicaid-enrolled children experience as good as or better care than their commercially enrolled counterparts. However, there are other possible explanations for our findings. It may be that families of Medicaid-enrolled children hold their care providers to a lower standard than families of commercially enrolled children, given historic inequities in care between poor and nonpoor families. In addition, some degree of selection bias may be present in our sample, although that is true for both the Medicaid and commercial populations. Conclusions.—Our findings suggest that large commercial HMOs are capable of eliminating the access barriers and stigma traditionally associated with the Medicaid program. However, this conclusion must be tempered with the knowledge that other explanations for our findings may also be at play. KEY WORDS: Medicaid

health care access; health care satisfaction; health care utilization; health maintenance organizations;

Ambulatory Pediatrics 2001;1:28 35

M

edicaid managed care has experienced explosive growth in recent years. For example, between 1991 and 1998, the number of Medicaid beneficiaries enrolled in managed care increased sixfold from 2.7 million to 16.6 million.1 Although managed care en-

rollment statistics are not compiled separately for children, children and their parents have been the principal target of Medicaid managed care initiatives. The expansion of managed care enrollment in state Medicaid programs has both proponents and critics. Proponents argue that managed care offers a systematic method of controlling health care costs while providing enrollees a dedicated source of health care. By providing access to mainstream provider networks, managed care holds the potential for equalizing access between Medicaid and privately insured children. In addition, prepaid managed care offers added predictability and stability to Medicaid spending.2 However, critics argue that the builtin financial incentives to control costs and utilization in

From the Institute for Health Policy Studies, University of California, San Francisco, Calif (Dr Newacheck); Department of Ambulatory Care and Prevention, Harvard Pilgrim Health Care, and Harvard Medical School, Boston, Mass (Dr Lieu); Kaiser Permanente Division of Research, Oakland, Calif (Mss Kalkbrenner and Chi and Mr Ray); and Agency for Healthcare Research and Quality, Rockville, Maryland (Drs Cohen and Weinick). The views expressed in this article are those of the authors, and no official endorsement by the Department of Health and Human Services or the Agency for Healthcare Research and Quality is intended or should be inferred. Address correspondence to Paul W. Newacheck, DrPH, Institute for Health Policy Studies, University of California, San Francisco,

AMBULATORY PEDIATRICS Copyright q 2001 by Ambulatory Pediatric Association

3333 California St, Suite 265, San Francisco, CA 94118 (e-mail: [email protected]). Received for publication May 9, 2000; accepted August 21, 2000.

28

Volume 1, Number 1 January-February 2001

AMBULATORY PEDIATRICS

HMO Experiences of Medicaid and Commercially Enrolled Children

some managed care plans, especially health maintenance organizations (HMOs), can lead to restricted access to certain services and potentially compromise the quality of care provided. Also, special concerns are raised for vulnerable populations, because primary care physicians who are financially at risk may face disincentives to refer their chronically ill patients to specialists.3–5 During the early 1980s, a series of demonstration projects involving child Medicaid beneficiaries were conducted to examine the efficacy of HMOs for low-income children. The literature that examined these demonstration projects has generated mixed findings. For example, whereas one study found increased physician visits among Medicaid children enrolled in HMOs compared with feefor-service Medicaid plans,6 other studies have found reductions in physician visits among children enrolled in group- and staff-model HMOs.7 Freund and Lewit8 reviewed studies on the impact of publicly and privately sponsored managed care on children and pregnant women. They found no consistent evidence for a decrease in hospital use for children in HMOs and little evidence that health outcomes or quality of care differed in HMOs and traditional plans. However, they did find that HMOs were associated with reduced emergency department use and lower rates of referral to specialists among Medicaid-enrolled children. In addition, localized studies have examined service use by Medicaid-enrolled children with chronic and disabling conditions and found that these vulnerable children tend to experience greater difficulty obtaining specialty services in HMOs.9–11 Taken together, early studies of managed care for Medicaid-enrolled children have yielded mixed findings but suggest that certain subpopulations, namely those with significant health problems, may experience barriers in obtaining needed services. A more recent comprehensive review of the literature on Medicaid managed care and children concluded that systematic differences in access and utilization are present, especially between capitated managed care plans and fee-for-service arrangements.12 Specifically, capitated managed care was associated with decreased use of preventive care and emergency department care and reduced rates of specialty referrals when compared with fee for service.12 Although growing and improving over time, the existing literature on managed care and children has several limitations. Many of the most rigorously designed studies of Medicaid managed care are now dated. Moreover, few studies have attempted to assess whether differences in access and utilization exist between Medicaid and commercially sponsored children who are enrolled in the same managed care plan. Indeed, if access and utilization disparities between privately insured and Medicaid-insured children remain after enrollment in the same managed care plan, the goal of mainstreaming will remain elusive. Recently collected survey data from Kaiser Permanente of Northern California provide a new information source for shedding light on differences between Medicaid-enrolled and commercially enrolled children’s health care experiences. We use these data to systematically answer

