SHADAC Staff

p 612.624.4802
e shadac@umn.edu

An Updated Brief Examines State Health Compare Estimates on State Funding for Public Health

As states have urgently strategized responses to the sudden arrival of the initial wave of the coronavirus pandemic and begun to prepare for the possibility of a second wave in the fall, states’ per-capita public health funding is a metric that has garnered new attention. Estimates for this measure not only reveal a state’s prioritization of investment in public health infrastructure, but also levels of preparedness for health emergencies such as COVID-19. With this understanding, a number of media outlets have used the public health spending rankings chart provided by SHADAC’s State Health Compare to publish articles which highlight the challenges that states on the lower end of the public health spending spectrum now face in attempting to respond to the COVID-19 crisis.

In light of this information, and with the release of the latest estimates for public health spending for 2019, SHADAC has just recently updated our brief on estimates of state-provided funding for the State Health Compare measure, “Public Health Funding.” Though this updated brief highlights state public health funding data from 2015–2019, data is available on the State Health Compare website for every year, starting in 2005 (with the exception of data year 2006 for which no estimates are available).

The Big Picture: Wide Variation between States
There is a wide gap between state public health funding among states, with 2019 state-provided funding ranging from $7 per capita in Missouri to $363 per capita in the District of Columbia (D.C.). In 2019, the median state, Kentucky spent $35 per capita in public health funding.

Per Person Public Health Funding, 2019

Top States for State Public Health Funding
Eight states occupied the top five funding spots over the course of five years, from 2015 through 2019 (the most recent five years for which data are available): Alaska, D.C., Hawaii, Idaho, New Mexico, New York, North Dakota, and West Virginia. Alaska, D.C., and Hawaii were among the top five states for each of the five years. In 2019, the top five states for per-capita state public health funding were Alaska, D.C., Hawaii, Idaho, and New Mexico.

Bottom States for Public Health Funding
Eight states also occupied the bottom five funding spots at various times over the five-year period (2015 through 2019). Arizona, Indiana, Kansas, Mississippi, Missouri, Nevada, North Carolina, and Ohio moved among the bottom five positions for state-provided public health funding during this time period. Both Missouri and Nevada consistently ranked among the bottom five states for each year of this period. In 2019, the bottom five states with regard to public health spending were: Kansas, Missouri, Nevada, North Carolina, and Ohio.

About the Estimates
Estimates of state funding for public health come from data collected by Trust for America's Health (TFAH), which obtains its calculations using budget documents that are publicly available through state government websites. Depending on which information is available, TFAH uses executive budget documents listing actual expenditures, estimated expenditures, or final appropriates; appropriations bills enacted by the state's legislature; and documents from legislative analysis offices to calculate public health spending. TFAH defines "public health" broadly to include most state-level health funding with the exception of Medicaid, CHIP, or comparable health coverage programs. SHADAC standardizes these estimates of state public health funding to the estimated population of each state to create per-capita estimates.

Why the Variation among States on this Measure?
According to Trust for America's Health, comparisons of public health funding levels across states are difficult because every state allocates and reports its budget in different ways, and states vary widely in the budget details they provide. Non-methodological sources of inter-state variation in state public health funding may also include the relative performance of individual state economies (since state public health funding is often cut during economic downturns) as well as the relative tax bases of individual states along with state population counts.

Download a PDF of the brief.

Explore Additional Public Health Data at State Health Compare
Visit State Health Compare to explore national and state-level estimates for the following public health indicators:

• Weight Assessment in Schools
• School Nutrition Standards Stronger than USDA
• School Required to Provide Physical Activity
• Smoke Free Campuses
• Cigarette Tax Rates

State Health Compare also features a number of other indicator categories, including: health insurance coverage, cost of care, access to and utilization of care, care quality, health behaviors, health outcomes, and social determinants of health.

Andrea Stewart
Research Fellow

p 612-624-4850
e astewart@umn.edu

Studying the Impact of COVID-19: State-Level Data Resources on State Health Compare

Though the spread of disease caused by the novel coronavirus (COVID-19) has impacted the lives of individuals across the United States, the magnitude of the virus' effects has varied greatly across different states. Access to current, accurate state-level data is important in order to inform state researchers, analysts, and policymakers’ efforts to understand and respond to the disparate impacts of COVID-19 on their resident populations.

