2. Select a primary data source:

The records kept by your EMS program may be sufficient, but often they are missing some important data. If you cannot clearly establish what happened or exactly where, you may be able to get that information from other local responders. Local people may also be able to help you by remembering details of events and they can refer you to other people with information. Take notes.

Useful primary data should contain the following information:

• Date
• Time
• Exact location
• Type of event
• Dispatch information
• Patient identifiers
• Mechanism of injury
• Observable injuries
• Treatment rendered
• Patient’s response to treatment
• Date of death
• Probable cause of death

3. Look at the records of fatalities and sort them by cause:

Sort fatality records into piles based on the primary cause of injury (motor vehicle crash, violence, suicide, etc.) As you sort, pay special attention to any identical matches within each pile. Once you have the records broken into primary cause of injury, review each pile for similarities.

• Make copies of each fatality report to use in your investigation. These are your working database
• Tag all original fatality reports for future reference
• File the originals for safekeeping
• Make piles of charts for crashes that occur in the same location. (Suspected cluster sites) Watch closely for cluster sites. If you find any, make a separate file just for them
• When you find obvious matches, mark them for future reference and keep them together
• Select the piles that have obvious similarities (particularly if they occur at the same site) for further investigation
• By the time you have gone through your primary fatality data two or three times, you should have a very good idea of where fatal events are occurring and some ideas on why they happen there. Remember, cluster sites are where you have the best chance of making a significant improvement, so always watch for them.

4. Identify a second matching data source:

Matching data sources add power and accuracy to your research. Every research project benefits from corroborating evidence. There are many possible additional sources of data that can be matched to your reports that will enhance your understanding of the events and add strength to your contention that something must be done to correct the hazards you find.

A good secondary data source will be accessible, relative to your district and have additional information specific to the injury problems you have identified in your primary data source. It will provide you with all the necessary information that your primary reports do not contain. Compare the different sources to find the reports that provide the most relevant information that is missing from your primary database. When you have selected a second source, obtain matching reports for the particular site or injury event that you are interested in. Clip the matching reports together and keep them together to avoid double counting. One good matching data set is usually sufficient, but when relevant and easily obtained, more is better.

Contact your local Department of Transportation (DOT) and the Highway Patrol. The DOT keeps records of every collision that happens on the State roadways. They will usually provide you with a big computer print out of the specific stretch of road you want to look at. The printout has lots of information that is useful to establish exactly what happened. Unfortunately the DOT may purge their records after five years or so, so the sooner you get the reports, the better!

The Highway Patrol will probably provide you with copies of their own investigations for a nominal copying fee. In California it costs $5.00 per report and you must provide them with precise information that identifies the reports you want. This is money well spent, because Highway Patrol reports will contain a thorough investigation of direction of travel, skid marks, speed, percipient cause, what was hit, tire wear, alcohol use, etc. In addition, the report will probably have a sketch of the scene with skid marks, position of the vehicle, exact location on the highway etc. It saves a lot of time to have a professional reenactment of the crashes in your hand when you go to the site.

The County coroner or medical examiner will have records of every death occurring within the county by date and name of deceased. You can usually review their records and make notes or copies that establish time and cause of death and next of kin, (in case you wish to interview them as witnesses or in order to gain some support for your efforts). Going through death certificates, you may find a few patients who were alive when they arrived at the hospital and died later. In these cases, it is generally accepted that if they expire within 30 days of the injury event, the event itself was the principal cause of the death.

The local library also usually has large microfiche files of local newspapers that provide additional information along with analysis. The advantage to using the newspaper clippings is that they often can be traced back farther than other data sources. Once you know where something is happening, simply scan backwards in time from your first record and find other similar events at the same site. Some microfiche viewers also have a copy function so you can print out any items of interest.

• Highway Patrol and local police reports
• Local emergency services agencies
• State Department of Transportation motor vehicle crash statistics reports
• Coroner and Medical Examiner
• Dispatch logs
• Local newspapers, especially if they accurately quote Highway Patrol and police reports.
• Emergency room and hospital records.
• The Indian Health Service’s Injury Prevention-Severe Injury Surveillance System database
• State Office of Vital Statistics
• National Highway Transportation Safety Administration (NHTSA)/Fatal Analysis Reporting System (FARS)

Match documents from your secondary data source(s) with your EMS charts. Review all these documents. Select the ones that have the most information that is relevant to your study. Look for similarities and discrepancies. Look for patterns between charts. If you can’t find a pattern that repeats in a specific location, try making piles of records based on the cause of the injury (rollover, head-on, etc.). Attach each primary report to its matching secondary report. Keep the records together.

