BILL ANALYSIS Ó
AB 2670
Page 1
Date of Hearing: April 27, 2016
ASSEMBLY COMMITTEE ON APPROPRIATIONS
Lorena Gonzalez, Chair
AB
2670 (Roger Hernández) - As Introduced February 19, 2016
-----------------------------------------------------------------
|Policy |Health |Vote:|16 - 0 |
|Committee: | | | |
| | | | |
| | | | |
|-------------+-------------------------------+-----+-------------|
| | | | |
| | | | |
| | | | |
-----------------------------------------------------------------
Urgency: No State Mandated Local Program: NoReimbursable: No
SUMMARY:
This bill requires the Department of Health Care Services (DHCS)
to administer, on an annual basis, the Consumer Assessment of
Health Care Providers and Systems (CAHPS) Health Plan surveys,
as developed by the federal Agency for Healthcare Research and
Quality for all Medi-Cal managed care plan (MCP) populations.
Specifically, this bill:
AB 2670
Page 2
1)Requires DHCS to annually administer the CAHPS survey to all
MCP models, including county organized health systems (COHS),
two-plan models, and geographic managed care plans.
2)Requires DHCS to translate the CAHPS survey in all Medi-Cal
threshold languages and requires DHCS to administer the CAHPS
survey in all Medi-Cal threshold languages in each county.
3)Requires DHCS to stratify the results in order to identify
disparities in the quality of care provided to Medi-Cal
managed care enrollees based on all of the following factors:
a) Geographic region;
b) Primary language;
c) Race;
d) Ethnicity;
e) Gender; and,
f) To the extent data is available, sexual orientation and
gender identity.
4)Requires DHCS to annually prepare and make publicly available
a report on the results of the surveys on the DHCS Website.
Provides that the report will include all of the following:
a) Aggregated data on MCP results compared to national
Medicaid data;
b) Aggregated MCP results stratified by the factors set
forth as specified; and,
c) Plan results at the county level, including results for
all MCP models, as specified.
FISCAL EFFECT:
1)Costs, likely in the range of $500,000- $1 million annually to
conduct the survey every year in multiple languages. The
AB 2670
Page 3
survey is currently conducted every three years and costs
$500,000 to administer in English and Spanish.
2)Unknown, significant one-time costs to translate surveys into
thirteen threshold languages. As discussed further below,
costs would depend on whether validated translations exist or
would have to be developed, and the extent of field testing
prior to implementation.
3)Unknown, significant ongoing costs associated with the
requirement to stratify by a number of variables. Performing
an estimate of the increased sample size needed to ensure the
ability to stratify results to identify disparities in care is
beyond the scope of this analysis, but costs could be very
high if a high level of oversampling is required in order to
ensure results can be meaningfully stratified.
COMMENTS:
1)Purpose. The author states that this bill gives us the
opportunity to assess California's most diverse populations
and different health care needs. Furthermore, the author
states that Medi-Cal's quality measures must be consistent and
reflect everyone's health care experience, as one in three
Californians is now enrolled in Medi-Cal.
2)Background. The CAHPS survey assesses both patient
satisfaction and perceptions of access to care, and the survey
exists for hospitals, managed care plans, and other providers.
In Medi-Cal managed care, CAHPS measures assess the
experiences of surveyed managed care plan members in the prior
six months. Members are asked questions like, "How often was
it easy to get necessary care, tests, or treatment?" or "How
would you rate your health plan?" Currently, DHCS conducts
AB 2670
Page 4
the CAHPS survey once every three years in two languages,
English and Spanish. DHCS publishes the results on its
website and uses results as part of its Medi-Cal Managed Care
Program Quality Strategy and to recognize plan quality
improvement.
3)Staff Comments. Data collection can provide valuable insight
into program operation, but does come with a cost. The
requirements of this bill address several items:
a) This bill increases the frequency of the CAHPS survey.
This is consistent with LAO recommendations and appears
warranted given significant changes in enrollment,
networks, and overall impacts on Medi-Cal managed care plan
operations.
b) This bill also requires the survey to be translated into
all Medi-Cal threshold languages, and administered in each
county in all Medi-Cal threshold languages in that county.
This is beneficial because it allows the opinion of
Medi-Cal enrollees who do not speak English or Spanish to
be counted in the poll. However, it is unclear whether
validated translations exist in languages other than in
Spanish. Survey instruments must be validated through
psychometric testing, and surveys such as CAHPS have been
very well-tested. The agency for Healthcare Research and
Quality, which houses CAHPS, offers guidelines for
translating CAHPS surveys into other languages. These
include the use of two simultaneous translators and a
translation review to ensure the quality and cultural
appropriateness of the translated instrument, as well as
subject the translated instrument to qualitative analysis
and psychometric testing to gauge the reliability,
validity, and equivalence of the instrument in measuring
the health needs of various subgroups. Thus, while
translation is certainly possible, translating a survey is
AB 2670
Page 5
quite involved. Questions and responses translated into
each threshold languages are not guaranteed to directly map
to the English version without significant effort to
carefully translate and test the surveys. It should be
noted a small number of Medi-Cal enrollees would be sampled
for any CAHPS survey, making the chances small that some
languages would be represented.
c) Finally, this bill requires stratification of results in
order to identify disparities in the quality of care
provided to Medi-Cal managed care enrollees based on the
factors: geographic region, primary language, race,
ethnicity, and gender, as well as sexual orientation and
gender identity to the extent data is available.
Stratification is different than simply including a diverse
population in the sample, as addressed in (b) above.
Stratification would theoretically allow one to measure,
for example, how persons of a particular ethnic group
perceive care in a particular plan.
CAHPS is designed to measure enrollees' recent experiences
with health plans and their service, to support consumers
in choosing a plan, and identify areas of quality
improvement for plans. However, the LAO report notes that
DHCS has not even been able to stratify by health plan
given the sample size they've used, let alone by
subpopulations within a plan. Depending on the goals,
stratification by these sub-categories may likewise either
be impossible due to limited sample size or, if sample size
was increased in order to collect enough data to allow
stratification for the many variables of interest, data
collection could become prohibitively expensive. For
example, to compare perceptions about the timeliness of
care between Fresno County Hmong-speaking individuals and
persons speaking other languages, the survey would need to
include a sufficient number of Hmong speakers from Fresno
County. Without oversampling small populations of
AB 2670
Page 6
interest, it is difficult to collect enough data to report
anything meaningful about members of these small
populations.
For comparison's sake, the California Health Interview
Survey is conducted in English, Spanish, two Chinese
dialects, Korean, and Vietnamese. Other languages can be
added on a pilot basis, and Korean and Vietnamese, as
smaller language groups, are oversampled.
Given the potentially high cost of requiring sufficient
sample sizes to be able to stratify by the listed
variables, staff recommends the author consider clarifying
that stratification is only required if the sample size is
large enough to report meaningful results. Special one-time
surveys could be conducted whereby certain populations of
interest are oversampled, allowing for stratification if
there are particular areas of concern, instead of requiring
a costly level of data collection on an ongoing basis.
Analysis Prepared by:Lisa Murawski / APPR. / (916)
319-2081