To that end, this blog provides a high-level overview of a range of data measures currently housed in our online data tool, State Health Compare, as well as a number of related resources that provide more in-depth analysis for certain measures. Notably, estimates for each of the measures listed are available for subpopulations that often highlight persistent disparities in health—a feature that may prove especially useful as calls to further break down data related to COVID-19 into subcategories (e.g., gender, age, race/ethnicity, etc.) to track the impact of the disease on specific populations have recently increased.

Risk Factors
Chronic Disease Prevalence
This measure captures the percent of adults who reported having one or more common chronic conditions such as diabetes, cardiovascular disease, heart attack, stroke, or asthma. It includes breakdowns by race/ethnicity and educational attainment. (2005-2010, 2011-2019)*

Adult Smoking
This measure shows the percent of adults who reported smoking, defined as adults who have smoked 100 or more cigarettes in their lifetime and who currently smoke some days or every day. The measure is broken down by race/ethnicity and educational attainment. (2005-2010, 2011-2019)*

Related Resource: SHADAC researchers used data from the Behavioral Risk Factor Surveillance System (BRFSS) to produce analyses focused on several different health behaviors, including Adult Smoking and E-cigarette Use in the United States. This blog, part of a spotlight series including binge drinking and obesity, looks at national and state-level rates at which adults with different racial/ethnic backgrounds smoke and vape in 2018 and 2017.

State Health Funding
Medicaid Expenses as Percent of State Budget
This measure captures state and federal spending on Medicaid as a share of each state’s budget. (2000-2018)

Public Health Funding
This measure shows state public health funding per capita by fiscal year. (2005-2019; no data available for 2006)

Related Resource: SHADAC researchers produce an annual brief, Exploring Public Health Indicators with State Health Compare: State Public Health Funding, exploring the wide variation in per capita public health funding by state, and why public health funding data, in conjunction with state performance on other public health indicators, might signal which states are best suited to absorb a potential decrease in funding and which states might be hit hardest.

Access to Health Care
Had Usual Source of Medical Care
This measure counts the percent of people who had a usual source of medical care other than the emergency department (i.e., doctor’s office, clinic, health center, etc.) in the past year. Breakdowns by age and coverage type are also available. (2011-2017)

Adults with No Personal Doctor
This measure denotes the percent of adults without a personal doctor and offers breakdowns by education level and race/ethnicity. (2005-2010, 2011-2019)*

Related Resource: SHADAC recently produced a blog and infographic - Affordability and Access to Care in 2018: Examining Racial and Educational Inequities across the United States (Infographic) - focusing on racial/education inequities in access and ability to afford medical care, using this measure along with estimates for Adults Who Forgo Needed Medical Care.

Broadband Internet Access
This measure shows the percent of households that have a broadband internet subscription (i.e., that pay a cell phone or internet services provider for the service), which includes a cellular data plan, cable, fiber optic, DSL, or satellite internet service. (2016-2018)

Related Resource: A new SHADAC blog explores the way that Internet Access Measures the Impact of the Digital Divide and COVID-19. An initial analysis of the estimates shows variation in access to broadband across states, and reveals disparities by income, rurality, coverage, and disability status.

Insurance Coverage
Coverage Type
This measure shows the rates of different types of health insurance coverage, including Medicare, employer-sponsored insurance (ESI), Medicaid, and individual coverage, as well as no insurance coverage. Users can view this measure by a variety of breakdowns, including: age, citizenship, disability status, education, family income, health status, limited English proficiency, marital status, poverty level, race/ethnicity, sex, and work status. (2008-2018)

Health Care Costs and Affordability
People with High Medical Care Cost Burden
This measure highlights the percent of individuals in families where out-of-pocket health care spending, including premiums, has exceeded 10 percent of annual income. Breakdowns by employer coverage, income, and race/ethnicity are available for each state. (2010-2012, 2013-2017, 2017-2018)*

