Injury data coding

BLS began collecting additional information on the major injuries in the form of worker and incident characteristics. At that time, BLS also initiated a separate CFOI to review events more effectively than had been possible in the previous survey. The CFOl database can be used to do statistical analysis for fatalities by Standard Industrial Classification (SIC) Codes. The CFOl fatality data is presented in several different categories as shown below ... 

Five of the 15 forms received require entry of codes for causal factors, incident types, and injury t)q)es. When computer analysis programs first became available, I had been an aggressive promoter of the entiy of causal factor codes for later analysis. That proved to be inappropriate because accurate causal data are often not included in supervisors investigation reports or in insurance claims reports. Now, I recommend that computer-based analysis systems not include provision for causal data entry. They serve analysis purposes quite well for types of accidents, injury types, parts of body injured, and identification data (location, age, job title, etc.). 

Data should be easy to access and sort. Sort fields could include type of part or product, manufacturer, cost, size, use, types of accidents or failures, geographic location, time, date, temperature, other environmental conditions, types of injuries, energy sources, dollar loss, analyst, project name, project number, system, subsystem, risk assessment code, and any other relevant comments that may aid in analysis of data. 

There are several metrics at different levels for the evaluation of active safety. Depending on the research question and the method used, the quality of the assessment can vary. For example, injuries coded in an in-depth data base have a different reliability than injury probabilities given by probabiUstic models or economic costs, which use injury information as basis. Stating the protection of the human as key objective, a metric based on injury severity seems to be appropriate. 

The first of these issues refers to the choice of a metric to describe the injury scale. In the data sets considered here, pedestrian injuries were originally coded according to the Abbreviated Injury Scale (AIS), revision 90 [8-10] (for cases 2008 and newer GIDAS also includes AIS coding following the 2005 revision [11-13]). Table 5.1 gives the AIS levels as well as the lethality rate associated with each level. The maximum AIS value (MAIS) of a person is separately coded and serves as an indicator for overall injury severity. 

Fatalities were coded and investigated independently, as they are distributed over a range of ISS or MAIS values [6], Cases with missing injury or mortality data were excluded from analyses involving target variables. 

In both GIDAS and PCDS, additional variables were computed from existing ones squared impact speed of the vehicle (to account for possible non-linearities between impact speed and injury severity), kinetic energy of the vehicle, and body mass index (BMI) of the pedestrian. Ratios between anthropometric values and vehicle dimensions were also constructed. Since only basic vehicle profile characteristics were coded in GIDAS, additional data sources were used to reconstruct some of these characteristics for the purpose of analysis (see Sect. 5.2.4). The notation GIDAS and PCDS used for labeling the variables refers to the original data set. 

In the augmentation step, secondary data sources describing the vehicle profile were fused with the GIDAS data set, as the coded information is very sparse, as not many variables describing vehicle profile characteristics are included in GIDAS. Those characteristics are supposed to influence injury severity and thus should be... 

It is important to assess internal relationships between explanatory variables and impact speed, as these can distort the interpretation of the results, especially if no hypotheses about the causal connection of the variable and injury or fatality risk exist. The findings show that some variables are correlated with impact speed and those findings are interpreted. In addition, conclusions about the behavior of the driver or the hazard induced by specific maneuvers are possible by testing these variables against impact speed. This method allows for the discovery of aspects that are not coded explicitly in the data sets and thus enhance the knowledge about the genesis of the accidents as well as their course of events. 

The accuracy as well as the power of the models depend on the number of cases available resulting practical limitations to research were highlighted using the GHDAS data base as an example. Especially for the construction of injury probability models, data sources should be up to date and should include far more cases. In addition, the accuracy of the models depend on the quality of the data used continuous improvements in coding and reconstruction are thus strongly encouraged. Imputation procedures, as included, for example, in US accident data bases, could minimize loss of data due to list-wise deletion in a standardized way. 

However, an additional lesson was learned from the computer run sent to this author that included data on several hundred injuries. Detail entered into the computer system on each incident was extensive. But, it allowed only one causal factor entry, the instruction for which was Enter the unsafe act code. Unwittingly, the system encouraged an unwanted focus on the unsafe act of the worker. The data entry system was revised. 

Evaluations should assess prolonged work in any posture that may result in harm or injury. Assess offices, computer areas, and nursing stations. Evaluate force, duration, position, frequency, and metabolic expenditure of workers. Workers should be provided with good chairs that have arm and leg rests if required. Provide workstations that permit posture variations and have sufficient space for knees and feet. Workers such as admission personnel, appointment clerks, transcriptionists, medical coding personnel, and other data entry personnel that work on computers four hours or more each are at risk for developing hand, arm, shoulder, neck, or back disorders. 

The application of accident-concentration analysis is not meaningful for small data sets. There must be at least in the region of 50 accident cases. Useful types of data in accident-concentration analysis are location, activity, equipment, accident type, type of injury and part of body affected. The analysis is facilitated if some of the data is coded (i.e. presented on a nominal scale of measurement), especially if large quantities of data are handled. The coding should, however, not be done at the cost of the details in the information. The free-text description of the sequence of events should always be available. 

Medical records provide the input to registers on medical treatments by the health-care system. The registers include demographic data on the patient (such as sex, age and place of residence) and data on diagnosis and treatment. The external cause is also registered for injuries by using the World Health Organisation s so-called ICD-10 codes (International classification of diseases version 10). It is here possible to identify road-traffic causalities. 

Some large construction organizations have analysed occupational injuries by introducing computers. The coded data on circumstances and causes accompanying accidents are supplied to the... 

For risk assessment purposes, national data for injuries, illnesses and fatalities may be obtained from the Bureau of Labor Statistics (BLS) website at www.bls.gov/iif. The annual data, beginning in 2003, are grouped by the North American Industrial Classification System (NAICS) that assigns a numeric code for each t5T)e of work establishment. Prior to 2003, the Standard Industrial Classification (SIC) system was used to categorize the data instead of NAICS. 

To use the BLS data, you must know your establishment s injury rates (see below) and your establishment s NAICS code (see above). 

Once you know your establishment s NAICS code, go to httpy/www.bls.gov/iLfioshsmn.htm. Find the Summary Table called Table 1 - Incidence rates - detailed industry level. Scroll through the BLS table and find the corresponding NAICS code in the left-hand colmnn. (NAICS data are presented in numeric order in the BLS Table). To compare your injury and dlness rates, use the number listed in the Total recordable cases column for your NAICS as the industry average incidence rate. Use the number listed in the total from the Cases with days away from work, job transfer, or restriction column as the industry average DART rate. Note At press time, the most current data available were for calendar year 2011 injuries and illnesses 2012 should be available soon. 

---