Objective. State Medicaid programs are obligated by federal law to ensure that enrolled children have access to care and services to the same extent as other children in the same geographic area. Because most children in the United States receive health care from private primary care physicians, participation by private, office-based primary care pediatricians is critical to meeting this equal access obligation. The objective of this study was to document variations in Medicaid participation of private office-based primary care pediatricians across states and to examine the effects of payment levels, prevalence of capitated Medicaid payment, and paperwork concerns on participation.
Methods. Survey data collected from 3773 primary care pediatricians who practice in private office-based settings were analyzed with Medicaid physician payment data from other sources. Univariate analyses and a multiple regression were used to examine the effects of payment level, prevalence of capitated Medicaid payment, and paperwork concerns on private primary care pediatricians’ participation in state Medicaid programs.
Results. Results revealed substantial state-to-state variation in respondents’ participation in Medicaid. Univariate analyses found that participation increased with state Medicaid payment levels but decreased as the proportion of Medicaid enrollees with primary care capitated payments rose and as paperwork concerns increased. With physician workforce held constant, a regression analysis showed that pediatrician participation in Medicaid increased significantly with Medicaid payment but decreased as the proportion of capitated Medicaid patients increased and as paperwork concerns rose.
Conclusions. This study found that low payment, capitation, and paperwork concerns all relate to low Medicaid participation by primary care office-based pediatricians. It behooves state policy makers to address these 3 factors to ensure sufficient primary care physician capacity to serve appropriately children who are enrolled in state Medicaid programs.
Health care in the United States is financed and organized primarily through employer-based commercial insurance. However, 2 vulnerable populations—the elderly and children—generally do not participate in the workforce. The elderly have a federally financed universal program, Medicare, that does not require means testing, has uniform eligibility and benefits across all states, and has federally determined payments for physician services. Children have 2 joint federal-state financed public programs, Medicaid and State Child Health Insurance Program. Both of these programs target low-income families by requiring means testing for eligibility and have substantial variations in eligibility, benefits, and payments for physician services across states.1
Although states have the responsibility to design and implement their Medicaid programs for children, the enabling Title XIX federal legislation addresses the need for equity by requiring states to ensure that Medicaid-eligible children have access to care and services to the same extent as other children in the geographic area. The law as specified in 42 USCA §1396a(a)(30)(A) explicitly states that the care and services under the Medicaid program are to be available “at least to the extent that such care and services are available to the general public in the geographic area.” This “equal access” federal statute has special relevance for access by Medicaid beneficiaries to private primary care pediatricians and family physicians. The majority of children in the United States receive their preventive health care from private primary care pediatricians or family physicians. According to the National Ambulatory Medical Care Survey, >80% of ambulatory pediatric primary care visits took place in private physician offices in 1998.2 Compliance with the law necessitates that children who have their insurance paid by Medicaid have the same choice of primary care physicians as do children with private insurance. Thus, it is critical to have the participation of pediatricians in private office settings in Medicaid if states are to maintain sufficient capacity to ensure equal access to care. Access to private primary care physicians not only satisfies the equal access requirement but also is a cost-effective way to ensure that children receive the services they need. Research has shown that having a private practice physician as a usual source of care can decrease total Medicaid expenses by one third.3
The goal of this study was to document the current level of participation of private office-based primary care pediatricians in state Medicaid programs and to determine how variations in payment levels, prevalence of capitated managed care, and concerns with Medicaid administrative policies and procedures influence Medicaid participation rates across states. Previous studies have found private practice physicians unwilling to participate fully in the Medicaid program for reasons related to reimbursement, paperwork burden, practice capacity, and others.4–6 This study assessed the following 3 hypotheses while controlling for child health care physician workforce ratio, a factor underlying the supply of physician services and potentially affecting physician acceptance of patients independent of issues examined in this article:
Hypothesis 1: States with higher primary care-related physician payments will have a higher proportion of primary care private office-based pediatricians who accept all Medicaid children seeking care in their practice.
Hypothesis 2: States with more efficient administrative systems and less physician concerns with paperwork burdens will have a higher proportion of primary care private office-based pediatricians who accept all Medicaid children seeking care in their practice.
Hypothesis 3: States that have a lower proportion of Medicaid enrollees in managed care plans with primary care capitated payments will have a higher proportion of primary care private office-based pediatricians who accept all Medicaid children seeking care in their practice.
