Remdesivir, the first drug shown to be effective against the coronavirus, will be sold for $520 per vial, or $3,120 per treatment course. Drug pricing is a complex equation balancing R&D expenditure, marketing, manufacturing and distribution costs while balancing benefit delivered to patients and society.
Increasing healthcare costs and drug prices at the individual patient level and the overall economic level have been a concern for governments, insurers and payers all over the world.
In the UK, the leader of the opposition party, Jeremy Corbyn has vowed to redesign the system to serve public health — not private wealth.
“We will create a new publicly owned generic drugs manufacturer to supply cheaper medicines to our NHS”
-Jeremy Corbyn, former Labour Party Leader
President Trump in the United States has pledged to reduce drug prices
“One of my greatest priorities is to reduce the price of prescription drugs”
-Donald Trump, US President
Drug development is a long and expensive process with a high failure rate. The entire process from drug discovery to clinical trials to market access approval takes around 10 to 15 years and the reported cost of bringing new drug to the market is about $2.6 billion according to a recent study by Tufts Centre for the Study of Drug.2
With such high stakes, demonstrating true value of the product over long period of time, becomes imperative for the success for any new treatment.
The FDA in the US and EMA in the EU are responsible for clinical assessment of drugs but for cost-effectiveness and thus reimbursement decisions, the responsibility is with individual health authorities in the countries.
How these authorities are evaluating treatments and the acceptable thresholds, differ from one organisation to another.
Pharma manufacturers have to spend significant time and money in developing models for differing requirements.
Eg: for Zolgesma, the Swiss drug maker has reached coverage deals with over 20 commercial plans, Novartis is going through the medical exception process, which is more of an individualized, case-by-case decision.
Value based modelling framework can help pharmaceutical companies table this “fourth hurdle”. Value assessment frameworks are a relatively new and emerging field. Given the increasing emphasis on cost effectiveness in addition to clinical effectiveness pharmaceutical companies need to prove the value of a new intervention using probabilistic value based models that can be independently validated.
Such modelling frameworks should allow easy adaptation to the needs of various health-care decision makers, including regulators, payers, administrators and providers.
Models
Decision trees
The simplest type of model is a decision tree, which uses branches to describe the decision problem. In such models, the number of parameters to be estimated is small, however the number of assumptions required for describing a complex disease would be high.
Decision trees start with a root decision, for example, treatment 1 vs treatment 2, and extend the branches for each event or secondary decision. The sequential chance events and/or decisions are separated by chance nodes. At each chance node, probabilities of an event occurring (conditional on the previous event) determine the proportion of patients progressing down each unique path represented in the tree.
Markov Models
For modelling complex disease models and to simulate long term impact, Markov models can be considered. Markov models, define a finite number of disease states. Markov models provide more flexibility; they incorporate uncertainty and also model disease progression.
Markov model can be run as cohort or microsimulations.
· Cohort Markov models basically simulate the average experience of the patients in a cohort.
· Microsimulations models run patient-level simulations to model the transitions of an individual patient rather than a cohort. Which can be compute intenstive and time consuming.
Discreet Event Simulation
Another type of model is a discrete-event simulation (DES), in which the disease is represented as a chronological sequence of events (health states). In a DES model, the duration spent in a certain health state is directly drawn per patient from the corresponding cumulative survival distribution which has the advantage of limiting the distributions to the number of states, thus reducing computation time.
Micro-simulation Markov models and DES can include a greater level of detail as compared to cohort Markov models but they tend to consume significant compute resources and take much longer time.
Selecting the model depends on factors such as the disease itself, if disease is acute or chronic, disease progression, disease prevalence, country requirements.
Summary
The challenges faced by pharmaceutical industry in bringing innovative product to market are unprecedented, however with advances in machine learning and modelling techniques that can be developed over the cloud, thus fostering collaboration and validation, these challenges can be overcome. Pharmaceutical manufacturers need to ensure the integrity of data and validation of models being submitted to health authorities for approval. A robust yet adaptable model is a critical component of any comprehensive health economic evaluation.
References
1. CommonSpace. (2019). ‘This could be the beginning of the end of Big Pharma’: Campaigners react to Labour public drugs manufacturer policy. Available at: https://www.commonspace.scot/articles/14743/could-be-beginning-end-big-pharma-campaigners-react-labour-public-drugs-manufacturer.
2. Policymed.com. (2019). Available at: https://www.policymed.com/2014/12/a-tough-road-cost-to-develop-one-new-drug-is-26-billion-approval-rate-for-drugs-entering-clinical-de.html.
