SERVIC,
3 ah Ey,

3 "ty
=
2 SA ASPE OFFICE OF pee | OFFICE OF THE SECRETARY
= PATIENT-CENTERED OUTCOMES
Se C ssanaractasret | HEALTH POLICY | TF] RESEARCH TRUST FUND
"i

 

Yara

FINAL REPORT
State Data for Conducting Patient-
Centered Outcomes Research to Improve
Maternal Health: Stakeholder Discussions
Summary Report

Prepared for
the Office of the Assistant Secretary for Planning and Evaluation (ASPE)
at the U.S. Department of Health & Human Services

by
NORC at the University of Chicago and AcademyHealth

December 2020

 
Office of the Assistant Secretary for Planning and Evaluation

The Assistant Secretary for Planning and Evaluation (ASPE) advises the Secretary of the U.S. Department
of Health and Human Services (HHS) on policy development in health, disability, human services, data,
and science; and provides advice and analysis on economic policy. ASPE leads special initiatives;
coordinates the Department's evaluation, research, and demonstration activities; and manages cross-
Department planning activities such as strategic planning, legislative planning, and review of regulations.
Integral to this role, ASPE conducts research and evaluation studies; develops policy analyses; and
estimates the cost and benefits of policy alternatives under consideration by the Department or
Congress.

The Office of Health Policy

The Office of Health Policy (HP) provides a cross-cutting policy perspective that bridges Departmental
programs, public and private sector activities, and the research community, in order to develop, analyze,
coordinate and provide leadership on health policy issues for the Secretary. HP carries out this mission
by conducting policy, economic and budget analyses, assisting in the development and review of
regulations, assisting in the development and formulation of budgets and legislation, and assisting in
survey design efforts, as well as conducting and coordinating research, evaluation, and information
dissemination on issues relating to health policy.

Office of the Secretary - Patient-Centered Outcomes Research Trust Fund

The Office of the Secretary - Patient-Centered Outcomes Research Trust Fund (OS-PCORTF) was
established as part of the 2010 Patient Protection and Affordable Care Act (ACA) and is charged to build
data capacity for patient-centered outcomes research. Managed by the Office of the Assistant Secretary
for Planning and Evaluation on behalf of the Department, OS-PCORTF has funded a rich portfolio of
projects to meet emerging U.S. Department of Health and Human Services policy priorities and fill gaps
in data infrastructure to enhance capabilities to collect, link, and analyze data for patient-centered
research. The OS-PCORTF portfolio includes projects that are developing and testing standards that
improve data interoperability, piloting novel approaches to patient-provided data collection, using real-
world data (RWD) in evidence generation, and addressing challenges to data linkages.
This study was conducted by NORC at the University of Chicago and AcademyHealth under Contract No.
HHSP233201600020I, Task Order No. HHSP23337001T with the Office of the Assistant Secretary for Planning
and Evaluation - Patient-Centered Outcomes Research Trust Fund.

Contributing Authors

Susan Kennedy, MPP, MSW, Director, AcademyHealth

Jeff Schiff, MD, MBA, FAAP, Senior Scholar, AcademyHealth
Sunita Krishnan, MPH, Senior Associate, AcademyHealth
Adil Moiduddin, MPP, Senior Vice President, NORC

Denise Bellows, PhD, Senior Research Scientist, NORC

Mandar Bodas, PhD, Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human
Services

Violanda Grigorescu, MD, MSPH, Division of Healthcare Quality and Outcomes, Assistant Secretary for Planning
and Evaluation, U.S. Department of Health and Human Services

Danielle DeCosta, MPH, Senior Associate, AcademyHealth
Tess Bernhard, Research Assistant, AcademyHealth
Caitlin Otter, Research Assistant, AcademyHealth

Acknowledgements

Technical Expert Panel members

Amy Branum, CDC National Center for Health Statistics

Juanita Chinn, NIH Eunice Kennedy Shriver National Institute of Child Health and Human Development
Shanna Cox, CDC Division of Reproductive Health

Renee Fox and Kristen Zycherman, CMS Center for Medicare and CHIP Services

Lawrence Reid, AHRQ Center for Financing, Access, and Cost Trends

Catherine Vladutiu, HRSA Maternal and Child Health Bureau

Subject Matter Experts and Reviewers
Jennifer Smith and Scott Leitz (NORC)

Project officers and Project Leadership

Mandar Bodas, Susan Lumsden and Scott Smith (OS-PCORTF Project Officers)
Prashila Dullabh and Rina Dnopeshwarkar (NORC)

