Safety in Mines Research Advisory Committee Final Project Report The development of an occupational diseases database, to be managed by the Department of Minerals and Energy

3.8 Implementation

The DME’s draft implementation plan recognises that the Director: Occupational Medicine

is identified as the main user of the system, and that the system is now developed and in

its final phase of acceptance testing.

12

The draft implementation plan (still to be refined and approved) includes the following

steps:

•

 Obtain the approval of the main user.

•

 Make final corrections to the input form to the database.

•

 Finalise the manual for the mines, currently being drawn up by the Directorate.

•

 Draw up internal procedures for the Department.

•

 Target Medical Inspectors to distribute the manual to the mines in their specific

Region.

•

 Arrange a workshop for Medical Inspectors of Regions to explain the purpose and

functioning of the database.

•

 The Medical Inspector of a Region will be the contact person who will support

mines and co-ordinate actions in connection with the database.

13

4 Background to design of the database system

4.1 Relevant Parliamentary Acts

The Acts referred to in paragraph 2, on page 8, were referenced to define and categorise

the occupational diseases that will be captured and tracked by the SAMODD database.

This is illustrated in Figure 4.1.

Figure 4.1 

Parliamentary Acts

14

4.2 Disease record contents

Figure 4.2 illustrates the data associated to a person’s disease record in the SAMODD

database. A person can have more than one of these disease records during his working

career.

Figure 4.2 

Disease record attributes

Attribute

Class of Data

Description

Person

Personal

Name, employee number, etc.

Date of Birth

Personal

Date of Birth; age can be calculated for

statistical analysis.

Disease

Disease

Disease occurrence, e.g. Noise induced

hearing loss.

Severity

Disease

Indication of morbidity of disease occurrence.

Occupation

Occupational

history

Occupation at the time the disease was

reported. The full occupational history is also

recorded.

Mine

Occupational

history

Employer at the time the disease was reported.

15

Attribute

Class of Data

Description

Number of employees

Statistical

Number of persons employed by the mine.

Mineral

Statistical

Main commodity being produced by the mine,

e.g. gold.

Number of years in

the mining industry

Statistical

Number of years the person had worked in the

mining industry at the time the disease was

reported.

16

5 Overview of use of the database system

5.1 Data sources for disease data

The flow of information from the mines to the database is illustrated in Figure 5.1. The

mines submit data entry forms to the DME, giving details of occurrences of occupational

diseases. Information will also have been submitted to one of the staturtory bodies, for

evaluation to decide whether compensation should be awarded.

Figure 5.1 

Data sources

17

5.2 Disease record data input flow

The data entry forms from the mines are processed in the DME’s offices, by the Medical

Inspectors of Mines. The flow of work for data entry is illustrated in Figure 5.2.

Figure 5.2 

Flow of work for data input

18

5.3 Subsequent analysis of disease data

As illustrated in Figure 5.3, the SAMODD data can be processed by age, occupation,

mineral, etc. The occupational disease trends analysed can then be used to influence

various aspects of the working environment.

Figure 5.3

Analysis of disease data

19

6 Discussion on design of the Database System

The full design document, as supplied to the DME’s IT Section, is attached as an

Appendix, “Design Specifications”. The following notes are less formal, and are intended

for a wider readership.

6.1 Shared SAMRASS Database Tables

Where it was found suitable, the tables in the existing SAMRASS database system were

incorporated into the occupational diseases database. These are listed in Table 6.1.

Table 6.1

Shared SAMRASS Database Tables

Table

Contents

Occupation

This table contains a list of occupations in the industry.

Region-Codes

The DME has divided the country into regions for

administrative purposes.

Mines

This is the Minerals Bureau’s list of mines

Main Commodity

The main commodity, such as gold, produced by a

mine.

Technical Manager

The owner of a mine.

Unfortunately, it was not possible to use the Person table from SAMRASS, owing to

design considerations, as discussed below.

No provision has been made for the occupational diseases system to modify data in the

SAMRASS tables

3

.

6.2 Code Lists

The code lists are tables which hold static lists to control data entry and thus to allow

some subsequent analysis of data which has been accumulated in the database. An

example is the SAMRASS list of Occupations, mentioned above. In this section, though,

only the code lists “belonging” to the current database are considered.

