Predictors of Work Injuries in Mines

1
Predictors of Work Injuries in Mines A Case
Control Study

• Dr. J. Maiti
• Assistant Professor
• Indian Institute of Technology
• Kharagpur, India

2
Outline of Presentation

• Introduction
• Objectives of the Study
• Literature Review
• Determinants of Work Injuries
• Design of Questionnaires
• General Description of Mines Studied
• Data Collection
• Applications to Mines
• Conclusions

3
Introduction

• Mining is a hazardous profession associated with
  high level of accidents, injuries and illnesses
• For example, in Indian coal mines the fatal and
  serious bodily injury rates per 1000 persons
  employed for the years 2001 and 2002 were 0.30,
  0.28 and 1.14, 1.21 respectively
• Several causes starting from personal to
  technical factors are responsible for such high
  injury experience rates in mines
• There is a critical need of study to identify
  these factors and to evaluate their effects on
  accident/injury occurrences in a multivariate
  situation

4
Objectives of the Study

• Identification of the causative factors
  associated with work injuries in mines
  representing the social, technical and personal
  characteristics of the workers.
• Evaluation of the risk of injuries to the
  underground coal mine workers, controlling their
  social, technical and personal characteristics.
• Evaluation of sequential relationships amongst
  personal, social and technical factors and work
  injuries
• Implementation of the findings to case study
  mines.

5

• Literature Review
• Selected literature on quantitative analysis of
  mine safety studies (1970-2005)

6
The Salient Geological and Mining Related
Information for the Case Study Mines
7
Production, Employment and Injury Statistics of
the Sample Mine, for the Five Year Periods,
1998-2002

1 The frequency rate of occupational injuries
is the number of injury occurrences expressed as
a rate per thousand employees. Such rates were
calculated using the following formula
Number of annual occupational injury
cases   X   1,000
Number of employees
8
Data collection

• Data were collected through accident/injury
  reports available at the mines and through a
  questionnaire survey.
• Interview was taken for individual miners from
  different categories of workers from both the
  mines (Mine 1 and Mine 2).
• Two groups namely Non-Accident Group (NAG) and
  Accident Group(AG) of workers were identified to
  study the influence of different factors
  contributing mine accident/injury amongst the
  workers.

9
Data collection (Contd.)

• For most of the mine workers who were not fluent
  in reading and writing the questions were read
  out. It took 45-60 minutes to fill up the
  questionnaire forms for an individual
  participant.
• Out of 175 participants from case group, 150
  miners answer matched the inclusion criteria of
  the study. Inclusion criteria consist of proper
  identifying information and proper response to
  each of the questionnaires.

10
Data collection (Contd.)

• Through frequency matching 150 participants were
  chosen randomly from the participants in the
  control group whose answers matched the inclusion
  criteria of the survey.
• Overall, of the 375 participants, 175 miners
  participated from case group and 200 miners
  participated from control group with an overall
  response rate of 80.

11
Reliability and Validity Test of the Collected
Data
12
Models applied for this study
Logistic Regression Structural Equation Modeling
13
Logistic Model
Description of Variables Used in Logistic
Regression Model
(RC) indicates reference category
14
Multivariate Logistic Regression Results
Predicting Work Injury
Note. For Models 1 6, standardized regression
coefficients (ß) are reported. P lt 0.10, P
? 0.05, P ? 0.01. a category (0) represents
the reference group. For example, AGE (0) is the
reference group of age variable. The parameter
(ß) for AGE (1) is estimated with reference to
AGE (0).
15
Results

• Older age group is 2.14 times more likely to be
  injured than the younger age group.
• Negatively affected workers are 2.54 times more
  prone to injuries.
• Highly job dissatisfied workers are 1.71 times
  more likely to become injured.

16
Results

• Workers who perceive higher level of physical
  hazards are 1.69 times more likely to be injured.
  
• Highly impulsive and more risk taking workers are
  1.73 and 1.44 times more likely to be injured
  (but none of them were statistically
  significant).

17
Conclusions

• The national level accident/injury statistics in
  Indian mines showed that for the last 25 years,
  there is no apparent improvement in safety in
  mines.
• This has instigated the need for studies beyond
  engineering control of safety in mines.
• Studies on the effect of sociotechnical factors
  on work injuries is present day needs for Indian
  mines.

18
Conclusion (Contd.)

• The case study results showed that the
  accident-involved workers are
• aged (OR 2.14),
• negatively affected (OR 2.54),
• more job dissatisfied (OR 1.71), and
• perceive the physical hazards more harmful (OR
  1.69)

19
Conclusions (Contd.)

• The sequential interrelationships amongst factors
  reveals the keys factors as
• social support (total effect -0.14)
• work hazards (total effect 0.15)
• safety environment (total effect -0.16)
• job dissatisfaction (total effect 0.29)

20
T
H
A
N
K
U
21
(No Transcript)
22
Methodologies applied for this study (Contd.)
Logistic Model Specification
The analysis of the data is based on logistic
regression procedure. The logistic model allows
the estimation of the probability of a coal miner
with given characteristics (e.g. age,
experience, risk taking, negative affectivity,
job stress, safety training, safety practice,
etc) that will have an accident resulting in an
injury. Following the coding scheme of the
variables mentioned in previous Table, the
logistic model is thus specified as
follows Probability of an injury ?( age,
experience, risk taking, negative affectivity,
job stress, safety training, safety
practice, etc) The logistic regression
equation for this study can be expressed as P
( X1 , X2, .........., Xk ) 1/ 1 exp?
(?0 ?1 X1 . ?k Xk) where X1, X2,
. , Xk are the variables of interest (age,
experience, .., management worker interaction
with their categories) being used to provide
P(X1, X2, ., XK), the probability of an
accident/injury in question. ?1, ?2, ?3,.?k
corresponding parameters of Xj, for j 1, 2,
, k. The ? parameters of the logistic model
were estimated by maximum likelihood method
suggested by Cox (1970).