Biosolids case study, Drexel December 2013

1. QMRA on Municipal Solid Waste
Dumps in India
Anne Thebo
Bharathi Murali
Parnab Das
Peiying Hong
1

2. Problem Statement & Motivation
2

3. 3

4. Problem Statement: To
assess the risk due to
the microorganisms that
the rag pickers are
exposed to.
Motivation: To
determine if it is
possible to formally
legalize the rag picking
business without
Legend:
Dumping sites 1, 2, 3, 4 are near to the residential district
Dumping sites 5, 6 and 7 are located further away
Dumping site 8 is a control site (construction waste)
Assume 1 million population in this city, with 1% of
4
population being rag pickers

5. Goals of this QMRA
•
•
•
•
What is the magnitude of health risks faced by
rag pickers?
Which pathogen(s) pose the greatest risks to
rag pickers?
How does risk vary across dumping sites?
What interventions would be most
appropriate for:
1.
Mitigating health risks faced by rag pickers?
2.
5
Minimizing secondary transmission?

6. Hazard Identification
6

7. Hazard ID
•
Proteus sp.
•
Bacillus sp.
•
Klebsiella sp.
•
Salmonella sp.
•
Escherichia coli
•
Staphylococcus aureus
•
Pseudomonas aeruginosa
7

8. Hazard ID
•
Proteus sp.
•
Bacillus sp.
•
Klebsiella sp. (Klebsiella pneumoniae)
•
Salmonella sp. (S. enterica serovar Typhi)
•
Escherichia coli (Enteroinvasive E. coli)
•
Staphylococcus aureus
•
Pseudomonas aeruginosa
8

9. Exposure routes & assessment
9

10. @Risk Populations
10

11. Exposure Routes
Klebsiella pneumoniae
Salmonella Typhi
Escherichia coli (EIEC)
Staphylococcus aureus
Candida albicans
11

12. Data Provided
•
Dump Sites:
–
–
Bacteria and fungi (CFU/g)
–
•
Approximate location
Probability of transmission to rag pickers
Rag Pickers:
–
Microorganism observation frequency by site
–
Microorganism observation frequency by sample
type (stool/urine/nasal swab/sputum)
–
Blood/Immune/Liver function
12

13. Data  Exposure
Bacteria & Fungi isolated from different dump sites and control sites, CFU/g
1
2
3
4
5
6
7
-
1.2 x 106
-
1.1 x 106
-
-
0.1
1.2 x 106
1.2 x 106
1.4 x 106
1.1 x 106
1.4 x 106
1.2 x 106
1.1 x 106
-
-
-
1.1 x 106
-
-
0.2 x 106
-
-
1.9 x 106
1.3 x 106
1.1 x 106
1.2 x 106
-
1.4 x 106
1.1 x 106
1.2 x 106
1.2 x 106
-
BACTERIA/FUNGI
Candida sp.
Escherichia coli
Klebisiella sp.
Salmonella sp.
Staphylococcus aureus
2.1 x 106
1.4 x 106
1.4 x 106
1.2 x 106
8
Exposure factors from USEPA
Activities (adult male, 21-31 yrs)
Soil and dust ingestion
Hand to eye-lip-nostril contact
50
15.7
Units
mg/day
times/h
Dermal exposure - adherence of soil to face
0.024
mg/cm2/d
Dermal exposure - adherence of soil to arms
0.0379
mg/cm2/d
Dermal exposure - adherence of soil to hands 0.1595
mg/cm2/d
Dermal exposure - adherence of soil to legs
0.0189
mg/cm2/d
Dermal exposure - adherence of soil to feet
Working week
Working hours
0.1393
6
6
mg/cm2/d
d/wk
h/d
BACTERIA
/FUNGI
Candida sp.
Escherichia coli
Klebsiella sp.
Salmonella sp.
P (transmission to rag picker
from dump site)
0.002
0.004
0.002
0.001
Staphylococcus aureus
13
0.001

14. Data  Exposure
1. Ingestion dosage
Ingestion through mouth
contact
Klebsiella sp. conc in
MSW
Hands to mouth
1.1 x 106
15.7
P(transmission)
K. pneumoniae
Total per dose
0.002
0.1
4330.18
Accidental ingestion of
soil
CFU/g
times/h
Klebsiella sp. conc in
MSW
Soil and dust ingestion
Fraction
CFU
P(transmission)
K. pneumoniae
Total per dose
1.1 x 106
50
CFU/g
mg/day
0.002
0.1
11
Fraction
CFU
2. Subcutaneous dosage
Subcutaneous exposure
Klebsiella sp. conc in
MSW
Exposure
1100
0.3796
P(transmission)
K. pneumoniae
0.002
0.1
Subcutaneous dose
CFU/mg
mg/cm2
Fraction
0.83512
CFU/ cm2
14
Sum
4341.176 CFU/d