Medical Out-of-Pocket Spending
This measure estimates the annual median family “out-of-pocket” spending, including premiums, on health care costs that are typically not covered by health insurance but paid for out of an individual’s own resources. These costs can include copays for doctor and dentist visits, diagnostic tests, prescription medicine, glasses and contacts, and medical supplies. (2017-2018)

Related Resource: SHADAC used this measure to produce a brief entitled State-level Estimates of Medical Out-of-Pocket Spending for Individuals with Employer-sponsored Insurance Coverage, which shows state variation in medical out-of-pocket spending (with a brief section that also looks at high medical cost burden) for people with employer-sponsored insurance.

High-Deductible Health Plans
This measure reports the percent of private-sector employees enrolled in high-deductible health insurance plans. High-deductible health plans are defined as plans that meet the minimum deductible amount required for Health Savings Account (HSA) eligibility (e.g., $1,350 for an individual and $2,700 for a family in 2018); breakdowns by firm size are offered. (2012-2019)

Average Annual ESI Deductible and Average Annual ESI Premium
This first measure reports the average annual deductible for private-sector workers who receive a health insurance plan through their employer that has a deductible. The second measure reports the average annual premium for private-sector workers who get health insurance through their employer. Both measures can be broken down by plan type (family or single). (2003-2019; partial data for 2002)

Related Resources: SHADAC produces an annual report, the latest being State-level Trends in Employer-sponsored Insurance (ESI), 2014-2018, focused on the more than 62 million private-sector employees enrolled in Employer-Sponsored Insurance (ESI) coverage and their growing out-of-pocket costs. The report includes a narrative summary, interactive map, two-page state profiles, and data tables.

From the data used to produce our annual ESI report, SHADAC researchers also created A Deeper Dive on Employer-sponsored Health Insurance: Costs in Five States in Comparison with the United States; a focused, five state analysis of ESI trends in order to better understand which states are most affected by increasing premiums and deductibles or might have a population with high enrollment in high-deductible health plans (HDHPs), leaving them financially vulnerable in case of an unexpected health crisis that leads to large medical bills.

Adults Who Forgo Needed Medical Care
The measure indicates the percent of adults in each state who could not get needed medical care due to cost. Breakdowns by education level and race/ethnicity are available for all states. (2005-2010, 2011-2019)

Related Resource: SHADAC recently produced a blog and infographic - Affordability and Access to Care in 2018: Examining Racial and Educational Inequities across the United States (Infographic) - focusing on racial/education inequities in access and ability to afford medical care, using this measure along with estimates for Adults with No Personal Doctor. 

Trouble Paying Medical Bills
This measure tracks the rates of individuals who had difficulty paying off medical bills during the past twelve months or that were paying off medical bills over time. The measure can be broken down by age and insurance coverage type. (2011-2017)

Related Resource: In light of rising health care cost burdens, SHADAC produced a blog, Measuring Health Care Affordability with State Health Compare: Trouble Paying Medical Bills, assessing changes and patterns in health care affordability across the United States by tracking the percent of individuals that had difficulty paying off medical bills currently or over time.

Explore these measures on State Health Compare!

Notes
For a full overview of all available state-level measures, please visit the “Explore Data” page on State Health Compare, or take a look at our one-page guide to State Health Compare measures and their data sources.

All measures marked with an “*”: This indicates a break in series due to survey changes in either data processing or implementation methodology.

SHADAC Staff

p 612.624.4802
e shadac@umn.edu

Expert Perspective: State COVID-19 Data Dashboards (State Health & Value Strategies Cross-Post)

The following content is cross-posted from State Health and Value Strategies. It was first published on April 9, 2020.
Authors: Emily Zylla and Lacey Hartman, SHADAC

Accurate, timely data is a key tool in states’ efforts to understand and respond to the impact of the coronavirus (COVID-19) outbreak at the local level. There have also been increasing calls to further break down COVID-19 data into subcategories (such as by gender, age, and race and ethnicity) in order to track the impact of the disease on specific populations. As of April 6, all 50 states and DC are publicly reporting some type of data related to COVID-19, such as the number of positive tests and/or the number of deaths. Furthermore, some states have recently begun to utilize innovative dashboards in order to visualize and track reported cases of coronavirus disease as well as monitor additional related key indicators. These dashboards are designed to organize complex data in an easy-to-digest visual format, allowing the audience to easily interpret key trends and patterns at a glance (e.g., see SHADAC’s COVID-19 dashboard template, which is currently under development using mock data).

