Title of the PhD proposal:
AN INVESTIGATIONOF RISK MANAGEMENT PRACTICES FOR DRUG SHORTAGES IN THE CANADIAN PHARMACEUTICAL SUPPLY CHAIN
Research Questions:
First Question: How can Canada’s pharmaceutical supply chain be better managed to reduce the shortage of drugs? This question leads to the followings:
- How can drug shortages in Canada be predicted?
- What are the causes of drug shortage in Canada?
Second Question: What are the internal risks and external to the SC?
This research also seeks to answer these sub-questions:
- Which supply chain measures can help keep prices of prescription drugs under control in Canada?
- What is the nature of the relationships between drug shortages, overpricing and sub-standard quality in Canada’s pharmaceutical supply chain?
Research Objectives:
This research seeks to critically evaluate supply chain risk management practices and their influence on the successful performance of Canada’s pharmaceutical supply chain. The following specific objectives are pursued:
Main Objectives:
• To investigate the supply chain risks related to process caused drug shortages.
• To develop an optimization model to test the hypotheses of the research using the results of the survey tools.
• To apply an adaptation of the model to better understand the processes in the Canadian pharmaceutical SC.
• To determine the external monetary relationships with the SC drug shortages.
• To implement two small risk models to evaluate the internal and external SC risks.
Sub Objectives:
• To recommend supply chain measures that can help reduce drug shortage in Canada’s pharmaceutical supply chain.
• To investigate the relationship between drug shortage, overpricing and sub-standard quality in Canada’s pharmaceutical supply chain.
Methodology:
First Objective, I will collect the data from different pharmaceutical companies in Canada in order to determine the causes of the problem of the drug shortages, then it will be analyzed. Based on the result, the research will create a model with its constraints.
Second Objective: Deterministic Supply chain model OR mathematical programming approach:
Literature Review Introduction:
This objective aims to model the Canadia Pharmaceutical Supply Chain that will optimize the whole Supply Chain System.
Model description:
The goal is to choose the manufactures to produce each drug, and then the quantity delivered to each wholesaler to minimize the sum of production costs and inventory holding costs over a horizon of T periods. In doing so, we must meet demands on time at the end- customers.
Our problem will most likely be formulated as a linear integer programming. The name linear integer programming refers to the class of combinatorial constrained optimization problems with integer variables, where the objective function is a linear function and the constraints are linear inequalities (Genova et al. 2011).
A formulation will be as follows:
Indices:
• Drug type, are the drugs available in the market to be optimized, for instance, VALSARTAN, APO-TERAZOSIN, and IRINOTECAN,
• Manufacturing type, this index includes all the drug manufacturers like Sandoz, Apotex and Pfizer,
• Front-line user (such as hospitals and pharmacies) directly, or through group purchasing organizations, wholesalers and distributors like McKesson Canada, Distribution Pharma Plus Inc, Shoppers Drug Mart, and JCB Laboratories
• Time, is a period of time that we want to optimize. For example, a period of a year or a month.
Parameters: You need to describe the parameters of my problem???????
The data is separated by the stakeholders in the supply chain.
For each manufacturer, the following data are necessary to build the model:
• Production cost of each drug,
• Production capacity of each drug,
• Use of total capacity when manufacturing each drug,
• Production capacity of the manufacturer, and
• Estimated number of shortage of each drug.
• Working time required for each drug in every company,
For each wholesaler of the supply chain the following data are needed to solve the problem:
• Inventory holding cost per each drug,
• Storage capacity,
• Distance in time from each manufacturer and each costumer, and
• Delivery cost.
Finally, for each retailer the following parameter is needed:
• Demand of each drug,
• Inventory capacity.
Decision Variables:
• Quantity of each drug produced in each manufacturer,
• Number of each drugs delivered from each manufactured to each wholesaler,
• Number of each drugs delivered from each wholesaler to each retailer,
• Inventory remaining in manufacturer/wholesaler/retailer.
Objective function:
This work considers mainly the minimizations of drug shortage. Therefore, the demand must be satis?ed and the production in manufacturers with high shortage must be reduced.
Moreover, we can consider that each manufacturer has a different production cost (or drug cost). Thus, from a customer point of view the minimization of cost (price) is also desirable.
The main constraints:
The model of the supply chain in the Canadian pharmaceutical companies has the following restrictions:
Capacity: These kinds of constraints must consider the production and storage Capacities for all the manufacturers, wholesalers and retailers. Moreover, this constraint will assure that we respect the capacity of resources as raw materials, and time.
Balance: these constraints will guarantee inventory balance in retail and wholesalers.
Demand: These constraints will assure that the demand is satis?ed in each retail. In this part, we will consider a deterministic demand.
Solution approach:
Once our variables are mostly integer we can use a mixed integer solver like the commercial CPLEX or we can develop some exact methods like Branch-and-bound (cut), Column generation or Lagrangian methods.
Exact methods Branch and Bound (B&B) is by far the most widely used tools for solving large scale N P-hard combinatorial optimization problems. B&B algorithm seeks to a different aspects of all solutions in order to provide any problem with the optimum solution (Clausen et al. 1999).
With column generation, we can separate each sub problem by each drug or by each manufacturer. And another interesting approach is the Lagrangian approach. Lagrangian relaxation also partitions the problem, but by relaxing constraints into the objective function with a penalty for their violation (Desaulniers et al 2006).