Background. Gatekeeping has been a central strategy in the cost-containment initiatives of managed care organizations. Little empirical research describes the impact of switching into a gatekeeping plan on health care expenditures and utilization for children.
Objective. To determine the likelihood of a parent with a chronically ill child enrolling in a health plan with gatekeeping, as well as the effects of gatekeeping on health care expenditures and utilization for children, especially those with chronic conditions.
Design. We followed a cohort of 1839 children who either voluntarily switched to a gatekeeping plan or remained in an indemnity plan from 1991 through 1994. Study participants were children of employees of a large hospital. The gatekeeping plan was virtually identical to the previous indemnity plan except for lower monthly employee contribution and the requirement for a primary care physician to preapprove subspecialty referrals. We determined the likelihood of a household containing a child with a chronic condition enrolling in the gatekeeping plan, as well as mean annual total, subspecialty, and primary care expenditures and utilization for all children and children with chronic conditions.
Results. Households switching to gatekeeping were less likely to have children with chronic illness (8% vs 15%). Total and subspecialty expenditures for all children decreased more in the gatekeeping group (53% and 59%, respectively) than in the indemnity group (11% and 6%, respectively). For children with chronic conditions, mean visits to subspecialists decreased 57% in the gatekeeping group but increased 31% in the indemnity group. Mean visits to primary care physicians decreased 23% in the gatekeeping group compared with 13% in indemnity group.
Conclusion. Parents of children with a chronic condition were much less likely than other parents to switch to a gatekeeping plan. Switching to gatekeeping was associated with reduced visits to specialists but did not increase the involvement of primary care physicians in the management of children with chronic conditions. The implications of these findings for the health of children are unknown.
- managed care
- child health services
- primary care
- chronic illness
The requirement for a designated primary care physician to preapprove subspecialty referrals is commonly referred to as gatekeeping. Gatekeeping has been a central component of managed care strategies to contain health care costs.1 An assigned primary care physician is expected to contain health care costs by managing most conditions and referring patients to more expensive subspecialists only when circumstances warrant. In addition, although having a primary care physician increases use of preventive health services and improves coordination of care when multiple problems require subspecialist services,2 evidence that gatekeeping ensures these benefits is lacking.
Critics charge that gatekeeping restricts access to appropriate services, and patients have objected to this barrier placed between them and subspecialty care.3–5 Patients with chronic conditions who have an established relationship with a subspecialist may particularly object to gatekeeping.
Previous studies have shown that adults with chronic illness were less likely to switch into a managed care plan and that expenditures and utilization were usually lower for those who did switch compared with those who did not.6–9 Although 1 case-control study suggested that managed care plans enroll fewer children with chronic illness,10 no previous study has addressed whether or not parents of children with chronic conditions avoid gatekeeping per se. In addition, no previous study has isolated gatekeeping from other managed care cost containment strategies to assess the independent effect of the gatekeeping characteristic on the likelihood of switching plans, or on costs or use of care for children.
Medicaid gatekeeping demonstration projects in the 1980s showed decreased likelihood of a child seeing a subspecialist, but little change in the use of primary care.7 This study by Hurley and colleagues7 did not identify children with chronic conditions and therefore could not assess physician utilization for this population. The report by Martin and colleagues11 of a randomized trial of gatekeeping found modest reductions in costs and ambulatory subspecialty use associated with gatekeeping, but this trial was too small to consider effects on children independently. Forrest and colleagues12 recently addressed gatekeeping in pediatric practice through a multisite study of referrals. Although this study suggested higher rates of subspecialty referral in gatekeeping plans than indemnity plans, the design used physicians to identify and enroll patients and therefore could not assess patient utilization of subspecialists that bypassed the primary care physician. Furthermore, that study examined referrals rather than actual utilization.
In this study of an employed population, we sought to determine 1) the likelihood that a household with a chronically ill child would switch to a gatekeeping plan, and 2) the effect of switching to a gatekeeping plan on expenditures and utilization for all children, as well as children with chronic conditions. We hypothesized that 1) parents of children with chronic conditions would be less likely than other parents to choose a plan with gatekeeping, 2) overall utilization of subspecialists would decrease for children in gatekeeping, and 3) children with chronic conditions who switched to the gatekeeping plan would have lower but continued utilization of subspecialists and increased use of primary care physicians.