y~

we N O RC panes AcademyHealth

December 2020 FINAL REPORT i

 
TABLE OF CONTENTS

EX@CUtiVe SUIMMATLY .....cccsesescseneccccccesersnenancrcccenssenonanssonsuasessacanscecseasesccenassnsneasesscccasessneasesesseasessnenassnens 1
INtrOCGUCTION ..........:csceenceceeseronenensncnsesoreneroneuansuensesoscneroucuauauensonorcnereuensnunansesoresessescussononenoresesesecuesonones 2
Meethods...........ceceseseseneneccoeenerensuanenensocoeseerensuancuecnosoesusreueuanenacuonorsuereneuanscacceserssesssesesscucsesecesesssesesscacees 2
FUMGIINGS.....c.cccsesesccccsssccccanansscneuaceccccansssneuansescnensessonansuonauasessanansssnauansssccecssssnsasesenecasessneasessoceasensnenassnsns 3
1. State Initiative Characteristics 0.2.2... cecseeseeesceeseseseeesecseeeeesesesenenneneneesenesenensnessnentaeesseesssensentanensoes 3
2. Funding of State Maternal Health Initiatives.............:ccsssscssssccssssscssnecsssrecsssescsserssanesensasersneeseaasersn 4
3. State Data Infrastructure and Stewardship ..........:cssccsssccssssccssscesssccecsecsssesssseeesseecesseessnssenseeeeeses 5
4, Data Linkage Capacity ..........ccsssccssssccssssccsssessssccceseecssseseneeessssecssseessucecesauecsssaeessceseseuesssseeesaceneseeeess 7
5. Data Use for Quality IMproveMent.............:::cccssscsecesssesesessserenessseeeeeessseeesessseeenensaesesessserensssseneesesss 8
DISCUSSION ...........scsesesesessesesnsesencnesenenneneneneneneneneneeneneneaeaeaeaeaeauanauacececacananeasasasasacacasacacasoosessosoucncacaeues 11
Conclusion and Considerations for Future Actions .........ccccscsscccccecesesescsssssssssssssssssssssssssssesesescccesenes 12
Appendix A. State Birth and Maternal Mortality Rates............s.ssscccssssscscsssscccssssscccsssscccnsnssescnesssccaenes 13
Appendix B. List of DiscUSSAnts.........sssscescssssneseneesessscneneneneneransonenenenesesaucnuenanenesesccceueceaesesessoseueaenenenes 14
Appendix C. List of TEP Members .........ccccsssssescscssorcnsneneneneserasaouenensnasesaucuuoussesasasananaecenssesasasanenasessenes 15
REFEENCES ........ceseseseccseccoeererensnensccccosoesnerenenanccenosoreraneuauanaceenossnereneuansccansnecesesesesasscacaeseresesesesasscoses 16
TABLES
Table 1: Prominent Funding Mechanisms for State Maternal Health Initiatives" ...............::cssscsssesserrerees 5
Table 2: Examples of PQC Maternal Health Interventions...............::ccccsssccccesssssecesssssccesssceeeessseeeessssseeesens 10
Table 3: Future Goals Mentioned by Stakeholders...........cccscssscsceseeseessssssresereseeesesesssessensseresesesssesseesses 10

 

December 2020 FINAL REPORT
EXECUTIVE SUMMARY

This report presents a summary of discussions with stakeholders conducted to identify key components of
successful state-based maternal health data linkage initiatives, to recognize the gaps in implementing such
initiatives including data use for research and finally to understand the current status of existing linkages of
state-level data with other data sources and with electronic health records. Discussants were selected from
each of the following stakeholder groups: 1) Researchers working at the state-level to identify quality
improvement opportunities, including academic and non-academic institutions, individuals that use data from
a particular state or group of states, and networks of Medicaid researchers; 2) State Medicaid Medical
Directors; 3) Data stewards, such as individuals or organizations that are involved in collecting, processing and
transmitting diverse state-level data to data repositories or to researchers or research networks; 4) State
Maternal Mortality Review Committee (MMRC) members; and 5) Perinatal Quality Collaborative (PQC)
members.

This report was produced by AcademyHealth and NORC at University of Chicago under a contract with the
Office of the Secretary's Patient-Centered Outcomes Research Trust Fund (OS-PCORTF) which is administered
by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) for the United States Department of
Health and Human Services (HHS).

The stakeholder discussions highlighted several advanced state data infrastructure characteristics such as
having the capacity to blend multiple data resources, iterative development of data processes, ability to
overcome technical challenges and the facility to explore the use of health information exchanges (HIEs) to
access clinical records on women's pre-pregnancy health. Stakeholders also provided information on funding
mechanisms of state data initiatives on MM/SMM. Furthermore, the discussions revealed several data-related
challenges faced by state initiatives such as concerns about the quality of data on race/ethnicity, challenges in
formulating data use agreements (DUAs) for data linkages and having limited ability to link data on mothers
with their babies. Stakeholders provided insight on future goals of their state's initiatives on MM/SMM. Finally,
stakeholders mentioned several considerations on future actions that may contribute to improving the health
and well-being of women nationwide.

 

December 2020 FINAL REPORT 1
INTRODUCTION

Maternal mortality (MM) rates in the United States are higher than in other developed nations.! Within the
U.S., maternal mortality rates differ by state, influenced by care quality structures and Medicaid coverage
policy among other factors", which can explain the increasing trend of severe maternal morbidity (SMM)
rates.? Further, states vary in their data capacity including the ability to collect, link, and surveil relevant data.
To understand their own MM/SMM rates, states use a variety of federal, state, and private resources. Despite
challenges encountered, states have continued to play an essential role and serve as the catalyst for prior
federal initiatives to combat the maternal health crisis through cross-agency collaborations at the national and
state (e.g., between state public health and Medicaid agencies) levels.

In early 2020, the U.S. Department of Health and Human Services (HHS) announced the "Improving Maternal
Health in America" initiative with a focus on reducing poor pregnancy and postpartum outcomes. One of the
goals outlined under this initiative is to "improve data and bolster research to inform future interventions."*
HHS's Office of the Assistant Secretary for Planning and Evaluation (ASPE) commissioned NORC and
AcademyHealth to explore state-level maternal health initiatives focused on improving maternal health
outcomes and understanding the data systems required to conduct such efforts. There is a plethora of
literature on MM/SMM; however, there are gaps in our understanding of how data relevant to MM/SMM are
used in research and quality improvement, and the challenges encountered therein. Furthermore, strong
evidence on promising practices to identify and address MM/SMM disparities through data linkage is currently
lacking. This report helps to address those gaps by providing multi-stakeholder perspectives on state data
systems, state data capacity, data linkage, timeliness of data, and quality of data.

METHODS

In collaboration with ASPE and NORC, AcademyHealth developed semi-structured discussion guides? for hour-
long conversations with five stakeholder' groups with maternal health expertise: Medicaid Medical Directors
(MMDs), Maternal Mortality Review Committees (MMRCs), Perinatal Quality Collaboratives (PQCs),
researchers, and data stewards. Informed by our project's literature review of current state maternal health
initiatives, these stakeholders were selected based on the unique perspective each group brought to state-
based maternal health care delivery and/or research. MMDs and data stewards provided perspectives based
on their roles in maternal health-focused state programmatic and administrative responsibilities. MMRC and
PQC members represented a more diverse composite, ranging from community providers to state officials,
which taken together had a deep interest and evolving understanding of the issues impacting maternal
morbidity and mortality. Maternal health researchers, as the ultimate translators of data to useful information,
contributed valuable insights on the challenges and opportunities existing data sources and infrastructure.
MMbDs and PQCs were most often involved in applying the information gained from these analyses to
comprehensive maternal health initiatives. In addition, feedback was sought from Technical Expert Panel (TEP)
members', including internal HHS experts who work on MM/SMM prevention and related quality
improvement initiatives. Our partners and stakeholders reviewed and provided input on the discussion guides
that aimed to: 1) capture key components of state maternal health data initiatives, 2) understand use of
existing state-level data, 3} identify challenges in data system implementation, and 4) capture the association
of data systems with quality improvement efforts.