29

the following research question: Within the same large, nonprofit HMO, how do children enrolled in Medicaid compare with children enrolled commercially across the domains of access, utilization, and satisfaction with care? METHODS We compared access, satisfaction, and utilization of services between Medicaid and commercially sponsored children enrolled in Kaiser Permanente of Northern California through use of a telephone survey and administrative data. Kaiser Permanente is a nonprofit, integrated, group HMO that serves 2.8 million members in more than 15 counties in northern California. During the study period, approximately 4% of Kaiser Permanente’s membership younger than 18 years were Medicaid recipients. California’s Medicaid managed care initiative is county based. In most counties, enrollment in managed care is mandatory for all children in the Medicaid program except those who are blind or disabled. Because of the added potential for selection bias, children whose Medicaid eligibility was based on being blind or disabled were excluded from our analysis. Although enrollment in managed care is mandatory for the remainder of the child Medicaid population, there is a choice of at least 2 plans in each county. Hence, all Kaiser Permanente Medicaid enrollees in our study selected Kaiser Permanente as their health plan. Our sampling universe consisted of all children enrolled in Kaiser Permanente at any time between July 1, 1998, and September 30, 1998, who were younger than 18 years on September 30, 1998. All children who were covered under a Medicaid contract for some or all of this period were eligible for the study. Our comparison population was drawn from all commercially insured children who had no Medicaid coverage during this time. A final sample of 500 Medicaid and 500 commercially insured children was targeted for the study. To obtain 500 completed interviews with each group, 1449 cases were drawn from Kaiser Permanente’s enrollment database by an age-stratified random sampling technique to ensure adequate age representation in the final sample. Of these, 195 participants (13%) were deemed ineligible for the survey because of a variety of reasons, including not being a Kaiser Permanente member when contacted for the survey (n 5 158), language barriers (n 5 19), or other reasons (n 5 18). Of the remaining eligible sample of 1254, interviews were completed for 1022 children, resulting in an overall response rate of 82%. The response rate was slightly higher for the commercially insured group (512 completed interviews for an 87% response rate) than the Medicaid group (510 completed interviews for a 77% response rate). To assess the potential for response bias, we compared the age and sex distributions of respondents and nonrespondents separately for the Medicaid and commercial samples. No statistically significant differences were found in the distributions of respondents and nonrespondents. All of the interviews were conducted in English or Spanish by 1 of 4 trained interviewers. Our preferred re-

30

Newacheck et al.

