I was invited to give a keynote presentation for the German languaged Epidemiology meeting which was held last week in Zurich, Switzerland. My presentation gave an overview of the decision problem in animal health and gives some examples of economic analyses that have been made at different levels of decision making. Specific items were: dry cow therapy, Q fever and BSE
2. Who am I
Born on a dairy farm (1966)
Animal science at Wageningen University
● Epidemiology (simulation model of management
regarding cystic ovaries)
● Economics (long term effects of herd health
management programs)
PhD at Fac. Veterinary Medicine (AI to diagnose mastitis)
Professor in Animal health management
In between Wageningen University and Faculty of Vet. Med.
(since 2001)
@henkhogeveen
animal-health-management.blogspot.com
3. Outline
Decision making on animal health
● The decision problem
● The levels of decision making
Some examples of analyses
● Dry cow therapy
● Q fever outbreak
● Slaughterhouse measures to reduce the BSE load
Final words
4. Economic effects of animal disease
Output
Milk
Meat
Eggs
Draft power
…….
Human
benefit
(utility)
After: McInerney, 1996
Input
Land
Labour
Capital
5. The field: Economic effects of animal
disease
Most economic work
Output
Milk
Meat
Eggs
Draft power
…….
Disease
1. Lower efficiency
2. Lower suitability for
consumption
3. Lower human well-being
Human
benefit
(utility)
After: McInerney, 1996
Input
Land
Labour
Capital
1.
2. 3.
7. The management problem
Veterinary knowledge of diseases
Consequences
animal health
Epidemiological
consequences
8. The management problem
Consequences
animal welfare
Consequences
human health
Consequences
animal health
Epidemiological
consequences
Knowledge about externalities
9. The management problem
Consequences
animal welfare
Consequences
human health
Costs of
intervention
Consequences
animal health
Epidemiological
consequences
10. Decisons become increasingly complex
Decision maker
Objectives
Available resources
Consequences
animal welfare
Consequences
human health
Costs of
intervention
Consequences
animal health
Epidemiological
consequences
11. Outline
Decision making on animal health
● The decision problem
● The levels of decision making
Some examples of analyses
● Dry cow therapy
● Q fever outbreak
● Slaughterhouse measures to reduce the BSE load
Final words
12. Levels of decision making
Individual animals
● Treatment
● Culling
● Interaction
Groups of animals (herd/farm)
● Prevention
● Eradication
Sector
● Control
● Eradication
Region
● Control
● Eradication
13. Levels of decision making
Individual animals
● Treatment
● Culling
● Interaction
Groups of animals (herd/farm)
● Prevention
● Eradication
Sector
● Control
● Eradication
Region
● Control
● Eradication
Type of disease
Production diseases
&
Endemic contagious
diseases
Contagious nofiable
diseases
14. Levels of decision making
Individual animals
● Treatment
● Culling
● Interaction
Groups of animals (herd/farm)
● Prevention
● Eradication
Sector
● Control
● Eradication
Region
● Control
● Eradication
Decision maker
Farmer, supported by
advisor
Farmer’s organisation
Processors
Government
15. Basic approach
Normative modelling
● Relate costs of intervention
with animal health and
epidemiological consequences
● Cost-benefit analysis (alternative: cost effective or
cost utility analysis)
● Assuming profit maximising behaviour of farmers
● Basis for on-farm decision support tools
Empirical modelling
● Use data to compare farms/animals/groups of
animals with and without intervention
● Experiments or existing datasets (accountancy data)
17. Outline
Decision making on animal health
● The decision problem
● The levels of decision making
Some examples of analyses
● Dry cow therapy
● Q fever outbreak
● Slaughterhouse measures to reduce the BSE load
Final words
18. Dry cow therapy
Individual cow decision
Two modes of action:
● Cure of existing (chronic) intramammary infections
● Prevention of new infections during dry period
Often herd decision (blanket dry cow therapy)
Debate on selective vs blanket dry cow therapy
19. Stochastic model (Huijps et al., 2007)
Cow as basic unit
Dynamic around dry period
Results summarized for whole herd
Accounting for differences between pathogens
Dutch circumstances
22. New discussion on
antibiotic resistance
Resistance of mastitis pathogens
● Self-interest
● No increase seen (Hogan, IDF-factsheet)
Antibiotic resistance in humans
● Externality
● Dairy cattle has very minor contribution (Oliver et al., 2011)
Decision of government
In the Netherlands (self) regulation
● Maximum amount of antibiotics to be used (< 50 %)
23. Optimizing: linear programming
(Maas, 2014, MSc thesis)
Farm level
Cows with high SCC are treated
● Primiparous > 150.000 cells/ml
● Multiparous > 250.000 cells/ml
Other cows selective
Categorized at SCC level
Optimization to minimize total costs of treatment and
mastitis around dry period
Based on: Maas, 2014, MSc thesis, in
preparation
25. Constraining antibiotic use has economic
effects
€53
€51
€49
€47
€45
€43
€41
€39
Costs per lo w SCC cow
Percentage allowed antibiotics (%)
Average farm
Low BTSCC farm
High BTSCC farm
26. Outline
Decision making on animal health
● The decision problem
● The levels of decision making
Some examples of analyses
● Dry cow therapy
● Q fever outbreak
● Slaughterhouse measures to reduce the BSE load
Final words
27. Q fever outbreak
In 2005 Coxiella burnetii diagnosed in the
Netherlands as cause of abortion problems on a
dairy goat farm
In 2007 the first Q fever outbreak in humans was
diagnosed
Since then thousands of people got infected,
which reached a climax in 2009
Year and week of notification
30. Government involved
Control measures
• Vaccination programme
• Culling of (pregnant) goats from infected
farms
• Animal movement restrictions
• Breeding ban
• Bulk milk monitoring -> no good confirmation
• Extra hygiene programmes
Around 62,500 dairy goats were culled
significant drop in milk production
31. Economic impact (Gonggrijp et al., 2014)
How large was the negative economic impact for
affected farmers?
