ARTICLE |
Urban Institute, Washington, DC
| ABSTRACT |
|---|
|
|
|---|
METHODS. We used children who were aged 3 to 17 and living in families with incomes of <250% of the federal poverty level in the Medical Expenditure Panel Survey data from 2001 to 2003 to estimate linear probability models on receipt of preventive advice. The main outcome measures were receipt of clinician's advice about healthy eating, physical activity, the harmful effects of smoking in the home, proper safety restraints in a car, and use of a bicycle helmet. We also examined 2 related barriers to receiving clinician advice: whether the child had any preventive care visits in the past year and whether the child had a usual source of care other than a hospital emergency department.
RESULTS. Publicly insured children were more likely than privately insured, full-year–uninsured children, and part-year–uninsured children to have had a preventive care visit in the past year, but regardless of health insurance group, many children went without preventive care. Even conditional on having had a preventive care visit, 48% did not receive clinician advice in any of the areas measured, and 41% of the overweight children were advised about neither healthy eating nor exercise in the past year.
CONCLUSIONS. Enrolling more uninsured children in Medicaid and State Children's Health Insurance Programs could improve the chances that families receive advice about health behaviors and injury prevention; however, nearly half of the children who were insured for the entire year did not receive important advice from their clinicians.
Key Words: Medicaid child health services health promotion preventive health care visits
Abbreviations: FPL—federal poverty level SCHIP—State Children's Health Insurance Program MEPS—Medical Expenditure Panel Survey AAP—American Academy of Pediatrics USC—usual source of care ED—emergency department
In recent years, the relative prevalence of child health problems that are caused or exacerbated by health behaviors such as unhealthy eating, lack of exercise, and parental smoking has been growing.1–3 In addition, injuries remain the leading cause of medical spending for children aged 5 to 14.4 Physician counseling may be a first step toward combating these trends, because studies have found that counseling is associated with improved parenting skills, especially in the area of injury prevention.5–10 The research examining the effectiveness of physician counseling in preventing childhood obesity is thinner, but best practice guidelines continue to suggest that physicians should spend time educating their patients about healthy eating and exercise.11,12 Despite this evidence and these recommendations, many parents do not routinely receive advice in these areas from their children's physicians.8,13–18
Previous work18–22 showed both that the deficit in physician counseling is particularly large for low-income and minority children and that lack of insurance is an important barrier to receiving physician counseling; however, past literature did not analyze the extent to which having public insurance increases the probability of receiving physician counseling during preventive care visits, both through increased access to preventive care and differential rates of counseling during preventive care, particularly for this population of low-income children. We examined receipt of clinician advice about health behaviors and injury prevention for children who lived in families with incomes of <250% of the federal poverty level (FPL). We focused on children who were covered by Medicaid or the State Children's Health Insurance Program (SCHIP), because during the course of 1 year, as many as 46% of all low-income children participate in 1 of these 2 public programs (authors' calculations based on 2005 Current Population Survey).
We compared the care provided by these public insurance programs with the care received by privately insured, full-year–uninsured, and part-year–uninsured children. Because providers are expected to convey to their patients advice in many areas in the course of preventive care,23 we also examined whether practitioners seem to be targeting advice to children who would benefit most from it by examining whether parents of children who are overweight or at risk for becoming overweight are more likely to receive advice about healthy eating and exercise and parents who smoke are more likely to be advised about the harmful effects of parental smoking in the home.
| METHODS |
|---|
|
|
|---|
Our key analytic variable of interest was the child's insurance group during the past year. We defined 4 insurance groups: (1) publicly insured (through Medicaid or SCHIP) for the full year, (2) privately insured for the full year, (3) uninsured for the full year, and (4) part-year uninsured (with part-year private or public coverage). We excluded the small group of children who were insured for the full year but with a mix of public and private insurance (<2% of the sample).
Our multivariate analyses included controls for personal characteristics (age, gender, race/ethnicity, metropolitan statistical area, and geographic region) and family characteristics (lives with mother, lives with father, family income as percentage of the FPL, number of working parents, parents' highest education, survey answered in English only, and at least 1 parent smokes). Because children who are in poor health might be more likely to receive physician advice, we controlled for health status in our multivariate regressions with variables indicating whether physical or mental health was fair or poor as assessed by the parent, the parent reported that the child consistently had poor health, the child had received a diagnosis of asthma, the child had any physical limitations, the child was overweight (BMI for age > 95th percentile), or the child was at risk for becoming overweight (BMI for age between 85th and 95th percentiles). All regression analyses controlled for health status measures, except for the last 2 measures because BMI for age is a measure that is collected only for children who are aged 6 to 17. So as not to reduce our sample for all analyses, these measures were not included as standard control variables.
To examine whether parents were receiving advice from clinicians about preventing injury and encouraging healthy behaviors in their children as recommended by the AAP,24 we analyzed the following 5 questions about receipt of advice:
In the MEPS, respondents were first asked whether they had ever been advised in these areas; when they responded in the affirmative, they were then asked whether they were advised in the past year. We also examined whether families reported having ever been advised but focused on the results for the past year because this is the period during which we had a record of insurance group, our key analytic variable. Families who replied that they did no know whether they received advice in a given category (
1% of the sample) were excluded from that particular analysis.