A secondary data source for a fatal MVC investigation must include information on environmental, vehicular, and human factors that caused the crash or aggravated the injuries such as:

• Exact location of the event
• Roadway description including grade, curvature etc.
• Road surface at the time of the crash: Wet, icy, dry, etc.
• Weather
• Visibility
• Direction of travel
• Type of crash: Angle, frontal, side, rollover, fixed object, immersion, etc.
• Ejection
• Speed of travel
• Alcohol present
• Safety equipment used
• Events leading to the crash
• Cause of fatal injuries
• Drawings of the crash and the scene

5. Compare both data sets and identify cluster sites:

Compare the piles of fatal reports that are related to specific locations. Which pile has the most reports? Obviously the biggest pile represents an area with a lot of fatal incidents. Start with the biggest piles and begin the investigative work. Look for patterns that repeat from incident to incident at each site.

6. Narrow down the search and define the exact nature of the problem:

Most of the rest of the job so far has been to prepare you for the real investigation, and this is where it begins. This is the really interesting and productive part. Detective work is a process of gathering and comparing evidence and looking for patterns. Injury Epidemiology works exactly the same way. Below are some of the more common clues to watch for as you begin your investigation.

Are fatal incidents related to particular location? Where?

• Intersection
• Unlit residential area
• Curve ("Deadman’s Curve")
• Milepost marker
• Prominent land feature
• Particular stretch of highway ("Blood alley")
• Community or neighborhood
• Embankment, cliff, ditch or body of water
• Public swimming hole or diving area
• Tree, pole, fixed object
• Near a bar ("The Knife and Gun Club")

Does a particular time of day or night have an unusually high death rate? When? Why then?

• After dark
• Noon
• Rush hour
• Bar break

If you have milepost markers as location identifiers, tour your district and note the milepost marker number on your map along with an identifiable feature on the scene. A list or map of milepost markers and landmarks (creeks, bridges, turnouts, passing lanes, etc.) may be available from your roads department or the State Department of Transportation. Use prominent land features and milepost markers on your map to find the exact location of each fatal motor vehicle collision. Place a pin for each event. It’s a good idea to mount your pin map solidly on a piece of corkboard, and to put a bit of glue on each pin before you push it in. If the pins start falling out, you’ll have to go back and do it all over again!

To help keep track of incidents, you can use flags on the pins to indicate the type of fatality, use different colored pins for the different types of incident, or a identifier numbers and a table that refer to the chart. Remember to look at all traumatic fatalities. You may be looking for automobile collisions and find a cluster of drownings at a popular swimming hole, an auto/pedestrian cluster or perhaps an area over-represented for homicides.

Pin maps work for more than just trauma. In "Merry Olde England," before anyone knew about germs, sanitation, or the spread of illness, there was a severe outbreak of cholera. People were beginning to panic. A physician credited with being the first epidemiologist, made a map precisely locating each ill person. The map showed a pattern. There was only one factor common to every person with the disease; every affected person was using the same water supply. Rather than trying to warn the residents about the dangers of that particular water supply, trying to convince everyone that they should boil their water before drinking it, or pushing through a law against drinking un-boiled water, he merely took the handles from each of the pumps connected to the contaminated water source. The epidemic stopped. Note that this was an environmental solution. He literally attacked the problem at its source. That is the power of having accurate data and one person who’s not afraid to do something with it.

I once did a quick pin map for fatal cancer incidence on the Hoopa Reservation. As a preliminary investigation, I interviewed a group of tribal elders and senior caregivers from the local clinic. From them I obtained a list of names of people they knew who had died from cancer. Then I asked where the decedents had lived most of their lives and made a map. No surprise, there were clusters of patients who had died from cancer in particular areas of the reservation which was itself a hotspot for cancer within Humboldt County which is a hotspot in the state of California. Hoopa not only had an epidemic of Over-the-Bank motor vehicle crash fatalities, it also had an epidemic of cancer.

There were three places on the reservation that appeared on the pin map as obvious cluster sites. The first were in the vicinity of old lumber mills that made pressure treated lumber with pentachlorothiazine that had standing areas of black water during rainstorms. There were also cancer clusters near the old "Copper Mine," one of the ten worst "superfund" cleanup sites in the nation at that time. We also found cancer clusters in particular HUD housing projects with no apparent connection to any contaminant. The HUD locations puzzled me until I was informed by a knowledgeable local that when those homes were built, fill rock, used to level the foundations and the yards, was obtained from the copper mine’s tailings! These tailings were found to be highly contaminated with heavy metals from open pond leach fields. It is a good demonstration that when you find an area of increased incidence, there is usually a reason for it, whether you are investigating injuries or illness. To my knowledge these secondary sites of contamination have never been investigated further.