spondent was the parent who usually brings the sampled child for his or her physician’s visits. When that person was not available, interviews were completed with another guardian who took the child to the physician or with the child if he or she was 15 years or older. Interview times averaged 17 minutes for both Medicaid and commercial enrollees. Interview quality (evaluated by the interviewers after each interview) averaged 4.6 on a 1–5 scale for both study groups. All interviews were completed between December 1998 and September 1999. Informed consent was obtained from all respondents. The telephone survey instrument was designed to collect information on access-related facets of the following topics: 1) place of health care, 2) obtaining a personal physician, 3) frequency and reasons for use of emergency departments, 4) use of and satisfaction with urgent visits, 5) use of and satisfaction with preventive visits, 6) barriers encountered during the last visit, 7) ability to see a specialist if needed, 8) obtaining appropriate care for specific symptoms, 9) satisfaction with clinical staff, 10) overall ratings of personal physician and health care received, 11) use of interpreters, and 12) use of telephone advice nurses. Many of the access and satisfaction survey questions were derived from the Consumer Assessment of Health Plans Study, the National Health Interview Survey, and the Medical Expenditure Panel Survey. Demographic variables included the following: family composition, parents’ work status, mother’s education and age, birth order of child, childcare arrangements, language spoken at home, race/ethnicity, and income. Copies of the questionnaire can be obtained directly from the authors. Additional administrative information on use of outpatient and inpatient services, including outpatient visit counts and frequency and length of hospitalizations, was obtained from Kaiser Permanente’s electronic data systems. Clinic visits, emergency department visits, and hospital days in the year before the interview were analyzed for those sample children having a full year of membership during that time (n 5 408 for Medicaid enrollees, n 5 438 for commercial enrollees). We excluded birth-related hospitalizations from our analyses because it was extremely uncommon for Kaiser Permanente members to have Medicaid coverage at or shortly after birth. We used bivariate and multivariate analysis to assess whether Medicaid-enrolled children differed from commercially insured children across several measures of access, satisfaction, and utilization. The multivariate analyses were conducted to identify the independent effect of sponsorship (Medicaid or commercial) on the outcome variables while controlling for a variety of confounding factors. Variable selection for the multivariate analysis was based on the behavioral model of health care access developed by Andersen and Newman13 and Aday and Anderson.14 This model seeks to explain variation in access and utilization based on predisposing, enabling, and need characteristics. In our models, predisposing variables included age, race, sex, family size and composition, maternal age, and language; enabling variables included family income and maternal education; and need variables

AMBULATORY PEDIATRICS

included the child’s perceived health status, presence of a special health care need, and average annual days spent in bed because of illness. Our multivariate analyses used ordinary least squares regression for continuous outcome variables and logistic regression for dichotomous outcomes. In these analyses, when less than 5% of cases had missing responses for an independent variable, we dropped those cases from the analysis. Income and race were the only variables where more than 5% of cases had missing values; to avoid bias in our regression analyses, we substituted the mean sample income value (calculated separately for the Medicaid and commercial samples) for cases with missing income data and added an indicator variable for cases with missing race data. We used the results of the bivariate and multivariate analyses to calculate unadjusted and adjusted means and odds ratios, which are presented in the text and tables. All results presented in the text and tables have been weighted to reflect the age distribution of Kaiser Permanente’s Medicaid and commercially insured child populations. RESULTS Demographic and Health Characteristics The demographic and health characteristics of Kaiser Permanente’s commercially insured and Medicaid-insured enrollees are shown in Table 1. Substantial demographic differences are apparent between the 2 groups of enrollees. When compared with their commercially sponsored counterparts, the Medicaid enrollee group had a higher representation of younger children, blacks, lower family incomes, single parents, and mothers younger than 21 years (P , .01 for all comparisons). The commercially insured group had a higher representation of Hispanics (P , .01) and non–English-speaking parents or guardians (P , .01). There is some evidence that Medicaid-enrolled children are more likely to have special care needs than those who are commercially insured (27% vs 22%), although the difference is not statistically significant at the .05 level (P 5 .06). No significant differences were found between Medicaid and commercial enrollees across the other measures of health and functional status shown in Table 1. Getting Needed Care Several questions were used in the survey to ascertain problems in obtaining needed care. Although most respondents reported no access problems (Table 2), a significant minority of Kaiser Permanente member children were reported to experience problems in finding a personal care provider (12.0%), getting care overall (18.2%), getting specialty care (27.6%), getting telephone advice (14.5%), getting urgent care (13.4%), and getting preventive care (22.8%). For most of these measures, there were no statistically significant differences between the Medicaid and commercial populations. However, Medicaid enrollees were found to have substantially fewer problems than commercial enrollees in finding a personal care pro-

AMBULATORY PEDIATRICS

HMO Experiences of Medicaid and Commercially Enrolled Children

31

TABLE 1. Demographic and Health Characteristics of Children by Type of Insurance Coverage Distribution of Characteristics, %†

