Supply Chain Coordination and Influenza Vaccination David SimchiLevi

Supply Chain Coordination and Influenza Vaccination David Simchi-Levi Massachusetts Institute of Technology • Joint work with Stephen E. Chick (INSEAD) and Hamed Mamani (MIT) March 2007

The Influenza Pandemic ¡ Globally, annual influenza outbreaks result in 250, 000 to 500, 000 deaths l 20, 000 deaths and 100, 000 hospitalizations in the US Social costs of influenza vary between $1 M-$6 M per 100, 000 inhabitant yearly in industrialized countries ¡ The “Spanish flu” (H 1 N 1) of 1918 killed 20– 40 million people worldwide ¡ Source: Report by the World Health Organization, 2005

Influenza Vaccination ¡ Vaccination is a principal tool for controlling influenza l l ¡ Reduces the risk of infection to exposed individuals (Longini et al. , 1978) Reduces the probability of transmission from a vaccinated individual infected with influenza (Longini et al. , 1978) Vaccination is cost effective l l Immunization in elderly saved $117 person in medical costs (Nichol et al 1994) Systematic children vaccination can result in significant population-wide benefits (Weycker at al 2005)

The Production and Delivery Process Flu season Growing viruses in millions of fertilized eggs Northern hemisphere Immunity takes About 2 weeks

Flu vaccine supply chain challenges ¡ Operational challenges l Beginning of the value chain ¡ l End of the value chain ¡ l Strain selection Vaccine allocation and delivery logistics Middle of the value chain ¡ Align incentives of the different parties involved

Flu vaccine challenges ¡ Change of virus over time l Antigenic drift ¡ l Seasonal epidemics Antigenic shift ¡ ¡ strain selection Global pandemics Wu et al. develop an optimization model of antigenic changes l l Current vaccination policy is reasonably effective Develop some heuristics to improve selection process

Flu vaccine supply chain challenges ¡ Operational challenges l Beginning of the value chain ¡ l End of the value chain ¡ l Strain selection Vaccine allocation and delivery logistics Middle of the value chain ¡ Align incentives of the different parties involved

Influenza vaccine challenges allocation and delivery ¡ Vaccine allocation to different subpopulations l l (Hill and Longini 2003): mathematical model of optimally allocating vaccine to different subpopulations (Weycker et al 2005): stochastic simulation model to illustrate the benefit of vaccination of certain individuals (children)

Flu vaccine supply chain challenges ¡ Operational challenges l Beginning of the value chain ¡ l End of the value chain ¡ l Strain selection Vaccine allocation and delivery logistics Middle of the value chain ¡ Align incentives of the different parties involved

The different players and their objectives ¡ Governments (CDC in US), State health departments l Balance the public health benefits and the vaccination program costs Focus on high-risk individuals. ¡ In the US, in 1999, 66. 9% of individuals of age 65 and older were vaccinated (GAO -2001). ¡

The different players and their objectives ¡ Manufacturer l l Production volume and the need for profitability Highly uncertain production yield due to biological nature of production process ¡ Considerable shortage of flu vaccination in 2000 -01. According to the US GAO l l ¡ Considerable shortage in 2003 -04 l ¡ Unanticipated problems growing the new influenza strains Quality control issues raised by FDA Early break of the epidemic Significant shortage in 2004 -05 l Chiron’s manufacturing plant in the U. K. was shut down due to bacterial contamination

Research on Supply Contracts ¡ Focus on supply chain with l l ¡ Single supplier and single retailer Order Quantities; Production levels Coordinating contracts l l l Global optimization Nash equilibrium Flexible

Supply Contracts Fixed Production Cost =$100, 000 Variable Production Cost=$35 Wholesale Price =$80 Selling Price=$125 Salvage Value=$20 Manufacturer DC Retail DC Stores

Demand Scenarios

Distributor Expected Profit

Distributor Expected Profit

Supply Contracts (cont. ) Distributor optimal order quantity is 12, 000 units ¡ Distributor expected profit is $470, 000 ¡ Manufacturer profit is $440, 000 ¡ Supply Chain Profit is $910, 000 ¡ –IS there anything that the distributor and manufacturer can do to increase the profit of both?

