What is the most comprehensive source of data on health care utilization and costs for Americans? - Westat, Inc. (2023)

Client

Health Research and Quality Agency (AHRQ)

Challenge

Policy changes in the US health care and funding system that affect how Americans use and pay for health care services and prescription drugs require continuous, large-scale data updates. These updates not only project how much the government will need to spend on current and future federal health programs and drug benefits, but also alert policymakers to pressure points in the private sector. These could include, for example, an increase in the proportion of Americans without health insurance and its impact on emergency room utilization. All this important data must be available in a timely and continuous manner as the country's healthcare system is constantly changing.

Collecting this data efficiently and accurately via the Home Component (HC) of the Medical Expense Panel Survey (MEPS) requires a highly skilled survey company with extensive experience in large-scale surveys and an extensive team of experts, including in data science and field data Collectors, cutting-edge technology and multi-mode approaches to make data collection and processing efficient and create archives for public use.

Westat conducts MEPS-HC to support the Agency for Healthcare Research and Quality (AHRQ)'s mission to provide evidence to make healthcare safer, better quality, more equitable, more accessible and more affordable.

Solution

Each year, Westat compiles a new panel of approximately 10,000 households who responded to the National Health Interview Survey in the previous year. These families will be invited to participate in a series of five interviews over a 2.5 year period and allow MEPS to obtain data from their health care providers' records. This design results in a series of overlapping panels, allowing us to combine data from two panels each year to create annual estimates of the US civilian population. To manage this vast amount of data, Westat created 4,000 class variables to impute healthcare expenditure and reduce bias in MEPS expenditure estimates.

Our interviewers typically conduct the main interview in person, assisted by Computer Assisted Face-to-Face Interviews (CAPI), Computer Assisted Telephone Interviews (CATI), and Computer Assisted Video Interviews (CAVI). In addition, we used printed and web-based self-administered questionnaires (SAQs) to collect additional information about individual adults in each household.

When the COVID-19 pandemic disrupted in-person interviews in 2020 and 2021, we responded with creativity and flexibility, looking for other ways to resume data collection without jeopardizing future estimates. We immediately switched to phone interviews and built on an existing respondent website to provide families with materials and "showcase cards" so they could see response categories for specific items. For a new SAQ 2021 on social determinants of health, we created both web and paper versions of the questionnaire and invited respondents via email and SMS to complete the web version. In 2022, we launched CAVI for essential interviews, allowing our interviewers to connect with respondents through eye contact and a dedicated window to display business cards, eliminating the need for the respondent to visit a website to view them.

MEPS-HC data collectors are equipped with advanced technology to support Westat's data collection process. Our advanced field service system supports case assignments, local and national travel, and all time and expense reporting. These systems run on laptops and mobile devices, and we use GIS data to help data collectors navigate and contact respondents. The systems also capture a wide range of operational data, allowing field workers to provide rapid feedback, reporting to managers and customers, quality control, and exploration of research methods.

MEPS uses data science methods, including natural language processing (NLP), machine learning (ML), and deep learning techniques, to improve query processing and imputation of missing data. For example, our interviewers may make candid comments to clarify respondents' answers. About 20,000 comments are entered into the CAPI system by interviewers each year. We trained a classification model to automatically classify comments into 10 predefined classes. This increases processing efficiency while maintaining high data quality standards. In addition, our extensive team of experts looks for anomalies in the data throughout the data collection process to ensure its quality. As part of our quality control, our data science department uses artificial intelligence (AI) to scan transcripts of thousands of computer-assisted recorded interviews (CARI) to determine if there are any biased responses that could affect the quality of the data. . Increasing the efficiency of data management through innovative data science techniques is critical as MEPS requires rapid recovery of field cases for the next research cycle.

We're also working with AHRQ to develop web SAQs for MEPS respondents with specific health conditions, such as diabetes, cancer, or heart disease. We implement continuous improvements using multi-mode and advanced technology to reduce under-reporting and perceived burden on respondents. This allows us to collect data on health events more continuously, streamline in-person interviews, improve data quality, and reduce perceived burden on respondents.

Results

Since 1996, we have conducted nearly one million MEPS-HC face-to-face interviews, providing the government with annual nationwide estimates of Americans' use and cost of healthcare, how they pay, their health insurance coverage, their health status, and other items related to healthcare use .

Each year we create three main data files and a number of topic-specific files by combining data reported by households with data from their healthcare providers and imputing missing data. This data is continually used to inform the government on how best to improve the quality of federal health programs and expand access to health care.