PEDIATRICS Vol. 118 No. 6 December 2006, pp. e1766-e1778 (doi:10.1542/peds.2006-0286)
ARTICLE |
Compliance With Well-Child Visit Recommendations: Evidence From the Medical Expenditure Panel Survey, 2000–2002
Division of Modeling and Simulation, Center for Financing, Access, and Cost Trends, Agency for Healthcare Research and Quality, Rockville, Maryland
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
OBJECTIVES. This study examines national compliance rates with well-child visit recommendations using the Medical Expenditure Panel Survey. The Medical Expenditure Panel Survey provides nationally representative information on preventative care for children, combining visit-level data over a 2-year period with a rich array of socioeconomic and health status measures.
METHODS. Visit-level data from 2000 to 2002 were used to construct a well-child visit "compliance" measure equal to well-child visits as a percentage of age-specific recommendations from the American Academy of Pediatrics. Compliance was examined across age, gender, race/ethnicity, health status, poverty, insurance coverage, eligibility for public coverage, family structure, parent education, insurance, citizenship and country of origin, language, urbanicity, and census division.
RESULTS. On average, 56.3% of all children aged 0 to 18 years had no well-child visits during a 12-month period, and 39.4% had no well-child visits over a 2-year period. The average compliance ratio was 61.4%. Large differences in compliance exist among children. High compliance rates were observed among infants (83.2%), children with special health care needs (86.6%), children with college-educated parents (74.3%), children with family incomes >4 times the poverty level (71.6%), and children in the New England (94.6%) and Middle Atlantic (83.2%) census divisions. Low levels of compliance were observed among uninsured children (35.3%) and especially uninsured children simulated to be eligible for public coverage (28.4%). Other groups with low compliance rates include teenagers (49.2%), noncitizen children (43.9%), and children in the West South Central (44.9%), East South Central (48.8%), and Mountain (49.7%) census divisions.
CONCLUSIONS. Well-child visit compliance in the Medical Expenditure Panel Survey is less than found in other househould surveys, yet consistent with or above results based on data from provider and claims data. Although experts dispute the optimal frequency of well-child visits, the disparities observed in compliance rates among population subgroups raise important public health concerns.
Key Words: well-child care • MEPS
Abbreviations: AAP—American Academy of Pediatrics • MEPS—Medical Expenditure Panel Survey • NHIS—National Health Interview Survey • CTS—Community Tracking Survey • HEDIS—Health Plan Employer Data and Information Set
Well-child care plays an important role in the provision of quality health care for children. Well-child visits help promote timely immunizations and screening for health conditions and normal development. They also offer an opportunity for providers to answer parents’ health-related questions and provide anticipatory guidance. The quality of well-child visits can vary greatly,1 and there is a paucity of evidence on the efficacy of recommendations regarding the number, timing, and content of visits.2 Nevertheless, researchers have found associations between increased well-child visits and reductions in avoidable hospitalizations, reductions in emergency department use, and improved child health.3–7 Moreover, differences in well-child visit compliance across subgroups of the population may serve as a marker for quality differences in other dimensions of health care.
Although there remain unanswered questions regarding the exact value of well-child visits, it is a fact that many children have far fewer well-child visits than are recommended by the American Academy of Pediatrics (AAP).8 Moreover, there are large differences in compliance across subgroups of the population. Previous research has identified a wide range of factors influencing compliance with AAP guidelines, including age, race/ethnicity, insurance coverage, family income, parent education, parent insurance, family structure, and location by region and urbanicity.9–17 This article seeks to contribute to the literature by examining well-child visit compliance using nationally representative data from the 2000–2002 Medical Expenditure Panel Survey (MEPS). The MEPS data are a valuable but relatively untapped resource for examining well-child care, combining visit-level data over a 2-year period with a rich array of socioeconomic and health status measures.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The MEPS is a household survey of the noninstitutionalized civilian population sponsored by the Agency for Healthcare Research and Quality and the National Center for Health Statistics. It is a stratified and clustered random sample, oversampling minorities by factors of 2 to 1 for Hispanics and 1.5 to 1 for blacks. Survey weights can be used to produce nationally representative estimates for insurance coverage, medical expenditures, and a wide range of other health-related and socioeconomic characteristics.18–20
The MEPS has an overlapping panel design, collecting visit-level data over a 2-year window. Data for all family members are provided by a single respondent, typically the person deemed to have the most knowledge about the family. To reduce problems associated with long recall periods, data are collected in 5 rounds of interviews, and respondents are asked to keep a calendar of medical events and to supply supporting paperwork regarding those events. Data were pooled from the 2000, 2001, and 2002 full-year MEPS files, yielding a sample of 20423 children age 
18 years who supplied 
1 complete year of data. The total number of annual observations is 29460, because many children provide 2 years of data. These sample sizes exclude children who died or otherwise drop from the survey during the year (lest they be counted as undercomplying with visit recommendations). Analyses of visits over a 2-year window entailed subsetting to children entering the MEPS in 2000 and 2001 and removing cases that drop from the MEPS during the second year, yielding a sample size of 8894. All of the SEs and statistical tests have been adjusted to account for the complex design of the MEPS, the possibility of intrafamily correlation across siblings, and the fact that many children supply >1 observation for the 1-year analyses.