We obtained all medical claims for a cohort of employees of Massachusetts General Hospital (MGH) who were continuously enrolled in Massachusetts Blue Cross Blue Shield (MBCBS) medical insurance benefit plans from 1991 through 1994. Participants were enrolled in either of 2 plans: an indemnity plan or a gatekeeper plan. The indemnity plan, offered throughout the 4 years of the study, provided unlimited access to any physician contracting with MBCBS. The gatekeeping plan was initially offered to employees in November and December of 1992, with an activation date of January 1, 1993. The gatekeeping plan had a lower employee premium contribution (average difference approximately $30/month).
The gatekeeping plan required each enrolled member to select a primary care physician from a list of practitioners provided by the insurer. Primary care physicians had to preapprove all subspecialty visits. The gatekeeping plan provided automatic eligibility for all pediatricians, internists, and family practitioners as primary care physicians, but generally prohibited designation of subspecialists as primary care physicians. Both the gatekeeping and indemnity plans reimbursed physicians on the same per visit rates; thus physicians had no financial incentive to limit costs of care in the gatekeeping group.
MGH employed 10 284 people who obtained health insurance through their hospital affiliation between 1991 and 1994. Before the initiation of the gatekeeping plan in 1993, most employees (8617) were covered by 1 of 2 MBCBS indemnity benefit plans: Master Health Plus (6656), or Master Medical (1961), a more basic plan with significant deductibles and coinsurance provisions. We excluded employees with Master Medical because this plan was not comparable to the gatekeeping plan. We also excluded employees for whom there was disagreement between MGH and MBCBS records concerning enrollment status (2173). From the set of continuously enrolled subscribers in the MBCBS Master Health Plus indemnity plan or the gatekeeping plans (3311), we identified all subscribers with family or dependent coverage (1061) and included the claims of the benefit holder and all of their children age <18 years on Jan 1, 1991 (1839) in this study. The majority of households had either 4 years of claims in the indemnity plan (565) or 2 years of indemnity and 2 years of gatekeeping (496; Fig 1). We excluded small numbers of households that either switched into gatekeeping in 1994 (20 households, 73 children) or switched back from gatekeeping to indemnity in 1994 (11 households, 42 children).
Preparation of Data Files
MBCBS and MGH identified all employees and their spouses/dependents who were covered through MGH by MBCBS plans in 1991 through 1994. Information on job category and salary range from MGH was merged with MBCBS claims using alias numbers, and all unique identifiers were removed. Blue Shield (professional) claims included total payments, place of service (inpatient, outpatient, nursing home, office), physician specialty, type of service, procedure code, and a single diagnosis code. Blue Cross (institutional) claims data included total payments, ancillary expenditures, room and board expenditures, and place of service. Extended benefits claims included total payments and location.
Definition of Variables
The principal outcome in the analysis of health plan switching was a dichotomous variable defined as a household (enrollee plus at least 1 dependent under age 18) which either 1) enrolled in the gatekeeping plan by January 1, 1993, or 2) remained in the indemnity plan.
An expenditures outcome variable was defined as the payments made by the insurer (MBCBS), adjusted to 1993 dollars. Annual mean total expenditures were determined by adding all professional, institutional, and extended benefits expenditures for each child for each year of the study and then dividing by the number of study participants. Mean annual primary care and subspecialty expenditures were determined by summing total professional expenditures for primary care and subspecialists and then dividing each sum by the number of children.
Utilization outcome variables included mean annual number of visits to primary care physicians and subspecialists. Well child visits were identified using International Classification of Diseases-Ninth Revision (ICD-9) codes and current procedural terminology codes for well-child examination or general medical examination. Hospital admissions identified from institutional claims were summed independently for indemnity and gatekeeping plans and adjusted to account for differences in the size of the groups.
Primary predictors of switching included children and adults with chronic conditions. We identified children as having a chronic condition by a previously developed algorithm using specificICD-9 codes.13,,14 We excluded laboratory and radiology claims to avoid the inclusion of ICD-9 codes used to “rule-out” a diagnosis. Adults with a chronic condition were identified by using ICD-9 codes for the 10 most common chronic conditions in the study population. Any child or adult with a claim for a chronic condition during the period of the study was identified as having a chronic condition.