 

4 Discussion guides are available upon request
5 See appendix B for a list of discussants
© See appendix C for a list of TEP members

 

December 2020 FINAL REPORT 2
 

We sourced recommendations for discussants from
State maternal health initiatives ASPE, NORC, the TEP, AcademyHealth's Evidence-
bring together individuals and Informed State Health Policy Institute (ESHPI) contacts

_ . from the Medicaid Medical Directors Network (MMDN)
organizations that otherwise would and the State-University Partnership Learning Network

not meet or share improvement (SUPLN), and PQC and MMRC member lists published by
ideas. These include medical the Centers for Disease Control and Prevention (CDC).
providers, hospitals, public health The goal was to seek a diverse set of perspectives along

multiple dimensions, including geographic location,
Medicaid non-expansion and expansion states, and

officials, Medicaid officials, data

specialists, university researchers, relative development stage of maternal health initiatives
and individuals representing patients in the state. Of note, this project used a convenience
and communities. sample of stakeholders and was not a complete survey of

states.

We facilitated 26 semi-structured discussions (representing 21 states and 59.8% of births in the US°) over the
course of five weeks in summer 2020. We then coded the discussion transcriptions using qualitative data
analysis software, NVivo 10, to extract pertinent information, which aligned with the key themes addressed in
the specific stakeholder discussion guides. Stakeholder responses were synthesized to develop this report.

FINDINGS

These findings are a snapshot of statewide maternal health initiatives from stakeholder discussions. Revealed
are promising and best practices that leverage data infrastructure, data linkages, and data capacity to address
MM/SMM to inform ASPE and support patient-centered outcomes research (PCOR) on maternal health. The
project team used a set of five domains to describe and compare state maternal health initiatives: 1) state
initiative characteristics; 2) funding structures; 3) data infrastructure; 4) data linkage processes; and 5) data
use for quality improvement. The discussants who participated in this project were instrumental in launching
their state initiatives and calling the attention of state leadership. They represented perspectives from
university-based perinatal researchers, clinicians, and state officials.

1. State Initiative Characteristics

The genesis of maternal health quality improvement initiatives and use of state data systems to address
MM/SMM varies widely by each state's administrative structure and policy focus. While some states, like
Delaware and Colorado, created initiatives to review maternal mortality as far back as the 1960s, most state
programs are less than 20 years old, with the last decade seeing an increased focus by states due to the rising
rates of MM/SMM. In many states, Governors and legislatures recognized MM as a priority through patient
experiences and preliminary data.

When exploring the founding of their state's programs, discussants explained that building internal trust within
the program and with external agencies/partners was vital for quality improvement. State maternal health
initiatives bring together individuals and organizations that otherwise would not meet or share improvement
ideas. These include medical providers, hospitals, public health officials, Medicaid officials, data specialists,
university researchers, and individuals representing patients and communities.

We found that advanced or rapidly developing initiatives shared characteristics such as: 1) creating and
maintaining coalitions with a variety of stakeholders; 2) focusing on locally important maternal health issues
such as substance use disorder (SUD) treatment or depression screening for mental illness; 3) appreciating the
value of effective and timely data feedback to front line providers to improve care; 4) understanding data

 

December 2020 FINAL REPORT 3
system development; and 5) communicating effectively with decision makers (e.g., submitting MMRC and PQC
annual reports to legislature).

While focusing on using data to address MM/SMM, state data initiatives vary in nature in terms of where the
data are housed, how the data are collected, the ability to link data, the clinical focus of the data system, and
the level of support received from policymakers. In many states, discussants had developed or were forming
partnerships with one or more major state universities. State-university partnerships can draw in expertise and
provide data capacity that is often not otherwise available to state agencies. These partnerships support data
quality, timely input of data, linkage, analysis, and provision of data to providers. Additionally, PQCs are
primarily located within universities with close ties to state government and provider organizations, though
some are located within state agencies. For example, Kentucky's PQC program is funded by state and federal
sources and housed in the Department of Public Health (DPH), while Illinois's PQC program is grant-funded and
housed outside of a state agency. State Medicaid programs, like California's Medi-Cal, support PQCs with
funding for operations, but also through other quality improvement initiatives (e.g., managed care
organization (MCO) incentives).

In addition to state-university partnerships, technical support from federal agencies has been instrumental in
improving maternal health quality initiatives. Most states commented that their use of the CDC Maternal
Mortality Review Information Application (MMRIA) database for compiling MM data enabled them to track
metrics and evaluate MM efforts more easily. Smaller states commented that sharing their MM data with the
CDC MMRIA system allowed them to enhance their own maternal mortality reviews, which were limited in
size, by being able to compare their cases with a larger set of cases from other states. A few states also noted
their active participation in the Centers for Medicare and Medicaid Innovation (CMMI) technical assistance
grant® to support data linkage of vital records to Medicaid claims. Finally, a few states participate in CMMI's
Maternal Opioid Misuse (MOM) model and described how it provides supplemental assistance to their quality
improvement efforts by targeting additional resources to improve care delivery for those beneficiaries, both
pre- and postnatal, who receive treatment for opioid use disorder (OUD).

2. Funding of State Maternal Health Initiatives

States utilize various funding streams for their maternal health initiatives; the most common funding streams
discussed by participants were Title V block grant funding', Medicaid funds, and state appropriations based on
legislative support. Table i illustrates the prominent funding mechanisms states use for maternal health
initiatives, as cited by discussants. Most states used Title V block grant funding as the principle funding to
establish initiatives. Some states were able to leverage federal Medicaid funding to develop effective data
quality systems. States receive federal funds, referred to as Federal Medical Assistance Percentage or FMAP to
match state Medicaid funds for program administration including data systems and quality improvement. The
FMAP for each state is based on a formula that provides higher reimbursement to states with lower per capita
incomes relative to the national average®.