Potential facilitators of the receipt of recommended advice about health behaviors are the frequency of interactions that the child has with a health care provider and the continuity of care that comes from regular interactions with the same health care provider. To examine whether there were differences between children's access to preventive care by insurance group, we examined whether the child reported a usual source of care (USC) other than the emergency department (ED), whether the USC was reported to be the ED, whether the child had at least 1 visit with a provider in the past year, and whether the child had any visits for the purposes of preventive care (immunizations, general checkups, or well-child care). To assess whether continuity of care and receipt of preventive care had a direct effect on receiving advice, we analyzed receipt of clinician advice for the subsets of children who (1) had at least 1 preventive care visit with a clinician in the past 12 months, (2) reported a USC that was not the ED, and (3) were in both categories 1 and 2.
Although the AAP recommends that all parents be counseled about all of these items on an annual basis, time constraints might lead clinicians to target their advice to those who are most in need of it. In particular, practitioners might target advice about the harmful effects of smoking in the home toward families in which parents smoke; likewise, practitioners might target advice about how to eat healthfully and exercise toward children who are overweight or at risk for becoming overweight. For this reason, we also analyzed these 2 groups individually, looking specifically at the types of advice that would benefit each.
The multivariate models were estimated as ordinary least squares models. To test the sensitivity of our results, we estimated logistic models, as well as models that were restricted to younger (aged 3–6) and older (aged 7–17) children; the results presented in the "Results" were robust with respect to these alternative specifications. All analyses were adjusted to account for the complex survey design of the MEPS by using Stata 9.2 (Stata Corp, College Station, TX).
| RESULTS |
|---|
|
|
|---|
|
|
In our low-income sample, only 22% of children received advice about exercise from a clinician in the past year. Moreover, even among children who were at risk for being overweight or who were overweight, only 26% received this advice (data not shown), suggesting little targeting of this advice. The percentage of families who received advice in the other areas measured was similarly low, even conditional on having had a preventive care visit, with fewer than one third of children with a preventive care visit being advised to use a bike helmet and fewer than half being advised about eating healthfully, using a proper safety restraint while riding in a car, or the negative effects on child health of parental smoking in the home.
When the analysis was broadened to include whether the children and their families were ever advised in the measured areas (data not shown), the overall percentages were slightly higher but remained far below recommended levels. For example, fewer than half reported having ever been advised about healthy eating, and fewer than one third reported having ever been advised about exercise. Children who were privately insured in the past year were less likely to have ever received this advice than children who were publicly insured in the past year in 3 categories instead of 4 (healthy eating: 45% vs 49% [P = .02]; exercise: 28% vs 31% [P = .08]; and harmful effects of smoking: 36% vs 48% [P < .01]); differences between children who were publicly insured for the past year and children who were uninsured for all or part of the past year persisted when measured as having ever having been advised rather than having been advised in the past year.
Multivariate Analysis of Access to Care
Table 3 shows that controlling for child and family characteristics did not eliminate the access differences shown in Table 2. Children who had Medicaid and SCHIP were more likely to report having a non-ED USC than all other insurance groups. For example, relative to children who were publicly insured for the full year, children who were uninsured for the full year were 30.2 percentage points less likely to have a USC other than the ED. They were also less likely to have had at least 1 visit to a health professional in the past year. Full-year privately insured children were 4.4 percentage points less likely than full-year publicly insured children to have had a preventive care visit in the past year. Children with part-year uninsurance did not do nearly as poorly as children with full-year uninsurance in terms of having at least 1 preventive care visit in the past year and having a USC other that the ED, but they were still 14.2 percentage points less likely than those with Medicaid/SCHIP coverage to have a USC that was not an ED and 4.1 percentage points less likely to have had any preventive visits.
|
|
We explored 2 factors related to health insurance group that could affect receipt of advice from a provider about health behaviors and injury prevention: whether the child had any preventive care visits in the past year and whether the child had a USC other than the ED. When the sample was restricted to children in each of these categories in Table 4, the effect of full-year uninsurance continued to have a strong negative effect on receipt of each of the 5 categories of advice. For children who reported having a USC and at least 1 preventive care visit in the past year, the estimates were much less precise; there remained a statistically significant difference, however, between children who were publicly insured for the year and other children in receipt of advice about harmful effects of smoking in the home even among this sample of children who had a USC and had a preventive visit in the past year.