We tested both employee income and job category (physician, nurse, service/other, and clerical) in our models of switching to the gatekeeping plan and found that job category performed better as a predictor of switching as judged by the amount of variance explained. We include here only the data for job category. For purposes of comparison, mean physician income was $102 786, mean nurse income was $57 764, mean clerical income was $32 367, and mean income of other employees was $37 135.
Primary care physician visits were defined as any visit to a physician without subspecialty certification, or to physicians listed by BCBS as a primary care physician. Subspecialist visits were defined as any visit to a physician with subspecialty certification.
In an effort to separate the effects of changing plans from the effects of changing primary care physician, we attempted to identify patients who switched their primary care provider as a result of their change in plan. New primary care physicians were identified by the existence of >1 new claim from a previously unencountered primary care physician after January 1, 1993, and no claims from a primary care physician after January 1, 1993 who had submitted claims during 1991 or 1992.
We initially tested child and household predictors of switching into the gatekeeping plan using the χ2 and Student's t tests. A multivariable logistic regression model was used to identify independent predictors of switching where the dependent variable was household enrollment in gatekeeper or indemnity after January 1, 1993. Statistical significance was set atP < .05 for a 2-tailed test.
The influence of gatekeeping on expenditures and health services use was evaluated by comparing mean expenditures and visits using repeated measures analyses with generalized estimating equations techniques.15 Linear regression models were used for the expenditure variables because the sample size was sufficiently large albeit the data were not strictly normally distributed. Poisson regression models were used for the utilization variables. These models compared the mean change of expenditures and utilization between the 2 groups from 2 time periods: the 2 years before the intervention (1991 and 1992 combined) and each of the 2 years after the intervention (1993 and 1994). These models permitted us to account for time trends in the control group, adjust the means for known differences between the groups, and determine both immediate (1993) and delayed (1994) effects of switching to gatekeeping. We used separate models for the analysis of all children in the study population and the analysis of just children with chronic conditions.
Covariates entered into the models included whether or not the child changed the usual source of primary care in 1993, age and gender of child, marital status, and job category of parent. We included chronic condition status of the child as a covariate in the models of expenditures and utilization for the entire sample (all children). None of the adjustments included in our analysis had a large effect on our results. We present unadjusted means in the tables, whereas theP values are from the regression models. The significance of differences between means for each group in 1991 and 1992 are presented along with the significance of differences between groups from before and after the intervention (difference in differences analysis) derived from the interaction terms of year and group in the regression models.
The handling of outliers in the analysis of claims data are problematic, particularly in pediatrics where a few high cost individuals may account for a substantial fraction of total expenditures. We excluded 8 extreme outliers (6 in indemnity and 2 in gatekeeping), all with annual expenditures >10 standard deviations from the mean in any one year. This boundary was chosen to minimize the influence of trimming on the mean costs and utilization presented in our results and simultaneously minimize the influence associated with extreme outliers.
The children in study households had an average age of 9 years, and over half had at least 1 sibling. Parents were usually married and were on average 40 years old. The parent with the health insurance was as likely to be male as female. Children with a chronic condition were identified in 12% of households. Conditions identified included asthma, affective disorders, seizure disorders, hearing loss, congenital heart disease, mental retardation, cerebral palsy, diabetes, thyroid conditions, hereditary blood dyscrasias, and congenital musculoskeletal anomalies.
The households in gatekeeping and indemnity plans were similar in several respects, including the mean age of the employed parent, the mean age of the children, and employment status as a nurse or service worker (Table 1). Gatekeeping and indemnity households differed in some important respects as well. Indemnity households had a higher proportion of children with a chronic condition (15%) than gatekeeping households (8%, P < .001). Indemnity households were also more likely to have adults with a chronic illness (44%) than gatekeeping households (23%,P < .001). Similarly, indemnity households were more likely than gatekeeping households to include physicians (29% vs 12%), and clerical workers were more prevalent in the gatekeeping plan (41% vs 26%, overall P < .001).
Results from the multivariate analysis confirmed that parents of children with a chronic condition were significantly less likely to enroll in the gatekeeping plan (odds ratio [OR] 0.5; 95% confidence interval [CI]: 0.3–0.7) than other parents (Table 2). The presence of a chronically ill adult in the household also was independently associated with a decreased likelihood of switching to the gatekeeping plan (OR 0.4; 95% CI: 0.3–0.5). Compared with nurses, physicians were less likely to switch to the gatekeeping plan (OR 0.4; 95% CI: 0.2–0.5).