Characteristics Age, y 0–,1 1–,2 2–,5 5–,11 11–,15 $15 Mean (SD) Sex Female Male Race Black, non-Hispanic White, non-Hispanic Other, non-Hispanic Hispanic Language English Non-English Income, $ ,10,000 10,000–,20,000 20,000–,30,000 30,000–,40,000 40,000–,50,000 50,000–,80,000 $80,000 Family size Mean (SD) Family composition Two parents One or no parent Maternal education Some high school or less High school graduate or higher Maternal age, y ,21 $21 Mean (SD) Perceived health status Excellent Very good Good Fair/poor Have a special healthcare need§ Yes No Bed days in the past 12 months Mean (SD)

All Children (n 5 1022)

Medicaid Insured (n 5 510)

Commercially Insured (n 5 512)

3.1 5.5 18.1 37.5 21.0 14.8 8.80 (4.91)

2.8 6.2 21.1 40.5 18.5 11.0 8.06 (4.70)

3.3  4.8 15.0 34.6  23.6 18.7  9.54 (5.00)

47.2 52.8

45.9 54.1

48.5   51.5 

18.7 42.7 15.8 22.8

30.4 37.5 12.5 19.6

7.1 47.8 19.2 26.0

    

.00*

83.8 16.2

91.8 8.2

75.8   24.2 

.00*

11.8 21.4 14.5 10.6 8.6 19.3 13.7

22.9 37.0 16.0 11.2 4.9 6.5 1.6

0.4  5.4 13.0 10.1  12.5 32.5 16.1 

 

.00*

4.70 (1.84)

4.80 (2.06)

4.59 (1.59)

0.07

57.6 42.4

35.8 64.2

79.3   20.7 

0.00*

13.8 86.2

15.0 85.0

12.7   87.3 

.30

2.7 97.3 36.11 (9.07)

4.8 95.2 34.69 (10.82)

40.1 35.9 18.9 5.1

39.5 36.3 19.3 5.0

40.9 35.7 18.6 4.8

    

.93

24.4 75.6

27.0 73.0

21.8   78.2 

.06

2.98 (11.93)

3.46 (15.78)

2.39 (5.52)

.15

 

0.7   99.3  37.53 (6.64)

P Value‡

.00*

.00* .41

.00* .00*

†Data are presented as percents unless otherwise indicated. ‡x2 Tests of independence for children’s characteristics and type of insurance coverage; P-values also presented from t-tests for continuous variables (age, family size, maternal age, bed days in the past 12 months). §A child is defined as having special care needs if he or she 1) is restricted in doing normal activities for at least 12 months; 2) has a special need that has been going on or is expected to go on for at least 12 months; or 3) needs elevated services for at least 12 months. *P , .01.

vider and in getting care overall, even after adjustment for confounding (P , .01 for comparisons for both variables). Barriers to Care As part of the survey, respondents were read a checklist of potential barriers to receiving care for the sample child.

More than two thirds of respondents cited at least one access barrier, with the most prevalent barrier being telephone ‘‘hold’’ times (Table 3). For the most part, there were few statistically significant differences between Medicaid and commercially insured children. However, Medicaid enrollees were less likely than commercial enrollees to experi-

32

Newacheck et al.

AMBULATORY PEDIATRICS

TABLE 2. Access Characteristics of Children by Type of Insurance Coverage Distribution of Characteristics, %

Access Characteristics Problem Problem Problem Problem Problem Problem

finding personal care provider (n 5 763)§ getting care overall (n 5 1011) getting specialty care (n 5 262)\ getting telephone advice (n 5 689)¶ getting urgent care (n 5 659)# getting preventive care (n 5 670)††

All Children (n 5 1022)

Medicaid Insured (n 5 510)

Commercially Insured (n 5 512)

