Analyzing accident data

By analyzing your accidents/incidents, you are in a better position to compare apples to apples rather than apples to oranges. You will be able to identify not only the types of injuries and types of accidents and causes, but you will also be able to intervene and provide recommendations for preventing these accidents/incidents in the future. You will be able to say with confidence that I do or I do not have a safety and health problem. If you find that you have a problem, your analysis and data will be essential if you try to elicit advice on how to address your health and safety needs. 

Gathering and analyzing accidents/incidents data do not comprise your entire safety and health program, but are single elements that provide feedback and evalnative information as you proceed toward accomplishing your safety and health goals thns, they are important elements. 

The two most frequent statistical pieces of information, which are designed to allow you to compare your company s safety and health performance with others, are the incident rate and the severity rate. These two rates, respectively, answer the qnestions of How often or frequently do accidents occur and How bad are the injnries/illnesses that occur The number of times that occupational injuries/illnesses happen is the determinant of the incident rate whereas the number of days (lost workdays) is the prime indicator of the severity rate. Both of these rates provide unique information regarding your safety and health effort. 

How can you compare your company of 15 workers with a company that has 250 workers This can be accomplished only by using a statistic or formula that allows a standardized numerical value to be developed, which takes into acconnt the differences and places your company on the same playing field so that each company s front line appears to be the same weight. 

To find the incident rate, count the nnmber of distinct events that resnlted in injuries/illness. To compare your incident rate with that of other companies, you must normalize your data. This is accomplished by using a constant of 200,000 work hours, which was established by the Bureau of Labor Statistics. The 200,000 work hours are the number of hours that 100 full-time workers would work during 50 weeks at 40h per week. Thus, you can calculate your incident rate in the following manner  

Tracking should not just be a count of recordable incidents. Companies should track near-misses, equipment damage, and first-aid events. 

Tracking means that company officials must pay close attention to workers compensation cases as well as their associated costs. The longer an individual is off the job, the more the cost and the greater the likelihood that the worker will not return to work. 

---

The frequency analysis step involves estimating the likelihood of occurrence of each of the undesired situations defined in the hazard identification step. Sometimes you can do this through direct comparison with experience or extrapolation from historical accident data. While this method may be of great assistance in determining accident frequencies, most accidents analyzed by QRA are so rare that the frequencies must be synthesized using frequency estimation methods and models. 

The insurance industry as a whole has a substantial self interest in analyzing accidents, sponsoring research and issuing publications in the methods tc prevent accidents or mitigate their effects. Hence their data is the... 

Equation of state and in particular, vapour pressure data on nuclear fuel materials at very high temperatures (2500-5000K) are especially required to analyze accident conditions. A laser induced vaporisation mass spectrometric (LIV-MS) facility has been developed to measure these high temperature vapour pressures of fuel materials. Initial experiments are carried out on UO2 and the vapour species observed are U, UO, UO2, UO3 and O and their partial pressures are measured over a temprature range of 3,300 K to 5,500 K in which the total pressure changes from 0.05 MPa to 6.5MPa. 

Accident Trending (4) All employees fully aware of incident trends, causes, and prevention methods. (3) Trends analyzed and displayed, common causes communicated, management ensures prevention. (2) Data collected and analyzed, and common causes communicated to concerned supervisors. (1) Data collected and analyzed but not widely communicated for prevention purposes. (0) No consistent effort to analyze incident data for trends, causes, and prevention. 

Frequency and severity data from accidents can help identify risks. A review of accident records and classes of accidents can help. Various statistical methods applied to accident data can reveal trends in losses and factors that contribute to accidents and injuries. Analyzing claims, such as worker compensation claims or customer claims against products, will help isolate factors associated with losses. 

Like p-charts, np-charts are used to analyze accidents and other nonconforming events or deviations over a constant period however, the np-chart uses a constant sample size and plots the nt/mher of nonconformities (np) rather than the percentage of nonconformities (p) (Griffin, p. 439). The steps to follow for constructing an np-chart are the same as for a p-chart (Training Resources and Data Exchange (TRADE) 1995,2-1). The np-chart is used when ... 

Gathering and analyzing accident/incident data is not the company s entire safety and health program, but a single element. Data provides feedback and evaluative information as companies proceed toward accomplishing their safety and health goals thus, data contributes an important component in the analysis process. 

There are five phases in accident investigation preparing for the investigation, gathering the evidence and information, analyzing the data, discussion of the analysis... 