Expenditures: All Children
We determined mean health care expenditures for each year from 1991 through 1994 for all children in the indemnity and gatekeeping groups (Table 3). Mean total expenditures for children in the gatekeeping group decreased 53% from $486 before switching (average of 1991 and 1992) to $180 in 1994. Mean total expenditures in the indemnity group decreased 13% over the same time period, from $623 (average of 1991 and 1992) to $554 in 1994. Mean total expenditures decreased 42% in the gatekeeping group compared with the indemnity group (Fig 2,P < .001 for difference). In the first year following the change in health benefit plan (1993), mean total expenditures increased 20% in the indemnity group, but decreased dramatically (80%) in the gatekeeping group (P < .001 for difference).
Inpatient expenditures accounted for a significant proportion of total expenditures for both groups (68%) and remained consistently higher in the indemnity group than in the gatekeeping group (data not shown). However, changes in outpatient expenditures did differ between the indemnity and gatekeeping groups. Mean annual subspecialty expenditures decreased 6% in the indemnity group and 59% in the gatekeeping group (Fig 2, P < .001 for difference). Outpatient expenditures for primary care physician services declined for both the indemnity (5%) and gatekeeping (53%) groups over the period of the study (P = .004 for difference).
Utilization: All Children
Changes in utilization paralleled changes in expenditures (Table 3). The mean number of outpatient visits to primary care physicians before 1993 was similar in the indemnity (1.9 visits/year) and gatekeeping groups (1.6 visits/year). In 1993, children in indemnity continued to see their pediatrician at the same rate (1.8 visits/year), but primary care visits in the gatekeeping group dropped dramatically in 1993 (0.5 visits/year) and remained significantly lower than indemnity in 1994 (0.9 visits/y). The difference between the indemnity group and the gatekeeper group in the change in visits to primary care physicians was highly significant for both years (P < .001 for difference). Similar changes in the gatekeeping group occurred in visits coded as well-child visits.
Visits to subspecialists also declined dramatically in the gatekeeping group after the switch to gatekeeping (1.6/year in 1991 and 1992 to 0.5/year in 1994), whereas visits to subspecialists in the indemnity group did not change (P < .001 for difference).
Hospital admissions were uncommon in our study population and decreased over time. Nonetheless, admissions were consistently higher in the indemnity group (P = .02 for 1991 and 1992). The difference in the decreased rate of admissions between the gatekeeping (76%) and indemnity plans (49%) was not statistically significant (P = .18 for difference).
Children With Chronic Conditions
Before 1993, children with chronic conditions in the indemnity group had higher mean annual total expenditures and subspecialist expenditures than children with chronic conditions who later enrolled in the gatekeeping group (Table 4). These differences were primarily attributable to a higher rate of hospitalization for patients in the indemnity group. Although mean total expenditures for chronically ill children in the indemnity group remained relatively stable throughout the 4 years, total expenditures dropped dramatically for the gatekeeping group in 1993 (P = .03 for difference) and remained lower than the indemnity group in 1994 (P = .32 for difference). The expenditures for primary care physicians and subspecialists followed similar patterns, although the difference in the changes in expenditures between the 2 groups was not statistically significant.
Mean visits to subspecialists for children with chronic conditions decreased 57% in the gatekeeping group while increasing 31% in the indemnity group (P = .005 for difference). Mean visits to primary care physicians for children with chronic conditions decreased more in the gatekeeper group (23%) than in the indemnity group (13%), although this difference was not statistically significant (P = .71 for difference).
The characteristics of the health plan changes described in this study allowed us to examine the effects of the introduction of gatekeeping on plan choice with few potential confounding elements. This study found that parents of children with chronic conditions were more likely to pay an additional premium to maintain direct access to subspecialists rather than switch to a less expensive plan with an administrative barrier to subspecialist care. In addition, we found a dramatic decrease in subspecialty expenditures and utilization for children with and without a chronic condition associated with switching to a gatekeeping plan. Contrary to our expectations, we found a decrease in the use of primary care physicians among all children who switched to the gatekeeping plan, including children with chronic conditions.