Conduct a site visit. Once you have clearly established the exact location of your fatal cluster, take a video camera, a still camera or two, plenty of slide-film and videotape, spare batteries and some material for note taking. Visit the site in uniform and in your rescue unit if possible.

Have another EMT drive. If you visit the site in the rescue unit, you have a driver, another EMT to comment on tape on the aspects you may not be seeing, their stories of what they saw when they responded to calls at that location, and you have the added advantage of being visible in the community. This creates significant leverage on the responsible party without being overtly threatening. Gentle persuasion to fix a problem that they already are aware of can come in the form of visually letting them know that you are also aware of the problem and that you have publicly documented it.

The grapevine in a small community can work for you. ("Did you see the rescue squad out at dead mans curve yesterday?" "It looked like they were filming skid marks!") A few people talking about this in the community will go a long way toward bringing the issue to the attention of the responsible party without any overt threats. ("Fix it or else.") Sometimes this is all that is needed to move a known hazard off the back burner and closer to a resolution. People in the community will be relieved that someone is finally doing something about "Deadman’s curve. It is a good first step towards short cutting the hazard resolution process, and excellent public relations as well.

Adjust the camera’s field of view and try a wide-angle approach, then narrow it down and try it looking at particular aspects of the site as you approach. The camera sometimes can capture what the eye cannot.

• What is different about this site from the roadway one-half mile away where there are no fatal crashes?
• What jumps out at you when you first approach the scene slowly? At highway speeds?
• How does the roadway look?
• Smooth or in poor repair?

When you feel you have captured the approach aspect of this site completely, set the camera up on a tripod alongside the road in a safe, visible location filming the approach of vehicles and capturing the impact site as well. You don’t want vehicles to take undue notice of you and thus react differently, nor do you want them to pay attention to you and not their driving. Be discreet, and be safe! Once you set the camera up and it has the field of view that you want in clear focus, let it film by itself and watch the traffic approaching. Avoid the temptation to touch the camera, just let the camera capture what is happening.

• How do the vehicles move into and through the site?
• Is there a substantial weight shift from rear to front indicating hard braking?
• Do some of the vehicles show an off balance or swaying appearance?
• Is there some shift in the attitude of the vehicle that would tend to cause a loss of control if it was exaggerated?
• Does the shift in attitude appear to throw the vehicle toward the hazard?
• How close are their tires to the right side of the road, and to the centerline?
• Are you hearing any tire howl?

Next, film vehicles as they pass the hazard and proceed down the road away from the site. If you are filming on a curve, try to find the camera angle that best shows the attitude of the vehicle at the apex of the curve. You want to capture the hazard in the shot as well. Again, let the camera do the work, the less you fiddle with the machine, the better. Be patient.

Take some still pictures, focusing on any features you want to emphasize. (I once photographed a mile post marker that had been hit multiple times, complete with multiple skid marks going straight into it and smashing the mile post marker into a log barrier. It was a very convenient way to document exactly where and how the crashes were occurring.) Consider using slide film, slides display a much bigger and clearer picture when you are presenting your findings. Take stills of any skid marks as well as videotape, because they are some of the most telling physical features at any crash site. Try to capture skid marks by shooting repeatedly from different angles and in different lighting. Take the tape back home and review it. Did you catch what you wanted? If not, it’s worth doing again.

Graphic photographs and videotape strengthens your case dramatically. If you can take a disposable camera along on the ambulance units, it’s often useful for snapping shots of the scene as you enter. Footage with the wreckage on scene is some of the most valuable evidence you can collect. Videotape and photograph extensively after any crash incident. The period of time immediately following a serious crash is a "teachable moment" when the information that you are trying to impart is graphically represented by a real incident. Teachable moments result in an increase in the likelihood that the information you are trying to convey will result in an improvement or change in behavior. Take advantage of every teachable moment, they are extremely effective motivators. Remember to be sensitive to the feelings of survivors and be selective about what and how you record, what you show to the public, and how it’s presented.

9. Focus on environmental hazards at the site. Try to find environmental strategies.

Effective environmental modifications are "equal opportunity solutions" to fatal events. The most effective motor vehicle collision injury prevention solutions are those that don’t require any behavioral change and that protect everyone equally. Environmental solutions are often the best and simplest option, so look for them first. At cluster sites, there are hazards that threaten everyone.

