Could a 'Default Option' Improve Cancer Care?

Zosia Chustecka

March 03, 2016

When clinicians planning cancer treatment are faced with a choice of options, a "default option" that is compliant with evidence-based guidelines could improve care and could also reduce costs, suggests a group of researchers at the University of Pennsylvania.

Outlining their proposal in an article published online February 16 in the Journal of Clinical Oncology, they give three examples in which a default option could be useful clinically.

One example involves choosing antiemetics to go with a chemotherapy regimen. At present, a physician chooses a chemotherapy regimen, and then goes on to choose various supportive care medications, which would include antinausea and antiemetic medications.

With the default scheme being proposed, the computer system would be programmed in accordance with evidence-based clinical guidelines, so that when a physician chooses a particular chemotherapy regimen, the program would offer a particular antiemetic regimen, based on published guidelines, to accompany that chemotherapy choice.

"The beauty is that it would nudge clinicians to make the choice that is based on evidence," commented lead author Eric Ojerholm, MD, from the Department of Radiation Oncology at the University of Pennsylvania, in Philadelphia.

"Defaults are a method of steering people in one direction," he explained, but he emphasized that the practitioner could reject the default and choose another option. This is important, he explained to Medscape Medical News, because this "preserves the freedom to choose otherwise or go a different way, which is appealing compared to a heavy-handed approach."

One criticism of such proposals is that they are paternalistic, in that they dictate from above how people should act. Dr Ojerholm described the use of "smart defaults" as "libertarian paternalism, in which you do steer people, but they do have the freedom to choose something else instead."

Some of these "smart defaults" need to take into account patient-specific characteristics, so "it's not a cookie-cutter or one-size-fits all default," he added.

Two More Examples

Two other examples of the use of defaults in cancer care are outlined in the article.

One concerns use of a "hidden option default" in the electronic order entry system. With this option, the system would suggest the use of a cheaper drug, such as a generic or a biosimilar agent, instead of a branded drug.

Such a system could also be used to suggest a cheaper option when several drugs are available for a particular use. As an example, the authors notethat denosumab (multiple brands) costs twice as much zoledronic acid (multiple brands). These agents are used in the prevention of skeletal-related events associated with solid tumors. Although denosumb is superior in reducing such risks, its use does not affect survival or disease progression, nor does it offer clinically meaningful benefits for quality of life or pain outcomes.

In cases such as these, a search of the system "would return only the cost-effective or guideline-concordant medication," the authors write. "The alternative drug would still be available, but only after additional data base look up."

Another example outlined in the article concerns the use of radiotherapy. A short-course hypofractionated regimen would be the default in what could be "considered as the radiotherapy version of a generic prescription."

"Makes a Lot of Sense"

An expert in choice architecture and decision research, Eric Johnson, PhD, professor of marketing at Columbia University, in New York City, told Medcape Medical News: "These are some of the best applications of defaults to decision making that I have seen."

 
These are some of the best applications of defaults to decision making that I have seen. Dr Eric Johnson
 

Dr Johnson, who is codirector for the Center for Decision Sciences at Columbia University and was approached for comment, said: "I knew a little of this work but read this review in greater detail.

"The thing I particularly like about the article is that it talks about using different defaults, not 'one size fits all,' and uses defaults (what we call smart defaults) to customize the response to the case at hand," he said.

"We know that defaults matter and that every choice has a default (usually nothing gets done), so setting the default to the standard of practice makes a lot of sense," Dr Johnson added.

 
Setting the default to the standard of practice makes a lot of sense. Dr Eric Johnson
 

Blaise Polite MD, MPP, from the University of Chicago Medical Center, who was also approached for comment, told Medscape Medical News: "In general, setting up systems that steer physicians to guideline-concordant care is the way to go."

He added that "the question of whether it should be done through strict defaults or through looser pathways which then measure adherence and provide feedback is dependent on the situation.

"For example, it makes complete sense to build chemotherapy order sets with built-in antiemetics and clearance guidelines so that the physician has to make an active choice to change them. The guidelines here are clear, and deviation is the exception rather than the rule," he commented.

However, for other situations, such as choosing a chemotherapy regimen, there is more nuance, Dr Polite said. In this case, "Physicians should be presented with a preferred pathway based on efficacy, toxicity, and cost but should have a clear path to deviate from that pathway, based on individual patient characteristics," he said. In addition, "physicians who deviate a large percentage of the time should be asked to justify this.

"Bottom line: medicine is and always will be part art and part science, which is why we will never allow, in my opinion, artificial intelligence to decide how we treat the patient in front of us," Dr Polite commented.

"However, deviating from the science should be done with clear knowledge and justification for why we are doing so," he said. "Sometimes that deviation should be made very difficult, when the science is clear and exceptions are rare. In other situations, it should be expected to happen, but we should still be monitored in how often we are doing it and for what reasons."

Next Step ― Test in Clinical Trials

In an interview, Dr Ojerholm explained that his team envisaged that default options would be incorporated into electronic health record systems. Some would be easier to insert than others, but it should be fairly easy to incorporate these defaults into the computer systems, he said.

"But first we need to test them, and test them rigorously, to check that they work the way we think they will work," he added. His team is now considering how to move forward with testing the proposals in clinical practice.

Probably the best way to test would be in a randomized trial, and the easiest example to test would be with regard to antimetics being added to chemotherapy, he said. Testing could be conducted in two clinical practices, one that uses the regular computer system, and one in which the default has been introduced. The idea would be to use both systems for a period and to then see whether the use of the default option improved adherence to clinical guidelines, he explained.

"Hopefully, this paper will spark some interest," Dr Ojerholm commented, and he hoped that other groups will be inspired to test these proposals.

Dr Ojerholm and Dr Polite have disclosed no relevant financial relationships.

J Clin Oncol. Published online February 16, 2016. Full text

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....