Unadjusted

Adjusted‡

12.0 18.2 27.6 14.5 13.4 22.8

5.6 13.3 24.4 13.4 12.1 21.1

19.2 23.0 30.6 15.7 14.7 24.6

0.25* 0.51* 0.73 0.83 0.80 0.82

0.26* 0.52* 0.60 1.14 0.96 1.09

Odds Ratio†

†The odds ratios are expressed as the odds of each outcome among Medicaid-insured children compared with commercially insured. ‡Adjusted for the following variables: age, sex, race, language, income, family size, family composition, maternal education, maternal age, perceived health, having a special care need and bed days during the past 12 months. §Among those who have had personal providers in the past 12 months. \Among those who have needed to see a specialist in the past 12 months. ¶Among those who have called to get advice in the past 12 months. #Among those who have needed an urgent clinic visit in the past 12 months. ††Among those who have made an appointment for preventive care in the past 12 months. *P , .01.

ence barriers related to missing work (24% vs 32%, P , .05) and were also less likely to experience at least one type of barrier overall (62% vs 72%, P , .01). Receipt of Care in Response to Symptoms Modeled after a series of questions first used in the 1987 National Medical Expenditure Survey, respondents were asked to report on whether sample children experienced symptoms during the previous 12 months that normally require medical attention, including sore throat with high fever for at least 2 days, ear infections or earaches for at least 2 days, and asthma symptoms or wheezing. When any of these symptoms or conditions were reported, the respondent was asked whether treatment was obtained at Kaiser Permanente. The results, shown in Table 4, indicate that few children in the overall sample went without care for ear infections and earaches lasting 2 days or more. However, more than one quarter of sample children did not receive care at Kaiser Permanente for asthma

symptoms or wheezing, and nearly one third did not receive care at Kaiser Permanente for sore throat with high fever. Further analysis of the survey data (not shown) suggests that access barriers may have played a relatively small role in explaining why children did not visit Kaiser Permanente for care. For example, 61% of commercial respondents and 70% of Medicaid respondents who reported their child did not see a physician or nurse at Kaiser Permanente for sore throat with high fever indicated that the symptoms did not seem severe or important enough to warrant a visit. Similarly, 73% of commercial respondents and 67% of Medicaid respondents indicated that they successfully treated their child’s asthma symptoms or wheezing at home and thus did not seek care at Kaiser Permanente. Only one statistically significant difference in seeking care for symptoms was apparent between commercial and Medicaid enrollees before adjustment (sore throat with

TABLE 3. Barriers to Care Experienced at Last Visit by Type of Insurance Coverage Distribution of Barriers, %

Barriers to Care Distance Transportation Missing work Hard to find location Need special childcare Office waiting time Inconvenient hours Appointment waiting time Telephone ‘‘hold’’ time Any barrier above

All Children (n 5 1022)

Medicaid Insured (n 5 510)

Commercially Insured (n 5 512)

Unadjusted

Adjusted‡

7.4 6.4 27.9 2.1 12.4 15.2 15.5 14.4 35.9 66.9

6.9 9.5 23.6 2.2 13.2 14.6 14.9 14.0 33.8 61.6

7.8 3.2 32.1 2.0 11.6 15.8 16.0 14.8 38.0 72.1

0.87 3.17* 0.65* 1.06 1.16 0.91 0.92 0.94 0.83 0.62*

0.54 1.36 0.67** 1.91 0.88 1.22 0.79 1.03 0.85 0.58*

Odds Ratio†

†The odds ratios are expressed as the odds of each outcome among Medicaid-insured children compared with commercially insured. ‡Adjusted for the following variables: age, sex, race, language, income, family size, family composition, maternal education, maternal age, perceived health, having a special care need, and bed days during the past 12 months. *P , .01. **P , .05.