The code lists are specified in the Design Specifications (see Appendix); only those

related directly to diseases deserve mention here.

20

Diseases fall logically into groups, medically, and this has been found convenient for entry

of data into the database, also. The groups are shown in Table 6.2.

Table 6.2

Disease Groups

Code

Name of Group

NIHL

Noise Induced Hearing Loss

PNEU

Pneumoconioses

CRTB

Cardiorespiratory Tuberculosis

COAD

Chronic Obstructive Airways Disease

HEAT

Heat-related Illness

OTHR

Other Occupational Diseases

The first five of these groups encompass the great majority of occupational disease occur-

rences in the South African mining industry; the last one is a catch-all for the less frequent

diseases.

Selecting a suitable code for Diseases evoked much discussion. The International

Classification of Diseases (ICD10) was considered, for the purpose of classifying the

diseases within the occupational diseases database, but was found to be too

cumbersome for this purpose

4

. It was decided to try to aim for simplicity of design, by

using a simple 4-character code, as shown in the table, below.

The list of applicable diseases was taken from the Acts referred to on page 8. The first two

columns in Table 6.3 show Disease Group and Disease Code.

                                                                                                                                             

3

SAMRASS “owns” this data, and it is, in general, considered an unsafe practice to

allow different systems to modify each other’s data, as there may be undesirable

side-effects.

4

The ICD10 Classification fits a thorough clinical diagnosis. This leads to a very

complex code structure that is not suitable for the purpose of the Occupational

Diseases Database, which has disease prevention in view.

21

Table 6.3

Diseases

Group

Code

Disease Name

NIHL

NIHL

Noise Induced Hearing Loss

PNEU

SLCS

Silicosis

CWPN

Coal Workers Pneumoconiosis

ASBS

Asbestosis

HMPN

Hard Metal Pneumoconiosis

STAN

Stannosis

CRTB

CRTB

Cardiorespiratory Tuberculosis

COAD

COAD

Chronic Obstructive Airways Disease

HEAT

HEEX

Heat Exhaustion

HEST

Heat Stroke

OTHR

SITB

Silica-tuberculosis

SCLR

Scleroderma

PLAT

Platinosis

MESO

Malignant mesothelioma

PLPL

Pleural plaques or thickening

LCAN

Lung cancer

MNPO

Chronic manganese poisoning

BROB

Bronchiolitis obliterans

BARO

Barotrauma ... in miners

LEAD

Any disease ... due to lead

CADM

Any disease ... due to cadmium

CHRO

Any disease ... due to chromium

MERC

Any disease ... due to mercury

MNCY

Any disease ... due to manganese or cyanide

HVIB

Hand-arm vibration syndrome

ASFX

Asphyxiation due to carbon monoxide, etc.

The database supports the addition of other diseases within the existing groups, without

modification to the program; this is not true of disease groups, however, as each group

has its own data-entry screen.

6.3 Personal Identification Tables

The SAMRASS database does not seem to have been normalised

5

: the Person table is

really a Personal Accident table, and there is no table containing merely Person data

6

. It

                                               

5

See the Glossary.

22

was necessary, therefore, to design new tables for identifying persons in such a way as to

allow a reasonable probability of linking multiple disease records to the correct person,

bearing in mind that a person may have several identifiers, such as Passport Number and

Company Number.

This was done by using two tables: one to hold the personal data, and one to hold the

identifiers. This seemed to offer a reasonable compromise between keeping the design

simple and providing a good theoretical database design.

Five personal identifiers are accepted here, at least one of which must be provided when

a disease record is being entered:

•

 South African National ID;

•

 Passport Number;

•

 Industry Number;

•

 PF Number; and

•

 TEBA Number.

These identifiers are reasonably static. An employee’s company number is much more

changeable; it is, however, a valuable identifier, so it is recorded in the disease record,

where it can be linked with the issuing company.

Because there is no universally used identifier for a person (e.g. not every miner has an

SA National ID), it was necessary to create one for the purpose of uniquely identifying a

person in the database. This identifier is a simple sequential number, to which disease

records can be related.