Measures of Well-Child Visits
Well-child care was identified using visit-level information on office and hospital outpatient visits. The MEPS asks respondents the primary reason for the visit and prompts respondents with a list of possible reasons. For this analysis, visits were coded as well-child care if the primary reason given was "well-child examination," "general checkup," or "immunization or shots." Because the MEPS also provides detailed condition codes at the visit level, one can expand this definition to include all other visits for which neither a specific condition nor a specific primary reason was provided. Although this broader definition seems overly inclusive, it only raises the weighted average frequency of having 1 well-child visit during the year from 43.7% to 49.1%. Moreover, all of the patterns discussed in this article hold true with either definition. Thus, all of the results presented in this article use the narrower definition.
Three well-child visit measures are presented: (1) an indicator for whether the child had 1 visit during the past 12 months (or 24 months in the 2-year analysis); (2) a count of the number of visits; and (3) a count of visits expressed as a ratio of the age-specific number of visits recommended by the AAP.8 To construct the compliance ratio, the AAP guidelines were merged into the MEPS using the child’s age in months at the end of the year.
Well-child visits can (and often do) occur slightly before or after the recommended date with little or no reduction in quality of care. Consider an example of two 4-year-olds who are approximately in compliance with the AAP recommendation of 1 visit per year. The first child happened to have visits just before and just after the survey year, whereas the second child had 2 recorded visits, 1 at the beginning and 1 at the end of the year. The compliance ratio of the first child is 0%, and, because compliance ratios are allowed to exceed 100%, the value for the second child is 200%. Thus, the average across these 2 children is 100%, as seems appropriate. By allowing for a degree of overcompliance, the ratio used in this article minimizes the impact of small variations in the timing of visits. Because a handful of children report an implausibly large number of well-child visits, the number of visits was limited to the recommended number plus 2. This truncation affected 2.3% of the sample.
One important technicality involves the coding of compliance among children aged 7 to 10 years. Because the AAP does not recommend any well-child visits for 7- and 9-year-olds, some studies give 7- and 9-year-olds compliance scores of 100%, regardless of whether they had any visits or not. In an analysis of children aged 6 to 12 years, this practice would ensure an average 28.6% compliance rate even if none of the children ever had visits. Note also that some children who are aged 7 as of the end of the survey year may have had their 6-year checkup during the survey year. Arbitrarily coding all of the 7-year-olds as compliers thereby obscures valuable information about compliance. To avoid these problems, the AAP guidelines are coded for this analysis as 0.5 visits per year for ages 7 to 10, and the average number of visits for this age group is then compared with this norm.
Child and Family Characteristics
Well-child visits are examined across a wide range of child and family characteristics. Age is measured as of the end of the calendar year (or the end of the second calendar year in the 2-year analysis). Three health measures are indicators for whether (1) health status is judged "fair" or "poor" by the (adult) respondent versus "excellent," "very good," or "good" at the first interview during the calendar year, (2) the child is deemed to have a special health care need on the basis of responses to the MEPS Special Health Care Needs Screener,20,21 and (3) the child is "overweight" or "at risk of being overweight" versus being normal weight on the basis of the child’s BMI relative to the 85th and 95th percentiles of the Center for Disease Control and Prevention 2000 growth charts.22 The MEPS supplies BMI for children aged 3 to 17 years for 2001 and 2002 but not 2000. Also, children aged 3 to 5 years were dropped from the BMI analysis because of high rates of missing values for height and/or weight.
Poverty is constructed by comparing family income to federal poverty lines by year and family composition. The poverty measure relies on a family definition that is consistent with federal poverty calculations. All of the other family measures use a narrower definition of family, typically consisting of children and their parents (natural, step, or adoptive). Children are coded as having a parent in the household if they ever lived with that parent during the year. In cases where no parent can be identified within the household, "parent" characteristics pertain to the adult(s) deemed responsible (typically a grandparent or other adult relative). For children with 2 parents, education is coded as the highest level achieved by either parent/caregiver.
Health insurance status reflects coverage held at any point during the year. Thus, children classified as uninsured lacked coverage for the entire year. Children with both public and private coverage were counted as having public coverage (results are not sensitive to this assumption). In addition to actual insurance coverage, the analysis uses a simulated measure of eligibility for public coverage through Medicaid or the State Children’s Health Insurance Program.23 This simulated measure was constructed by using detailed program rules by state and year, together with information from the MEPS on child age, family composition, family income by type, state of residence, and citizenship. The simulation reflects eligibility during the first portion of the calendar year.
Children were coded as having a usual source of care if the identified provider was either office based or at a hospital outpatient department. Most children who report having a hospital outpatient department as a usual source of care also report this to be the location of their doctor’s office or the location where they obtain the highest quality of care.
Information on citizenship and country of origin were obtained by linking the MEPS to the National Health Interview Survey (NHIS). Because NHIS was administered 1 to 2 years before the MEPS, some persons identified as noncitizens may have become citizens by the end of the MEPS. The citizenship and country of origin of parents is hierarchically ranked as: 1 is US born, 1 is a naturalized immigrant, or both are noncitizens.