State-level data on 1) private office-based primary care pediatrician participation in the Medicaid program, 2) ratios of capitated Medicaid patients to fee-for-service or reduced fee-for-service Medicaid patients, and 3) pediatricians’ ratings of paperwork concerns were obtained from a survey of pediatrician participation in public and private health insurance programs. This survey was mailed to a state-stratified random sample of 13 212 pediatrician Fellows of the American Academy of Pediatrics (AAP) from December 1999 to April 2000. Respondents to the AAP Participation Survey were asked whether they accepted all, some, or none of the Medicaid and non-Medicaid patients who contacted them. Practice characteristics, ratings of factors that affect Medicaid participation, and demographic data were also collected as part of the survey. A response rate of 67% was obtained after a postcard reminder and 4 additional mailings to contact nonresponders.
To assess how representative the AAP Survey respondents were of the nation’s pediatricians, we compared characteristics of the 2000 AAP survey respondents to American Medical Association (AMA) published statistics of US pediatricians from 1998 to 1999.7,8 Results of the comparison (Table 1) suggest that respondents to the AAP Survey resemble pediatricians in the nation in terms of gender distribution (46.3% vs 47.9% female), employee status (50.0% vs 50.4% employee), and age distribution for pediatricians ages 35 through 64. There were fewer AAP Survey respondents in the youngest (<35) and the oldest (≥65) age categories. The underrepresentation of pediatricians younger than 35 likely accounted for a lower proportion of residents (7.8% vs 16.9%) and a smaller percentage of international medical graduates (22.4% vs 33.7%) among the AAP Survey respondents.
Two indicators of state Medicaid payment levels were obtained. One is an overall non-age-specific measure of Medicaid to Medicare fee ratio for primary care services published in a 1999 Urban Institute report.7 This measure includes 1998 Medicaid to Medicare fee ratios for primary care services in 41 states and the District of Columbia. The other measure focuses on pediatric primary care services based on Medicaid payment data collected for 45 states and the District of Columbia in a 1998/1999 AAP Survey of Medicaid Directors.8 This child-specific Medicaid to Medicare fee ratio includes payments for 3 frequently used pediatric primary care codes weighted by utilization rates pertaining to these services as reported in an AAP survey of service utilization in 2000.9
State child health care physician workforce ratios were obtained from an AAP workforce report based on US Census Bureau projections of the 1998 child population and the 1998 AMA Masterfile of pediatricians and family and general practice physicians.10
To examine Medicaid participation among private primary care pediatricians in this study, we excluded all pediatric subspecialists as well as primary care pediatricians who practice primarily in safety net settings, which include medical schools and local, state, or US government hospitals and clinics. Of a total of 7149 survey respondents in direct patient care, 4282 (59.9%, unweighted) are primary care pediatricians. Among this group, 3773 (88.1%) practice primarily in private office-based settings, including solo, 2-physician, pediatric group, and multispecialty group practices; health maintenance organizations; and private hospitals. Participation data used in this analysis are based on the last group only.
Outcome Measures: Program Participation
On the basis of the AAP participation survey, 2 alternative measures of private office-based primary care pediatrician participation in Medicaid are used in this study. The first, “full participation,” refers to the proportion of private office-based primary care pediatricians in the state who accept all Medicaid patients who request care. The second, “relative participation,” refers to the ratio of pediatricians who accept all Medicaid patients versus those who accept all non-Medicaid patients. A ratio of 1 suggests equal access by Medicaid and non-Medicaid patients to primary care pediatricians’ care. A ratio >1 indicates better access by Medicaid patients than non-Medicaid patients, and a ratio <1 suggests poorer access by Medicaid patients. Although the first measure reflects more directly how accessible providers are to Medicaid patients, the second measure provides a more accurate assessment of equal access by adjusting for those providers who cannot accept more patients because of full practices (Table 2).
Data were analyzed using the SPSS statistical package.11 Full and relative participation rates for each state were computed for each quartile of 1) state Medicaid to Medicare payment ratios, 2) the severity of pediatrician concerns with paperwork, and 3) the prevalence of Medicaid capitation payments. Mean estimates for the key subgroups were computed. Pearson’s correlation coefficients were computed for all variables examined in relation to participation. A multiple regression analysis examining factors that influence private primary care pediatricians’ willingness to accept Medicaid patients relative to non-Medicaid patients was performed on aggregated state level data, using as regressors the following predictor measures.
