Data

You are currently browsing the archive for the Data category.

The Department of Housing and Urban Development (HUD) is responsible for answering just that question.  To determine what level Section 8 vouchers should be set, HUD measures the rents for every county across the nation.  Specifically, they measure the 40th percentile and 50th percentile (i.e., median) rents in each area.  They choose to use the median so that high prices for luxury residences do not skew the measure of rent for a “typical” person in each area.  How does HUD calculate these Fair Market Rents (FMR)?  Today I will explain.

Read the rest of this entry »

Tags: , ,

Many researchers use household data sources to examine a variety of hypothesis.  The use of household data has many benefits including allowing for more detailed socioeconomic information (e.g., education, income) beyond what is contained in administrative claims files.  One drawback of household data is that extrapolations made from household survey data may not match national estimates.

For instance, this article examines how to align the Medical Expenditure Panel Survey (MEPS) to aggregate U.S. benchmarks provided in the National Health Expenditure Accounts (NHEA).  Today, I review some of these adjustments.

Read the rest of this entry »

Tags: , , , ,

Many health insurers (public and private) reimburse doctors based on the patient’s diagnosis. If you treat a patient for a more severe illness during a inpatient stay, Medicare pays you more money. Physicians use procedures to bill insurers for the care they provided.

How do insurers know the patient’s diagnosis and the procedures providers perform? The answer is the International Classification of Disease (ICD) taxonomy. Currently, this system is in its ninth iteration, but it will soon be replaced by ICD-10 (the tenth revision) codes. By January 1, 2012, CMS will mandate that all electronic health record transaction use the ICD-10 system and by October 1, 2013 providers will all have to use the ICD-10 diagnosis and procedure codes for their claim submissions.

What’s new about the ICD-10 compared to the ICD-9? Read more below to find out.

Read the rest of this entry »

Tags: , , , , , , , , ,

Health economists have always said that one of the main problems with health care is that no one knows the price of any services.  Thus, individuals have less of an incentive to shop for high value care. [Other experts claim that prices don't matter as much because demand for healthcare is inelastic.]

One step towards price transparency is the site Healthcare Blue Book.  Consumer Reports has recommended using Healthcare Blue Book to estimate cost for a variety of diseases such as ADHD.  The magazine has even recommended using Healthcare Blue Book to negotiate rates when a patient does not have insurance coverage.

When I asked Healthcare Blue Book representatives about the source of their data, this was their response: “The data is the average price that an insurance pays its provider for that service in your marketplace adjusted for additional information, market knowledge and a few other factors…The data comes from data warehouses which are storage centers that have claims and billing information.

Healthcare Blue Book provides their information for free.  How do they plan to make money?  Again, from Healthcare Blue Book’s media representative: “Our business growth is in creation of customized Blue Book databases for companies and other organizations. The Blue Book team will look across an organization’s health plans and what different providers are charging for the same service. The variations in price within the same health plan, for the same service, can be enormous because all the rates are negotiated with individual providers.

Sunlight may be the best disinfectant; transparent prices may be a cure for what ails the healthcare industry.

Tags: ,

The Agency for Healthcare Research and Quality’s (AHRQ)  Healthcare Cost and Utilization Project (HCUP) is a family of databases and tools intended to improve the quality, safety, efficiency, and effectiveness of the U.S.   health care system.  HCUP results from Federal-State-Industry partnership to build a comprehensive all payer data system.  A summary of the databases available from HCUP can be found here.  A summary is also provided in the table below.

Abbreviation

File Description States Participating

Available From:

SID State Inpatient Databases Data on 95 percent of community hospital discharges 44 1990
SASD State Ambulatory Surgery Databases Ambulatory surgical center data (hospital and free-standing) 29 1997
SEDD State Emergency Department Databases ED visits that do not result in hospitalization 29 1999
NIS Nationwide Inpatient Sample All-payer inpatient care database. A 20% stratified sample of U.S. community hospitals. varies by year 1988
NEDS National Emergency Department Sample All-payer inpatient ED database. A 20% stratified sample of U.S. community hospitals. 2006
KID Kids’ Inpatient Database All-payer inpatient database for children.  Contains 3 million discharges form 3,500 community hospitals. 1997

 

Tags: , , , , , , ,

One of the challenges of studying import issues in health economics is that the patients often come in and out of your data file.  For instance, beneficiaries often switch their private health insurance plan, or can become eligible for public insurance (e.g., Medicare or Medicaid) or the individual may lose their insurace coverage.  In each of these cases, it is difficult to track these patients over time.

