The Use of THOR data for benchmarking of companies work related incidence rates
Dr J McCaul
Many organisations use a single source to identify cases of work related ill health (commonly cases referred by management to OH) whereas in RWE npower we use several sources. As a result we believe our incidence rates are more likely to be accurate.
Our definition of “work related” is that work was the major cause (i.e. at least 51% contribution).
The 3 sources are as follows:
- the opportunity for the employee to self-declare that they believe work was the cause of their illness when they return after any spell of sickness absence. Such self-declaration is always followed by an OH review which verifies or rejects the role of work;
- employees seen as management referrals by the OH Nurse or Occupational Physician (for absence, fitness for work or performance/behaviour issues). The OH practitioner must decide on the role of work in every case;
- employees self referring to a First Aider or Occupational Health Nurse who considers the part played by work in the illness.
In theory we could miss a few cases, the obvious example being when an employee feels ill at work or shortly afterwards and sees their GP after work, but does not have an episode of sickness absence. We believe such cases will be very small in number.
We collect the data over 6 month periods and then present them as incidence rates per 100 employees and days lost per 100 employees. As we have data extending back over several years, trends in both indices can be examined.
Our greatest problem is benchmarking our rates against those of other organisations; colleagues elsewhere tend to consider only management referrals (and the criteria for such referrals vary greatly depending on the organisation’s absence procedures). Therefore in order to satisfy our Board’s requirement for the benchmarking of health and safety data, I use two external sources – the Labour Force Survey (as published by HSE) and the OPRA database.
Labour Force Survey (LFS - self reported work related illness)
I use pie charts to show the relative numbers of cases in the Labour Force Survey (LFS) for “all industries” and compare these with our absolute numbers. An advantage of the Labour Force Data is that musculoskeletal cases are split into categories so we can benchmark our low back pain numbers. However, there are considerable disadvantages - LFS cases are self declared, they are not specific to the Electricity Supply Industry (because numbers are too small) and they include diseases made worse by work.
We also compare incidence rates per 100 employees (npower) and per 100 of the population (LFS).
The THOR group provide me with estimated OPRA case numbers over a 12 month period which corresponds with the reporting period we use. I then use the pie chart format to show the relative numbers. I emphasise that there are considerable differences between the 2 sources: OPRA cases are “work related” (with no definition of this phase) and they are based only on physician diagnoses.
I am also provided with estimated incidence rates for a combination of “electricity, gas, steam and gas” industries which I compare with our rates. However, the denominator used in OPRA for calculating such data, is based on headcount numbers provided by reporting occupational physicians and such data may be inaccurate.
In summary, I value the opportunity to compare our raw numbers and incidence rates with those from OPRA, given the paucity of data from other organisations.
Chief Medical Officer, RWE npower