Preventing absences from work caused by occupational injuries is an important part of every manager's job. Your companies injury records contain information which can be used to prevent accidents and reduce losses. This procedure describes how injury incidence and frequency rates are calculated and shows how we can use them to help assess our performance over time, and compare our rates with those of others particularly in the food industry.

The Reporting of Injuries, Diseases & Dangerous Occurrences Regulations require us to keep records of injuries at work and report certain types to the appropriate enforcing authority, usually the Health and Safety Executive (HSE) or a local authority. Reportable injuries include fatal and major injuries to employees, self-employed people and members of the public, and injuries that cause incapacity from work for more than three days to employees and self-employed people ('over 3 day injuries').

Comparing reportable injury information is just one way of assessing our safety performance. In many of our companies, particularly those with fewer than 100 employees, reportable injuries represent only a small proportion of the total number of injuries to employees. Records of more minor, non-reportable injuries, and of 'near misses', may also be converted into incidence rates and used to monitor trends over time or between different parts of our operations. Analysis of the data to identify the main causes of injury, for example, can help to identify risks that need to be controlled and prevent further accidents.

HSE's formula for calculating an annual injury incidence rate is:

Number of reportable injuries in financial year X 100,000
Average number employed during year

This gives the rate per 100,000 employees or agency.

The formula makes no allowance for variations in part-time employment or overtime (see section 3). It is an annual calculation and the figures need to be adjusted pro-rata if they cover a shorter period. Such shorter term rates should be compared only with rates for exactly similar periods
not the national annual rates.

HYPOTHETICAL EXAMPLE:

A yoghurt producer had 1 fatal, 2 major and 42 over 3 day injuries in 1995/96. The average number employed by the firm during the year was 2300.

The fatal and major injury rate in 1995/96 was therefore:

(1 + 2) X 100,000
2300

= 130.4 per 100,000 employees

The all reported injury rate was:

(3 + 42) X 100,000
2300

= 1956.5 per 100,000 employees

The fatal and major injury rate for the yoghurt and dairy industry in 1995/96 was 150.7 - so this firm's figure was slightly less (i.e. better) than the national rate.
For all reported injuries that rate was 1877.1 - so this firm's rate was above (i.e.. worse than) the rest of the industry.

When injury incidence rates are calculated to compare performance between companies or parts of firms, it is essential that the injury numbers and the average number of employees are consistently counted and calculated. Firms that have few reportable injuries in a year may find that their annual injury rates are sensitive to small variations in numbers. This sensitivity may be reduced by calculating rates over a longer period of time. Accurate recording and monitoring both reportable and less serious injuries should enable any short term trends to be quickly identified.

While HSE calculates injury incidence rates per 100,000 employees, some parts of the food industry prefer to calculate injury frequency rates, usually per million hours worked. This method, by counting hours worked rather than the number of employees, avoids distortions which may be caused in the incidence rate calculations by part and full-time employees and by overtime working. Frequency rates can be calculated for any time period.

The calculation is

Number of injuries in the period X 1,000,000
Total hours worked during the period

Both incidence and frequency rates are valuable in that they provide simple indicators as to whether the safety performance of a firm (or department in it) is improving or deteriorating. Whatever series of data is being monitored over any given period, some random fluctuation between periods can be expected and may not be statistically significant. This is particularly the case with short-term (e.g.. monthly) comparisons.

The simplest way to monitor injury trends is by plotting the measured data on a graph. Targets can also be plotted (e.g.. based on the national figures for the food industry) and runs of consecutive points above or below target can then be identified. A number of statistical tests can be applied to estimate the significance of variations in the figures but these go beyond the scope of this procedure. Some companies now market computerised accident data analysis packages which can be of considerable value particularly for organisations where substantial quantities of injury related data are generated and need to be analysed. These are not considered as appropriate at this time for Greencore Group.

The accuracy of the nationally collated injury statistics depends on us complying with the legal reporting requirements. In some industries under-reporting of injuries by employers is a serious problem. Firms with good record keeping arrangements in an industry with a high level of under-reporting may therefore find that their injury rates compare unfavourably with the published rates for their industry. The figures in such cases obviously must be interpreted accordingly. Incidence rates can, however, still be used to monitor performance over time and between different departments.

Many factors affect injury rates and care is needed in interpreting the reasons why a particular company (or department) is above or below the national industry rate. There are substantial differences between injury rates for men and women in different parts of the industry, so the composition of the workforce can be a factor, as of course can the range of occupations and activities involved. HSE research also suggests that smaller establishments have higher injury rates than larger establishments.

At a meeting of Greencore Joint Consultative Committee (JCC) on 4th February, 1997, at Kiveton. Representatives of all the Greencore Group operating Divisions unanimously agreed to adopt the use of accident frequencies based on 100,000 person hours from 1 January, 1997.

Each quarter all Divisions will supply data to the Greencore Group, Health and Safety Office (Wrexham) in graphical form for each factory within that Division. A moving Annual Total and previous years data should be supplied to be incorporated. If the data is unavailable at this time the report will be annotated DNA (Data Not Available).

Those sites without computer support can provide reports to The Greencore Group. in hand written format on traditional graph paper.

Examples have been included as figure one. Those companies requiring copies of the software used to generate these examples will be provided with discs on request.