NOHSAC : Methods and Systems Used to Measure and Monitor Occupational Disease and Injury in New Zealand : Tech Report 2

1.1 Occupational Disease

Occupational disease accounts for greater mortality and morbidity than occupational injuries, but is harder to diagnose, measure, and monitor for a range of reasons, including long latency periods after exposure, difficulties distinguishing occupational diseases from non-occupational diseases, and a lack of awareness about the occupational origins of some diseases. Consequently, there are no comprehensive sources of routinely-collected data on mortality or morbidity due to occupational disease in New Zealand. The following potential sources of occupational disease data were reviewed.

Death Certificates and Coroners’ Reports

Medical certificates of causes of death and coroners’ reports are reported to the Births, Deaths and Marriages (BDM) electronic database of causes of death. The database is complete for all deaths occurring in New Zealand and records detailed information about direct, antecedent, and underlying causes of death. However, the medical certificate does not collect any information on whether or not the death is work-related, and the recording of occupation is unstructured.

Coroners’ reports include data obtained from many sources and represent a potentially rich source of information on the nature and circumstances of some deaths. The recording of occupation involves the use of a free-text field asking for “usual occupation, profession, or job” of the deceased. However, coroners are more likely to investigate deaths due to occupational injury than those due to occupational disease. Coroners’ findings are not directly recorded electronically, and this represents a major barrier to the efficient retrieval of surveillance information.

Mortality Collection

The New Zealand Health Information Service (NZHIS) maintains the Mortality Collection, which classifies the underlying cause of death for all deaths registered in New Zealand. It integrates data from a range of sources including BDM, coroners’ findings, and postmortem and toxicology reports. The National Health Index (NHI) number is used as a unique identifier on the Mortality Collection, facilitating linkage with other data sets.

Occupation and work-relatedness data are inadequate for the surveillance of occupational disease. The recording of occupation involves the use of a free-text field. Little guidance is given to funeral directors who complete the occupation details on the notification of death registration form. Although there is a work-relatedness flag on the database, it is only used where the cause of death was related to an accident at work.

New Zealand Cancer Registry (NZCR)

The NZHIS maintains the New Zealand Cancer Registry (NZCR), which records all primary malignant diseases diagnosed in New Zealand, excluding squamous cell and basal cell carcinomas of skin. It provides full coverage of cancer incidence since 1994, when compulsory reporting by laboratories was introduced, virtually complete (>90%) coverage since 1972, and partial coverage as far back as 1948. The NHI number is included as a unique identifier on the NZCR, facilitating linkage with other data sets.

Occupation details on the NZCR are inadequate for occupation disease surveillance, being sourced from the National Minimum Data Set (NMDS) of hospital discharges, where it is poorly recorded and uses free text. Moreover, since 1 July 1994, the occupational information from the NMDS has not been routinely imported into the NZCR. There is no indicator of work-relatedness on the NZCR.

National Minimum Data set (NMDS)

The NMDS is a national collection of public and private hospital discharge information, including clinical information, for inpatients and day patients. It is maintained by NZHIS. A particular strength in the data set, aside from its completeness for public hospital discharges and near completeness for private hospital discharges, is the strong coding of diagnosis and severity, using the International Classification of Diseases (ICD). The NMDS includes two unique identifiers, which would facilitate linkage with other data sets. These are the patient’s NHI number and, where applicable, their ACC claim number.

The recording of occupation is inadequate for the surveillance of occupational disease, being under-recorded and restricted to free text and an occupation code. The hospitals may report either (or both) the code and the free-text field. There is no work-relatedness indicator on the NMDS. However, the potential to identify work-related diseases and injuries using E codes has been improving with successive upgrades of the ICD-10-AM system, in particular the 3rd Edition, which is used in New Zealand from 1 July 2004.

Notifiable Occupational Disease System (NODS)

The Notifiable Occupational Disease System (NODS) is a voluntary reporting scheme whereby health professionals and other individuals can notify a health-related condition that is suspected to arise from work. NODS is administered by the Department of Labour (and formerly by the Occupational Safety and Health Service (OSH) within the Department).

NODS was designed to supplement the statutory requirement for employers to notify serious harm and fatalities, by providing a vehicle for voluntary notification of suspected occupational diseases. However, the notification card implicitly restricts data collection to those occupational diseases included in the legislative definition of “serious harm” used by OSH.