Public-release versions of variables were used whenever available, and in many cases these variables were thoroughly edited and imputed on the public-use files. For variables containing missing values, observations were dropped from the descriptive analyses in Tables 1 through 3. Missing data are primarily an issue for special health care needs, being overweight, and parent smoking. Published MEPS sampling weights that adjust for missing values were used in these cases as appropriate.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Table 1 presents weighted means for the 3 well-child visit measures. SEs are in parentheses. Significance levels pertain to comparisons with the mean of the reference category for each variable, using survey-adjusted Wald tests. All of the differences discussed in the test are significant at the 5% level unless otherwise noted.
Nationally, 43.7% of all children aged 18 years had 1 well-child visit during the sample year. The mean number of visits was 0.75, and the mean compliance ratio was 61.4%. As one would expect, there are large differences in visit prevalence by age group, with age being measured as of the end of the survey year (so that 1-year-olds, for instance, include children aged 12–23 months on December 31). Not surprisingly, the youngest children have the most visits, and children aged 6 to 12 years have the fewest. Dividing by the number of recommended visits to obtain compliance ratios, however, yields more similar results across age groups. The only 2 age groups with markedly different compliance ratios are infants (83.2%) and teenagers (49.2%). These results by age group highlight the value of using compliance ratios to make comparisons across subpopulations that may have different age distributions.
All 3 of the visit measures are very similar for boys versus girls. Differences by race/ethnicity are somewhat larger, with white compliance ratios 
10% higher than minority compliance ratios. Although Asian and Pacific Islander children compare favorably with white children on insurance coverage,24 their estimated compliance ratio is the lowest of the 4 groups (although not significantly lower than that of the other 2 minority groups).
Not surprisingly, children with health concerns are more likely to have 1 visit, have more visits on average, and have higher compliance ratios than healthier children. For instance, the compliance ratio for children with fair/poor health is 82.0% vs 60.8% for children with excellent/very good/good health. Similarly, children with special health care needs have 86.6% compliance vs 56.4% for other children. Of course, it should be noted that children who undercomply with well-child visit recommendations may be unaware of health problems, so that reported health status might well be a consequence of well-child care, rather than the reverse. This may be particularly troublesome with respect to the special health care needs indicator, insofar as this measure is constructed from questions about whether children need or use more medical care than average.
Being overweight is not strongly correlated with compliance. Overweight children aged 6 to 17 years are somewhat less likely than normal weight children in that same age group to have had a well-child visit during the past year (33.8% vs 37.7%), and the average number of visits is lower (0.45 vs 0.50). However, the compliance ratios are not appreciably different.
Not surprisingly, having health insurance is strongly correlated with well-child care. Children lacking coverage for the entire year have compliance ratios of only 35.3% vs 64.1% and 63.1% for children with private and public coverage, respectively. Of course, having health insurance and having adequate well-child care may both be driven by other child or family characteristics. This relationship should not, therefore, be interpreted as causal. Nevertheless, a more detailed analysis of coverage and access to care using these same data found strong insurance effects even using multivariate techniques that control for the potential that coverage and visits are spuriously correlated.25
In addition to asking about the source of insurance, the MEPS collects evidence about managed care. Among children with public coverage, those covered through a health maintenance organization have 7.4% greater compliance than those in fee for service. In contrast, there are no corresponding differences across types of private coverage. These results are intriguing but should be interpreted with caution because of the potential for systematic differences among children by plan type.4,26–33
One of the lowest compliance rates in Table 1 is among uninsured children who are simulated to be eligible for public coverage (28.4%). One straightforward explanation is that eligible but uninsured children are poorer than uninsured children who are deemed to be ineligible. Lacking coverage and low visit compliance may also reflect differences in child health and/or parent attitudes and abilities. Certainly the low compliance rates observed among eligible but uninsured children highlight the importance of outreach efforts seeking to increase enrollment in public coverage.
The next row of Table 1 highlights the close connection between having a usual source of care and receiving adequate preventive care. Among children lacking a usual source of care, compliance with AAP visit guidelines is only 26.9%. As with insurance coverage, it is likely that having a usual source of care and complying with AAP visit guidelines are both driven by common child and family characteristics. Nevertheless, it seems clear that having a usual source of care is an important marker for adequate well-child care.
Visit compliance did not vary much with income among children in families <400% of the federal poverty line. This is likely attributable at least in part to the widespread availability of public coverage among low-income children, especially given that poverty level is associated with compliance among uninsured children (see above). Children in the highest income group have compliance ratios nearly 15% higher than children in the 3 lower income groups. Family composition also matters, with children in large families having lower compliance ratios (50.7%) than children in smaller families (62.8%). Similarly, children living with 2 parents have 8% higher compliance ratios than children living with 1 or no parents.
Having a parent/caregiver who obtained a high school degree or General Educational Development degree is associated with an increase in well-child compliance from 50.3% to 56.1% compared with children whose parents/caregivers had no such a degree. There is a much larger education-related effect if the child has a parent/caregiver who attended college (74.3%). Parent/caregiver smoking is also a strong predictor of preventive care, with compliance being nearly 13% lower when a parent/caregiver smokes.