AMBULATORY PEDIATRICS

HMO Experiences of Medicaid and Commercially Enrolled Children

33

TABLE 4. Receipt of Care in Response to Symptoms Children Who Reported Symptoms but Did Not Get Care, %

Symptoms

All Children (n 5 1022)

Medicaid Insured (n 5 510)

Commercially Insured (n 5 512)

Unadjusted

Adjusted‡

31.2 6.2 26.5

25.1 5.3 24.8

36.6 7.1 28.5

0.58* 0.73 0.83

0.9 1.02 1.27

Sore throat with high fever for at least 2 days (n 5 273) Ear infection or earache for at least 2 days (n 5 303) Asthma symptoms or wheezing (n 5 251)

Odds Ratio†

†The odds ratios are expressed as the odds of each outcome among Medicaid-insured children compared with commercially insured. ‡ Adjusted for the following variables: age, sex, race, language, income, family size, family composition, maternal education, maternal age, perceived health, having a special care need, and bed days during the past 12 months. *P , .05.

high fever). However, after adjustment for confounding factors, there were no significant differences between commercially- and Medicaid-enrolled children in their propensity to seek care at Kaiser Permanente facilities. Utilization of Ambulatory and Hospital Care We analyzed likelihood of use and volume of care for users of clinic services, emergency departments, and inpatient hospital care based on encounter data from Kaiser Permanente’s electronic record system. The results, presented in Table 5, show that most sample children had at least one clinic visit in the past year (86%), with much smaller proportions using emergency department services (16%) and inpatient hospital stays (4%). With the exception of emergency department utilization, there were no statistically significant differences in utilization of inpatient or outpatient services between commercial and Medicaid enrollees. Although there was no significant difference in the likelihood of emergency department use between the 2 study samples after adjusting for confounding factors, Medicaid enrollees who used emergency department services had a slightly higher volume of use compared with commercially enrolled users (P , .05), with Medicaid children averaging 1.4 visits per year compared with 1.1 visits for commercially insured children. Overall Care Experience A series of questions was used in our survey to capture the overall care experience of sample children. The first 4 rows of Table 6 show that respondents were generally satisfied with the care provided by their child’s physician and the helpfulness of the office staff. Nearly 70% of respondents rated their child’s personal care provider as a 9 or 10 on a 10-point scale. However, only about half of respondents rated their child’s overall care as a 9 or 10. After adjustment for confounding, no statistically significant differences were found between commercially enrolled and Medicaid-enrolled children on the 4 measures of satisfaction and the 2 global ratings on care. DISCUSSION Our assessment of survey data for a sample of children enrolled in Kaiser Permanente of Northern California suggests that members generally enjoy good access and are satisfied with the care they receive. However, some prob-

lem areas are apparent. For example, more than a fifth of families reported problems in obtaining preventive care for their child, more than a quarter of families reported problems obtaining specialty care for their child, and more than a third reported telephone ‘‘hold’’ times to be a barrier to care. Of interest, we found few differences between commercial and Medicaid enrollees across the domains of access, utilization, and satisfaction. Where access differences were present (problems in finding a personal care provider, problems getting care overall, and experiencing 1 or more barriers to care), the differences favored Medicaid-enrolled children. That is, Medicaid enrollees were reported to experience fewer access problems and barriers than commercial enrollees. Only one small difference in utilization was found (volume of emergency department use), and no differences were found for the satisfaction and global rating of care measures. Taken together, our results suggest that Medicaid-enrolled children experience care as good as or better than their commercially enrolled counterparts. It may well be that Kaiser Permanente is doing an equivalent job serving Medicaid and commercially insured children. If so, Kaiser Permanente may represent an example of how a large commercial HMO can eliminate historical disparities in access and erase the stigma frequently attached to enrollment in Medicaid programs. That Medicaid enrollees are not identified as such on their membership cards or medical files is supportive of this view. Without any obvious indication of enrollment status, Kaiser Permanente medical and office staff would not be expected to treat Medicaid enrollees differently from commercial enrollees. However, there are other possible explanations for our findings that should be considered. First, it may be that families of Medicaid-enrolled children hold their care providers to a lower standard than families of commercially enrolled children. Historically, Medicaid beneficiaries have experienced difficulties in finding providers willing to accept the low payment levels typically offered when Medicaid services are provided on a fee-for-service basis. In contrast, with substantially higher incomes, families of commercially enrolled children may be accustomed to higher levels of service in many areas of life, including medical care, and consequently may hold higher expec-

20.3 0.65 0.66 4.0 1.6 4.0 1.8

5.0 1.9

1.35 1.51*

†The odds ratios are expressed as the odds of each outcome among children enrolled in Medicaid managed care plans compared with those with commercial coverage. ‡Adjusted for the following variables: age, sex, race, language, income, family size, family composition, maternal education, maternal age, perceived health, having a special care need, and bed days during the past 12 months. §Includes inpatient and same-day hospitalizations. \Same-day hospitalizations counted as a 1-day hospital stay. *P , .05.