6.4 Main Working Tables

6.4.1 Personal Disease Record

This records an occupational disease occurrence; it is uniquely identified by the person

(the patient), the type of disease and the date (examined or diagnosed). If the disease

recurs, a new disease record may (and should) be created, linked to the original disease

record.

                                                                                                                                             

6

This may, however, represent a conscious decision by the SAMRASS designers to

reduce the complexity of the SAMRASS system, as the keeping of personal data is

not a simple task, which the current study also illustrates.

23

A disease occurrence may be either Reportable

7

 or Compensable. (If it is Compensable, it

is automatically Reportable; if it is so mild as not even to be Reportable, one could

question whether it should really have been entered into the database.) By definition,

Compensable means that compensation has been awarded for a permanent disability,

and does not, for example, refer to compensation for loss of earnings during treatment of

tuberculosis.

Certain information is captured for later use in statistical analysis. An example of this is

the number of years the patient had been working in the mining industry when the disease

occurred.

Data may be entered in two stages:

•

 When the disease originally becomes reportable; and

•

 When the compensation submission is returned.

The reason that a disease occurrence is classed as Reportable or Compensable varies

with the disease group, and this is reflected in the data entry screens.

6.4.2 Occupational History

A patient’s occupational history is useful for analytical purposes, when investigating

problem areas needing remedial measures.

It would appear to be counter-productive to use the occupational history of patients for

apportioning levies or recovering compensation payments, as the opinion of the research

team is that this could result in lawsuits to repudiate employees’ claims, and workpeople

would thus be likely to suffer loss of legitimate compensation.

6.4.3 Annual Workforce

Denominator data is useful for analysing disease data, in order to highlight problem areas.

This consists of the average number of employees on a mine in a year, broken down by

place (underground, etc.) The number of risk workers is also recorded – these are all the

workers who have periodic medical surveillance examinations.

                                               

7

There was some debate over whether a disease should be called “Reportable” or

“Notifiable”. The word “Notifiable” is used in connection with certain non-

occupational diseases, and it was therefore rejected as having an undesirable

connotation.

24

6.5 Data Input

Printed forms have been designed for data entry. Initially, it is anticipated that these will be

used by all mines for submitting details of occupational disease occurrences to the DME.

Data-entry clerks in the DME’s offices will then capture this data into the system.

The forms have been designed to be similar to the data entry screens. Most of the data is

common to all disease groups. The data which is specific to the groups is split up on the

forms; the screen changes to show the correct data fields for the group being entered.

Later, the larger mining houses will probably acquire programs to capture the data for

electronic submission to the DME.

6.6 Data Volumes

An attempt was made to estimate the annual number of occupational disease occurrences

in the South African minng industry. The current lack of data makes such an estimate very

dubious. The DME has allowed for the capture of about 30 000 cases annually, but this

figure cannot be regarded as an authoritative indication of the number of actual cases.

6.7 Output from the Database

A number of standard reports were designed, based largely on existing SAMRASS

reports. It is anticipated that, as the data grows, reporting and analysis needs will be

recognised, and this will lead to the specification of further reports, as well as data

extracts for processing with analysis packages or spreadsheet programs.

7 Testing of Prototype System

Testing of the prototype system highlighted some problems which can be anticipated in

the final working system, especially in the data submitted to the DME.

•

 These problems revolve mainly around misunderstanding of the required data.

This showed that it is essential that the submitting bodies (usually Mines or

Compensation Commissioners) are made aware of what the different fields on the

data input forms mean. The DME is producing an instruction booklet for this

purpose.

•

 The name of the disease should not be regarded as a free-form entry; only names

from the approved list of diseases in the database may be used.

25

•

 When recording the reason that an occupational disease occurrence is classed as

Reportable, especially for the group “Other,” it must be borne in mind that clinical

and diagnostic details are not required, only a brief description telling why there is

an obligation to report the occurrence to the authorities. It might be helpful here to

refer to the case of noise induced hearing loss: the fact that there was a greater

than 15 dB hearing loss in, say, the 3kHz band would be recorded (as a simple

Yes/No); the actual amount of the loss is not required, and should therefore not be

given.