Medicaid Payment Levels
Two indicators of state Medicaid payments for primary care services were used in this study. The first is 1998 Medicaid to Medicare fee ratios for non-age-specific primary care services, which include services represented by the following Current Procedural Terminology codes: 99203, 99213, 99214, 99244, and 93000.7 The second is the 1998/1999 Medicaid to Medicare fee ratios for 3 frequently used pediatric services (99391, 99213, and 99214) compiled from the 1998/1999 AAP Survey of Medicaid Reimbursement8 and the 2000 AAP Utilization Survey.9 (See Table 3 for code descriptions; see Table 4 for Medicaid to Medicare ratios for non-age-specific and pediatric primary care services.)
Although neither indicator was fully representative of Medicaid fees paid for services to children and despite the gap that exists between the time the participation data (December 1999 to April 2000) and the payment data (1998 and 1999) were collected, they were significantly correlated (r = 0.58, P < .01). The largest discrepancy between the 2 payment indicators occurred with rates for New York and Texas. In New York, pediatric providers are paid substantially higher rates than regular Medicaid rates when they participate in the Preferred Physicians and Children Program.12 Texas Medicaid fees, ranked highest among all states at 146% of Medicare rates in the Norton study,7 were not confirmed by the Texas Department of Health. When these 2 states were excluded, correlation of the 2 payment indicators increased to 0.89. In the regression analysis reported in this study, the second payment indicator is used because it includes preventive visits for infants (99391), the second most frequently used code for pediatric providers.9
Pediatrician Concerns With Paperwork
Regardless of whether respondents to the 2000 AAP Participation Survey fully participate in the Medicaid program, they were asked to rate a number of reasons that they and their colleagues might decide to limit their Medicaid participation or not to participate at all. For the statement, “It takes too long to complete the necessary paperwork,” respondents chose 1 of the following ratings: “very important,” “somewhat important, ” or “not at all important.” The proportion of respondents in each state who rated the paperwork factor as “very important” was used as an index of the severity of paperwork concerns among pediatricians in the state (Table 5). For examining the possibility of a potential built-in relationship between participation and ratings of paperwork concerns, which would reflect an after-the-fact justification not to participate rather than paperwork burden as a cause in the first place, individual respondents’ rating of paperwork concerns as a participation barrier was cross-tabulated by whether he or she accepted all Medicaid patients and tested for statistical significance.
Ratio of Capitated to Traditional or Reduced Fee-for-Service Medicaid Patients
Percentages of Medicaid patients in 1) capitated managed care plan and 2) traditional or reduced fee-for-service plans reported in the AAP Participation Survey were averaged by state for all respondents who indicated that they accepted all Medicaid patients who sought care. The ratio of capitated Medicaid patients in the state (Table 5) was derived as follows: % patients in capitated payment systems ÷ % patients in traditional or reduced fee-for-service systems.
Physician Capacity for Child Health Care
A child health physician workforce ratio, expressed as the number of child health care physicians (including pediatricians, family physicians, and general practitioners) per 100 000 children for each state,8 was obtained and used as a control variable (Table 5).
AAP Survey Respondents and Their Representativeness of US Pediatricians
A comparison of the 2000 AAP Survey respondents to pediatrician statistics published by the AMA13, 14 (Table 1) suggests that respondents to the AAP Survey resemble pediatricians in the AMA Masterfile in terms of gender distribution (46.3% vs 47.9% female), employee status (50.0% vs 50.4% employee), and age distribution for pediatricians ages 35 through 64. There were fewer AAP Survey respondents in the youngest (<35) and the oldest (≥65) age categories. The underrepresentation of pediatricians younger than 35 likely accounted for a lower proportion of residents (7.8% vs 16.9%) and a smaller percentage of international medical graduates (22.4% vs 33.7%) among the AAP Survey respondents.
Broad State Variation in Medicaid Participation by Private Office-Based Primary Care Pediatricians
The findings of the AAP Participation Survey documented substantial state-to-state variation in 1) the proportion of private office-based primary care pediatricians who accept all Medicaid patients who request care (“ full participation”) and 2) the ratio of pediatricians who accept all Medicaid versus all non-Medicaid patients (“relative participation”).
Although 95.7% of respondents in North Dakota reported that they accept all Medicaid patients, only 19.6% of respondents in Tennessee reported the same, whereas the national average stands at 54.6%. Relative to non-Medicaid patients, respondents in 10 states reported that they were as or more open to Medicaid patients, whereas respondents in the remaining 40 states and the District of Columbia reported a tendency to be less open to Medicaid patients. State-specific findings for full and relative participation are displayed in Table 2.