20.5

0.3* 0.3*

20.2 0.78 1.18

85.0 4.8 14.0 1.1 87.0 4.5 19.0 1.4 86.0 4.6 16.0 1.2

Ambulatory care Percentage with at least 1 clinic visit (excluding emergency department) Annual No. of physician visits among users Percentage with at least 1 emergency department visit Annual No. of emergency department visits among users Hospital care Percentage with 1 or more hospital admission§ Average hospital days among hospital users\

Utilization Characteristics

AMBULATORY PEDIATRICS

20.3

Unadjusted Adjusted‡ Unadjusted

Odds Ratio†

Commercially Insured (n 5 438) Medicaid Insured (n 5 407) All Children (n 5 845)

Distribution of Characteristics

TABLE 5. Utilization for Children With Medicaid Coverage Compared With Those With Commercial Coverage

Adjusted‡

Newacheck et al.

Mean Difference

34

tations for care provided to their children. Although consistent with our findings on access and satisfaction, these arguments do not explain why few differences were found for the utilization indicators, where results are not subject to respondent reporting biases. Second, there is the possibility of selection bias. Although managed care enrollment was mandatory for Medicaid-enrolled children in our study, choosing Kaiser Permanente was entirely voluntary and required an active choice on the part of families. All Medicaid beneficiaries enrolling in Kaiser Permanente have a choice of at least one other managed care plan and the option of disenrolling on a periodic basis or anytime with just cause. Several studies have demonstrated that enrollees tend to report better access and higher satisfaction when they have a choice of health plans.15–17 If Medicaid beneficiaries had been randomly assigned to Kaiser Permanente, our results may have differed. However, it is important to point out that selection bias is also likely to be present for commercial enrollees, since many of them also have other choices of health plans. As a consequence, the results for both groups of enrollees are likely to be affected by selection bias. The implication is that the absolute results for both groups may appear more favorable than if selection bias was not present. Nevertheless, the comparative results between the 2 groups should remain valid as long as the degree of selection bias is equal in both groups. Third, it is possible that respondents for Medicaid-enrolled children were disproportionately reluctant to criticize Kaiser Permanente for fear of reprisal. All respondents to the survey were advised that their responses would be kept confidential and not appear on their child’s medical record. However, respondents were aware that Kaiser Permanente employees conducted the interviews, so it is possible that respondents for Medicaid enrollees may have been less critical out of a concern about losing their Kaiser Permanente membership. Again though, such concerns would not explain the near absence of utilization differences between commercially enrolled and Medicaidenrolled children. Finally, there is the issue of generalizability of our findings. Our study was limited to children subject to mandatory Medicaid managed care enrollment; children eligible for Medicaid because of blindness or disability were not subject to this same mandate and were excluded from our study to avoid selection bias. Hence, our findings cannot be generalized to blind and disabled beneficiaries. In addition, Kaiser Permanente of Northern California, although perhaps typical of large, nonprofit, group- and staff-model (closed-panel) HMOs, is not typical of all managed care systems where Medicaid-insured children have been enrolled en masse. For example, our results are not necessarily generalizable to the experiences of Medicaid beneficiaries enrolled in network-model HMOs, independent practice associations, and other managed care entities. Additional empirical studies are needed to evaluate the impact of other types of HMOs on the health care experiences of children enrolled in Medicaid.