Citizenship and country of origin seem to affect well-child care primarily through the child’s, not the parents’, status. Among native-born (citizen) children, it matters little for compliance whether parents/caregivers are native-born citizens, naturalized immigrants, or noncitizens. Instead, the main difference is between native-born citizen children (with 60% compliance) and noncitizen children (43.9% compliance). Language also seems to play a role, with children having 10% higher compliance ratios if the MEPS was fielded in English than if it was fielded in a foreign language (typically Spanish).
One of the most striking findings in this article concerns differences by census region and division. Children living in the New England division (Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and Connecticut) had a remarkable 94.6% compliance ratio. Children in the Middle Atlantic division (New York, New Jersey, and Pennsylvania) placed second at 83.2%. In contrast, compliance was at or below 50% in the East South Central division (Kentucky, Tennessee, Alabama, and Mississippi), the West South Central division (Louisiana, Arkansas, Texas, and Oklahoma), and the Mountain division (Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, and New Mexico). In contrast to these geographic differences, the MSA/non-MSA difference in compliance, whereas statistically significant, is much smaller in magnitude.
Compliance Over 2 Years
Table 2 presents results that exploit the fact that the MEPS follows each child over 2 years. Using a 2-year window reduces the potential impact of minor variations in the timing of visits and of decisions regarding the coding of visit recommendations, especially for children aged 7 to 10 years. However, using a 2-year window also greatly reduces the sample size, making it difficult to form estimates for certain policy-relevant subgroups. Thus, the primary motivation for presenting the 2-year results is to provide a consistency check for the 1-year findings. The average number of visits nearly doubles, as one would expect. More importantly, the mean compliance ratio (in the last column) is 58.6%, which is similar to the average compliance ratio of 61.4% in the 1-year analysis. The most important new result in Table 2 is that 39.4% of all children went 2 years without any well-child visits (ie, 100% – 60.6%). Among children who were continuously uninsured for both years, the percentage of children going without any well-child visits was 61.5%, and the compliance rate was only 31.7%.
|
Results According to Age Group
Table 3 presents compliance ratios according to child and family characteristics separately for children aged 0 to 5, 6 to 12, and 13 to 18 years. Despite using the larger 1-year sample, some categories from Table 1 are not shown because of small numbers of observations, and some estimates have large confidence intervals (such as young Asian and Pacific Islander children and children with fair or poor health). In general, the same patterns shown in Table 1 hold within each of these age groups. One implication is that whereas teenagers as a group have low compliance, some subsets of teenagers have very low compliance ratios indeed. For instance, teenager compliance is only 24.1% if uninsured, 32.1% if parents did not graduate from high school, 18.9% if lacking a usual source of care, 26.3% if a noncitizen, 26.5% if native born with noncitizen parents, and 32.7% if living in the West South Central census division.
Multivariate Analysis
Table 4 presents multivariate results from the weighted regression of the compliance ratio on a range of child and family characteristics. Coefficient estimates in the table measure the percentage change in the compliance ratio associated with each explanatory variable (relative to the omitted category). The first set of coefficients is from a model that excludes health insurance, whereas the second model controls for coverage. The first approach is appropriate if one seeks to measure, say, compliance differences by race/ethnicity that are inclusive of the large coverage differences across groups. The first model also avoids concerns that coverage and compliance are both being influenced by factors that are omitted from the model (such as parent attitudes and abilities). The second approach is appropriate if one seeks to understand, say, racial and ethnic differences in compliance beyond what one might expect on the basis of disparities in coverage. Indicators for having a usual source of care and for having special health care needs were excluded because of concerns that they might be spuriously correlated with well-child visits. The BMI measures were also excluded, because they were not available or were unreliable for children outside the 6- to 17-year age group.
|
Many of the same patterns found in the bivariate analyses in Tables 1 through 3 are also seen in the multivariate analysis. Infants and teenagers continue to be at opposite ends of the compliance spectrum. There continues to be little difference between girls and boys. Children in fair or poor health continue to have much higher compliance rates than other children. Having parents who are college educated or nonsmokers continues to be associated with substantially higher compliance, and coverage (in the second model) continues to have a large impact on compliance, especially among children with public health maintenance organization coverage. In some cases, differences seen in the bivariate analyses persist in Table 4 but are smaller in magnitude. Examples include differences by race/ethnicity, number of parents, poverty, citizenship, and survey language.
One of the most striking results from Table 4 is that differences across census divisions remain large even when one controls for a long list of other child and family characteristics that might plausibly have explained the geographic differences in Table 1. Children in the East South Central, West South Central, Mountain, and Pacific divisions continue to have compliance ratios that are 40% lower than children in the New England division.
Table 4 also presents regression results estimated separately for each age group. Many of the age-specific results are broadly consistent to those obtained by pooling across age groups, although estimate precision, not surprisingly, is less. There are, however, several noteworthy differences. First, gender differences appear among older children, with boys aged 6 to 12 years having slightly higher compliance than girls, with the reverse being true among teenagers. Second, although young children living apart from their parents seem to have 0.230 higher compliance rates than their peers living with 2 parents, this may be the result of a relatively small sample size (there are only 135 such cases in the data). Third, and perhaps most significantly, geographic differences in compliance by census region are more pronounced among older children than among those 0 to 5 years old.