Medicaid Payments and Pediatrician Participation
Overall, our univariate analyses indicate a positive relationship between state Medicaid payment levels and pediatrician participation. Figures 1 and 2 show the full and relative participation of primary care pediatricians who practice in private offices in relation to the quartile averages of 2 Medicaid payment indicators. In states with payments in the lowest quartile, only approximately half of their private primary care pediatricians are willing to serve all Medicaid patients who request care. In the lower paying states, for every 10 pediatricians who would accept all non-Medicaid patients, only approximately 7 would accept all Medicaid patients. In states with payments in the highest quartile, >65% of the primary care pediatricians are willing to serve all Medicaid patients who request care. By contrast, in the higher paying states, for every 10 pediatricians who would accept all non-Medicaid patients, >9 would accept all Medicaid patients.
Regardless of the payment indicator used, both full and relative participation rates are higher in quartiles of higher Medicaid payment levels, with the exception of full participation by the third (69.43%) and fourth (65.08%) fee quartiles based on fees for 99391, 99213, and 99214.
Pediatrician Concerns With Paperwork and Medicaid Participation
Figure 3 illustrates the association between participation and respondent ratings of paperwork concerns as a reason for limiting Medicaid participation or not participating at all. The states with average higher concerns about paperwork had lower participation rates. Meanwhile, the respondent level analysis showed that although 37.6% of full participants rated paperwork concerns as “very important,” 38.8% of limited and nonparticipants did the same. The difference was nonsignificant (Pearson’s χ2 = 0.71, P = .40, N = 5190), suggesting that ratings of paperwork burden are attributed to variations among states rather than the individual respondent’s attempt to justify a participation decision.
Prevalence of Capitated Payment and Medicaid Participation
Figure 4 depicts how participation varies with the ratio of capitated to traditional or reduced fee-for-service Medicaid patients. States with lower proportions of children in capitated Medicaid plans had higher full and relative pediatrician participation rates.
Effects of Payment, Paperwork Concerns, and Prevalence of Capitation on Relative Participation: Results of a Multiple Regression Analysis
The data on relative participation were analyzed in a multiple regression analysis, using as regressors state Medicaid to Medicare payment ratios based on 3 frequently used pediatric services (99391, 99213, and 99214), pediatrician concerns with paperwork burden, the ratio of capitated to traditional or reduced fee-for-service Medicaid patients, and child health physician workforce capacity. The regression (Table 6) was a moderate fit (R2 adj = 47.4%), and the overall relationship was significant (F4,41 = 11.142, P < .01). With the physician workforce factor held constant, participation increased with payment (P < .01) but decreased as paperwork concerns increased (P < .01) and as odds of seeing capitated Medicaid patients increased (P < .01). Correlations among the dependent, independent, and control variables are listed in Table 7.
The use of both full and relative Medicaid participation of primary care pediatricians who practice in private offices is the appropriate outcome measure in this study for several reasons. First, being able to have a primary care pediatrician is a fundamental component of equal access to care as defined in the federal statute. Second, the most efficient way to enhance pediatric capacity involves increasing the full participation of primary care pediatricians in private practice. These pediatricians are likely to be much more sensitive to factors such as payment rates, capitated Medicaid managed care, and administrative concerns than primary care pediatricians who practice in settings dedicated to caring for underserved populations such as public/teaching hospitals, public health clinics, and community health centers. Pediatricians in these safety net-type settings already accept all Medicaid patients according to organizational policy. Third, full and relative Medicaid participation both are useful ways of monitoring access. Each is superior to the more frequently used indicator of having a Medicaid billing number and having submitted a claim for at least 1 Medicaid physician during a year. Relative participation provides a more accurate measure of equal access to pediatric primary care services because it adjusts participation by whether the physician’s practice is full and closed to all new patients. However, full participation rates measure total pediatric primary care capacity regardless of whether pediatrician practices lack the capacity to take new patients of any type.
The findings of this study support hypothesis 1. States with higher pediatric primary care service payments have a higher proportion of primary care private office-based pediatricians who accept all Medicaid children who seek care in their practice. The data indicate a strong relationship between low payment and low participation rates, as states in the lower quartiles as identified with both Medicaid payments indicators have significantly lower participation rates than those in higher paying states. Among the lowest participating states, as measured by the proportion of primary care pediatricians in private office-based settings who accept all Medicaid patients, are California (33%), New Jersey (37%), and Michigan (39%)—all states in the lowest quartile by payment rate. The general findings are consistent using both full and relative measures of pediatrician participation and both Medicaid payment indicators.