Medicaid Information Technology Architecture (MITA) provides an enhanced (90%) federal match for state
data systems enhancements and was used by one state we interviewed. In most states MITA is being designed
to enhance enrollment and claims processing capacity with the downstream goal of using this architecture for
quality systems. States that used MITA match had quality linked to MITA and developed strong relationships
between perinatal quality initiatives and state Medicaid enrollment and claims systems. Most states did not
have these relationships and therefore, did not consider this option.

 

4 Title V is a key source of support for promoting and improving the health and well-being of the nation's mothers, children, including
children with special needs, and their families.
© Matching Rates, Medicaid and CHIP Payment and Access Commission, www.macpac.gov/subtopic/matching-rates/

 

December 2020 FINAL REPORT 4
In addition, several states participate in the Health Resources and Services Administration's (HRSA)-funded
Alliance for Innovation on Maternal Health (AIM), which assists state-based teams in developing and
implementing maternal safety bundles. State agency discussants referenced the CDC's Enhancing Reviews and
Surveillance to Eliminate Maternal Mortality (ERASE MM) program as a valuable resource in their states'
efforts to improve their overall data system infrastructure as well as data linkage and analysis capabilities.
State MM/SMM initiatives combine these funding streams with a variety of state and local resources to create
sustained and comprehensive initiatives.

Table 1: Prominent Funding Mechanisms for State Maternal Health Initiatives*

CDC Federal Private
(staffing, State

ERASE MM, (el pye CMMI appropriations
other Funded HRSA Medicaid NYKO) ha CMS via legislature or aa eK)
SEC arias grants) PQCc's® (AIM)+ Match" grant IAP** in agency budget funding

    

af AL xX

DW x x x x x
ER eX x x x x X
Ca | x x x x x

[SFL x x x x

Ce x x x x

Ds x x x x x
Ea | x X

Dax x x x
Ewe | x x x x

a a =x x X x x

a |X x

En =| x x x x
Ca |x x x X

ET =X x x x X
ay |x x x x

[17 sc_sS x x

a ee = x

Ca x x x X X
Cw x x x

 

*Defined as Maternal Opioid Misuse Model
**Defined as Medicaid Innovation Accelerator Program

3. State Data Infrastructure and Stewardship

Stakeholder discussions with data stewards! and researchers revealed that the agency location, function, and

responsibility of state data stewards vary markedly, which has an impact on state data capacities. Data
stewards range from state maternal and child health (MCH) directors or epidemiologists with primary
surveillance management responsibilities to university research institutions with full oversight of state data

 

 

December 2020 FINAL REPORT 5
warehouses, including data storage, quality assurance, and linkage responsibilities for a comprehensive
repository of data sources (i.e., vital records, hospital discharge data, emergency department records,
Medicaid claims and encounter data).

Notably, data steward roles vary with respect to data ownership versus data management, with one not
necessarily indicating responsibility for the other. Data stewards who oversee large data warehouses or
repositories noted they are not the data owners, but rather manage the data. For example, the University of
South Carolina's Institute for Society and Families manages a data warehouse comprised of key data sources
like hospital discharge data (HDD), vital records, and Medicaid claims. However, the warehouse does not own
the data, but rather acquires and manages different data sources either through a state contract or legislative
mandate.

Kentucky established a centralized office within the Cabinet for Health and Family Services (CHFS) to serve as
an "honest broker," which functions not as the data owner, but rather the provider of data quality oversight
who is principally responsible for storing and linking data based on agency requests. Other states, such as
Maryland, use a regional health information exchange (HIE) like the Chesapeake Regional Information System
for our Patients (CRISP), which primarily serves as a centralized data repository of several major data sources
in Maryland for participating providers. Data stewards such as MCH directors have responsibilities and may be
granted access to specific data (i.e., birth defects screening, Medicaid claims) based on their roles. Some data
stewards also assist with reviewing data requests. For example, researchers in South Carolina submit data
requests to a committee within the data warehouse that vets research applications. Once the committee
reviews the data request and makes its recommendation to the Medicaid agency, the data steward reviews
the recommendation with the Medicaid agency to provide guidance in the final approval decision. In most
states, data ownership and decisions about data use are maintained by the state agency responsible for the
particular data set (e.g., Medicaid claims with Medicaid, hospital discharge claims with facilities or aggregated
to the hospital association, vital records with the State Registrar or Vital Records office).

Some data stewards also provide key quality assurance responsibilities, by which they both explore the data
for inconsistencies and advise the agency on the integrity of the data. As an example, while Kentucky's
centralized data office outsources the collection of their entire state's HDD to the Kentucky Hospital
Association, it provides data quality assurance to ensure accuracy and submits the data to the Agency for
Healthcare Research and Quality (AHRQ) for the Health Care Utilization Project (HCUP)®. New Hampshire's
MCH epidemiologist shared that the state established a Birth Data Quality Committee where data quality
enhancements are made. For example, recognizing an increase in SUD-related maternal deaths, the committee
added items to the birth certificate related to prenatal substance exposure. The epidemiologist explained that
previously these maternal deaths were classified as pregnancy-associated deaths, but with this additional
information, more deaths could be counted as pregnancy-related deaths.

Discussants shared challenges regarding the collection of race, ethnicity, and other self-reported demographic
data. Medicaid discussants universally indicated their interest in improving their knowledge of race, ethnicity,
and language by improved collection of this data or linking Medicaid data to better-populated sources. Most
discussants reported that vital records were the most reliable source of information on race, ethnicity and
other demographic data. Title 42 §435.907 of the electronic Code of Federal Regulations (eCFR) only requires
Medicaid to collect information from an applicant that is necessary to make an eligibility determination. Race

 

® Currently, Kentucky's data center follows documented procedures by which an analyst conducts quality checks on the data, but the
office is currently transitioning to the use of commercial products to automate the process, including master data management (MDM)
and the use of unique identifiers to allow for improved integration across different data sources.

 

December 2020 FINAL REPORT 6
and ethnicity are optional fields during enrollment, and collection processes vary by front-line workers. As an
exception, Kentucky captures race and ethnicity data in the Integrated Eligibility and Enrollment System (IEES).