Source: SHADAC COVID-19 dashboard template under development using mock data.

This expert perspective reviews the key indicators currently being tracked by states via their COVID-19 dashboards and also provides an overview of “best practices” states can consider when developing or modifying these same COVID-19 dashboards.

Current Status of COVID-19 Dashboards

As of April 6, we identified 31 states with public-facing COVID-19 data dashboards (i.e., the information is displayed with charts and other graphics, not just in tabular form), and we anticipate that more states will publish COVID-19 dashboards in the coming days. 

States are reporting a wide number (ranging from 4 to 13) and type of indicators in their dashboards, most of which are updated at least daily.
Many states are also starting to show trends in these data points over time. The most common indicators reported on a state dashboard include:

• Number of total cases
• Number of total deaths
• Number of cases by county
• Map of cases by county
• Number of tests completed
• Number of cases by age group
• Number of cases by gender
• Number of deaths by county
• Number of hospitalizations

Other key indicators that some states are reporting that may be of interest include:

• Total number of recovered cases (i.e., cases that are no longer required to isolate)
• Number of hospitalizations that require ventilation
• Number of deaths by age/gender/race/ethnicity
• Case rate per 100,000 people by county
• Number of cases by race/ethnicity
• Number of cases by congregate living setting (e.g., long-term care, assisted living, dorms, jails, correctional settings, etc.)
• Number of tests completed by laboratory type (e.g., public vs. commercial labs)
• Number of tests completed by race/ethnicity  
• Number of calls to a state’s COVID-19 hotline or number of hits on a COVID-19 website

It is important to note that states may be defining indicators that appear initially similar in different ways. For example, some states report “hospitalizations” as the total number of cases who have ever been hospitalized, while other states report “hospitalizations” as the current number of hospitalizations on a certain day. As a result, users should be cautious about making comparisons across states. In most cases, states have not identified the sources of their dashboard data beyond indicating that data is maintained by the respective state’s health department (or equivalent) and that it comes from a variety of sources such as state and local public health surveillance data, lab data (including public health, hospital, and commercial labs), and hospital reporting systems, among others. While the quality of the data being reported is difficult to assess currently (and is therefore reported as provisional), many states have acknowledged that data on confirmed cases represent an undercount due to a lack of widespread testing. Similarly, states have suggested that data on the number of deaths from COVID-19 may also change as post-mortem testing expands and guidance on how to record COVID-19 deaths is established. As mentioned below, we encourage states to include information on data quality, such as levels of missing data, where possible.

COVID-19 Dashboard Best Practices

Audience: Before designing a dashboard, make sure to clearly identify who is the intended audience. Different levels of detail, explanation, or source information may be necessary depending on whether the intended audience is state agency leadership, political leadership, or the general public. It is also important to think about what medium you will be using to reach the audience. Will the dashboard only be published on a website? Will it be available on a mobile device? Or, might you want to print it as a handout or post it on social media?

Organization and layout: Prioritize key measures. Because timeliness of these data is so important, the dashboard needs to have enough data points to convey key information, but limited enough to update quickly. It is helpful to have a landing page that makes all indicators visible to users with limited scrolling, but also provides users with the ability to “drill down” to more detail—comparisons, methodology, etc. If it is not possible to show all indicators, there should be an obvious and intuitive option for the user to “hover” over a list and get an “at a glance” view of the available content. The following example shows Florida’s COVID-19 Dashboard landing page, which implements many of these best practices.

Source: Florida COVID-19 Data and Surveillance Dashboard. Accessed March 30, 2020.