The key strengths of NODS are that it was introduced specifically to record occupational diseases and that anybody can make a notification. However, NODS currently has a low potential to contribute to the surveillance of occupational disease. Key problems include poor diagnosis and under-reporting of occupational diseases to OSH, a system design that does not lend itself well to the aggregation of data for surveillance purposes, a low state of readiness of the data for integration with other data sets, and work practices that are intended to support efficient investigations and are not always consistent with the recording of high-quality data.

NODS notifications tend to contribute to the prevention of the recurrence of harm through the identification of learnings from individual cases rather than aggregated data.

OSH Panels

There are currently four panels, comprising medical and non-medical specialists, which were established to review and monitor specific occupational diseases and extend the evidence bases relating to the occupational origins of these diseases. These are the Cancer Panel, the Respiratory Diseases Panel, the Solvent Panel, and the Chemical Panel. These panels are linked to the NODS system, and cases of occupational disease identified by these panels are entered into the NODS system.

Cancer Panel

The Cancer Panel endeavours to review all cases of selected cancer sites reported to the New Zealand Cancer Registry, to identify possible occupational causes. Currently, the sites under review are bladder cancer, non-Hodgkins lymphoma, and leukaemia; it is proposed that, in 2005, the review of cases from these sites will end, and the focus will be on lung cancer for the next two years.

Unlike the other OSH panels, the Cancer Panel is not solely reliant on notifications made to OSH. Instead, it has the significant benefit of access to New Zealand Cancer Registry data, covering all new cancers diagnosed in New Zealand. This represents a significantly different approach in that the panel takes a “top down” approach and starts with all cases of the cancer sites under review and then determines which cases are work-related (and which are not). Thus, it can, in theory, identify all of the work-related cases for these sites. This differs from the other panels and the rest of the NODS system, which use a “bottom up” approach that is reliant on individual voluntary notifications.

The demographic and diagnostic information provided by the Cancer Registry is combined with detailed occupational and exposure histories gathered through interviews with individual patients. Approximately 60–70% of all new cases of these cancers have been reviewed in the last three years. The Cancer Panel has successfully demonstrated that the incidence of these cancers from occupational causes in New Zealand is similar to that in other Western countries and has published a study on bladder cancers, with non-Hodgkin’s lymphoma and leukaemia being the subject of two future reports currently in preparation.

Respiratory Diseases Panel

The Respiratory Diseases Panel was established to review and monitor occupational respiratory disease notifications, including asbestos-related diseases, occupational asthma, and other respiratory diseases. It was formed out of two previous OSH panels in 2001: the Asbestos Panel and the Asthma Panel. The Asbestos and Asthma Panels contributed to the body of knowledge about asbestos-related diseases and occupational asthma, through annual reports and ad hoc studies. However, notifications to the Respiratory Diseases Panel have declined in recent years, and staff turnover and vacancies at OSH are said to have resulted in the panel being poorly supported by head office. It is understood that few respiratory diseases, other than asbestos-related diseases, have been reported to the panel. The Asbestos Diseases Register covers an estimated 30% of mesotheliomas and a very small proportion of other asbestos-related diseases. Electronic recording of asbestos exposure and disease reports is understood to have ceased in the mid 1990s, when OSH transferred to a new computer system and problems were encountered transferring the data to the new system. It is understood that the data from the previous databases was subsequently lost or destroyed.

Solvent Panel

The Solvent Panel has been successful in demonstrating the existence and importance of chronic organic solvent neurotoxicity as an occupational illness, particularly within certain industries. Although only a small proportion of all cases are believed to be reported to OSH, they are probably among the more severe cases. The number of notifications has declined in recent years, however, and this is believed to be principally due to a decline in notifications from GPs. There has been no analysis of solvent neurotoxicity data in recent years.

Chemical Panel

The Chemical Panel was established to review and monitor notifications relating to diseases originating from chemical toxicity. The Chemical Panel has not convened in the last two years, due to non-reporting of cases. In contrast, ERMA New Zealand data recorded 57 cases of poisonings or toxic effects as a result of workplace exposure to hazardous substances in 2003/04, with chemicals and chemical products being the most common substances associated with such incidents.