Other states noted efforts that they are undertaking to improve demographic data. Massachusetts' Medicaid
program, MassHealth, explained that the state's DPH extracts race and ethnicity information from the HDD for
reference when calculating their SMM rates. Due to the poor reporting of race and ethnicity in Medicaid
claims, MassHealth also needs access to the state's HDD for conducting SMM analyses specific to Medicaid
beneficiaries. This step requires a data request to their state's sister agency, the Center for Health Information
and Analysis (CHIA), which serves as the data warehouse for the state's providers and payers and can conduct
the data linkage.

4. Data Linkage Capacity

Both researchers and data steward discussants reported conducting data linkage, often a key responsibility of
the entity in charge of data quality assurance. Discussants varied in the data sources that they currently link for
MM/SMM, but commonly cited linkages between HDD and vital records. Discussants mentioned that linkages
between vital records and Medicaid claims are preferred because they show a more complete picture of health
care access (e.g., prenatal care) than HDD alone, but are also more challenging due to time lags and limited
analytic capacity.

Discussants identified several challenges with data linkages, such as the requirement of separate data use
agreements (DUAs) for each state agency, which delays data sharing. As an example, one discussant explained
that their state's DPH includes both the vital records office and the MCH office. The State's Medicaid and DPH
agencies have a formal interagency DUA; however, in order for the DPH's MCH epidemiologist to review linked
birth certificates with Medicaid claims, each agency requires their own DUA. This results in an onerous process
for both DPH's MCH epidemiologist as the requestor and Medicaid as the claims data owner, which then needs
to request access to the data for any review of linked claims. Other data linkage challenges relate to the data
owners and any changes they may make to their own data sets and/or infrastructure. For example, a few
discussants noted that hospital mergers often lead to delayed HDD submissions due to a change in electronic
health record (EHR) vendors. Data linkage may also be more complex due to a limited number of identifiers to
facilitate accurate linkages.

Many stewards cited the necessity of DUAs and memorandums of understanding (MOUs) with data owners to
authorize data linkage. Significant time and resources (i.e., legal staff) are dedicated to establishing these DUAs
and MOUs between state agencies and between state agencies and external partners (usually universities).
States with more robust maternal health initiatives tend to create MOUs and/or DUAs for ongoing linkages
(rather than discrete projects) through cooperative agreements between the partner agencies. Mature data
sharing agreements are long-term contracts (usually five years) and based on agreed-upon procedural steps
for data linkages (i.e., the order of operations). For example, Maine's MCH director found that an MOU
between her Office of Child and Family Services and the state's Medicaid program, MaineCare, was somewhat
streamlined because her role as Title V director already required her to have an MOU with the state's Medicaid
office. Her office leveraged this requirement to establish its ability to conduct data analyses and data linkages
for the next five years.

South Carolina's data warehouse relied on its longstanding relationship with Medicaid to establish MOUs
across the vital records and HDD entities, enabling a three-way linkage. Through these data linkages, a more
complete maternal health care profile is captured. Such a linkage reveals information on prenatal care,
pregnancy history, chronic conditions, and patients' residential location. These collective linkages enable South
Carolina to build and now apply the Palmetto small area deprivation index' to assign risk scores based on the
mother's geographic community. Kentucky has also overcome common data linkage obstacles through strong
state executive-level buy-in. Because linked data are critical for reviewing the extent of the OUD and SUD-

 

December 2020 FINAL REPORT 7
related MM rates, Kentucky's Cabinet
for Health and Family Services (CHFS)
designated linking these data as "a
standard process."

Timeliness of data feeds for record
linkage is also a common challenge. The
timing varies depending on the data
infrastructure and completeness of the
data sources, but on average, HDD and
birth certificate data are available
monthly for data stewards and/or

researchers responsible for data linkage.

For MCH officials, access to vital records
varies from daily to monthly depending
on the dataset (i.e., birth certificates,
death certificates). With respect to
timeliness in fulfilling data requests
some states like California can provide
feedback to clinical providers through
their rapid-cycle data center and others
like Ohio have developed capacity to
respond promptly to specific queries.
Kentucky shared that they can usually
respond to a request from their PQC
lead for information within two days to
two weeks, depending on the
complexity of the request. New
Hampshire can expect a cut of Medicaid
claims data in two to four weeks.
Maine's MCH director can access MCH
department data (e.g., blood spot data,
newborn screening) daily, birth
certificate data weekly, and death data
monthly.

5. Data Use for Quality
Improvement

In states that have documented
improvement in MM/SMM outcomes
such as California and Illinois, local data
are integral to implementation of
maternal reviews and quality
improvement. Advanced data programs
are able to feed data back to providers
and other systems at the local level to
improve care. Discussants spoke of
local-level process measures that are
associated with decreased morbidity
outcomes (e.g., time to hypertension

 

STATE SPOTLIGHT: BEST PRACTICE

The Commonwealth of Kentucky's Cabinet for Health and Family
Services (CHFS) formed the Office of Health Data and Analytics'
Division of Analytics in 2019 "to provide oversight and strategic
direction and be responsible for coordinating the data analysis
initiatives for the various departments that regulate health care and
social services..."

RESPONSIBILITIES

The Division is principally responsible for:

e facilitating data integration by "initiating and maintaining data-
sharing agreements in order to improve inter-agency and cross-
cabinet collaboration."

e reviewing data analyses conducted by the CHFS departments
within the cabinet "to ensure the consistency, quality, and validity
of the analysis,"

e reviewing and responding to all public data requests received by
the CHFS, and

e assuring the privacy of individual data via the chief privacy officer.

FUNDING

The Division is funded through state appropriations, Medicaid
(including FMAP), Medicaid Management Information System
(MMIS), and other grant resources.

