Preventive Service Utilization by Minnesota Children

1
Preventive Service Utilization by Minnesota
Children

• Minnesota Health Services Research Conference
• January 26, 2004
• April Todd-Malmlov
• Health Economics Program
• Minnesota Department of Health

2
Background Cover All Kids (CAK)

• In early 2001, the Cover All Kids Coalition was
  formed to promote health care coverage and
  preventive care for MN children.
• The Coalition is comprised of various public and
  private partners including
• MN Dept of Health - MHHP (now MHA)
• MN Dept of Human Services - Neighborhood
  Health Care Network
• Childrens Defense Fund - MN Community Action
  Association
• Academy of Pediatrics-MN - Minneapolis
  Healthy Learners Board
• MN Council of Health Plans - Legal
  Services Advocacy Project
• MMA - Joint Religious Legislative
  Coalition
• MNA - Congregations Concerned for
  Children
• Natl Assoc. of Pediatric NPs - MPHA
• Associates

3
Background Information Needs

• In order to measure progress towards the
  Coalitions goals of increasing health insurance
  coverage and use of preventive services for
  children, baseline information was needed for all
  children in the state.
• No State level survey exists that measures health
  insurance coverage and use of preventive services
  for all children in Minnesota
• MN Health Access Survey provides information on
  health insurance for all MN children, but not
  preventive services.
• HEDIS provides information on preventive services
  for certain age groups of children enrolled in
  HMOs or public health insurance programs. Only
  provides information for 30-40 of children in MN
  ignores children with commercial health
  insurance, self insured employer coverage, and
  children who are uninsured.

4
Background CAK and BRFSS

• In 2001, CAK worked with the MN Department of
  Health to add questions to the 2002 Behavioral
  Risk Factor Surveillance System (BRFSS) for
  children.
• Questions in the BRFSS Child Health Module
  provide information on health insurance, dental
  coverage, well child visits, and dental visits
  for a random sample of children ages 0 through
  17. Data provided allows for detailed analysis of
  factors related to preventive service
  utilization.
• The Child Health Module was funded by MDH, DHS,
  the MN Council of Health Plans, and the St.
  Paul-Ramsey County Dept of Public Health.

5
Background Relationship to Prior Research

• Many studies have looked at the relationships
  between demographics, insurance status and health
  care utilization in general for children (Flores
  et al 1999 Newacheck et al 1998 Newacheck,
  Hughes, and Stoddard 1996 Short and Lefkowitz
  1992).
• Some studies have looked at the impact of parent
  and child characteristics on general health care
  utilization for children (Davidoff et al 2003
  Hanson 1998 Newacheck and Halfon 1986).

6
Background Relationship to Prior Research
(Continued)

• Other studies have looked at the impact of parent
  and child characteristics on preventive service
  utilization by children (Davidoff et al 2003
  Bates et al 1994 Byrd, Hoekelman, and Auinger
  1999 Ronsaville and Hakim 2000 Yu et al 2002).
• However, these prior studies that have looked at
  the impact of various factors on preventive
  service utilization for children are limited in
  scope. No study has included all of the
  following
• various parent, child, and household
  characteristics
• all children, regardless of age, insurance
  status, or income
• well child visit guidelines as opposed to at
  least one visit

7
Research Questions

• What percentage of Minnesota children are meeting
  various well child visit frequency guidelines?
• What percentage of Minnesota children over age
  three had at least one dental visit in the past
  year?
• What are the characteristics of children who did
  not meet well child visit guidelines or did not
  have a dental visit?
• What is the impact of various parent, child, and
  household characteristics on whether or not
  children met well child visit guidelines or had a
  dental visit?

8
Methods Data Source

• The BRFSS is a collaborative telephone survey
  conducted by the states and CDC.
• The BRFSS is designed to measure behavioral
  health issues and risk factors in the adult
  population over the age of 18.
• All states ask a certain set of questions in
  order for the data to be comparable from state to
  state however, states have some flexibility in
  adding their own questions.
• In Minnesota, the BRFSS is conducted by the
  Center for Health Statistics at the Minnesota
  Department of Health.

9
Methods Data Source (Continued)

• The questions in the Child Health Module added to
  the 2002 MN BRFSS were asked of all adult
  respondents who reported that there were children
  in the household.
• The adult respondent was asked to respond to the
  Child Health Module questions for one randomly
  selected child in the household.
• In 2002, 4500 adults responded to the MN BRFSS
  and the interview completion rate was 88.
• Of the 4500 adults responding, 1600 completed the
  Child Health Module or 99 of adult respondents
  reporting children in the household.

10
Methods Data Source (Continued)

• Due to the length of time allotted for the Child
  Health Module, questions related to
  race/ethnicity, geography, household income, and
  family composition were not asked directly of the
  child or of the adult respondent on behalf of the
  child. The analyses assume that the responses of
  the adult to these questions also apply to the
  child.
• In addition to data collected through the Child
  Health Module, data relating to household
  characteristics and parental demographics, health
  care access, utilization, and coverage from the
  BRFSS was used in the analyses.

11
Methods Data Source (Continued)

• The analyses assume that the adult respondent is
  a parent of the randomly selected child.
• Some adult responses were eliminated from the
  analyses comparing adults and children. In
  instances where the age of the child and the age
  of the adult suggest that the adult is not the
  parent of the child, the responses of the adult
  to some demographic, family composition, health
  insurance, dental coverage, and health care
  utilization questions were eliminated.
• The responses of approximately 80 adults were
  dropped from the descriptive and regression
  analyses.