Health Plan Choice
Previous studies have shown a strong preference among adults with chronic conditions to avoid managed care organizations,6but no study has demonstrated this selection effect for families whose children have chronic conditions. In our study, parents of chronically ill children had a strong preference for the indemnity plan. Furthermore, the considerable difference in expenditures for hospitalization before 1993 between the groups of children with chronic conditions suggests that only children with chronic conditions at little risk of hospitalization enrolled in the gatekeeping plan. Although our results indicate the importance of a child's health status in parental health plan decisions, additional investigation is required to determine the relative importance of health status among all members of a household on health plan choice.
The costs of care for children with chronic conditions are low compared with their adult counterparts and have little impact on individual plan costs or local markets. Nonetheless, these data suggest that the bias against gatekeeping among parents of children with chronic conditions can create an unequal distribution of children with chronic conditions among health plans with different features. Plans with lower rates of chronic conditions seem to perform better on health status measures and per member health care expenditures. Managed care organizations use these performance measures to market their plans to employers, consumers, and shareholders.
Perceived access to subspecialty care may be among the most important factors in health plan choice for populations with chronic conditions and their caregivers. In previous studies, the disinclination for adults with chronic conditions to enroll in managed care was similar to our results, suggesting that attitudes toward gatekeeping are important in affecting choices for and against managed care.6 As employers attempt to limit their costs, the differences in employee contribution between gatekeeping and indemnity plans are likely to increase dramatically, forcing those patients with chronic conditions to pay higher premiums for the plans without gatekeeping. The poorest parents are faced with the most difficult choices between financial savings and direct access to subspecialists. The segregation of relatively well and relatively ill segments of the population into different insurance pools also threatens to subvert the notion of spreading risk over a wide population base to make care more affordable for all.16
We found physicians were less likely than other types of hospital employees to switch into the gatekeeping plan. Physicians' negative attitudes toward managed care in general17 and gatekeeping in particular18 may influence their own health benefit plan choices, as may physician's higher incomes.
Expenditures and Utilization
Health care expenditures decreased 42% more for children switching into gatekeeping compared with children in the indemnity plan. This dramatic difference in costs may have several explanations.
First, parents who chose to change plans may have known that they were unlikely to use health care in the near future. We controlled for the presence of chronic conditions and focused our attention on comparisons in the second year after switching (1994) to separate as much as possible the effects of low anticipated utilization from the effects of gatekeeping. The increase in total costs between 1993 and 1994 for the gatekeeping group suggests that low anticipated health care utilization could well be an important determinant of the likelihood of switching plans. The absence of primary care substitution for specialty care and the absolute reduction in primary care in the gatekeeping plan is additional evidence that selection based on propensity to use health care may be an important explanation for the decreased costs and utilization in those children who switched to gatekeeping.
Second, the act of changing plans may itself create barriers to care.19,,20 Burstin and colleagues19 found delays in patients seeking help when ill were associated with changing health plans and attributed this finding to confusion about where to seek care and loss of a relationship with a primary care physician. Consistently lower expenditures in a newly introduced plan may create an incentive for insurers to introduce new plans continuously, with potentially disruptive effects on patient-physician relationships.21,,22 A recent study has found that patients were nearly twice as likely to change their usual source of care as a result of changing insurance plans if their new insurer was a managed care organization.23 This difference may be attributable, in part, to the requirement for a gatekeeper in managed care.
Third, gatekeeping may have accounted for the persistent decrease in subspecialist visits. Unlike total costs, which increased dramatically between 1993 and 1994 for the gatekeeping group, subspecialist visits in the gatekeeping group remained significantly lower in 1994. The 67% decrease in subspecialty visits in our study is nearly twice the reduction in subspecialty use found by Martin and colleagues11 in their randomized trial of gatekeeping. Their study did not isolate children and had a low rate of subspecialist visits (mean = 0.6 visits). The large magnitude of the decrease in subspecialist utilization in our study may relate to the issues discussed above, as well as local market factors and the population we studied (children of hospital employees). Specifically, children in the indemnity plan averaged nearly 2 subspecialist visits per year, significantly more than published reports of other child populations.7
To the extent that the decrease in subspecialist use represents more appropriate health care use, the gatekeeping plan may have been effective in reducing inappropriate care.24 Gatekeeping may be more effective in decreasing costs in populations with relatively high subspecialty use. Nonetheless, the large majority (60%) of costs for children are from hospitalizations,25and these costs are unlikely to be affected by having an assigned primary care physician preapprove subspecialty referrals.