DATA

The Division currently manages the integration of six data sets, with
additional sources anticipated shortly:

e MMI|S claims and encounters, all beneficiaries except the 1915c
waiver population (claims) are in management care (encounters)

e Office of Vital Statistics (OVS) birth and death certificates

e Kentucky All Schedule Prescription Electronic Reporting (KASPER),
the state's Prescription Drug Monitoring Program (PDMP)

e Hospital Discharge Data System

e Kentucky Cancer Registry (KCR)

e Neonatal Abstinence Syndrome (NAS) Registry

SUSTAINING AND EXPANDING THE STATE INTEGRATED DATA
STRUCTURE

This strategic infrastructure has enabled Kentucky to adopt data
linkage processes as standard procedures, allowing them to quickly
operationalize cross-agency data initiatives, including responding to
the COVID-19 public health emergency. Kentucky's COVID-19
response has hastened the addition of the Kentucky Health
Information Exchange (KHIE) to the integrated data structure. In
turn this new data source, can be utilized for their state's MM/SMM
project to identify many more births and access EHRs to capture
prenatal and/or postnatal care.

 

December 2020

FINAL REPORT 8
 

treatment under one hour associated with
decreased risk of stroke). Several state STATE SPOTLIGHT: BEST PRACTICE
discussants pointedly spoke of data use in these
quality improvement cycles as part of public
health practice and data-driven quality
improvement. For example, Ohio associated
maternal health process measures to infant
mortality by linking Medicaid claims to vital
records (birth and death certificates) to better
understand these correlations. All state
discussants had an understanding of quality
improvement processes and the need for timely
linked data to contribute to improvement cycles.

The Delaware Fetal and Infant Mortality Review (FIMR)
recently began reviewing all fetal deaths annually. This
has allowed them to take a closer look at SMM based on
FIMR cases using CDC criteria over the past two years.
The data for these cases come from the National Center
for Fatality Review and Prevention, although the
Delaware FIMR does its own quality assurance to flag
SMM, as it is not easily identifiable or standardized in the
database. The Delaware MMERC is considering expanding
its SMM definition to better understand the severe
complications that arise during pregnancy that may be
due to behavioral issues, chronic conditions, and other

When asked about specific data measures used ' : ~ a
factors not included in the traditional definition of SMM.

or needed to analyze and improve MM/SMM,
many discussants spoke in depth about measures
related to obstetric hemorrhage, hypertension, and SUD. Some states, such as Texas and Colorado, are
working to address these issues through participation in the AIM Program with the American College of
Obstetricians and Gynecologists (ACOG). States that are part of AIM report on outcome and process measures
within four categories: readiness, recognition and prevention, response, and reporting/systems learning.® For
instance, the Illinois PQC (ILPQC) measures treatment of hypertension within 60 minutes, patient education at
discharge on warning signs of preeclampsia, and follow-up visits scheduled within seven to 10 days. The ILPQC
also includes a process measure on debriefing (e.g., a huddle) after birth.

Some states focus on improvement directed by the findings of their MMRCs. The Missouri Pregnancy-
Associated Mortality Review (PAMR) looks at cases of maternal mortality to identify which deaths could have
been preventable, in particular they reviewed cases of maternal hypertension-treatment plan, associated
SUD, location and timeliness of treatment, if medication is prescribed, and other health care quality measures
to make recommendations for improved treatment protocols.

According to the National Institute for Children's Health Quality (NICHQ), many PQCs receive neonatal data
that can be used to measure MM/SMM from hospitals through elective delivery, breastfeeding, and preterm
labor metrics. Multiple discussants acknowledged this challenge-that MM and SMM analyses have
traditionally focused on obstetric complication measures, rather than factors such as behavioral health, social
determinants of health, and chronic health conditions such as cancer and hypertension. Some MMRCs have
difficulty assigning a cause of mortality due to a lack of linked or timely data for pregnant and postpartum
patients, and limitations on access to records that identify chronic diseases and social risks. The Colorado
Department of Public Health and Environment has addressed this issue through a program called Health
eMoms,' which incentivizes mothers to take an annual survey until their child's third birthday. The survey
covers a range of questions regarding postpartum health and wellness, which the Colorado MMRC then uses
to look back and understand why SMM may have occurred.

Similar to MMRCs, depending on their data structure and data validity, PQCs have different priorities for
interventions to improve maternal health. See Table 2 for examples of PQC maternal health interventions.
Most PQCs disclosed how their initial quality improvement work on MM/SMM was an addition to their work to
reduce infant mortality. Quality improvement projects were chosen based on local identification of SMM
challenges. Most states had a focus on pregnancy-related morbidity and mortality, specifically focusing on
medical management of hypertension and postpartum hemorrhage. The Illinois, California, and Florida PQCs
adapted the AIM maternal safety bundles and reported significant improvements in maternal health when
using these bundles with an integrated data feedback process. Texas reinterpreted the AIM bundles to create

 

December 2020 FINAL REPORT 9
an initiative to identify maternal early warning signs overall.' States more deeply affected by the opioid crisis
were more likely to choose initiatives that aimed to help pregnant mothers with OUD and SUD, while a
number of states that identified maternal mental health as a key issue targeted depression screening. Virtually
all states expressed interest in expanding their understanding of racial and ethnic disparities in MM/SMM in
their communities, often beginning with better data collection. MMRC analyses were used as part of the
decision-making for specific state POC efforts; however, due to the lag in the collection and reporting of data,
MMRCs' recommendations more often supported PQC quality improvement programs and the enhancement
of data systems.

Table 2: Examples of PQC Maternal Health Interventions

Intervention Type ela y(t

® Hypertension bundle
Medical Health Interventions e@ Postpartum hemorrhage bundle
e@ Use of PP LARC

e Depression screening
Treatment of SUD/OUD during and after pregnancy

Mental Health Interventions

Community health worker referral

Housing support

Prenatal doula programs

Programs targeting specific racial/ethnic populations

Social Risk Factor Interventions

 

At the conclusion of all discussions, we asked discussants about their aspirations and goals for using data to
understand their state's MM/SMM rates and to support programmatic improvements. Most discussants
identified improvements in data reporting and quality to better inform non-medical services and actions
towards social and health care equity for the mother and child, which can influence follow-up care. In addition,
discussants underscored the importance of a broader picture of the lifestyle and risks a mother might face. See
Table 3 for shared data use goals for MM/SMM by stakeholders.