Don’t be side tracked into blaming the injured people or thinking, "It’s just another damned drunk driver." Think instead, "how can that drunk driver (and everyone else) be best protected from another fatal crash at this site?" Drunk, sober, poor, rich, Native American, non-Native, intelligent or foolish, an effective environmental solution protects everyone equally. Effective environmental modifications work around the clock in any weather and under all conditions. Once an effective protection is in place, the fatality problem ceases to exist.

• Install a barrier, guardrail or energy-absorbing device to isolate traffic from the hazard
• Remove fixed objects within 50 feet of a highway
• Add breakaway supports for light standards and signs within 50 feet of the roadway
• Modify road surface to improve traction and smoothness, improve striping
• Improve signage, install lighting
• Install traffic signal or other flow controls in high traffic areas
• Reduce speed limits, add speed bumps or "roundabouts" in areas with high traffic and pedestrian flow
• Build pedestrian walkways and crossings physically separated from the traffic
• Add rumble strips or "Bott’s dots" along highway shoulder where run off road crashes and rollovers occur
• Light residential areas where auto vs. pedestrian crashes occur at night
• Separate opposing traffic and widen the road in areas with high frontal impact incidence
• Correct "off camber" curves where rollovers are common
• Install pedestrian and bicycle pathways remote from the traffic lanes
• Repair missing or worn road edge striping, center lines and no-passing stripes
• Install turning lanes and widen shoulders in areas with a high incidence of auto vs. auto crashes
• Install improved smoke detectors with 10 year lithium batteries and annoyance reset buttons in residences with cigarette smokers and those that use wood heat, especially if there are elderly or very young children in residence
• Check the length of the yellow phase of stoplights at any intersection that is over represented for broadside collisions. Lengthen it if it is shorter than those at non-crash locations
• Install pool fencing that is not attached to the home and has an automatically locking gate to prevent toddlers from drowning
• Install emergency call boxes in remote highway areas to reduce EMS notification time
• Inspect popular swimming/diving holes and remove any submerged hazards, divert hazardous currents
• Build comfortable lifeguard-lookout areas at public beaches
• Install hand railing in showers and tubs in the homes of the elderly to help prevent serious fall injuries

10. Document your findings

Make a report linking your review of the data with the information you gained from the site visit. Take time to document each cluster site you’ve identified during your research. Include a summary of your data, make a short video or photographic presentation and include the most promising strategies for solving the problem. An accurate, one-page executive summary with facts and logical conclusions including an environmental strategy is a very effective tool.

Accurately documenting your findings forces you to clearly state exactly what you have found, what conclusions you have reached and what you propose to do about it. This helps you focus your thinking and may point out issues that you have overlooked. Having this information in a concise form will help you during negotiations with the agencies that are responsible for fixing the problem.

Your report is only as valid as the data you collect. Your conclusions must be supported by accurate, specific data. Make sure to include only data that is specific to the particular problems you want to correct. Don’t over reach or draw wide-ranging conclusions about other similar situations. Don’t speculate. If you don’t have sufficient data on a problem, get appropriate supporting data before you put it in your report.

Don’t add confusion. A fatal vehicle crash near but not at the cluster site you’re researching is not important to the study of your specific site. Unrelated information will confuse people and weaken your report. Save any data that isn’t specific to the problem for a future investigation.

Use anecdotal examples carefully or not at all. An anecdote is a story told by someone that is not substantiated by data. It is defined as, "an illustrative story or fable." Using anecdotes can backfire on you. They can damage your credibility and bring into question the accuracy of your work. However, anecdotal evidence supported by data can be very powerful. For example, telling the story of a family’s struggle with the aftermath of a death or spinal cord injury can graphically illustrate the severity of a problem and bring emotional pressure to bear on the agency responsible for correcting a problem. The downside of using this kind of pressure is that it can harden the resistance of any agency that feels threatened by their potential liability in the incident or by your assertion that they are at fault. If your data is accurate and thorough, it has plenty of power by itself, I recommend saving the anecdotes for the courtroom.

Litigation stalls your project so try to avoid it. Many organizations and lawyers assume that if they correct an obvious problem, they are accepting responsibility for all the injuries caused in the past. If lawyers get involved before the problem gets fixed, things usually do not progress as quickly and they may come to a complete halt. Intelligence, logic and the "public good" are not always at work in these matters. Your interest in protecting the public from fatal injuries may be very different from the responsible agency’s interest in protecting itself from possible liability and from their lawyer’s interest in a nice long case with a resultant juicy fee.