Health and Safety Accident Recording Database (HASARD)

The Health and Safety Accident Recording Database (HASARD) records serious harm notifications made to OSH. The legislative definition of serious harm includes certain occupational diseases which, when notified to OSH, are recorded in the NODS system as described above. The instances of serious harm recorded by the HASARD system are principally occupational injuries. Therefore, HASARD has a low potential to contribute to the surveillance of occupational disease.

Accident Compensation Corporation (ACC) Claims Database

ACC administers New Zealand’s Accident Compensation Scheme, which provides personal injury cover for all New Zealand citizens, residents, and temporary visitors to New Zealand. ACC collects levies, determines eligibility of claims, and provides entitlements, including compensation payments, cover for medical fees and other care, and rehabilitation services. The ACC claims database provides a record of all cases that meet the criteria for compensation and for which compensation is claimed.

The ACC scheme provides cover for occupational diseases specified in legislation. Coverage is very specific and includes certain occupational diseases, linked to specific exposures, and other diseases that meet the legislative definition of a “personal injury caused by a work-related gradual process, disease, or infection”. Although there is a financial incentive for individuals to submit claims, it is unclear how comprehensively the database reflects the true incidence of the diseases covered by the scheme.

The structure and coding systems of the ACC database are, in many respects, well-suited to the surveillance of occupational disease. For example, the database records the ACC claim number and, where available, the claimant’s NHI number, facilitating record linkage to NZHIS databases. Occupation and industry are coded according to the standard Statistics New Zealand classification systems. There is a specific indicator for work-relatedness, although its usage is incomplete. Latterly, work-related injuries (excluding motor vehicle traffic crashes) can be identified, since they are paid from the Employer and Self-Employed accounts. The diagnosis field accommodates both ICD-10 and Read codes, and ACC routinely maps Read codes to ICD-10. The database records robust and objective cost information, including time off work.

However, the overriding functions of the ACC database have been administrative, such as determining eligibility for a claim, determining which ACC account should fund the claim, facilitating case management, and providing data to inform the setting of levies. These administrative objectives are not always consistent with surveillance imperatives. In particular, these objectives do not always require complete and accurate data on occupation.

EpiSurv <

EpiSurv records all notifiable diseases that have been reported to Medical Officers of Health. Reporting of these diseases is mandatory, and there is a strong emphasis on completeness and accuracy of data entry. Coverage is variable by disease type, depending upon the proportion of total cases that result in a GP visit. Detailed information is recorded on the diagnosis and the basis of diagnosis. Exposure information is also captured where available. There is a free-text field for recording occupation and, although it is under-utilised overall, it is fairly well completed for work-related cases. The database includes patients’ NHI numbers, facilitating linkage to NZHIS data sets.

Environmental Risk Management Authority (ERMA)

ERMA New Zealand’s mission is to achieve effective prevention or management of risks to the environment, public health and safety associated with importing or manufacturing hazardous substances and introducing new organisms, and their use.

ERMA has a specific mandate to measure and monitor impacts of the legislation governing hazardous substances and new organisms on health and the environment. The workplace is a major source of exposures to hazardous substances. Substances are still being transferred into the new legislative regime from previous pieces of legislation, and ERMA is in the process of developing its surveillance capability.

Current data analysis is based on aggregated and confidentialised data from a range of sources, including the NMDS, Fire Service, HASARD, and some directly received reports of hazardous substance incidents. The Ministry of Health is also working with the Institute of Environmental Sciences and Research to facilitate progress toward the development of a Chemical Injuries Surveillance System. Technical and process issues identified to date suggest that implementation of such a system may be a few years away and not necessarily on a full national scale.

Industry-Specific Surveillance Systems

This review included data collections maintained by the Civil Aviation Authority (CAA), Maritime New Zealand (MNZ), the Land Transport Safety Authority (LTSA), and the Forest Research Centre for Human Factors and Ergonomics (COHFE). The surveillance systems operated by these organisations are principally applicable to the surveillance of occupational injury, although they may detect cases of stress and fatigue to the extent that these were contributing factors to an injury.

Strengths, Weaknesses, and Gaps

The key strengths across the systems are:

the complete capture of fatalities, cancers, and hospital admissions (notwithstanding the inability to identify work-related disease cases within these data sets)
the existence of two unique identifiers (NHI number and ACC45 claim number) that can facilitate the linkage of all of the five most promising data sets in terms of coverage (Mortality Collection, Cancer Registry, NMDS, ACC, and EpiSurv)
the use of ICD-10 as a common coding standard across four of the five most promising data sets (although some concordance issues between different versions of ICD-10-AM must be acknowledged)
ready access to timely data and advice on its interpretation.