The findings of the multivariate regression analysis support hypotheses 2 and 3 as well. The greater odds of receiving capitated payments for serving Medicaid patients and high paperwork concerns both contribute significantly to provider tendency to accept Medicaid patients less than they accept non-Medicaid patients after controlling for physician supply.
With regard to hypothesis 2, however, it is not clear whether the observed lower levels of participation associated with greater odds of receiving capitated payments for serving Medicaid patients are attributable to 1) generally lower physician payments in capitated systems or 2) state managed care contracting policies, which may restrict the provider pool. Although some may argue that pediatricians who want to participate more fully can attempt to join those plans where Medicaid patients are enrolled, any restriction presents additional barriers and potentially hampers provider participation. Managed care payment and contracting practice is a complex area that merits examination beyond the scope of the current study.
Given the importance of the participation of private primary care physicians in state Medicaid programs, data recently compiled by the AAP and reported in this article raise significant concerns about the level of participation of primary care pediatricians in private practice settings. These data indicate a strong relationship between low payment and low participation rates. The findings hold true using 2 alternative measures of pediatrician participation: 1) the proportion of private office-based primary care pediatricians who accept all Medicaid patients who request care (“full participation”) and 2) the ratio of pediatricians who accept all Medicaid versus all non-Medicaid patients (“relative participation”). The relative participation measure has the advantage of adjusting for practices that lack the capacity to take new patients of any type. By both measures (full and relative participation), providers in states in the lower quartiles of Medicaid payments in terms of 1) overall payment for primary care services7 and 2) child-specific payments for 3 frequently used primary care pediatric codes8,9 have significantly lower participation rates than those in higher paying states.
This analysis found that pediatricians’ concerns about the administrative inefficiencies of the Medicaid program as measured by paperwork concerns and the prevalence of Medicaid capitated payments also affect pediatricians’ willingness to participate. Payment rates, paperwork burden, and managed care contracting and payments all are factors under the control of state public policy makers. State decision makers would be wise to address these 3 factors that contribute to private primary care pediatricians’ willingness to participate in Medicaid if the state is to achieve the equal access to health services guaranteed to each child enrolled in Medicaid.
- Received July 30, 2001.
- Accepted April 12, 2002.
- Reprint requests to (S.B.) Children’s Hospital, 1056 East 19th Ave, B032, Denver, CO 80218. E-mail:
- ↵Holahan J, Liska D. Variations in Medicaid Spending among States. Assessing New Federalism: Issues and Options for States Series (Number A-3). Washington, DC: Urban Institute; 1997. Available at: newfederalism.urban.org/html/anf_a3.htm. Accessed June 26, 2001
- ↵National Ambulatory Medical Care Survey. Hyattsville, MD: National Center for Health Statistics; 1998
- ↵Cohen JW, Cunningham PJ. Medicaid physician fee levels and children’s access to care. Health Aff.1995;Spring:255– 262
- ↵Norton S. Recent Trends in Medicaid Fees, 1993–1998. Washington, DC: Urban Institute; 1999. Available at: newfederalism.urban.org/pdf/discussion99–12.pdf. Accessed June 26, 2001
- ↵1998/1999 Medicaid Reimbursement Reports. Elk Grove Village, IL: American Academy of Pediatrics; 1999. Available at: www.aap.org/research/medreim.htm. Accessed June 26, 2001
- ↵Utilization Survey. Elk Grove Village, IL: American Academy of Pediatrics; 2000
- ↵Physician Workforce: Ratios for Child Health, 1988, by County, HSA and State. Elk Grove Village, IL: American Academy of Pediatrics; 2000. Available at: www.aap.org/research/complete.pdf. Accessed June 26, 2001
- ↵SPSS for Windows, Version 9.0.0. Chicago, IL: SPSS; 1998
- ↵Child Health Financing Report. Elk Grove Village, IL: American Academy of Pediatrics;1990;7(3) :1 and 4
- ↵Physician Socioeconomic Statistics, 1999–2000 Edition. Chicago, IL: Center for Health Policy Research, American Medical Association
- ↵Physician Characteristics and Distribution in the US, 2001–2002 Edition. Chicago, IL: Center for Health Policy Research, American Medical Association
- Copyright © 2002 by the American Academy of Pediatrics