12
Methods Descriptive Statistics

• The American Academy of Pediatrics (AAP), Child
  and Teen Checkup (CTC), and the Institute for
  Clinical Systems Improvement (ICSI) guidelines
  are all used to analyze the percent of children
  in Minnesota who met well child visit frequency
  guidelines in 2002.
• Children were classified as meeting the well
  child visit frequency guidelines if they received
  the recommended number of well child visits in
  the past year(s) based on their age (in months
  for children under age two). The survey did not
  attempt to assess whether all of the recommended
  services were provided during the well child
  visits reported by parents.

13
Methods Descriptive Statistics (Continued)

• Due to question wording, an exact estimate of
  children meeting ICSI guidelines is not possible.
  It is assumed that a child over the age of two
  meets ICSI guidelines if they had a well child
  visit in the past two years. This assumption may
  underestimate those meeting ICSI guidelines.
• Question wording does not allow for an analysis
  of children meeting dental guidelines. It only
  allows for an analysis of children having at
  least one dental visit in the past year.
• The data was analyzed using Stata survey commands
  to account for complex survey design and
  weighting.

14
Methods Logistic Regression

• Logistic regression is used to quantify the
  relative impact of various parent, child, and
  household characteristics on dental and well
  child visit utilization.
• The dependent variables include
• Whether or not children age three or older had a
  dental visit in the past year
• Whether or not children met the CTC or ICSI well
  child visit frequency guidelines CTC
  guidelines are applied to uninsured children and
  those with public coverage and ICSI guidelines
  are applied to children with private coverage.

15
Methods Logistic Regression (Continued)

• Conceptual Framework Aday and Andersons model
  of health services utilization (1974, 1981).
  Dental and well child visit utilization are
  modeled as a function of predisposing, enabling,
  need, and availability factors.
• Analyses were conducted using svylogit commands
  to account for complex survey design and
  weighting.
• Odds ratios are reported for the logistic
  regression analyses.

16
Methods Logistic Regression (Continued)

• Interactions between parent and child dental
  coverage and parent and child health insurance
  coverage were detected.
• To deal with these interactions and account for
  the importance of coverage for both parents and
  children, three logistic regression models were
  conducted for both dental and well child
  utilization
• One model was constructed including coverage for
  parents and children to obtain odds ratios for
  other variables in the model excluding parent and
  child coverage.
• In a second model, an odds ratio for parental
  coverage was derived by excluding child coverage
  from the model.
• In a third model, an odds ratio for child
  coverage was derived by excluding parental
  coverage from the model.

17
Methods Independent Variables in Dental and Well
Child Models
18
Results Percent of Children 3 With Dental Visit
in Past Year
Indicates statistically significant difference
(90 level) from all children Indicates
statistically significant difference (90 level)
from White, Non-Hispanic
19
Results Characteristics of Children With and
Without a Dental Visit
Indicates statistically significant difference
(90 level) from children who had a dental visit
20
Results Percent of Children Meeting Well Child
Visit Guidelines
Indicates statistically significant difference
(90 level) from all children Indicates
statistically significant difference (90 level)
from White, Non-Hispanic Indicates
statistically significant difference (90 level)
from Greater MN
21
Results Characteristics of Children Who Did and
Did Not Meet Well Child Guidelines
Indicates statistically significant difference
(90 level) from children who met guidelines
22
Results Logistic Regression Dental Visit Model
23
Results Logistic Regression Well Child Visit
Guideline Model
24
Conclusions

• Over 87 of Minnesota children ages three or
  older had at least one dental visit in the past
  year.
• Depending on the guideline, roughly 72 to 83 of
  MN children met well child visit guidelines in
  2002.
• A childs age, parental dental care, household
  income, and type of child and parental dental
  coverage had a greater impact on dental
  utilization than other factors.
• A childs age, parental and child health
  coverage, household income, parental age,
  continuity of care, and geography had a greater
  impact on meeting well child visit guidelines
  than other factors.

25
Conclusions (Continued)

• Similar themes for dental and well child
  utilization
• Younger children and children of young parents
  were less likely to have a dental visit or meet
  well child guidelines
• Insurance coverage for parents and children is
  associated with having a dental visit and meeting
  well child guidelines
• Different themes for dental and well child
  utilization
• Income higher income important factor for dental
  visit and lower income important factor for
  meeting well child visit guidelines
• Geography no impact for dental visit, but Twin
  Cities location important factor for meeting well
  child visit guidelines
• Parental preventive actions important for dental
  visit, but not important for meeting well child
  visit guidelines

26
Limitations and Future Work

• Limitations
• Health status not asked of children in Child
  Health Module
• Type of insurance coverage for adults not asked
  on BRFSS
• Self-reported information from parents
• Limitations mentioned earlier regarding question
  wording and assumptions of parent and child
  characteristics
• Future Work
• Rewording of some questions in Child Health
  Module
• Addition of some questions to Child Health Module
  
• More research on the impact of geography and
  parental preventive actions on use of preventive
  care by children.
• SLAITS National Survey of Childrens Health
• Similar questions to 2002 MN BRFSS Child Health
  Module
• Data collected during 2003 and available sometime
  in 2004
• 2000 sample size per state

27
Take Home Message

• Various strategies in addition to increasing
  insurance coverage for children, could
  potentially increase the use of preventive
  services for Minnesota children. Other strategies
  include
• Improving continuity of care for children
• Increasing preventive service utilization and
  insurance coverage for parents
• Increasing outreach efforts in Greater MN and for
  young parents with young children
• Outreach efforts may be more effective and easier
  for providers and parents to understand and
  remember if one set of guidelines is used for all
  kids