Step 3 Analyzing the data. Once the information has been gathered (and even before all the information becomes available), data analysis begins. Whenever possible, it is best to analyze the actual system measurements (e.g., motor rpm) rather than pseudomeasuranents. The purpose of Step 3 is to understand the accident based on the available data. 

One of the most obvious means of collecting and analyzing error data in complex, safety-critical systems is through the retrospective investigation and analysis of accidents and incidents. Accident investigation is used to reconstruct accidents and... 

Monitor and assess safety performance. Organizations should continually monitor, assess, and learn from their own performance with regard to safety. Incident and accident data should be collected and analyzed accordingly, and appropriate lessons should be drawn from such endeavors. 

The present paper is divided into two main areas current status, laws of Romania and OHS (Occupational Health and Safety) European directions, on one hand, and analyzing the causes of work accidents in Romania and the model of influence, on the other hand. By analyzing the data and presenting a model, a presentation of future directions and proposed improvement of the current situation is required in the end. 

In this paper, the description of accidents occurred in the Baltic Sea during the winters 2002-2003 2009-2013 are detailed analyzed. The data analysis presented in this paper covers accidents reported on Finnish waters and/or to Finnish vessels registered by the Finnish Transport Safety Agency during the mentioned time periods. The quantitative and qualitative analyses presented in this paper have carefully assessed the accidents occurred under sea ice conditions. Thus, the aim of this study is to present which are the most common... 

Based on the accident data analyzed in this paper, the accident severity in most cases has been reported as less serious. Comparing this finding against previous studies (e.g. Tuominen, et al., 2010), the severity of the accidents occurred under ice conditions has also been considered as less serious in most of the cases. From the total amount of data analyzed in this paper, 75% of 68 accidents reported during the winters 2009-2013 were reported a less serious, 24% as serious, and only 1% as very serious accidents. And about the accidents occurred under ice conditions, 84% were reported as less serious, 16% reported as serious, and there were not accidents reported as very serious. Thus, these indicators may imply that the severity of the accidents may be better estimated and perhaps even controlled by the correct implementation of the current winter navigation system, a correct provision of the different sea ice services, and specially an adequate assistant by ice breakers. 

Transportation Research Record 1987 pp. 81-97. Hollo P., (2008) Methods and tools of analyzing road accident data. PI ARC International Road Safety Seminar Beijing, China. 

Analyzes pipeline safety and reported accident data... 

For this study, we analyzed all records of injirry acciderrts from the natiortal Austrian accident database, where cyclists were involved. The data covers the years 2002 until 2011 since data for 2012 was not available at the time of writing. While studies [ELV 99, LAN 03] recommend the use of hospital data for safety assessments, police data for accidents was used in this study since hospital data is not available for analysis in Austria. While accident insurance providers do collect data from hospitals, current data privacy laws do not allow for a consolidation of accident data from police and hospital sources. For the city of Vienna, the resulting dataset contains a total of 6,287 accidents. The database consists of several tables which describe the accidents. The accident table contains information about the location of the accident, the weather conditions, as well as the date and type of the accident. The participants table contains information about the participants such as age, degree of injury and type of vehicle. Table 10.4 shows the trends of the yearly nnmber of accidents. Fignre 10.1 presents a comparison of the trends of accident counts and bicycle counts based on the initial values for 2002. In the analysis time frame between 2002 and 2011, the data show no correlation between trends of accident counts and bicycle counts (R = 0.03). This suggests the validity of the concept of safety in numbers , which states, that an increase in the modal share of bicycles leads to a decrease in the number of accidents per cycled kilometer. 

The accident data also contain location characteristics, e.g. intersection, cycle path, bridge, pedestrian zone, etc. An accident can be described with up to five location characteristics. The most common location characteristics in the analyzed records are cycle path or lane (2,106 accidents), intersection with yield-sign (2,109 accidents), normal intersection (2,062 acciderrts), T-shaped intersection (1,594 accidents) and one-way street (1,583 accidents). Note that one accident location can share mnltiple of these characteristics, for example in the case of a location at a normal intersection with yield signs. Table 10.5 shows the development of accident cormts for the most common location characteristics. The graphical representation in Fignre 10.2 shows the differences in development of accident counts for these location characteristics. While the number of accidents at intersections is relatively stable, accidents on bicycle routes seem to be on the rise. 

From 1970, ITOFF has analyzed oil tanker spill-related accidents data and since 1974, it has maintained a database of accidental oil spills from tankers, combined carriers, and barrages. 

---