2015 NHIS Justification for Removal of Health Care Reform Questions

Attachment 3i - 2015 NHIS Removal of Health Care Reform Questions
NHIS 2015: Recommendations for Retention/Removal of Health Care Reform Questions
Summary:
Based on analyses of the NHIS 2013 Inhouse dataset as well as discussions with Anjel Vahratian (Chief, DAQAB),
Robin Cohen (analyst, DAQAB), and Whitney Kirzinger (analyst, DAQAB), what follows is a list of 25 questions
that have been identified as candidates for deletion from the Health Care reform supplement in 2015. (All
analyses use inhouse weights except where noted.)
In the Family Questionnaire, these include:
1) detailed information on persons covered by an insurance plan outside of the family roster (PRCTOHPRAGOH2), (5 questions, more for larger families)
2) insurance costs paid by the employer (EMPPAY-EMPCSTP) (4 questions, more for multiple plan types)
In the Sample Child Questionnaire, these include:
3) reasons that a child does not have a usual source of care (CNOUSLPL) (1 question)
In the Sample Adult Questionnaire, these include:
4) reasons that an adult does not have a usual source of care (ANOUSLPL) (1 question)
5) questions on wait time for appointments and to see doctor (AVISLAST-AWAITRMT) (6 questions)
6) questions about long term care (LTCFAM-LTCWHO) (3 questions)
7) questions on denial or higher costs for those who tried to buy insurance on their own (AINDENY1- AINDENY
3) (3 questions)
8) questions on reasons why a plan was not purchased for those who tried to buy insurance on their own
(AINDNOT-AINDNTSP) (2 questions)
Specifics:
1) Detailed information on persons covered by an insurance plan outside of the family roster
To date, these variables have not been used in health care reform analyses by DHIS. Analyses by Robin and
Whitney on insurance coverage among young adults aged 19-25 did not use these questions. It is unclear how
estimates based on these items would be used (estimates of nationally representative sample of parents whose
children do not live at home but have coverage?). Robin thinks that we should keep the filter question (“Does
this plan cover anyone who does not live here?” PRCOOHH), but none of the follow-up questions.
2) Insurance costs paid by the employer
For EMPPAY1 (“How much your employer or union pays for the plan,” Plan 1), only about 14% of persons know
how much their employer pays for the plan. Another 13% gave a “don’t know” response, while 72% said that
“No, they don’t know how much the employer pays.” Slightly more than 6% for the second insurance plan
reported that they knew how much the employer pays for the plan. Data from the follow-up questions on
detailed cost-sharing are not likely to be missing at random and may not be useful for generation of national
estimates. To date, these variables have not been used in health care reform analyses by DHIS.
3) Reasons that a child does not have a usual source of care
Approximately 4% of sample children are in the universe for this question. Of these, more than half (58%) do not
have a usual source of care because they do not need a doctor. Another 20% do not have a usual source of care
because they do not have insurance/because of cost. Prevalence of all other reasons is low (<1% of all U.S.
children, 6% conditional on no usual source of care). Even before subgroup analyses (e.g. by race/ethnicity,
insurance coverage) ,three of the individual reasons should be used with caution (RSE > 30%). To date, these
variables have not been used in health care reform analyses by DHIS.

4) Reasons that an adult does not have a usual source of care
Approximately 15% of sample adults are in the universe for this question. Of these, more than half (53%) do not
have a usual source of care because they do not need a doctor. Another 36% do not have a usual source of care
because they do not have insurance/because of cost. Prevalence of all other reasons is low (<1% of all U.S.
adults, 6% conditional on no usual source of care). To date, these variables have not been used in health care
reform analyses by DHIS.
5) Wait time for appointments and to see doctor
Preliminary unweighted analyses find that the median number of days between making the most recent
appointment (non-recurring) and seeing the doctor or other health care professional is 7 days (IQR: 2-14).
Preliminary unweighted analyses find that the median number of minutes that adults waited in the waiting
room to see the doctor or other health care professional at the last visit was 15 (IQR: 10-30).
To date, attempts to use the waiting-room variables have been unsuccessful because of concerns that the
questions does not adequately measure the actual waiting time (the question specifies “wait in the waiting
room” rather than a total wait time in the office or clinic). Anjel notes that the questions may have limited
application because they are specific to the last visit, and without detailed information about the nature of the
visit, it may be difficult to understand any of the variability in wait times. One suggestion from the group was to
remove the current set of questions, and to develop new questions that better tie into the existing questions on
physician accessibility in 2016 or 2017.
6) Long term care
While the prevalence of informal, or unpaid, long-term care is an important topic, it is not clear that these
questions on long-term care appropriately address the topic. These questions are hypothetical and do not speak
to the data needs on number of people currently providing or receiving such care, nor the demographics (e.g.
age, gender) of those providing or receiving such care. While the LTCFAM question asks whether respondents
have “a parent, spouse, sibling, or adult child who has needed help for at least a year with everyday needs like
bathing, dressing or eating due to a long term condition,” the follow-up is not about where or how such care is
provided. Additionally, LTCFAM is only asked of sample adults aged 40-65, while a 2009 survey conducted by
AARP notes that many older adults have caregivers between the ages of 18-49.
7) Denial or higher costs for those who tried to buy insurance on their own
Approximately 5% of sample adults are in the universe for this question. Prevalence of each issue is low (<1% of
all U.S. adults, 8-15% conditional on trying to purchase coverage directly). To date, these variables have not
been used in health care reform analyses by DHIS, and Robin has indicated that these questions are less relevant
since the implementation of the ACA.
8) Reasons why a plan was not purchased for those who tried to buy insurance on their own
Approximately 3% of sample adults are in the universe for this question. Of these, the majority (78%) did not
purchase coverage because of cost. Prevalence of all other reasons is low (<1% of all U.S. adults, 10% conditional
on having tried to purchase coverage directly). To date, these variables have not been used in health care
reform analyses by DHIS, and Robin has indicated that these questions are less relevant since the
implementation of the ACA.