15. Dose- response assessment
(modeling methods)
15

16. Analysis Steps
1.
Evaluate data availability
2.
Identify availability of dose response curves
3.
Select “planned” model
–
Exponential or β-Poisson
4.
Estimate dermal and ingestion exposure
5.
Estimate k for Klebsiella sp. and Candida sp.
6.
Model using “planned” model
7.
16
Evaluate modeling results/appropriateness

17. Data Availability & Background
Exposure Route
Ingestion
Ingestion
Ingestion
Ingestion
Dermal
Pathogen
Escherichia coli
Salmonella sp.
Proteus sp.
Bacillus sp.
Staphylococcus aureus
Klebsiella sp.
Dermal/Ingestion (Klebsiella pneumoniae)
Candida sp.
Ingestion
(Candida albicans)
Pseudomonas
Dermal
aeruginosa
Dermal
Streptococcus sp.
Airborne
Aspergillus sp.
Observed in
study sites?
Y
Y
Y
Y
Y
Observed Have Dose
in rag
Response Have Best Fit
pickers? Relationship?
Model?
Y
Y
Y
Y
Y
Y
N
N/A
N/A
N
N/A
N/A
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
N
N
No Data
Y
Y
Y
N
Y
N
N
Y
N
N
Frequency of microorganism occurrence:
Dump Site: S. aureus = E. coli > Salmonella spp.
Rag Pickers: S. aureus > E. coli > Salmonella spp.
17

18. Modeling Approach
Pathogen
Status
Escherichia coli
enterotoxigenic
enteroinvasive
enteropathogenic
Salmonella Typhi
Staphylococcus aureus
Klebsiella pneumoniae
Candida albicans
# of Available
DR Models
Modeled
Modeled
Modeled
Modeled
Modeled
Modeled
(w/derived DR)
Modeled
(w/derived DR)
14
7
4
3
3
1
N/A
N/A
Planned Model
β-Poisson
β-Poisson
β-Poisson
β-Poisson
Exponential
Developed Exponential
Dose Response
Developed Exponential
Dose Response
Final Model
Reference
Exponential
Exponential
Exponential
Exponential
Exponential
QMRAwiki
Haas et al. 1999
Du Pont et al. 1971
Powell 2000
Hornick 1970
Rose et al. 1999
Exponential
See next slide
Exponential
See next slide
•
Evaluated available dose response models
•
Built Excel model
–
•
1000 iterations
18

19. Approximation to exponential
model
1. Gather LD50 available in published literature
2. Substitute known values to determine k
LD50
Strain KP1-O
Response k
7.20E+00 Death
Ave
Read off the
LD50 dose,
Proceed same
way as before
Stdev
9.63E-02
6.50E+00 values 1.07E-01 1.01E-01 0.007331
3. Use determined kDeath for Monte Carlo
19

20. Risk characterization
20

21. Risk Characterization: Risk across
different sites
S. Typhi
C. albicans
K. pneumoniae (ingestion)
K. pneumoniae (subcutaneous)
E. coli (EIEC)
S. aureus
•
Population at risk
1.00E+04
8.00E+03
6.00E+03
•
4.00E+03
2.00E+03
0.00E+00
1
2
3
4
5
Dumping sites
6
7
8
•
Sites 4 and 6 face risk
from >2 pathogens (i.e.,
S. Typhi, E. coli, C.
albicans & S. Typhi, E.
coli, and K.
pneumoniae)
Sites 1, 2, 3, 5 and 7
face risk from 2 types of
pathogen
Site 8 (control site) has
no known pathogenic
21

22. Risk Characterization: Risk across
different pathogens
Dumping sites/ pathogens
1
E. coli (EIEC)
6.56E-03
S. Typhi
6.07E-01
S. aureus
1.36E-05
Risk per year
(Maximum allowable risk = 10-4 or 1 in 10,000 population)
2
3
4
5
6
7
6.56E-03
8
7.63E-03 6.02E-03 7.63E-03 6.56E-03 6.02E-03
5.47E-01 4.23E-01 3.76E-01 4.15E-01
1.36E-05
1.16E-05 1.36E-05 1.07E-05 1.16E-05 1.16E-05
K. pneumoniae (ingestion)
6.37E-01
1.72E-01
K. pneumoniae (subcutaneous)
9.22E-01
3.88E-01
C. albicans
•
1.00E+00
1.00E+00
6.95E-06
Authorities should be more concerned of the
pathogens in descending priorities of:
C. albicans = K. pneumoniae = S. typhi > E. coli (EIEC) > S. aureus
22

23. Sensitivity Analysis
•
Sensitivity Analysis
–
3 strains of pathogenic of E. coli
–
3 different ratios of Pathogenic EC:EC
–
3 different concentrations of E. coli in MSW
23