AMBULATORY PEDIATRICS

HMO Experiences of Medicaid and Commercially Enrolled Children

35

TABLE 6. Overall Care Experience by Type of Insurance Coverage Children, %

Care Experience Reporting ‘‘never/sometimes’’ when asked about how often: Office staff were helpful (n 5 903) Physician listened carefully (n5 902) Physician explained things (n 5 904) Physician spent enough time (n 5 905) Rating ,9 on a 0–10 scale for: Personal provider (n 5 831) Overall care (n 5 936)

Medicaid Commercially All Children Insured Insured (n 5 1022) (n 5 510) (n 5 512)

Odds Ratio† Unadjusted

Adjusted‡

13.5 11.9 10.9 15.4

12.8 11.9 10.5 16.0

14.1 11.9 11.4 14.7

0.89 1.00 0.91 1.11

0.73 1.10 0.85 0.90

30.1 47.2

31.0 42.3

29.1 52.3

1.09 0.67*

1.50 1.01

†The odds ratios are expressed as the odds of each outcome among Medicaid-insured children compared with commercially insured children. ‡Adjusted for the following variables: age, sex, race, language, income, family size, family composition, maternal education, maternal age, perceived health, having a special care need, and bed days during the past 12 months. *P , .01.

CONCLUSION Our findings suggest that large commercial HMOs may be capable of eliminating the access barriers and stigma traditionally associated with the Medicaid program. However, this conclusion must be tempered with the knowledge that other explanations for our findings may also be at play. Additional studies that use prospective designs and follow up populations in different managed care models are needed to definitively address whether Medicaid managed care is appropriately meeting the health care needs of child beneficiaries.

6. 7.

8. 9.

10.

ACKNOWLEDGMENT The authors appreciate the insightful comments of the anonymous reviewers. Those comments helped strengthen this article. Supported by a grant from the Community Services Fund, Kaiser Foundation Research Institute.

11.

REFERENCES

13.

1. Kaiser Commission of the Future of Medicaid and the Uninsured. Medicaid and Managed Care. Washington, DC: The Henry J. Kaiser Family Foundation; 1999. 2. Holahan J, Zuckerman S, Evan A, Rangarajan S. Medicaid managed care in thirteen states. Health Aff (Millwood). 1998;17:43– 63. 3. Newacheck PW, Hughes D, Stoddard J, Halfon N. Children with chronic illness and Medicaid managed care. Pediatrics. 1994; 93:497–500. 4. Hughes D, Luft H. Managed care and children: an overview. Future Child 1998;82:25–38. 5. Simpson L, Fraser I. Children and managed care: what research

12.

14. 15.

16.

17.

can, can’t, and should tell us about impact. Med Care Res Rev. 1999;56:13–26. Laufenberg H. Medicaid patients’ use of HMO services. Wis Med J. 1989;88:19–21. Hurley RE, Freund DA, Paul JE. Managed Care in Medicaid: Lessons for Policy and Program Design. Ann Arbor, Mich: Health Administration Press; 1999. Freund DA, Lewit EM. Managed care for children and pregnant women: promises and pitfalls. Future Child. 1993;3:92–122. Fox HB, Wicks LB, Newacheck PW. Health maintenance organizations and children with special health care needs: a suitable match? Am J Dis Child. 1993;147:546–552. Karlson TK, Sumi MD, Braucht SA. The Impact of Health Maintenance Organizations on Accessibility, Satisfaction and Costs of Care for Children With Special Health Care Needs. Madison, Wis: Center for Health Systems Research and Analysis; 1990. Horwitz SM, Stein R. Health maintenance organizations vs. indemnity insurance for children with chronic illness: trading gaps in coverage. Am J Dis Child. 1990;144:581–586. Szilagyi PG. Managed care for children: effect on access to care and utilization of health services. Future Child. 1998;8:39–59. Andersen R, Newman J. Societal and individual determinants of medical care utilization in the United States. Milbank Q. 1973;51:95–124. Aday L, Anderson R. A framework for the study of access to medical care. Health Serv Res. 1973;9:208–220. Schur CL, Berk ML. Choice of health plan: implications for access and satisfaction. Health Care Financing Rev. 1998;20: 29–43. Gawande AA, Blendon R, Brodie M, Benson JM, Levitt L, Hugick L. Does dissatisfaction with health plans stem from having no choices? Health Aff (Millwood). 1998;17:184–194. Ullman R, Hill JW, Scheye EC, Spoeri RK. Satisfaction and choice: a view from the plans. Health Aff (Millwood). 1997;16: 209–217.