Avoid the pitfall of blaming, think win/win. Blaming only raises defenses against you and may cause legal logjams. In my view, taking the direct, confrontational approach with any agency or individual that is responsible for fixing a problem is an option, but should be your last option. Keep in mind that your primary goal is to effect change, not to establish fault and place blame. Try to find a way to provide some benefit to the responsible party as well as to the public, just be careful not to make promises that you can’t keep.

Consider using a computer database program to organize all of your data. If you are only planning to investigate motor vehicle fatalities, you probably don’t need a database program. In that case, sorting and comparing by hand is the simple low-tech alternative. However, if you have a very busy or large district, or if you choose to look at many different fatal events, it’s useful to build a database program so that once the information is inputted, you can run quick analyses on the data using any variable in the database. Some surprising and interesting results can come from looking at data that is sorted by many different variables.

Consider having a professional design your database to your specifications. Building the database and inputting data can be time consuming, but if you have a lot of data to sort through, it will save you time and effort in the long run. Most modern database software will be useful, but some are easier to use than others. If you are facing a major data inputting and sorting job, the fee for a few hours of design can help you avoid a lot of headaches later on.

Be accurate when inputting all data. Either do it yourself, carefully, and triple check yourself, or have someone who is conscientious do it under your supervision. Remember, "Garbage in, garbage out." Incorrect "facts" in your report may be seized on by responsible parties to discredit all of your otherwise accurate work. Make sure that what goes into your report accurately represents (and doesn’t over-represent) what has happened. Be factual.

11. Identify responsible agencies, individuals and resources to help implement the solution.

Present your findings to the appropriate party or organization. Be professional and courteous and avoid confrontation. Wear the other person’s shoes for a while. It’s not fun to have people from outside your agency confront you with problems that you are responsible for, especially when they have resulted in death and disability. They will initially feel threatened and defensive, so try to reassure them that your only interest is in getting the problem fixed before it causes another death. Try to establish a working relationship with them, not an adversarial one. Remember to think win/win. Keep in mind that everyone’s budgets are tight and your site may very well have been recognized previously and shelved while waiting for the money.

Play the videotape and show the slides. Graphic visual representation of the hazard, along with accurate, well-documented research, and a reasonable plan for an environmental intervention provides the best chance for success. The informed, cooperative approach is the path of least resistance and often the quickest one. Try that one first.

12. Build a consensus on the problem and on the best possible solution.

Building support for your plan in the community will raise the visibility of the problem and it costs nothing. If you’ve gone the cooperative route and reached an impasse, you will need some people on your side. A good place to start building consensus is with is other local emergency services. They have a stake in the success or failure of your project. They will be responding to that site again and again until it is fixed. They will relate to your concerns and probably be easy to convince, especially if you have a well thought out solution. Physicians, nurses, firemen, police and EMT’s have clout in that people trust them with their lives. When these people say something is hazardous, people tend to listen. Get them on your side.

Be prepared with a relatively inexpensive, well thought out plan of action that can be implemented quickly. Things can get tricky when you bring in other individuals, agencies, groups or committees. Well-meaning people coming in late, who have no injury prevention background will tend to immediately suggest better law enforcement, educational campaigns, and beefed up emergency services (particularly if you’re talking to police, fire, and emergency personnel). While there’s nothing wrong with any of these, the plan most likely to succeed will be proactive and environmentally based, so be sure to come prepared with a plan that clearly identifies and addresses the specific hazard at the source. Avoid getting side tracked. If you’ve done your homework, you know what the problem is and have a pretty good idea what to do about it. Take the reigns and head for your goal so that people will follow you and hopefully not stampede off in their own directions.

Once you have the support of a group, return to the responsible person and present your case again. Advise them that you have built a consensus among community members and that you all agree that something needs to be done now. Having the backing of the community greatly strengthens your case.

13. Evaluate the effectiveness of your intervention.

Follow up on the success or failure of your efforts. Usually once a hazard is removed and an effective protection is in place, the hazard is no longer an issue. However, some new unforeseen issue at the site may arise, or the intervention may not have the desired outcome. The only way to know is to follow the data at that site for a few years afterward. This is usually pretty easy because there won’t be any data. The problem will have ceased to exist.

Reviewing the data at an ex-cluster site is a rewarding exercise because you will notice the lack of fatalities at that site within a year or two. Every year after that is another year without fatalities. Documenting the success of your project is almost as important as documenting the problem itself. A successful project can provide you some much-needed credibility when you take on a new hazard

Be a model. A truly effective and innovative solution that has been properly evaluated has a value far above and beyond the correction of one specific problem. If your solution is effective, it may work at other sites in your district, or it may serve as a model for other programs to copy.