One effort to solve this problem is AHRQ’s All-Payer Claims Databases (APCD).  AHRQ does not create the APCDs themselves but facilitates States efforts to create them.  According to a 2010 fact sheet:

Payers include insurance carriers, third party administrators (TPAs), pharmacy benefit managers (PBMs), dental benefit administrators,state Medicaid agencies, CMS (Medicare), Federal EmployeesHealth Benefit (FEHB) and TRICARE administrators.APCD systems collect data from existing claims transaction systemsused by health care providers (facility and practitioners) andpayers.

The information typically collected in an APCD includes patient demographics, provider demographics, clinical, financial,and utilization data. Because of the difficulties involved with thecollection of certain information, most states implementing APCD systems have typically excluded a number of data sources, such asdenied claims, workers compensation claims, and, because claimsdo not exist, services provided to the uninsured.

One concern with these data is that the APCD must maintain beneficiary confidentiality. On the one hand, one of the key benefits of APCD is that it allows researchers to examine patterns in the cost and quality of care for beneficiaries who change insurers.  The institution administering the APCD, however, must institute a strong institutional review board (IRB) or data use agreement (DUA) policy.

Also, the APCD must determine whether or not the claims files are updated.  Oftentimes, claims payment amounts can change or the claims can be later denied.  The APCD must determine a policy for updating the data files over time.

Although there are a number of challenges, some states are making progress.  ”Oregon and Tennessee willhave live systems in 2010. Hawaii and Colorado havecurrently submitted legislation for their 2010 session toauthorize development.”

Hopefully, the APCD will be a high-quality, useful data resource that can be used to answer a variety of research questions.

Tags: , ,

The year 2010 marks the last for the decennial Census.  Although you might miss the ad campaigns every 10 years, the Census hasn’t really disappeared completely.  Instead, it’s being replaced by the American Community Survey (ACS).  Although the goals of the ACS are similar to those of the Census, its approach is very different.  Rather than survey all individuals each year, the ACS surveys about 250,000 addresses per month (i.e., 3 million addresses/year).  This amounts to a survey of about 2.5% of households each year.

The Missouri Census Data Center notes some of the pros and cons of the change.

Pros

  • The ACS provides a more current picture of the country.  Currently, the latest Census data is available for the year 2000 while in December the ACS will release figures for 2009.
  • New questions can be added to the survey without having to wait for the decade to change
  • Researchers can calculate statistics at the national, state, MSA, large city,  counties, and even PUMAs every year.

Cons

  • The sampling error associated with the decennial census long form data is much lower (in general) than that of the ACS.
  • Data for smaller geographic areas (especially those under 20,000 as well as for all ZIP codes, even those with populations over 20,000), are only released as 5-year period estimates.
  • Sampling methodology relies on the accuracy of Census population estimates.
  • Since using data for small area requires the use of a sample taken over a 5-year period, it will be impossible to use the ACS data to pinpoint areas that may be undergoing significant changes over the period.

Overall, the key benefit of the ACS is that it provides more timely responses while the drawback is that the pool of household sampled in a given year is much less than is the case for the decennial Census.

I have also made this spreadsheet to compare the ACS to the decennial Census.

Tags: ,

How is the health care labor market in your area?  HWS Enterprises put together a gauge of the healthcare workforce labor demand throughout 30 large metropolitan regions in the United States.  The results for Q4 are available here.  The strongest healthcare labor markets are Sacramento, Riverside, Pittsburgh, Cleveland and Dallas.  The weakest is New York City.  My home, San Francisco, ranked 18th out the 30 large metro areas evaluated.

According to the Health Workforce Solutions’ press release:

The HWS Labor Market Pulse® Index (LMPI) provides a quarterly barometer of local market health care workforce expansion and contraction. Patterned loosely after the Case-Shiller home index and based on a proprietary algorithm, the LMPI identifies and enables comparison of 30 health care labor markets by tracking elements including temporary health workforce shortages and surpluses, facility and bed closures, announced layoffs and expansions, and local economic trends. The LMPI will be published quarterly as part of Labor Market Pulse® and distributed nationally to health care executives, the media and other interested parties.

Tags: , , ,

Survey estimates of Medicaid enrollment are 43 percent lower than raw Medicaid program enrollment counts.  Why is this the case?  Roebuck and Liberman (HSR 2009) find that many people are not reporting that they have Medicaid coverage.  “43 percent of Medicaid enrollees answering the CPS as though they were not enrolled and 17 percent reported being uninsured.”

One reason for the underreporting could be that the poor may only enroll after they get sick.  Further, if they do not pay for Medicaid, they may feel that they are just receiving government assistance rather than “insurance.”

Tags: , ,

Google has created easy to use charts on U.S. population and employment from public-use data from the U.S. Census Bureau and the Bureau of Labor Statistics.  Instructions of how to use this data can be found here (video).  Here are some examples I created:

Tags: , , , ,

« Older entries