Other Visits for Undercomplying Children
The analysis above highlights large gaps in well-child visit compliance among subpopulations of children in the United States. Of course, many undercomplying children receive other health care services throughout the course of the year. In some cases, children with other visits may be receiving elements of well-child care despite lacking formal well-child visits. Among children aged 0 to 18 years with no well-child visits during the year, the percentage with 1 office or hospital outpatient visit is 48.8% (SE: 0.8%; results not shown in Table 4). Including emergency department visits raises this percentage to 52.6% (SE: 0.8%). Among children aged 0 to 5 years with no well-child visits, the percentage with 1 office or hospital outpatient visit is 51.8% (SE: 1.3%) or 56.7% (SE: 1.4%) if emergency department visits are included. Thus, approximately half of all children lacking well-child visits also lack any health care visits whatsoever.
Expanding the time frame to 2 years considerably raises the percentage of children with other visits. Among children aged 0 to 18 years with no well-child visits during a 2-year window, the percentage with 1 office or hospital outpatient visit is 63.1% (SE: 1.2%). Among children aged 0 to 18 years with some well-child visits but fewer than half the recommended number, the percentage with 1 other office or hospital outpatient visit is 75.0% (SE: 3.1%). Thus, it seems clear that many chronically undercomplying children do have other provider contacts if one uses a time frame that is sufficiently wide. The extent to which these visits offset the lack of well-child visits is an important question. Another important question is whether these other provider contacts offer a potential pathway for screening and outreach to improve well-child visit compliance.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
MEPS estimates of well-child care in 2000–2002 are somewhat greater than comparable estimates from the 1987 National Medical Expenditure Survey, the precursor to the MEPS, which used very similar methodology.9 In 1987, the percentage of children aged 0 to 5 years with
1 well-child visit during the year was 60.0% vs 66.7% (SE: 1.0%) for the same age group in the MEPS. The percentage in full compliance with AAP guidelines (as distinct from the compliance ratio used in this study) was 41.8% in 1987 vs 45.8% (SE: 1.0%) in the MEPS.
Although well-child visit compliance seems to have risen slightly over time, the overall prevalence of well-child visits in the MEPS falls far short of the AAP norms. It is important to note in this regard, however, that MEPS estimates are substantially below results from 3 contemporaneous household surveys. In the 2000–2001 NHIS and the 2000–2001 Community Tracking Survey (CTS), the percentages of children aged 0 to 17 years who had 1 well-child visit over a 12-month period are 71% and 65.8%, respectively (calculation using tabulated numbers in ref 15, and calculations using the Center for Studying Health Systems Change online data analysis facility for the CTS [excluding missing values], available online at: http://hschange.com). The 2000–2001 MEPS estimate for this age group is 43.2%. In the 2002 National Survey of American Families, 68.6% of children aged 0 to 18 years had 1 well-child visit versus 45.9% for the same age group and year in the MEPS.34 Interestingly, the MEPS and NHIS align quite closely with respect to the fraction of children under age 2 with 1 well-child visit: 85.5% in the 2002 MEPS (SE: 1.3%) vs 87.8% in 2002 NHIS.35 However, the gap grows rapidly at higher age groups: 62.9% (SE: 1.9%) vs 84.3% for ages 2 to 3 years and then 52.6% (SE: 2.0) vs 81.5% for ages 4 to 5 years in the 2002 MEPS and 2002 NHIS, respectively.35
One potential explanation is that NHIS measures well-child visits using a single question with a 12-month recall period, so that some respondents may include well-child visits outside of this window, a phenomenon known as "telescoping." Similarly, CTS and National Survey of American Families ask about the total number of visits over the past 12 months, and then ask respondents to identify how many were for well-child visits or checkups. In contrast, the MEPS provides respondents with calendars to help reduce recall error, and the recall period is shorter (typically 4–5 months).
Although the MEPS finds fewer children receiving well-child visits than other household surveys, there is a broader question of whether households can accurately report well-child visits. One nationally representative study of matched provider and household data found that mothers systematically report far more well-child visits than are recorded in provider records, with the bias in the reporting of sick-child visits working in the other direction.36 Whether this discrepancy arises from provider coding decisions or household misreporting is an open question. In contrast to the wide disparities found in that study, well-child estimates in the MEPS are in line with or only slightly above national visit totals from provider surveys. The total number of office-based and hospital outpatient "preventive care" visits from the 2002 National Ambulatory Medical Care Survey and the 2002 National Hospital Ambulatory Medical Care Survey is 48.0 million (SE: 3.3 million) for children under age 15 years in 2002 versus an MEPS national total of 53.7 million (SE: 1.8 million) for the same age group and year.37,38 The difference is small and not statistically significant. Moreover, even if one includes National Ambulatory Medical Care Survey office and National Hospital Ambulatory Medical Care Survey outpatient visits of "unknown/blank" type, the 2002 total only rises to 54.2 million (SE: 3.5 million). This is only slightly above the MEPS (again, the difference is not statistically significant).