Key weaknesses include:

the inability to identify (either positively or probabilistically) most occupational disease cases, due to poor coding of occupation and work-relatedness and limited capture of exposure to known risk factors
under-reporting of occupational disease, including the under-reporting of work-related cancers
poor coding of industry and ethnicity in most databases
very little information on the costs of work-related disease
the low potential of New Zealand’s only purpose-built occupational disease recording system to contribute significantly to occupational disease surveillance, due to under-reporting and shortcomings in system design.

Gaps have been considered along three dimensions:

Vertical – cases not captured by the systems: Cases of occupational disease that are non-cancerous, have not resulted in a hospital admission or death, are not covered by ACC, and are not a notifiable communicable disease, will only be captured if they are reported to OSH and recorded in the NODS system. Occupational diseases affecting volunteers and bystanders will also be under-recorded or under-identified across the systems.
Horizontal – missing information about cases: The principal gap in data collection is the lack of standardised coding of occupation in the three NZHIS databases, the NODS system, and EpiSurv. Other important gaps include work-relatedness, industry, and ethnicity.
Temporal – coverage and consistency over time: Data should be relatively consistent across the main data sets since 2000. Key changes in the data sets over time are identified within the review.

Opportunities For Improvement

A range of opportunities for improvement have been identified. Similar findings can be seen in a range of previous reports over the last few years. The following opportunities have been identified purely with regard to their desirability from an occupational disease surveillance perspective and without consideration of legislative, policy, or budgetary implications:

Administration and Management

Put someone in charge at a national OSH level with the appropriate qualifications and support staff, to coordinate and improve the quality of all-cause occupational disease surveillance, working in collaboration with a range of agencies and independent researchers.
Ensure someone is responsible within each of the relevant organisations, with a specific mandate and appropriate resources to ensure the capture of high-quality occupational disease data.

Data Collection

Take deliberate steps to increase reporting of suspected cases to NODS, including targeted work with employees, unions, employers, and GPs, to raise awareness and fill knowledge gaps; and provide feedback to promote the value of NODS data.
Continue and enhance those aspects of NODS that are currently working well, including the work of the OSH Cancer Panel.
Improve recording and investigation of the work-relatedness of disease, including modifying death certificates to capture the certifying doctor’s opinion as to whether the disease was work-related; amending the BDM database to capture this information; broadening the definition of work-relatedness associated with the corresponding indicator in the Mortality Collection, to include suspected occupational disease; expanding the role of the Coroner’s Court to include inquiry into suspected cases of occupational disease; and establishing a coronial database to enable surveillance and systemic learnings from coronial findings, with a specific module for work-related deaths.
Improve recording of occupation in NZHIS data, including introducing the standard Statistics New Zealand coding system into the Mortality Collection, NZCR, and NMDS; coding the backlog of occupation using methods recently recommended in another study; including occupation in the NHI database; and opportunistically recording longitudinal occupation information on the NHI database each time a patient presents.
Extend and improve the coding of industry in NZHIS and OSH data, in order to add depth to the information captured in the occupation field, by including industry as a variable in the Mortality Collection, NZCR, and NMDS using the standard Statistics New Zealand classification system; and aligning the industry classification system in NODS/HASARD with the Statistics New Zealand system.
Collect additional information on work-relatedness, occupation history, and exposure history by developing a common definition of work-relatedness for OSH and NZHIS; a staged indicator for work-relatedness (e.g. confirmed/suspected/not); common definitions and fields for recording current occupation, usual occupation, and occupation history; common definitions and identify or develop a classification system for coding exposure history from patient case notes; and an exposures database to collect denominator information on exposures to risk factors by integrating data from a range of primary and secondary sources.

Accuracy

Improve the use of the work-related E codes in ICD-10-AM 3rd Edition.
Code work-relatedness and occupation for all ACC claims.
Ensure ACC claim data is updated on the database (e.g. following a change in diagnosis).
Review the classification and field structures in NODS/HASARD for recording occupation, industry, diagnosis, agent, and mechanism, with a view to aligning these with standard classification and coding systems and using hierarchical pick-lists for data entry.
Ensure NODS data is updated upon completion of each investigation.