Search for Relevant Data

[jeffjmei]

Let's try share useful datasets and highlight their useful columns. Our biggest challenge is finding a reliable way to link patient outcomes to hospital ownership.

I've found that this hospital dictionary is really useful for identifying CMS datasets that can be linked through useful identifying variables (e.g. Facility Name, Facility ID).

Here are some potentially interesting datasets:

• TIMELY_AND_EFFECTIVE_CARE-HOSPITAL.CSV
• MEDICARE_HOSPITAL_SPENDING_BY_CLAIM.CSV
• HCAHPS-HOSPITAL.CSV
• HVBP_SAFETY.CSV
• HVBP_EFFICIENCY_AND_COST_REDUCTIONS.CSV
• HVBP_PERSON_AND_COMMUNITY_ENGAGEMENT.CSV
• HVBP_CLINICAL_OUTCOMES.CSV
• FY_2024_HOSPITAL_READMISSIONS_REDUCTION_PROGRAM_HOSPITAL.CSV
• VA_TE.CSV
• VA_IPF.CSV
• MATERNAL_HEALTH-HOSPITAL.CSV

[jeffjmei]

Hospital All Owners

The Hospital All Owners Information dataset provides information on all owners of the hospitals. This data includes ownership information such as ownership name, ownership type, ownership address and ownership effective date.

Useful Variables:

• ORGANIZATION.NAME
• TYPE...OWNER
  • I: independent
  • O: organization
• CORPORATION...OWNER
  • Y: if owner's organization is a corporation
  • N: if owner's organization is not a corporation
• INVESTMENT.FIRM...OWNER
  • Y: if owner's organization is a investment firm
  • N: if owner's organization is not a investment firm
• FOR.PROFIT...OWNER
  • Y: if owner's organization is a for-profit organization
  • N: if owner's organization is not a for-profit organization
• NON.PROFIT...OWNER
  • Y: if owner's organization is a nonprofit
  • N: if owner's organization is not a nonprofit

Thoughts:

• this dataset is at the heart of our project thesis
• it's not clear to me which of these ownership types (e.g. investment firm, corporation, etc) makes most sense for our project

[jeffjmei]

Patient-Reported Outcomes

Patient-reported outcomes are reports from patients about aspects of care that matter most to them, like pain management, functional ability (like their ability to walk, think, see, hear and remember), and overall quality of life. Through surveys or questionnaires, patients self-report the effectiveness of the care they got from their provider.

• Facility ID (can be used to link to other datasets)
• Facility Name (can be used to link to other datasets)
• Measure Name: type of survey
• Score: patient satisfaction

Thoughts:

• this dataset has the potential to link to the hospital owners
• this is from patient self-reported surveys, so there is immediately some bias

[jeffjmei]

Hospital General Information

These are the official datasets used on Medicare.gov provided by the Centers for Medicare & Medicaid Services. These datasets allow you to compare the quality of care provided in Medicare-certified hospitals, Veterans Administration (VA) medical centers, and Department of Defense (DoD) hospitals nationwide.

• Facility Name
• Hospital Type
• Hospital Ownership
• Hospital Overall Rating
• Count of Safety Measures Better
• Count of Safety Measures Worse

Thoughts:

• this can be linked by facility name, which makes this a really useful dataset
• also has some information on hospital quality

[jeffjmei]

Hospital Change of Ownership

The Hospital Change of Ownership (CHOW) dataset provides information on the hospital ownership changes that occurred on or after January 1, 2016. This data includes information on the buyer and seller organization’s legal business name, provider type, change of ownership type (CHOW, Acquisition/Merger, or Consolidation) and the effective date of the change.

• CCN - BUYER
• CCN - SELLER
• ENROLLMENT ID - BUYER
• ENROLLMENT ID - SELLER

Thoughts:

• I found out that Facility ID in many of the CMS databases is the same as CCN (source: hospital_data_dictionary from the first comment in this thread). Therefore, this can help make important linkage:
• hospital_info: contains Facility ID (same as CCN) and gives access to hospital quality information
• owner_change: contains both CCN and Enrollment ID
• owners: contains Enrollment ID and gives information about the owner

[john-j-park]

Patient Survey

A list of hospital ratings for the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS). HCAHPS is a national, standardized survey of hospital patients about their experiences during a recent inpatient hospital stay.

Thoughts

• The data has about 450k rows, and looks to be broken out by individual survey item? It will be annoying to turn this into something we can work with.
• It contains the Facility ID so we can at least link it back to ownership data.

[john-j-park]

The Private Equity Stakeholder Project has a table called "PE Hospital Tracker" that seems to contain much of the ownership data we want.

I think this also provides some opportunities for other projects if we don't find too many interesting relationships.

• Make a Shiny map?
• Some basic data analysis on this dataset (where are the PE-owned hospitals? what are the demographics of these states? etc.)

[jeffjmei]

This is a great find! Taking a glance through the data (PE Hospital Tracker), it also uses the Facility ID (CCN) code, so we should be able to easily link it with the CMS database.

Compared with our own linked data, this is slightly more comprehensive (there are roughly a thousand more rows), and the data is probably cleaner than ours. Overall, it's probably better to use this dataset than the one we've linked in the data_exploration.Rmd file.

[jeffjmei]

We talked about some datasets for evaluating hospital performance. It turns out I can't push them to GitHub because the files are too large. I'm not sure what other way to do it, but here are the links to some more datasets: Timely and Effective Care Complications and Deaths Healthcare Associated Infections