Yet another data source for analyzing well-child visits is the 2002 Health Plan Employer Data and Information Set (HEDIS), which contains reports from public and private managed care plans for 2001. In this database, the percentages of children aged 3 to 6 years with 1 well-child visit are 58% (public) and 57% (private).39 The corresponding frequencies in the 2001 MEPS are 54.5% (SE: 2.9%) and 60.1% (SE: 2.3%). Similarly, the HEDIS percentages of adolescents with well-child visits are 37% (public) and 33% (private), whereas the corresponding MEPS estimates are 37.9% (SE: 2.8) and 39.1% (SE: 2.0).39 These comparisons should be interpreted with caution. The administrative data on which HEDIS scores are often based can undercount use, although the increasing use of HEDIS measures to evaluate performance may be helping to improve data accuracy.40
A final benchmark is provided by studies of administrative data for children with public coverage. One recent study of Medicaid data in 3 states found far lower well-child visit compliance than in the MEPS, with the percentage of children having 5 of the recommended 9 visits by the age of 24 months being 31% in California, 27% in Michigan, and only 15% in Georgia.3 Studies that match administrative claims to provider records, however, indicate that claims data can greatly underestimate the prevalence of well-child visits.17,41 Here again, questions arise regarding the accuracy of provider coding.
Gathering results, MEPS estimates are (1) slightly above estimates from the 1987 National Medical Expenditure Survey, (2) substantially below household estimates based on 12-month recall, (3) approximately in line with estimates from provider surveys and HEDIS, and (4) above estimates based on public claims data. One other noteworthy piece of evidence is that compliance in the MEPS within some subgroups of the population is close to 100% (eg, children living in the New England census division), suggesting that if the MEPS is undercounting well-child visits, then either the bias is uniform and groups apparently complying in the MEPS are in fact overcomplying or any underreporting of well-child visits in the MEPS is concentrated among other subgroups of the population. Clearly, one salient implication of this discussion is that there is a need for additional methodologic research into the true prevalence of well-child visits, with different sources and methods yielding a wide range of estimates for this key public health marker. Given the care with which the MEPS collects household-provided data on this subject, the estimates in this article should provide a valuable contribution to the ongoing debate.
MEPS results regarding compliance differences amongpopulation subgroups largely mirror previous findings based on other data sets. As in previous studies, high compliance rates were observed among infants, children with special health care needs, children with college-educated parents, and children with family incomes >4 times the poverty level. Low levels of compliance were observed among uninsured children, teenagers, and noncitizen children. Compliance rates were particularly low (28.4%) among uninsured children who were simulated to be eligible for public coverage through Medicaid or State Children’s Health Insurance Program, highlighting the importance of outreach efforts to enroll eligible children. Children with fair/poor health and children with special health care needs had greater compliance rates than other children. This makes intuitive sense, insofar as children with health concerns are more likely to have provider contacts of all types. The same is not true for overweight children or those at risk of being overweight, who had compliance ratios that were the same or lower than those of normal weight children. Because visit compliance is not linked to need along this key dimension of child health, the results in this article highlight the importance of other outreach efforts to counsel children and their parents about eating and exercise.42
Another striking set of findings concerns geographic differences in well-child visit compliance. Previous research has shown marked differences among census regions.15 Results presented here indicate that differences across census division are even more pronounced than regional differences. Not surprisingly, the pattern of well-child visits found in this article mirrors geographic differences in the supply of pediatricians.43 This is not surprising, because one would expect a rough equivalence between the supply of physicians and the use of (demand for) their services. The more intriguing result is that geographic differences remain large even after controlling for a wide array of child and family characteristics. In other words, geographic differences in well-child care are not strongly related to geographic differences in key demand-side variables, such as income, insurance coverage, prevalence of managed care, race/ethnicity, citizenship and country of origin, family structure, and education. There is some evidence that compliance may be related to the prices paid to providers for well-child care.44 However, more research is clearly needed into the factors driving geographic variation.
The primary finding of this article is that well-child visit compliance is low on average among US children, with very low rates observed within important subpopulations. On a somewhat more optimistic note, the study also finds that a substantial percentage of undercomplying children have other contact with health care providers over a 2 year time frame. For instance, among children going without any well-child visits over a 2-year period, 63.1% had 1 office or hospital outpatient visit. An important area for further research is the extent to which these other provider contacts offer an avenue for identifying undercomplying children and helping to increase their preventive care.
|
ACKNOWLEDGMENTS |
|---|
I appreciate the helpful comments of Jessica Banthin, Cindy Brach, Fran Chevarley, Joel Cohen, Denise Dougherty, Julie Hudson, and Merrile Sing.
|
FOOTNOTES |
|---|
Accepted Jun 26, 2006.
Address correspondence to Thomas M. Selden, PhD, Division of Modeling and Simulation, Center for Financing, Access, and Cost Trends, Agency for Healthcare Research and Quality, 540 Gaither Rd, Rockville, MD 20850. E-mail: tselden{at}ahrq.gov
The author has indicated he has no financial relationships relevant to this article to disclose.