Supply Contracts Fixed Production Cost =$100, 000 Variable Production Cost=$35 Wholesale Price =$80 Selling Price=$125 Salvage Value=$20 Manufacturer DC Retail DC Stores

Retailer Profit (Buy Back=$55)

Retailer Profit (Buy Back=$55) $513, 800

Manufacturer Profit (Buy Back=$55)

Manufacturer Profit (Buy Back=$55) $471, 900

Industrial supply chains ¡ Supply contracts: l l l Wholesale price Buyback Revenue sharing Options … Linear cost models ¡ Deterministic production ¡

Flu vaccine supply chain features nonlinear cost function • Nonlinear effect of infection dynamics Nonlinear cost value

Flu vaccine supply chain features uncertain production ¡ Inactivated virus vaccine l eleven day old embryonated eggs ¡ l Prediction of number of eggs well in advance egg yields are stochastic based on the strain and eggs ¡ Uncertain production yield

Introduction Epidemic Modeling Industrial Supply Chains

Outline ¡ Model description ¡ Current challenges ¡ Effective supply contracts

Infection dynamics ¡ Key components in epidemic modeling l l Initial infected fraction introduced to the population (I 0) Basic reproduction number (R 0): expected number of secondary infections caused by one infected in an otherwise susceptible, unvaccinated population

Infection dynamics ¡ Vaccine role: l Decreases the probability of infection for a susceptible person by Φ ¡ l Probability of getting the infection will be multiplied by 1 - Φ If fraction f of population vaccinated R 0 decreases to Rf ¡ If Rf ≤ 1 outbreak is prevented ¡ Critical vaccine fraction: f 0 = min { f : Rf ≤ 1} ¡ l

Infection dynamics f 0

Supply chain costs ¡ Social costs of the disease l ¡ ¡ On The Counter meds (OTC) Outpatient visit Hospitalization Indirect costs: ¡ work days loss Vaccination costs l Direct costs: ¡ l ¡ Vaccine purchase Administrative costs Production costs

Model Description Government & Healthcare provider Manufacturer

Model description assumptions A single manufacturer ¡ Homogeneous population ¡ Perfect information ¡ Government is the purchaser of vaccine ¡ l determines how many people to vaccinate

Game Setting Government & Healthcare provider Manufacturer

Model Description ¡ system problem System setting l l l Ignores the transaction between the different parties Optimizes the system wide cost Might not be beneficial for one of the parties

Model Description system problem Government & Healthcare provider Manufacturer

Game Setting vs. System Setting (convex case) Assumption: Manufacturer under produces production risk Potential Insufficient order by the government

Supply chain coordination ¡ Wholesale price contract: l ¡ supply contracts Proposition: There is no wholesale price contract that coordinates the supply chain Payback contract: l l Government agrees to buy any excess production, beyond the desired volume Shifts some of the risk of excess production from the manufacturer Proposition: There is no Payback price contract that coordinates the supply chain Problem: Payback contracts are based on the manufacturer output not on its effort

Supply chain coordination cost sharing + (convex case) wholesale discount ¡ Wholesale discount / cost sharing contract: l l ¡ Incentive for government to order more ¡ Wholesale discount pr(f) Incentive for manufacturer to produce more ¡ Cost share pe(f) Theorem: The contract defined above coordinates the supply chain: l l The optimal government action is f S while the manufacturer production volume is n. Es The contract is flexible, that is, it allows any split of the cost benefit within a certain range

Summary Uncertain production yield is an important reason for insufficient supply of vaccine ¡ Cost sharing contracts can have a major impact on the influenza vaccination supply chain ¡ Production risk taken by the manufacturers maybe the reason why only a small number of manufacturer exists ¡