Although we were not able to assess quality of care in this study, the results of the utilization analysis indicate some potential problems. Visits to primary care physicians decreased for both the gatekeeping group and the indemnity group over the course of the study. This is to be expected from the aging of the cohort; older children are hospitalized and visit a physician less often than younger children. Nonetheless, the decrease in overall primary care physician visits and well-child visits in the gatekeeping group compared with the indemnity group is concerning insofar as one of the goals of having an assigned primary care physician is to increase the use of preventive health services. The introduction of the gatekeeping plan likely disrupted ongoing patient-physician relationships with primary care providers as well as subspecialists.
Children With Chronic Conditions
Among the most striking findings of this study is the decrease in expenditures and physician visits among children with chronic conditions in the gatekeeping group. The children with chronic conditions in our study were relatively high consumers of subspecialty care (mean subspecialist visits = 3.3/year) before January 1993, and are therefore likely to have had an established relationship with a subspecialty provider. Changing to a gatekeeping plan clearly modified this relationship with a principal provider of care. It is worth noting that mean subspecialist visits for children with chronic conditions increased in the indemnity group after the introduction of gatekeeping. We can speculate that that this increase in subspecialist visits may have been because of a decline in health status for this population and/or a compensatory increase in visits used by subspecialists to fill empty appointments slots formerly occupied by children who switched to gatekeeping.
The right mix of subspecialist and primary care physician involvement in the care of children with chronic conditions is unknown and will vary by condition.26 Nonetheless, we expected that the primary care physician would provide a certain proportion of the care previously provided by subspecialists for children with chronic conditions who switched to the gatekeeping plan. However, we did not observe a significant increase in visits to primary care physicians among children with chronic conditions in the gatekeeping group, indicating that gatekeeping is unlikely to have improved the delivery of primary care services to this population.
This study has several limitations. The employed population we studied may be atypical of the employed population generally, and our results are set in the context of a particular health care market. These factors limit the generalizability of our results. We did not adjust for differences in case mix between the indemnity and gatekeeping groups because we were more interested in the actual differences between the groups. Case-mix adjustment would no doubt decrease some of the differences observed, but also distort real costs within each system. Our expenditures data suggest that parents of sicker chronically ill children demonstrated an even stronger aversion to gatekeeping than parents of children with less costly chronic conditions. The small numbers of chronically ill children in our cohort did not permit us to identify sub-groups among the chronic conditions, such as adolescents with mental illness, who may have been affected differently by the change in health plans. We followed children for only two years after the switch in health plan, and therefore could not assess differences between gatekeeping and indemnity arrangements beyond that point.
Parents of children with a chronic condition chose to pay higher health plan contribution rather than switch to a health plan that required a primary care physician approve all subspecialty referrals. Children's health status appears to be an important factor in parental health plan choice. Switching plans resulted in decreased costs, but we are unable to determine to what extent the lower costs stem from the disruption of care that follows a change in health plans or gatekeeping. Although it remains unclear if patients or physicians benefited from the introduction of gatekeeping, our study found switching to gatekeeping was associated with potentially significant disruption in care for children. The full effects of switching to a gatekeeping plan on the health of children remain unknown.
This study was provided by the Robert Wood Johnson Foundation. Dr Ferris received support from the Pediatric Scientist Development Program (AAP/NICHD: K12-HD00850).
- MGH =
- Massachusetts General Hospital •
- MBCBS =
- Massachusetts Blue Cross Blue Shield •
- ICD-9 =
- International Classification of Diseases-Ninth Revision •
- OR =
- odds ratio •
- CI =
- confidence interval
- ↵Starfield B. Primary Care: Concept, Evaluation and Policy. New York, NY: Oxford University Press; 1992
- Forrest CB,
- Glade GB,
- Starfield B,
- Baker AE,
- Kang M,
- Reid RJ
- Kuhlthau K,
- Perrin JM,
- Ettner SL,
- McLaughlin TJ,
- Gortmaker SL
- Liang KY,
- Zeger SL
- Newhouse JP,
- Sloss EM,
- Manning WG,
- Keeler EB
- ↵Reed C. Why People Change Their Health Care Providers. Data Bulletin, Results from HSC Research. 2000;1–2
- ↵Kuhlthau KA, Ferris TG, Beal AC, Gortmaker SL, Perrin JM. Who cares for Medicaid-enrolled children with chronic conditions?Pediatrics. 2001. In press
- Copyright © 2001 American Academy of Pediatrics