Skip to main content


Videos

Charles Saunders: OCM Drug-Cost Targeting and Predictive Analytics

October 31, 2019


Journal of Clinical Pathways spoke with Charles Saunders, MD, chief executive officer, Integra Connect, at the Oncology Clinical Pathways Congress (October 11-13, 2019; Boston, MA) regarding strategies that practices succeeding in OCM have used to decrease drug costs at the Oncology Clinical Pathways Congress.


Strategies that practices succeeding in OCM use to drive down drug costs...

Dr Saunders: One of them is to switch from high‑cost branded pharmaceuticals to biosimilars or generics where there's one available. In cases where there are multiple branded pharmaceuticals, one is generally going to be potentially better than the other. All things being equal in terms of total cost.

I would say the most powerful opportunities in front of them are probably the emerging biosimilars. The impact on the total cost per episode can be significant. It can be in the thousands of dollars per episode overall or more. Those practices that do employ those techniques generally have better results and improvement in performance against CMS targets.

How predictive analytics can be used to identify and address high-cost patients...

Dr Saunders: In our experience, if you break your cancer population down into deciles, the first two deciles drive about 50% of the costs. Those are because of excess hospitalizations, ER visits, consumption of resources at the end of life, and high drug costs. You don't have enough case managers that you can hire to manage every patient intensively.

What you want to do is focus your precious resources on the patients who are likely to be those high utilizers, who are likely to be the ones that have complications, so that you can intervene early with them and make a difference.

For example, if somebody has a high risk of hospitalizations, it's generally going to be through neutropenian sepsis, maybe nausea, vomiting, dehydration. Those are side effects of chemotherapy.

What you want to do is to make sure that they're on appropriate supportive therapy and on bone‑marrow stimulating factors. They call you first if they get a fever because not all patients with a fever need to be admitted. If they go to the ER, they will. You can manage those patients, that they've got good home‑supportive therapy, etc. For end of life, just make sure that they're introduced appropriately early to palliative and hospice care.

We find that that's the least and most poorly managed phase of the patient journey with a very high rate of chemotherapy being delivered in the last 2 weeks of life. It gives them not only no hope of survival but also a horrible finish.

How predictive analytics can substantially decrease the time to process patient data...

Dr Saunders: If you do a randomized prospective clinical trial, first of all you'd have to dream up a study. Then you have to get it reviewed and approved by the IRB: Institutional Review Board. You've got to get your principle investigator. You've got to get the data‑collection forms and everything, the whole apparatus put in place.

Then you have to start recruiting and accumulating patients, and patients in one center. Let's say you're dealing with multiple myeloma or something. Maybe it's an uncommon disease. You might only see one of those patients once a month or once every 2 months. It takes a long time to accumulate those. If you're looking at the last 12 months, you're looking across a million cancer patients. Literally, within hours you can identify the cohort, match it, and then run the trial. 


Back to Top