Data Mining
The VA lends itself to both national stories and strong local pieces. While the VA headquarters is a couple blocks from the White House, the VA’s work really happens far outside Washington – at the collection of regional offices (which handle disability compensation) and hospitals and clinics (which handle health care).
The structure leads to plenty of rankings and other comparisons, by state, hospital, or region. The efficiency in the regional offices is particularly scattershot; wait times and errors by regional office, for example, vary widely. At one time, the waiting time to have an appeal decided in some regions was 100 days and in others, 1,400 days. The VA says it is correcting these problems. But the VA has been saying that for 15 years.
Here’s what you need to know about mining data at the VA, and how to turn the numbers into stories for your readers or viewers.
The VA is really two different agencies:
- The Veterans Health Administration oversees the health system
- The Veterans Benefits Administration manages the disability compensation system.
Each spends more than $50 billion a year. But while they report to the same secretary, they have different leaders, different cultures, different reputations, and – most important for reporters – different FOIA offices and data policies.
The Veterans Benefits Administration (the VBA) has at times had a dysfunctional, secretive FOIA office resistant to releasing information, and indeed hasn’t gotten many requests to do so over the years.
The Veterans Health Administration, by contrast, has more of a culture that respects data, regularly works with medical researchers and economists to quantify the great amount of information it maintains, and until recently had one of the best FOIA officers I have ever experienced.
The umbrella organization overseeing both is the Department of Veteran’s Affairs. (PDF of the department’s organizational structure; a PDF of the VA’s organizational “briefing book.”).
Expect significant FOIA battles.
In dealing with the VBA, we faced FOIA requests that were not logged in, that were routed from office to office, or that were simply ignored. Perhaps that’s par for the course among federal agencies. But you must plan accordingly.
- Spend time before making a request to determine the office or person who actually maintains the information in question and give as detailed a roadmap as possible in your FOIA letter to pinpoint where your information resides.
- Keep detailed logs of when you make a FOIA request and follow up regularly – every few weeks or so — for status updates. You can expend mental energy griping about how long things are taking, but on some level you’re at their mercy, so it’s usually best to file your FOIA and then switch your focus to another story.
The VA is a vast data warehouse.
At its core, the VA exists to give veterans checks or, in the health system, medical care. So it maintains vast collections of data. And it does a huge amount of sorting and ranking and analyzing data itself.
Among the most important databases we have used:
- The Comp and Pen Master File is the record of each veteran disability claim, detailing region, disability and the “rating” (a severity indication, from 0 to 100); there are dozens of other fields. This was the most pivotal, and also the hardest to get, as the VA was extremely hesitant to give it up. It’s more than 3 million records, and growing rapidly. (Search a McClatchy database of that data, by zip code or state).
- Monday Morning Report, which details the current backlog of cases and is updated each week. It’s a pivotal source to look for workload trends.
- Annual Benefits Report, which is put out each year by the VBA. It details a wealth of information about disabilities and other benefits, broken down by state. Our best stories took these easily getable indicators and, after obtaining the raw data underlying them, refined them in far more meaningful ways. Just spend time poking around the VA’s website – you’ll find plenty of interesting things.
- The health system has several research institutions around the country affiliated with VA hospitals and local med schools. One main site includes a link to the various centers and their research. One center deals with the cost of VA health care, and has a lot of research on what treatments and hospitals are cost effective (the Health Economics Resource Center).
- VACOLS is the appeals database and is handled by the quasi-independent Board of Veterans’ Appeals (part of VA but with its own authority). This database showed a case’s start date, end date, regional office, and resolution, among dozens of other fields. Most cases go through multiple appeals – they are appealed, sent back to the regional office, decided again, get appealed again, get sent back… Thousands of cases linger for between 5 and 10 years, during which time many of the elderly veterans die.
- The Pending Issue File showed all claims now in the works, as well as the regional office and type of claim. It was pivotal for showing how long claims were taking, particularly since delays vary widely by region and by type of claim.
- National Survey of Veterans – Conducted every five or six years, this is a wide-ranging survey that gets into veterans’ disabilities, incomes, etc. The most recent survey questioned 11,000 people.”. If you are having problems, the VA’s survey department has sometimes been helpful.
- Medical patient files – Those include the MedSAS and DSS files, which track patient visits (or “clinical stops”) and expenditures. These are massive files (tens or hundreds of million records per fiscal year), and are generally used only by medical researchers. But they can allow for in-depth breakdowns of the care given by location.
- Vital Signs database – This database details hospital-by-hospital statistics: average wait time for mental health patients, average wait time for orthopedic patients, percentage of patients who had certain tests conducted, etc. It’s very detailed, but there are limitations on what sorts of comparisons can be made. Also, the VA’s inspector general has reported in the past that waiting time data is manipulated and unreliable
My most recent veterans’ story used many of these resources, often for the third or fourth time since I began covering veterans’ issues. My analysis of the compensation database showed a wide spread from region to region in the awarding of PTSD and other disability claims. The data also pinpointed hotspots across the country where PTSD and other disabilities were a particular problem. Among the veterans returning from Iraq and Afghanistan, one hot zone was the tiny town of London, Ky.
There, I talked with the families of local soldiers who had died in Iraq or Afghanistan, and with the veterans who survived but would live with mental or physical reminders for the rest of their lives. My story was their stories. The numbers and the data provided a backbone to the story, but it was the humans who told the tale.
→ Next section: RESOURCES
