When WTVT-TV in Tampa, Fla., got a tip that its county's emergency medical system was taking too long to reach people in need, reporters used CAR and vans to illustrate the problem. With CAR, the reporters used spreadsheets of response times, maps of response zones and ambulance run reports to show how long ambulances were taking to reach people who had dialed 911. In unmarked vans, the station raced ambulances to the scenes of incidents they heard on a scanner. Reporters often arrived before paramedics and used a stopwatch to determine how long paramedics took to arrive. The reporters found that Hillsborough County's EMS failed to meet national response-time goals in one out of three cases. In one such case, Rafael Gonzalez died of a heart attack. It took paramedics 11 minutes to get to his home. The house was one mile from an EMS station. In one out of every four calls, it took paramedics nine or more minutes to reach people in the most densely populated parts of the county. In the rural areas, the response times were even longer. In more than one out of every three calls, it took paramedics 11 minutes or more. Chris Ford, senior investigative producer, said this series, which was broadcast in November 1995, marked the first time the station used computer-assisted reporting in all aspects of an investigation. It paid off. Ford said the statistics proved to the county's officials and residents that these poor response times were not isolated cases. "Unquestionably, the computer concept and ideas it exposed allowed us to do this story in a way that was comprehensive and hard to challenge," he said. "The computer really enabled us to make it hard for them to get any wiggle room." After receiving a tip from a paramedic, reporters at the station began investigating. The county's EMS officials did not want to release time records for the whole year, so reporters accepted response-time printouts for a month and started building an Excel database of their own. After much prodding from the station's lawyer, Hillsborough's EMS eventually released response records for the year. Reporters added those figures to their database. Ford said the key was in the field section. "We were able to sort by different variables, and that really made a difference," Ford said. "We were able to show not just that it was a problem, but where it was a problem, specific neighborhoods." One of those neighborhoods proved to be an area where the service was needed the most—a retirement community named Sun City. One paramedic, who did not want to be named, told reporter Kevin Kalwary and photographer Phil Celli that any resident of the retirement community who had a heart attack would be in danger because of the slow response times. "You're basically gonna die," he said. "We probably should give the pre-arrival instructions to hop in your car and drive like hell to the hospital because your chances are gonna be much better." The station was quick to remind viewers that they were not questioning the paramedics' quality of care, only their response times. The station talked to residents who had to wait for the ambulance, and some blamed the response times for the death of a family member. But Ford said reporters did not want to show up at people's doorsteps without an idea of what had happened. They again turned to computers. The station used Autotrak, an online public records database, for background information on victims and their families. They also used the Datatimes clip service to read about the incidents with long response times.
tals are searching for ways to generate funds - and obstetrics, especially neonatal intensive care, is a good way to do it. In Ohio, the average cost of treating a premature baby, in 1994, was about $40,000 and could hit upwards of $400,000.
So many Level II hospitals, in an effort to "get a bigger piece of the baby market," Davis says, were squeezing by state restrictions and presenting themselves to parents as able to provide the most advanced neonatal care.
However, when it comes to the treatment of premature babies, all hospitals are not equal. The Plain Dealer series, which ran on Nov. 27 and 28, 1994, shows that the death rate of premature babies was significantly higher at the Level II hospitals than at nearby Level IIIs. The articles highlights what Davis calls "another example of the business side of hospitals dictating how health issues are handled."
The data
Mazzolini and Davis requested Ohio's birth and death records for 1992 and 1993, although, at the time, only eight months of the 1993 records were available.
Each nine-track magnetic tape cost $100. Getting the records was relatively easy, Davis says. "In most cases, we had the data a few days after we filed our request" with the state health department. And, the data were clean.
The reporters transferred the data from the tapes to an IBM 486, then used FoxPro for their analysis.