Think about any potential layout in terms of a story. It is helpful to group indicators into high-level categories with headings (e.g., overview, demographics, hospitalizations, etc.). This provides additional context for interpreting the data without the need for lengthy text descriptions. In addition, many dashboards are modular in nature so that visual elements can be replaced as information relevancy changes over time (e.g., information on likely source of exposure may become less relevant over time, while information on health care workforce exposure may become more relevant).

Health equity: In order to understand the potentially disproportionate impact COVID-19 may have on communities of color, low-income, and other populations that face health disparities, it will be important for states to track both COVID-19 related cases and health outcome data disaggregated for key subpopulations, such as gender, race and ethnicity, geography (e.g., urban vs. rural), and insurance status. Several states are reporting data by race/ethnicity, which is critical as early reports suggest that the crisis is disproportionately impacting communities of color. For states that do report data on race, we recommend including detail about the scope of the missing data (and the reason, if possible) to help users interpret the findings. In the example below, North Carolina’s dashboard reports confirmed cases and deaths broken down by race and ethnicity. North Carolina also clearly states the data’s limitations—i.e., the number of cases for which race and ethnicity data are missing.   

Another breakdown important for monitoring equity is geography. Nearly all states are reporting data at the county level. It may also be helpful for states to present information that compares metrics in urban versus rural areas, as the unique challenges of the virus (e.g., overcrowding in densely populated areas vs. more limited hospital resources in rural areas) differ significantly by these factors. There are several approaches to defining urban versus rural areas and each have advantages and disadvantages, but given that states are already collecting COVID-19 information at the county level, it may be most straightforward to disaggregate information using the Census definition that classifies counties as “completely rural”, “mostly rural”, and “mostly urban”.

In addition to providing data by race and geography, it would be ideal if states could provide additional subpopulation breakdowns such as primary language, socioeconomic status (e.g., education, income, occupation), and disability status, if the data is available. Due to the rapid emergency response required to address the COVID-19 outbreak, we realize states may not initially have the time or bandwidth to produce a broad range of subpopulation analysis or to conduct additional analyses of their demographic data, such as looking at the intersectionality of data (e.g., by race and gender). However, those types of analyses will be increasingly important as states seek to understand disparities in COVID-19 treatment access, morbidity, and mortality.

Date and time-stamps: Because these indicators are subject to change so rapidly, it will be important to date and time-stamp any dashboard updates. In addition to date-stamping the entire dashboard, also consider adding the date (and source information) to any graphic that could potentially be used as a stand-alone item in another report or on social media. For example, the graphic below represents the age distribution of a state’s COVID-19 cases and is labeled “As of 3/31/2020” so that it’s clear what time period this represents, even when the image is viewed separately from the overall dashboard. 

Data labels, definitions, and sources: Provide clear data labels and documentation. Although you should avoid “cluttering” a dashboard with extensive text, it is also important to provide the audience with information about data definitions and sources. Below is an example from North Dakota’s data dashboard showing how they included definitions for each of their six key indicators below their visualization. If space is limited, it is fine to provide hyperlinks to more detailed information on these factors. However, the links should be tested regularly to ensure they are still “live” and taking users to the correct information.  Source: North Dakota COVID-19 Dashboard. Accessed March 30, 2020.

Time-series data: Visually displaying time-series data is an effective way to track changes. In order to improve readability, try to ensure that all time-trended data on the dashboard starts with the same date and covers the same time period, if possible. For example, although deaths and hospitalizations began ramping up at different times, these two time-trended graphs on Ohio’s dashboard start on the same date and cover the same time period. States may also choose to have a dual-axis marking both the date and the week (as shown in the first figure at the top of the page). This helps users understand the broader context of the trends being displayed.               Source: State of Ohio COVD-19 Dashboard. Accessed March 30, 2020.

Visualizations: Choose visualizations that are clean and compliant with a range of browsers. Simple visualizations can also help users interpret more complex data “at a glance.” For example, many dashboards use up or down arrows to indicate whether most recent data show improvements or declines. Make sure visualizations require limited manual data manipulation. For example, the visual to the right was created so that it links to a back-end Excel spreadsheet, which is easily refreshed. 