24. Strains of Pathogenic E. coli
and Model Type
β-Poisson
Exponential
Strain
Path
EC:EC
Dose
N50
α
Annual
Risk
K (estimated)
Annual
Risk
Enteroinvasive
(EIEC)
0.001
96.73
2.1*106
0.155
2.6*10-3
2.2*107
5.4*10-3
Enterotoxigenic
(ETEC)
0.001
96.73
8.6*107
0.178
3.7*10-5
2.2*108
5.3*10-3
Enteropathogenic
(EPEC)
0.001
96.73
6.9*107
0.221
1.0*10-5
2.2*108
5.3*10-3
EIEC – Exponential Model
EIEC - β-Poisson Model
f(x) = NaNx
15
R² = NaN
Risk (P(I)/d)
Risk (P(I)/d)
15
10
5
0
0
2
4
6
Dose(mg/d)
8
10
12
10
5
0
0
2
4
6
Dose(mg/d)
24
8
10
12

25. Ratio of Pathogenic E. coli to E. coli
Strain
Path EC:EC
Dose
(CFU/d)
K
(estimated)
Annual
Risk
enteroinvasive
0.001
96.73
2.2*107
5.4*10-3
enteroinvasive
0.01
967.26
2.2*108
6.5*10-2
enteroinvasive
0.1
9672.60
2.2*108
1.0
f(x) = NaNx
R² = NaN
12
Risk (P(I)/d)
10
8
Linear ()
6
4
Linear ()
2
0
Linear ()
1
10
Dose (mg/d)
25

26. Study Limitations
•
Quality of input microorganism data questionable
–
•
Detection methods – described methods not sufficient for listed microorganisms
Dose response relationships
–
Some extrapolated from animal models
–
Dose-response models for three of the pathogens are not available in QMRA Wiki
(Candida, Staphylococcus aureus, Klebsiella spp.)
•
–
•
Estimated from LD50
Switch from β-Poisson to exponential model
Need to assume strains present (based on available dose-response curves)
–
–
Salmonella spp. reported
–
•
Indicator E. coli reported
Most other MO reported as genus (generic) rather than specific species
Precise dumping site locations unknown
26

27. Risk management
27

28. Intervention measures
•
Microbial risks:
Sites 4 and 6 > Sites 1, 2, 3, 5
and 7
•
Proximity to residential
district
Sites 1, 2, 3, 4 nearer than Sites
5, 6, 7 (human flux to residential
district is impeded by the
presence of a forest reserve)
•
Legend:
Dumping sites 1, 2, 3, 4 are near to the residential district
Dumping sites 5, 6 and 7 are located further away
Dumping site 8 is a control site (construction waste)
Approach 1: Encourage rag
picking activities in Sites 5
28
and 7, with a localized

29. Intervention measures
•
•
C. albicans = K. pneumoniae = S. Typhi > E. coli (EIEC) > S.
aureus
Treatment removal efficiency of following should be achieved:
K. pneumoniae
S. Typhi
C. albicans
1.30E+02
6.50E+04
2-log removal
4.60E-01
1.52E+05
5.70E-03
7.67E+04
6-log removal
7-log removal
0.00E+00 5.00E+04 1.00E+05 1.50E+05 2.00E+05
29

30. Other Management Strategies
Sites
Strategies suggested
1
●
Usage of PPE (gloves,
● shoes, mask)
Periodic immunization,
◊ medical screening
Risk communication/
□ education at basic level
Installation of a collection
ᴥ center
2
3
4
● ◊ ● ◊□ ● ◊
v
v
v
5
●
6
●□ᴥ
7
●
# 2, 3, and 4 has high annual risk due
to Candida albicans and K.
pneumoniae.
Usage of PPE at all the sites would
reduce the dermal exposure
significantly
Risk communication strategies could
minimize secondary transmission
across sub-populations
30

31. Future Work
•
Conduct additional sampling
–
–
•
•
Concentrations on rag pickers
Proportion of pathogenic strains and identity
Check modeling assumptions
Better characterization of input variable
uncertainty
•
More rigorous sensitivity analysis
•
Evaluate additional dose response models
31

32. References
•
QMRA Wiki (CAMRA)
•
USEPA Handbook of Exposure
•
•
Wachukwu C.K., Mbata C.A., Nyenke C.U.
(2010) The Health Profile and Impact
Assessment of Waste Scavengers (Rag Pickers)
in Port Harcourt, Nigeria. Journal of Applied
Sciences 10: 9168-1972.
Chandramohan A., Ravichandran C. (2010)
Solid waste, its health impairments32 role of
and

33. References
•
•
•
Wennerås C, Erling V (2004) Prevalence of
enterotoxigenic Escherichia coli-associated
diarrhoea and carrier state in the developing
world. J Health Popul Nutr 22: 370-382
Wingard J.R., Dick J.D., Merz W.G., Sandford
G.R., Saral R., Burns W.H. (1982) Differences in
virulence of clinical isolates of Candida
tropicalis and Candida albicans in
mice. Infection and Immunity 37: 833-836.
33
Cryz Jr S.J., Furer F., Germanier R (1984)