Mazzolini and Davis discovered that birth records are a CAR gold mine, with more than 60 fields of information. The records included when and where the baby was born, as well as such details as the gestation age of the baby, the mother's ZIP code, the number of prenatal doctor visits; and, most important, whether the mother was transferred from one hospital to another, either during labor or afterwards. "Basically, we were able to track the flow of babies," Davis says. The birth records allowed Mazzolini and Davis to identify the hospitals that were treating extremely premature babies, Davis says. Then, linking this information to the death records using the hospital code number, they calculated the death rates at various Level II hospitals vs. those certified as Level III.
Finding the names of babies who died wasn't as easy. Birth records do not contain the names or an unique identifier, such as a Social Security number, so they roughly matched the birth records by linking the birth date and hospital code to these same fields in the death records, which do contain names.
"We had good luck; out of 20 names, 15 would be correct matches," Davis says.
They also downloaded data from the Census Bureau site (www.census.gov), which provided the median income level for various ZIP codes. This data did need some cleaning, Davis says. The Census database listed nine numbers in the ZIP code column, so they wrote a program in FoxPro that removed the last four digits.
After the series ran, the state legislature issued a moratorium on hospitals adding neonatal intensive care units-and, in the years since, have closed the loopholes that allowed Level II hospitals to skirt state restrictions. But, Davis says reporters should revisit the issue. "Hospitals routinely thumb their noses at the state health department."
Mazzolini and Davis discovered that many of the babies were from "families who were relatively well off" and the majority of these families had private insurance, Davis says. "Those are the cases that the hospitals wanted."
www.airportnet.org/DEPTS/federal/gao/safety.htm
For a federal court opinion on another tiff between the FAA and the NTSB, check out http://www.ll.georgetown.edu/Fed-Ct/Circuit/dc/opinions/94-1428a.html at Georgetown University.
NASA
The National Aeronautics and Space Administration offers several sites that will help you analyze its Air Safety Reporting System database. Begin at http://olias.arc.nasa.gov/ for the aviation operation's homepage. Check out http://olias.arc.nasa.gov/ASRS/ASRS.html for the database's homepage. The site offers information such as the monthly newsletter Callback, which excerpts some reports on studies that use the data. Also look at ASRS Directline, at http://www-afo.arc.nasa.gov/ASRS/drctlin.html, where analysts highlight key information from the database. You can order the data here, though another shameless NICAR plug - we sell the data to journalists for less.
Human factors
Many air crashes are blamed not on the planes but on the pilots. Two sites that offer information on human errors are the NASA Ames Research Center at http://olias.arc.nasa.gov/zteam/fcp/FCP-home.html and the FAA's http://www.tc.faa.gov/hfl/hfl.html
A group of psychologists funded by NASA published their paper - including some specific accident information - at http://ww.psy.utexas.edu/psy/helmreich/acdntlst.htm
International sources
The aviation safety Web pages at http://Web.inter.nl.net/users/H.Ranter/index.html offer a variety of information on international and U.S. air crashes. It's an intriguing Web site including things like a movie of the Ethiopian 767 crash, pages highlighting recent disasters and a listserv that promises to tell about accidents within a day or two. Much of the information is pulled from news accounts and others' data. This is a good attempt at gathering information on international air disasters.
The United Kingdom Air Accidents Investigation Branch at http://www.open.gov.uk/aaib/aaibhome.htm reports on accidents and makes safety recommendations.
The Commercial Aviation Resource Center at http://w3.one.net/~flypba/AIRLINES/carc.html offers a variety of links to airlines and airports around the world, identifying codes that may be useful for reporting certain stories, and symbols that indicate whether the sites are official or not.
Boeing offers summary information on the world commercial jet airplane fleet at http://www.bts.gov/smart/cat/461.html
Go to http://WWW.CAM.ORG/~icao/sites.htm and http://www.airparts.com/ for a vast list of airports, airlines and more.
Andrew Lehren can be reached at (573) 882-0684, or send e-mail to andy@nicar.org