The views expressed in this article are those of the author, 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.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Zuckerman B, Stevens GD, Inkelas M, Halfon N. Prevalence and correlates of high-quality basic pediatric preventive care.
Pediatrics. 2004;114
:1522
–1529
[Abstract/Free Full Text] - Moyer VA, Butler M. Gaps in the evidence for well-child care: a challenge to our profession.
Pediatrics. 2004;114
:1511
–1522
[Abstract/Free Full Text] - Hakim RB, Bye BV. Effectiveness of compliance with pediatric preventive care guidelines among Medicaid beneficiaries.
Pediatrics. 2001;108
:90
–97
[Abstract/Free Full Text] - Piehl MD, Clemens CJ, Joines JD. Narrowing the gap: decreasing emergency department use by children enrolled in the Medicaid program by improving access to primary care.
Arch Pediatr Adolesc Med. 2000;154
:791
–795
[Abstract/Free Full Text] - Hakim RB, Ronsaville DS. Effect of compliance with health supervision guidelines among US infants on emergency department visits.
Arch Pediatr Adolesc Med. 2002;156
:1015
–1020
[Abstract/Free Full Text] - Hakim RB, Boben PJ, Bonney JB. Medicaid and the health of children. Health Care Financ Rev. 2000;22 :133 –140[Web of Science]
- Vogel RL, Ackermann RJ. Is primary care physician supply correlated with health outcomes? Int J Health Serv. 1998;28 :183 –196[Web of Science][Medline]
- American Academy of Pediatrics, Committee on Practice and Ambulatory Medicine. Recommendations for preventive pediatric health care.
Pediatrics. 2000;105
:645
–646
[Free Full Text] - Short PF, Lefkowitz DC. Encouraging preventive services for low-income children: The effect of expanding Medicaid. Med Care. 1992;30 :766 –780[CrossRef][Web of Science][Medline]
- Freed GL, Clark SJ, Pathman DE, Schectman R. Influences on the receipt of well-child visits in the first two years of life.
Pediatrics. 1999;103
:864
–869
[Abstract/Free Full Text] - Long SK, Coughlin TA. Access and use by children on Medicaid. Inquiry. 2001–2002;38 :409 –422
- Yu SM, Bellamy HA, Kogan MD, Dunbar JL, Schwalberg RH, Schuster MA. Factors that influence receipt of recommended preventive pediatric health and dental care. Pediatrics. 2002;110 (6). Available at: www.pediatrics.org/cgi/content/full/110/6/e73
- Davidoff A, Kenney G, Yemane A. The effect of parents’ insurance coverage on access to care for low-income children. Inquiry. 2003;40 :254 –268[Web of Science][Medline]
- Gifford EJ, Weech-Maldonado R, Short PF. Low-income children’s preventive services use: implications of parents’ Medicaid status. Health Care Financ Rev. 2005;26 :81 –94[Web of Science][Medline]
- Olson LM, Tang SS, Newacheck PW. Children in the United States with discontinuous health insurance coverage.
N Engl J Med. 2005;353
:382
–391
[Abstract/Free Full Text] - Fairbrother G, Kenney G, Hanson K, Dubay L. How do stressful environments relate to access and use of health care by low-income children?
Med Care Res Rev. 2005;62
:205
–230
[Abstract/Free Full Text] - Chung PF, Lee TC, Morrison JL, Schuster MA. Preventive care for children in the United States: quality and barriers. Annu Rev Public Health. 2006;27 :491 –515[CrossRef][Web of Science][Medline]
- Cohen JW, Monheit AC, Beauregard KM, et al. The Medical Expenditure Panel Survey: a national health information resource. Inquiry. 1996;33 :373 –389[CrossRef][Web of Science][Medline]
- Cohen SB. Sample Design of the 1996 Medical Expenditure Panel Survey Household Component. Rockville, MD: Agency for Healthcare Policy and Research; 1997. MEPS methodology report 2, publication 97–0027
- Agency for Healthcare Research and Quality. MEPS HC-070: 2002 Full Year Consolidated Data File. Rockville, MD: Agency for Health Care Policy and Research; 2004: C-77–78. Available at: www.meps.ahrq.gov/PUFFiles/H70/H70doc.pdf. Accessed July 15, 2005
- Bethel CD, Read D, Stein REK, Blumberg SJ, Wells N, Newacheck PW. Identifying children with special health care needs: Development and evaluation of a short screening instrument. Ambul Pediatr. 2002;2 :38 –48[CrossRef][Web of Science][Medline]
- National Center for Health Statistics. Centers for Disease Control and Prevention Growth Charts: United States. Hyattsville, MD: National Center for Health Statistics; 2005. Available at: www.cdc.gov/nchs/data/nhanes/growthcharts/bmiage.txt. Accessed July 15, 2005
- Selden TM, Hudson JL, Banthin JS. Tracking changes in eligibility and coverage among children, 1996–2002.