Documentation to support data updates: After constructing your customized dashboard, create an “instruction sheet” outlining all of the steps necessary to update the data on an ongoing basis, including:

• Which specific cells or inputs need daily updates
• What data sources are being used and where the data is located
• How and where to document what data was pulled and when

This detailed “instruction sheet” is especially important in the event that the individual who normally updates the data is absent or leaves—that way someone else can easily complete the update.

Support for the development of this expert perspective was provided by the Robert Wood Johnson Foundation. The views expressed here do not necessarily reflect the views of the Foundation.

by Unknown

Revised CPS Estimates Show less High Burden Medical Spending

March 2020:

Last year, the U.S. Census Bureau enacted a second and final round of changes in a planned, two-stage redesign process for the agency’s Current Population Survey Annual Social and Economic Supplement (CPS ASEC). The first revision, made in 2014, was a redesign of the CPS ASEC questions regarding health insurance, income, medical expenditure, and poverty data; the second, an implementation of a new data processing system that specifically takes into account the 2014 changes across each category in the questionnaire.[1]

Using the new processing system, the Census Bureau released 2018 estimates on health insurance, income, poverty, and medical expenditures, as well as re-processing collected data for 2017 in a bridge file meant to serve as the transition between the legacy and the new processing systems.

A new brief from SHADAC researchers Brett Fried, MPP, and Lacey Hartman, MPP, takes a look at estimates of medical out-of-pocket spending in both the legacy and new production files, comparing changes in the data between both. In particular, the brief highlights how the number of individuals with high burden medical spending decreased between the two files as a result of the new processing system, rather than from any changes in policy.

The brief also includes a more detailed exploration of the changes made to medical expenditure questions in the CPS, just how the data processing system changed from the legacy system to the new, calculations of the differing estimates of high burden medical spending at the national and state levels, and concludes with a discussion of potential implications for policymakers as well as for researchers in attempting to produce accurate estimates of medical burden.

---

---

[1] SHADAC has produced these resources regarding the two phases of CPS redesign:
Turner, J. (2016). Guide to using the 2014 and 2015 Current Population Survey public use files. Retrieved from https://www.shadac.org/publications/guide-using-2014-and-2015-current-population-survey-public-use-files
Fried, B. (2019, October 31). Understanding the new CPS processing system and new 2018 health insurance coverage estimates. Retrieved from https://www.shadac.org/news/understanding-new-cps-processing-system-and-new-2018-health-insurance-coverage-estimates

by Unknown

Leveraging American Community Survey (ACS) Data to Address Social Determinants of Health and Advance Health Equity

In a new issue brief for the Robert Wood Johnson Foundation’s State Health and Value Strategies, SHADAC Senior Researchers Lacey Hartman, MPP, Elizabeth Lukanen, MPH, and Colin Planalp, MPA explain how researchers and policymakers can leverage federal survey data to inform and target interventions that seek to address social determinants of health and advance health equity. 

About the Issue Brief
As state Medicaid programs increasingly seek to understand and address social factors that contribute to poor health, they need data in order to identify priority areas of unmet social and economic needs, to execute SDOH initiatives, and to monitor and evaluate the impacts of these programs.

While Medicaid programs can choose to utilize data from a broad array of sources, this brief highlights several advantages to using data from one federal survey, the American Community Survey (ACS), to inform and target interventions that seek to address social determinants of health and advance health equity. The ACS contains a vast range of variables relevant to social determinants of health including but not limited to; food insecurity, unstable housing, and lack of access to social supports, and has a large sample size that supports estimates for smaller subpopulations.

This brief provides an overview of select social determinants of health measures and examples of how the data could be used to address questions and evaluate interventions related to housing, transportation, and nutrition. It also provides examples of states that use SDOH and health equity measures from the ACS, including which measures are used and what they are used for.

Access the new “Leveraging American Community Survey (ACS) Data to Address Social Determinants of Health and Advance Health Equity” brief in full to read more about how researchers can use federal-level SDOH survey data to advance health equity. Readers can also download the brief, along with a companion toolkit that includes a detailed data dictionary.

Support for this project was provided by the Robert Wood Johnson Foundation. The views expressed here do not necessarily reflect the views of the foundation.