Targeting Advice
In Table 5, the sample was restricted to the children who would most benefit from the particular category of advice. Restricting the sample to children who were in the
85th percentile of BMI for age did not eliminate the previous finding that publicly insured children were more likely to be advised about healthy eating and exercise than other children. Likewise, in the last column, even among parents who smoked, those with children covered by Medicaid or SCHIP were more likely to be advised that smoking in the home is bad for their child's health compared with privately insured or full- or part-year–uninsured children. As already described, children who were insured for part of the year did not do nearly as poorly on these measures as children who were uninsured for the entire year. Because these problems may have existed for years in a family, we also examined whether the parents had ever been advised in these areas (data not shown). The magnitude of the differences between children who were publicly insured or part-year uninsured relative to publicly insured fell slightly and was no longer statistically different from 0 in that specification, but the differences between children who were uninsured for the entire past year and children who were publicly insured remained similar in magnitude and significance.
|
| DISCUSSION |
|---|
|
|
|---|
We also found that, regardless of insurance group, many children did not have preventive care visits; moreover, many clinicians did not seem to take the opportunity of preventive care visits to give advice on proper safety restraint while riding in a car, bike helmet use, healthy eating, exercise, and the dangers of parental smoking in the home. In our low-income sample, we found that fewer than half of children who had a preventive care visit received clinician advice in each of the 5 measured areas.
Publicly insured children were the most likely to have had a preventive care visit in the past 12 months, yet only 41% of these children had one. Under the Omnibus Budget Reconciliation Act of 1989, state Medicaid programs are required to provide preventive care through the Early and Periodic Screening, Diagnostic, and Treatment service. Health education is explicitly required to be a component of these visits.26 Despite this, we found that even when they had a preventive care visit, many children who had public insurance, too many to be explained by the SCHIP portion of the group alone, were not receiving advice that could help them to address emerging health problems and prevent injury. Fewer than one third of publicly insured children who had a preventive care visit were advised about exercise, and just more than half were advised about eating healthfully. Even among children who were in the
85th percentile of BMI for age, those who might especially benefit from clinician advice, these numbers were only slightly higher. This represents a missed opportunity for parents to get potentially important information from a trusted source.
This analysis did not answer the question of why publicly insured children seem to be more likely than privately insured and children who spend some or all of the year uninsured to receive advice in the areas that we measured. One possible driver is the differential cost of obtaining preventive care visits; privately insured children and uninsured children are likely to have some out-of-pocket costs associated with preventive care visits, whereas children who have Medicaid and SCHIP do not. Another possible reason for the difference is that Early and Periodic Screening, Diagnostic, and Treatment requirements have a real effect on the content of preventive care.
This study has a number of limitations. First, the timing of when the MEPS asked participants about whether their doctor gave advice did not exactly correspond to the timing about which participants were asked about their health insurance; therefore, in some cases, we assigned health insurance group on the basis of the period slightly after the period during which they were asked about clinician advice. This measurement error would likely bias the effects of health insurance coverage toward 0 if there were some cases in which people had periods of uninsurance that we were not capturing in our measure.
Second, although past literature27,28 demonstrated that use increases when insurance coverage is expanded, we cannot be sure that previously uninsured children would have similar use patterns in terms of preventive care as children who were currently enrolled in public insurance if they were to enroll.
Third, recall bias is a limitation in all studies that use self-report (or parent report) to determine outcomes. Past studies have not found a differential recall bias by insurance group, so if these biases are there, then they will not change the story across the insurance groups but may partially explain why the estimates for receipt of advice are low across insurance types.29 The reported rates of advice in our sample were broadly consistent with past literature, which has been done primarily using samples of very young children. These studies have found slightly higher rates of clinician advice and higher rates of preventive care visits than we found for children aged 3 to 17 in the MEPS, which is not surprising given the difference in ages across studies.17,18,30,31 One survey of pediatrician-reported counseling patterns15 found that the percentage of pediatricians who reported always counseling about areas similar to those that we analyze were higher than what the parents in the MEPS reported receiving. For example, 62.1% of pediatricians reported always counseling children aged 6 to 12 years about healthy eating, and 51.3% reported always counseling children in this age group about physical activity. A new study32 found that pediatricians were more likely than other providers to counsel in these areas. Given that MEPS does not restrict its questions of clinician advice to pediatrician advice, these other studies may be consistent with what we found in the MEPS data.
Finally, neither the specific content of the advice nor the parent's ability to understand it was measured in the MEPS. Although counseling about health behaviors may be an important first step, it might also be necessary to provide additional supports to parents to improve their ability to follow through on the advice.
Despite these limitations, this study identified large deficits in preventive visit receipt and counseling, particularly for low-income children who lack health insurance coverage. The reauthorization of SCHIP could present an opportunity to address preventive care deficits in both visits and in content of care and also to build more quality and access monitoring into both Medicaid and SCHIP. Addressing preventive deficits among low-income privately insured children could be accomplished by expanding their access to wrap-around benefits under Medicaid and SCHIP. Even with potential coverage expansions, however, the overall low rates of advice suggest that the well-child care system could benefit from reform. Some possibilities include increasing reimbursement to providers for time spent advising patients and moving the transmission of advice to a time outside the preventive care visit, such as by e-mail, postal mailing, or telephone.
| CONCLUSIONS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
Address correspondence to Cynthia D. Perry, PhD, Urban Institute, Health Policy Center, 2100 M St, NW, Washington, DC 20037. E-mail: cperry{at}ui.urban.org
The authors have indicated they have no financial relationships relevant to this article to disclose.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||