Were other actors of the industry also negatively
affected by the control measures?
Were the relations of the actors and their
behaviour in the industry still the same?
Objective:
Study the impact of Q fever control measures on
the Dutch dairy goat industry with the use of a
quantified value chain analysis
32. Value chain analysis
Mapping the value chain
Governance in the value chain
Upgrading in the value chain
Distribution of value in the value chain
33. Value chain analysis
Information on the structure, the trade flows and all
the relations between the involved actors of a
livestock sector
Often qualitative and descriptive
In this value chain analysis focus on quantification
37. Gross margins of the Dutch dairy goat
industry in 2009 - 2010
100
90
80
70
60
50
40
30
20
10
0
Goat farmers Milk collectors Prim. dairy
processors
Second. dairy
processors
(Feed) suppliers Meat
processing
Retail Total
2009
2010
Euro (€) x 10⁶
38. Gross margin results
Decrease of gross margins in 2010 of the total
industry of -12% and -23% for farmers compared
to 2009
Enormous difference in decrease between affected
farmers (-53%) and non-affected farmers (-12%)
Primary dairy processors, meat processing and
retail not negatively affected
39. Outline
Decision making on animal health
● The decision problem
● The levels of decision making
Some examples of analyses
● Dry cow therapy
● Q fever outbreak
● Slaughterhouse measures to reduce the BSE
load
Final words
40. BSE
1986 first described
1996 -> link with Creutzveldt Jacobs Disease (vCJD)
Since August 1989 measures against BSE in the
Netherlands
● Since 1990 feed ban (no animal protein)
● Since 2000 dead cattle older than 30 m tested
● Since 2001 slaughtered cattle older than 30 m
tested
● Disposal of BSE risk materials
● Culling of cohort of detected animal
Incidence of BSE is decreasing
41. Are preventive measures cost-effective?
(Benedictus et al., 2009)
Simulation modelling
● Static
● Stochastic
● Simulation
Monte carlo model
● 1 iteration = 1 year
● Baseline: no intervention
● Alternative: one or more interventions
42. Model
3 types of BSE
● Clinically affected
● Test detectable
● Non detectable (3 for every detectable)
Per BSE type of BSE load (from different organs) of the
food supply was calculated
Based on Infectious doses, risk of vCJD
Prevented case of vCJD -> life years saved (most likely
51)
Comparison: do nothing vs intervention
43. Costs
Removal of specific risk material (~60 kg): €/kg
slaughtered weight
Transport of specific risk material
Post mortem testing: € 90 per head
Costs of cohort culling
44. Results - retrospective
Year 2002 2005
Number of BSE cases (total, at slaughter) 24, 12 3, 2
BSE load of the food supply Mean 5th – 95th Mean 5th – 95th.
Baseline scenario 34,857 30,213-39,602 5,502 3,592-7,620
SRM removal 2,330 2,020-2,648 368 240-509
Post-mortem testing (PMT) 7,455 4,846-10,306 939 198-2,091
PMT and cohort culling 7,059 4,505-9,865 939 197-2088
SRM removal and PMT 498 324-689 63 13-140
SRM removal and PMT and cohort culling 472 301-659 63 13-139
Food risk (life years lost) Mean 5th – 95tb Mean 5th – 95th pct.
Baseline scenario 16.98 8.66-26.70 2.69 1.25-4.61
SRM removal 1.14 0.58-1.79 0.18 0.08-0.31
Post-mortem testing (PMT) 3.63 1.67-6.27 0.46 0.08-1.11
PMT and cohort culling 3.44 1.56-5.94 0.46 0.08-1.11
SRM removal and PMT 0.24 0.11-0.42 0.03 0.005-0.07
SRM removal and PMT and cohort culling 0.23 0.10-0.40 0.03 0.005-0.07
48. Outline
Decision making on animal health
● The decision problem
● The levels of decision making
Some examples of analyses
● Dry cow therapy
● Q fever outbreak
● Slaughterhouse measures to reduce the BSE load
Final words
49. Take home message
Animal health management decisions are taken daily
Economics are useful/necessary to support decisions
A first step are “cost of disease” studies
● General interest
● Supporting stakeholders (negotiations)
● Start for “economics of intervention” studies
Cost-effectivity, cost-utility and cost-benefit
Choose appropriate method for level of decision making
More importantly: choose appriate approach in model:
animal vs farm vs sector vs society
Combine economic modelling knowledge with domain knowledge