•

 The SAMODD occupational diseases system uses the same list of mines as

SAMRASS, the intention being to standardise data on accidents and disease

incidents, for administrative simplicity, as well as for statistical purposes. If the

body submitting the data is not aware of this list, mine names can be entered on

the input forms, which do not occur in the list, making data-capture difficult. Also,

the major changes which have taken place in the South African mining industry

recently and which will probably continue to take place, have two adverse effects

on the SAMRASS and SAMODD systems. Firstly, the list of mines may not be up-

to-date with the latest corporate structures, so data may be submitted for a mine

whose name is not reflected in the list. Secondly, it may be difficult to correlate

historical records when mines’ names change.

•

 The opening statement of the previous comment also applies to Occupations; it is

the SAMRASS list that must be used.

•

 Regarding confidentiality of personal data, including personal medical data, it

should, perhaps, be emphasised that normal medical ethics must apply. It is

necessary to supply some personal data in order to avoid duplication of records.

There is, in any event, a legal obligation to notify certain diseases, such as

tuberculosis, to the authorities for the protection of the community. In such a case,

the common good has to override personal interests. But, as mentioned elsewhere

in the documentation, access to the detailed data must be restricted in order to

prevent violation of personal privacy rights. The input documents must also be

handled in a confidential manner.

26

8 Conclusions and Recommendations

8.1 Conclusions

8.1.1 Outputs

The primary outputs of the project, as stated in the contract document, were to:

•

 Develop a centralised medical diseases database system, to be managed by the

Department of Minerals and Energy.

•

 The requirements for a national occupational diseases database for the

South African mining industry were analysed and documented.

•

 Develop and establish its format (also a hard-copy version), in collaboration with

the Task Group on the S A Mines Reportable Accident Statistics System

(SAMRASS) convened by the Department of Minerals and Energy.

•

 The occupational diseases database was designed to conform to the

SAMRASS database structures, also adhering to the requirements as

documented.

•

 Develop and establish computerised software (usage: interrogatory data capture

and retrieval system).

•

 The occupational diseases database was implemented on the DME’s

computer network, under a BTrieve database management system.

•

 The software was developed so that it could be maintained by the DME’s

IT Department. The MAGIC development tool was used.

•

 The software allows on-line data capture, and provides powerful data-

querying capabilities.

•

 A data format was specified, for electronic data transfers.

•

 Develop a user manual and test its suitability.

•

 Data input forms were designed, and a handbook is currently being written.

8.2 Recommendations

8

For the database system to be successful, commitment will be required from all involved

parties. A “road show” to advertise the system to the mines, organised labour and

compensation commissioners should be planned.

                                               

8

These recommendations are made at a system or management level. For

recommendations on the working of the system on a day-to-day basis, see the

paragraph: Testing of Prototype System (on page 24).

27

It will take some time—a year at least—for sufficient data to accrue for the database to

become useful. Any changes to the system during this period are likely to cloud the users’

view of the data, at a time when clarity is needed to perform a fair assessment of it. It is

therefore recommended that no significant changes be made to the database design

during the first year or so.

It is further recommended that a committee be set up, along the SAMRASS lines, to give

guidance regarding use of the data, as well as enhancements and modifications to the

system.

If there is a need to add a disease to the approved list, or to break a disease down into a

finer level of classification, this should be motivated through a responsible body, such as

the committee envisaged in the previous paragraph. In this regard, the temptation to turn

the database into a set of clinical records should be resisted, as the database has disease

prevention, rather than treatment, in view. If the classification of an occupational disease

is changed in the database, it should be because of the (environmental) cause of the

disease, not its effects, symptoms or precise clinical definition.

One point that needs to be stressed is that the system is not designed to record clinical

details of a disease occurrence, the main purpose of the data being to address the lack of

information on the prevalence of occupational diseases in the mining industry. The nature

of the data is designed to facilitate subsequent statistical analysis of masses of cases,

rather than an in-depth clinical analysis of any specific case. This is to enable corrective

measures to be prioritised and monitored.

The Gazette recognised the need for the capture of occupational hygiene data, so that

cause-effect relationships may be effectively analysed. It is therefore recommended that a

project be initiated to design and implement a compatible occupational hygiene database

system, so that the correlation between diseases and causative factors may be analysed,

with a view to improving the occupational hygienic condition of working environments

found in the mining industry, thus reducing the incidence and severity of occupational

illnesses.

28