Table 3: Future Goals Mentioned by Stakeholders"

Future Goals

 

Increase Non- Increase Workforce (Doulas

Improve Data Improve Data Medical Services Life Course and Community Health Care Transitions
Stakeholder Type Reporting Quality (SDOH)* Approach Workers) and Follow-Up
MMDs 5 4 3 2 3 2
Data Stewards 2 1 1 0 0 0
Researchers 5 4 2 3 1 3
MMRCs 0 1 2 3 3 1
PQCs 2 3 3 2 1 2
Total 14 13 11 10 8 8

 

* SDOH is an abbreviation for social determinants of health.

 

h This table represents specific goals that were mentioned the most by discussants based on stakeholder group. Other goals mentioned
by a smaller amount of discussants were extending postpartum coverage, increasing MM/SMM funding, and providing access to
OUD/SUD treatment.

 

December 2020 FINALREPORT 10
DISCUSSION

AcademyHealth conducted these multi-stakeholder discussions to identify more recent, less formalized state-
based MM/SM\M initiatives that otherwise may not be captured in grey literature and peer reviewed
publications. These targeted discussions addressed how states leverage data sources and analytic capacity
through collaboration with cross-state agencies, data warehouses, and university researchers, to study
MM/SMM and inform policy decision-making related to maternal health outcomes. Similar to the literature
findings, our discussions identified MM/SMM research efforts across states that focus on quality via the use
and linkage of vital records, hospital discharge data and Medicaid claims. University researchers were
instrumental in monitoring maternal health outcomes, managing data sources, and facilitating data linkages.
Findings from discussions demonstrated opportunities where states could align efforts across systems (e.g.,
Medicaid programs, health plans, hospitals) to understand the health and wellbeing of their perinatal
populations through data sharing and data quality improvements. We found variability in the strength of data
infrastructure and data stewardship. Improvements could facilitate better data standardization and integration
and help to ensure sustainability and continuity of MM/SMM data initiatives. States emphasized the value of
having a strong leader and team that are responsible for combining resources to produce robust integrated
state data systems. These states also had champions that were effective at building coalitions with clinical
leaders, policymakers, academia, and social services partners.

Discussions identified states' use of both federal agency technical support and federal and state funding
streams (i.e., ERASE MM, PQCs, CMMI IAP technical assistance, MITA, Title V, etc.) to bolster states' MM/SMM
monitoring and quality improvement infrastructure, which together influenced the robustness of MM/SMM
efforts. Most often, these initiatives were initially

ADVANCED DATA INFRASTRUCTURE federally funded and later supplemented with

state resources. Successful states had inter-agency

 

CHARACTERISTICS and cross-agency collaborations that better
States with advanced data infrastructure shared similar positioned them to align funding from non-
characteristics: competitive Title V block grants with other federal

; sources such as CDC grants, the National Institute
1. A capacity to blend resources; on Drug Abuse (NIDA) awards, and Medicaid

2. The iterative development of a data process that is
effective in data collection, analysis, and feedback to
users;

3. The ability to overcome technical, administrative,
and legal barriers to sharing data; and

4. The facility to explore the use of HIEs to access
additional clinical records regarding a woman's pre-
pregnancy and prenatal health, and/or incorporate
geographic and social service data (e.g., child
protection) to consider the full life course

matching funds. State official discussants
appreciated federal programs (i.e., ERASE MM,
PQCs, AIM) that support infrastructure and
capacity building. Discussants acknowledged the
value of federal surveillance products and support,
but sought more flexibility to give them the
opportunity to align federal products with existing
systems. Discussants also noted the potential
benefits of having longer-term federal funding
cycles to provide sustained support for
infrastructure. States varied in their capacity to
assume program operations after federal grant cycles were completed.

There was marked variation with respect to where the data are housed, the type of data sources collected and
linked, the clinical focus/primary use of data, and policymaker involvement. Most states cited linking HDD, vital
records, and Medicaid claims as essential for informing comprehensive MM/SMM initiatives. Legal and ethical
considerations, limited resources to navigate data sharing agreements, and the disparate timeliness of the
data sets were challenges that these states overcame to link data.

Our report has several limitations. First, we used a convenience sample of state maternal health stakeholders
in five unique roles (MMDs, PQCs, MMRCs, researchers, and data stewards). Second, our review of 21 states,

 

December 2020 FINALREPORT 11
while representative of 59.8% of births in the US, is limited in scope and depth. Individual discussants shared
their knowledge of their state, which was often comprehensive, but sometimes limited in perspective
depending on the question asked. Despite these limitations, this report offers ASPE and HHS a broader
understanding of the current status of state MM/SMM data initiatives and potential opportunities to advance
federal support of these efforts. This report is a distillation of over 25 hours of interviews and associated
background research, which is a strength. Another strength is the use of a semi-structured interview process to
assure discussion uniformity and the coding of all discussions via qualitative analysis software NVivo 10 to limit
bias.

CONCLUSION AND CONSIDERATIONS FOR FUTURE ACTIONS

Based on our discussions, states that have an integrated and sustainable data infrastructure and a trusted data
steward that bridges data owners and maternal health researchers are better positioned to align funding
resources and leverage data linkage capacity to improve maternal health. With heightened awareness of the
importance of data systems that provide valid, timely and linked information, the federal government has the
opportunity to further align its support to states to strengthen maternal health initiatives.

Our discussions revealed several considerations to develop the infrastructure to improve the collection,
analysis (including linkage) and use of maternal health data by states. First, in order to support sustainable
state programs, both funding and technical support opportunities could be considered. These considerations
include funding support for all states to innovate data systems from their current status to enhance
capabilities including data linkage and timely feedback for quality improvement. Such federal funding sources
could be aligned and enhanced to support program continuity over time. To assist states in building their
analytic capacity, both direct technical assistance to states and support via an interstate learning collaborative
can be established. This comprehensive approach could enable states to improve data quality, link data, use
data to inform quality improvement and address technical, administrative and legal barriers to data linkage.
Finally, specific technical support and funding to enhance the collection and use of data on race, ethnicity and
social determinants is a major consideration revealed by the discussants. Modernizing MM/SMM initiatives
through these and other considerations about data standardization, integration and interoperability will
contribute to improving the health and well-being of women nationwide.