Health Aff (Millwood). 2004;23
:39
–50
[Abstract/Free Full Text] - Simpson L, Owens P, Zodet M, Chevarley F, et al. Health care for children and youth in the United States: annual report on patterns of coverage, utilization, quality, and expenditures by income. Ambul Pediatr. 2005;5 :6 –44[CrossRef][Web of Science][Medline]
- Selden TM, Hudson JL. Access to care and utilization among children: estimating the effects of public and private coverage. Med Care. 2006;44 (suppl):I19 –I26[Medline]
- Sisk JE, Gorman SA, Reisinger AL, Glied SA, DuMouchel WH, Hynes MM. Evaluation of medicaid managed care: satisfaction, access, and use.
JAMA. 1996;276
:50
–55
[Abstract/Free Full Text] - Szilagyi PG. Medicaid managed care and childhood immunization delivery. J Public Health Manage Pract. 1998;4 :67 –72[Medline]
- Rosenbach ML, Irvin C, Coulam RF. Access for low-income children: is health insurance enough?
Pediatrics. 1999;103
:1167
–1174
[Abstract/Free Full Text] - Gadomski A, Jenkins P, Nichols M. Impact of a Medicaid primary care provider and preventive care on pediatric hospitalization. Pediatrics. 1998;101 (3). Available at: www.pediatrics.org/cgi/content/full/101/3/e1
- Long SK, Coughlin TA. The impact of Medicaid managed care on children. Health Serv Res. 2001;36 :7 –23[Web of Science][Medline]
- Cunningham P, Trude S. Does managed care enable more low income persons to identify a usual source of care? Implications for access to care. Med Care. 2001;39 :716 –726[CrossRef][Web of Science][Medline]
- Kirby J, Machlin S, Cohen J. Has the increase in HMO enrollment within the Medicaid population changed the pattern of health service use and expenditures? Med Care. 2003;41 (suppl):III-24–III-34
- Kaestner R, Dubay L, Kenney G. Managed care and infant health: an evaluation of Medicaid in the U.S. Soc Sci Med. 2005;60 :1815 –1833[CrossRef][Web of Science][Medline]
- Kenney G, Haley J, Tebay A. Children’s Insurance Coverage and Service Use Improve: Snapshots of America’s Families III. n. 1. Washington, DC: Urban Institute; 2003
- Child Trends. Well Child Visits. Washington, DC: Child Trends; 2005. Available at: www.childtrendsdatabank.org/pdf/93_PDF.pdf. Accessed September 22, 2005
- Ronsaville DS, Hakim RB. Well child care in the United States: racial differences in compliance with guidelines.
Am J Public Health. 2000;90
:1436
–1443
[Abstract/Free Full Text] - Woodwell DA, Cherry DK. National Ambulatory Medical Care Survey: 2002 Summary. Advance data number 346. Washington, DC: Centers for Disease Control and Prevention; 2004
- Hing E, Middleton K. National Hospital Ambulatory Medical Care Survey: 2002 Outpatient Department Summary. Advance data number 345. Washington, DC: Centers for Disease Control and Prevention; 2004
- National Association of State Medicaid Directors. Medicaid HEDIS 2000 Database Report. Washington, DC: National Association of State Medicaid Directors; 2003. Available at: www.nasmd.org/pubs/pubs.htm. Accessed September 22, 2005
- Sing M. Using encounter data from Medicaid HMOs for research and monitoring. Inquiry. 2004;41 :336 –346[Web of Science][Medline]
- Schneider KM, Wiblin RT, Downs KS, O’Donnell BE. Methods for evaluating the provision of well child care. Jt Comm J Qual Patient Saf. 2001;27 :673 –682
- Louthan MV, Lafferty-Oza MJ, Smith ER, et al. Diagnosis and treatment frequency for overweight children and adolescents at well child visits.
Clin Pediatr (Phila). 2005;44
:57
–61
[Abstract/Free Full Text] - Chang RKR, Halfon N. Geographic distribution of pediatricians in the United States: an analysis of the fifty states and Washington, DC.
Pediatrics. 1997;100
:172
–179
[Abstract/Free Full Text] - McInerny TK, Cull WL, Yudkowsky BK. Physician reimbursement levels and adherence to American Academy of Pediatrics well-visit and immunization recommendations.
Pediatrics. 2005;115
:1068
–1070
[Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Schonwald, K. Horan, and N. Huntington Developmental Screening: Is There Enough Time? Clinical Pediatrics, July 1, 2009; 48(6): 648 - 655. [Abstract] [PDF]
|
|
|
|
 |
|
 D. X. Vo and M. J. Park Racial/Ethnic Disparities and Culturally Competent Health Care Among Youth and Young Men American Journal of Men's Health, June 1, 2008; 2(2): 192 - 205. [Abstract] [PDF]
|
|
|
|
 |
|
 S C Hill and L Liang Smoking in the home and children's health Tob. Control, February 1, 2008; 17(1): 32 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Van Cleave, M. Heisler, J. M. Devries, T. A. Joiner, and M. M. Davis Discussion of Illness During Well-Child Care Visits With Parents of Children With and Without Special Health Care Needs Arch Pediatr Adolesc Med, December 1, 2007; 161(12): 1170 - 1175. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. L. Schor Should Children With Special Health Care Needs Have Separate Preventive Care Visits? Arch Pediatr Adolesc Med, December 1, 2007; 161(12): 1216 - 1218. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
