Responsibility for the enforcement of health and safety is shared between HSE, the Office for Rail Regulation (ORR) and Local Authorities (LAs). HSE collates information from all three regulatory bodies and publishes the data on its web pages.
The published statistics include total numbers of prosecutions and notices, the latter being divided into three types. Improvement notices require employers to put right a contravention of health and safety legislation within a specified time limit; prohibition notices stop work activity that gives, or will give, rise to a risk of serious personal injury; and deferred prohibition notices stop a work activity within a specified time.
Both HSE and ORR record all of their frontline activity using a database called COIN (Corporate Operational Information System). Every intervention with a duty holder is recorded on the system and the inspector is required to complete records detailing the outcome of their work, including any notices issued or prosecutions initiated. These records are only closed when the intervention has been completed.
Due to the importance of the data, HSE has detailed procedures and staff have clear roles and responsibilities to ensure good quality enforcement data on COIN. This data quality system is overseen by a dedicated central team, supported by data quality co-ordinators in each of the operational directorates. Together they routinely perform validation checks and query any apparent anomalies with the inspector who entered the relevant record. Statistics Branch receive downloads from the COIN database which are used to create the annual published statistics. Further validation is completed during this stage to ensure figures are accurate and consistent across all publications.
All data extraction procedures are discussed and agreed with data system owners, and repeated each year. Such procedures have been in place for at least 30 years, so are well established and understood by all relevant parties. Any deviations from these procedures are agreed beforehand with system owners.
Local authorities maintain their own records of their enforcement activity, each one using its own IT system. There is no standard quality assurance procedure therefore each LA is responsible for the accuracy of their records. At the end of each year local authorities return two reports to HSE (either electronically or by post). The first, a summary report, details the total number of notices served, with information on notice type along with tabulations of other demographics such as the industry sector of the organisations who have been served the notice. No list of specific companies is provided. The second report details all prosecutions brought by the local authority and does include company specific information.
Although there is no statutory requirement for local authorities to provide enforcement statistics to HSE the response rate is very high (typically 90%) as local authorities recognise their statutory duty to manage health and safety and the importance of bringing together the data from across all regulatory bodies. HSE is largely reliant on local authorities to audit their own data. However, HSE’s Local Authority Unit assesses the forms for any missing data and contacts the authorities directly to query this and input the correct figures. HSE statisticians then perform a series of validation checks on the return for other enforcement data (notices) – for example any figure which differs from the previous year’s figure for that local authority by more than 25% is checked by directly contacting the local authority to ensure there has been no transcription error. Furthermore since the data is published at a local authority level, this provides an incentive for local authorities to provide complete and accurate information and also a mechanism by which the health and safety community (including other local authorities) can challenge any figures which do not look plausible.
For the notices data, where there are non-responses HSE uses an established and consistent estimation procedure to adjust the figures upwards to give a more accurate estimate of the total number of notices served. To assist in delivering an accurate estimation, HSE categorises local authorities into ten different types according to their size, function and the urban/rural division in the region covered. Previous data shows that local authorities of the same type tend to issue similar numbers of notices and this information is used to drive the estimation process. For example if nine out of ten local authorities of a particular type were to respond, the figures would be multiplied up by one tenth to account for the missing response which experience suggests would have been a similar figure to the average of the other nine. Since the response rate from local authorities is so high, this estimation procedure has been able to be tested by comparing figures from years where estimation was required with figures from years where all local authorities of a particular type have responded. The procedure is demonstrably reliable.
A similar adjustment is not made to the prosecutions data. Since details of the companies prosecuted is publicly available it is not appropriate to compute an estimate which would then not match up to the count of prosecutions in the public register. Furthermore the prosecution data is more complicated in that there are issues to consider regarding number of breaches, legislation used and the timescale between case creation and hearing dates. It would not be appropriate to produce estimates for all these factors as this would seriously undermine the validity of the data produced. Conversely since notices are a simple count, such an estimation does not suffer from the same difficulties and will be more reliable.
All enforcement data for a given fiscal year is validated and published in October of the following fiscal year (i.e. six months after the end of the corresponding year). However, COIN is a live database meaning the HSE and ORR data is theoretically available for use as soon as it is entered on the system. HSE uses the validated published data where possible but it is sometimes necessary to use more up-to-date information, for example in answering urgent parliamentary questions. In this case, HSE will include clear warnings alongside the data that it has not been validated.
For the prosecutions data there is potential for confusion over the distinction between cases and breaches. A case may be brought against an organisation which will result in prosecutions for several distinct breaches of health and safety regulation. In HSE’s standard published statistics, offences prosecuted refer to the individual breaches of health and safety legislation – a prosecution case may include more than one offence. The precise definition of the counts is always provided alongside the data as footnotes in the data table.
Where enforcement statistics are allocated against a particular year, for notices this refers to the date the notice was issued; for prosecutions this relates to the date of the final hearing and where an outcome is known. As each year's statistics are updated, within the HSE 'offences prosecuted' figure, the latest year will always include some cases that have been adjourned.
The majority of enforcement activity relates to the general Health and Safety at Work Act which has been in operation for thirty years, so it is unlikely that changes to other legislation would have a significant impact on overall trends. However when interpreting trends over time it is important to bear in mind the potential impact of any changes in policy which may have occurred either in regard to particular pieces of legislation or the regulation of health and safety more generally and understand that trends may be subtly influenced by such changes.
Since HSE’s policy on enforcement is constantly evolving in response to current workplace health and safety concerns and will to some extent vary between topics and regions, it is not possible to evaluate the impact of every change and no attempt has been made to do so in the past. However should any major change in policy occur which HSE statisticians believed would significantly impact upon the statistics, this would be evaluated and footnotes would be added to the enforcement data tables, in the same way footnotes have been added to injury tables when there have been changes to the way injuries are reported or recorded.
HSE’s operational work is driven by its business plan and its commitment to investigating the most serious of injuries and health and safety risks through a set of incident selection criteria. This provides an important policy context for the enforcement data. HSE does not set targets for formal enforcement but takes action in response to the circumstances it finds during its interventions. The business plan can influence the amount of enforcement reported in each fiscal year. For example, HSE uses a variety of methods for encouraging better practice in the management of health and safety and the methods will vary according to topics and risks. If during a particular time period HSE takes the decision to focus a substantial amount of effort on a topic which it believes is best approached through softer approaches such as guidance or stakeholder engagement, this may result in a lower rate of formal enforcement during that period. At present no attempt is made to quantify the effect of HSE’s multi-faceted approach on enforcement statistics.
By providing enforcement data, local authorities are also able to demonstrate to HSE that they are fulfilling their statutory duty as enforcing bodies. HSE makes every effort to ensure the data provided by local authorities is accurate by giving the local authorities plenty of notice of the data they need to collect - there is an agreement not to change the recording mechanisms without less than 18 months notice. Furthermore, HSE makes checks on the data and contacts local authorities directly for clarification if figures look unusual, for example substantially different from figures provided in previous years or where subtotals do not add up to grand totals. Local authorities are asked to comment and explain any great year on year changes. HSE is reliant on the local authorities to collect and record information accurately and it is not practical to directly oversee this process so instead HSE ensures guidance is clear and that it retains a close relationship and clear communication routes with the local authorities so that any discrepancies or issues can be quickly resolved.