 

December 2020 FINALREPORT 12
APPENDIX A. STATE BIRTH AND MATERNAL MORTALITY RATES

2018 State Birth and MM Rates

 

AEN t Total Births Maternal Mortality Rates"
AL 57,761 36.4
CA 454,920 11.7
co 62,885 xX'
DE 10,621 X
FL 221,542 15.8
GA 126,172 27.7

IL 144,815 9.7
KS 36,261 X
KY 53,922 40.8
MA 69,109 17.4
MD 71,080 14,1
ME 12,311 X
MI 110,032 16.4
NH 11,995 X
OH 135,134 14.1
PA 135,673 14
sc 56,669 27.7
SD 11,893 X
TX 378,624 18.5
WA 86,085 15.1
WV 18,248 X
Total for discussant states 2,265,752 X
Total in U.S. 3,791,712 17.4

 

* Maternal mortality rates are per 100,000 live births. This is public data extracted from: https://www.cdc.gov/nchs/maternal-
mortality/MMR-2018-State-Data-508.pdf

" Data suppressed due to confidentiality restrictions

 

December 2020 FINALREPORT 13
APPENDIX B. LIST OF DISCUSSANTS

Stakeholder Groups

Medicaid Medical

WEN CTarl Mi teye ela
CCW ella key=13

Perinatal Quality

Researchers

DeLee MaMa ek

 

Directors (MMDs)

Mary Applegate, MD
(OH)

Monica Le, MD (MA)

Jim Becker, MD (WV)

Judy Zerzan, MD (WA)

Mary Carpenter, MD
(SD)

Elizabeth Brown, MD
(DE)

(MMRCs)
Brie Thumm, MD (CO)

Anne Pedrick (DE)

Leticia Hernandez, PhD
(FL)

Manda Hall, MD (TX)

Collaboratives (PQCs)

Patricia Heinrich
(NICHQ)

Elliott Main, MD (CA
Maternal Quality Care
Collaborative)

Patricia (Patti) Lee King,
PhD, MSW (IL)

Devika Maulik, MD (KS)

Connie White, MD (KY)

Martha Wingate, DrPH
(AL)

Bill Sappenfield, MD
(University of South
Florida)

Michelle Menegay
(Ohio State
Government Resource
Center)

Jessica Smith, PhD
(Georgia Health Policy
Center)

Marian Jarlenski, PhD
(University of
Pittsburgh)

Kara Zivin, PhD
(University of Michigan)

Maryann Harakall
(Maine MCH Director)

Maik Schutze (Kentucky
Office of Health Data
and Analytics)

David LaFlamme, PhD
(University of NH)

Ana Lopez DeFede, PhD
(SCBOI)

Marc Rabner, MD
(CRISP)

 

NOTE: The views expressed by discussants were personal not of the organizations they represent.

 

December 2020

FINALREPORT 14
APPENDIX C. LIST OF TEP MEMBERS

Technical Expert Panel (TEP): State Data for Conducting Patient-Centered Outcomes Research to Improve
Maternal Health

 

Name Organization Contact
Amy Branum CDC/DDPHSS/NCHS zvl5@cdc.gov
Catherine Viadutiu HRSA/ MCHB CVladutiu@hrsa.gov
Shanna Cox CDC/DDNID/NCCDPHP/DRH cio8@cdc.gov
Lawrence Reid AHRQ/CFACT Lawrence.Reid@ahrg.hhs.gov
Juanita Chinn NIH/NICHD juanita.chinn@nih.gov
Kristen Zycherman CMS/CMCS Kristen.Zycherman@cms.hhs.gov
Renee Fox CMS/CMCS Renee.Fox@cms.hhs.gov

 

 

December 2020 FINALREPORT 15
REFERENCES

1 Hoyert, D.L. & Minifio, A.M. (2020). Maternal mortality in the United States: Changes in coding, publication,
and data release, 2018. National Vital records Reports, 69(2):1-18.

2 Searing, A., & Ross, D. C. (2019). Medicaid expansion fills gaps in maternal health coverage leading to
healthier mothers and babies. In Center for Children and Families.(May 2019)(citing Eliason, E."Adoption of
Medicaid Expansion under the Affordable Care Act is Associated with Lower Rates of Maternal Mortality."
Academy Health 2019 National Health Policy Conference.(Presentation, Washington.

3 Creanga AA, Berg CJ, Ko JY, Farr SL, Tong VT, Bruce FC, Callaghan WM. Maternal mortality and morbidity in
the United States: where are we now? J Womens Health (Larchmt). 2014 Jan;23(1):3-9. doi:
10.1089/jwh.2013.4617. PMID: 24383493; PMCID: PMC3880915

"U.S. Department of Health and Human Services. (2020). Putting America's health first: FY 2021 President's
budget for HHS. https://www.hhs.gov/sites/default/files/fy-2021-budget-in-brief.pdf.

> Martin JA, Hamilton BE, Osterman MJK, Driscoll AK. Births: Final data for 2018. National Vital records Reports;
Vol 68, no 13. Hyattsville, MD: National Center for Health Statistics. 2019.

5 https://www.medicaid.gov/resources-for-states/innovation-accelerator-program/functional-areas/value-
based-payment-financial-simulations/index.html

7 Lopez-De Fede, A., Stewart, J. E., Hardin, J. W., & Mayfield-Smith, K. (2016). Comparison of small-area
deprivation measures as predictors of chronic disease burden in a low-income population. Int J Equity Health,
15(89). https://doi.org/10.1186/s12939-016-0378-9

8 Council on Patient Safety in Women's Health Care. (2015). Patient safety bundle: Obstetric hemorrhage.
https://safehealthcareforeverywoman.org/wp-content/uploads/2017/11/Obstetric-Hemorrhage-Bundle.pdf

° Colorado Department of Public Health & Environment. (2019). Health eMoms.
https://www.colorado.gov/pacific/cdphe/Health-eMoms

10 The Texas Collaborative for Healthy Mothers and Babies. Maternal early warning system.
https://www.tchmb.org/maternal-early-warning-system

 

December 2020 FINALREPORT 16