Is It a Recommendation or a Suggestion?

One of the most valuable jobs I held following fellowship was working as a full-time Deputy Editor at UpToDate. My “territory” was breast, gynecologic, and genitourinary oncology, and I helped launch cancer survivorship and palliative care. I also learned to really and critically read the literature, and how to summarize it quickly so that my audience—whether it be colleagues or my own patients—could understand what we learned, and the limitations of those results.

At UpToDate, we used terms with purpose—especially the words  “recommend” and “suggest”. At UpToDate, only those findings backed by the strongest evidence and results were called recommendations; these were meant to be interpreted as standards of care. So, when I wrote a topic on HER2-positive breast cancer, there was overwhelming evidence that trastuzumab was useful. Therefore, “For women with newly diagnosed HER2 positive breast cancer, we recommend adjuvant trastuzumab.”

In the absence of high-quality data we stayed away from “recommendations”. Where there was some evidence to point towards one course of action over another, yet a recognition that different clinicians may come to different conclusion, we strove to point out a reasonable “suggestions”. As an example, in treating ovarian cancer, I would write: “For women with recurrent platinum-resistant ovarian cancer, we suggest liposomal doxorubicin rather than paclitaxel.”

I think the lesson from UpToDate is especially relevant today, where new treatments are emerging, and patients are faced with different options on how to approach their own cancer. As they seek second opinions, it’s entirely common for them to walk away with a treatment plan from one provider that differs from the one they received previously. I believe the job of one who delivers a second opinion is to provide it honestly, based on their own review and on what they believe to be the best course of action. However, when one rendering a second opinion does so without acknowledging possible differences in an approach—that is, “suggesting” rather than “recommending” one course over another—a patient may be left with more confusion rather than clarity, and perhaps worse, the fear that someone is wrong.

Such was the case that occurred early on in my life as an attending as I left the halls of Memorial Sloan-Kettering (where I was frequently offering second opinions) to Women & Infants Hospital (where I often referred elsewhere for them). Mrs. Jones* had seen me after her surgery for a new breast cancer. She had been healthy all her life, and this diagnosis at the age of 42 was a shock to her. We talked about her pathology and I sought to reassure her—her lymph nodes were not involved, despite the size of the tumor (2.5cm). It was also hormone-receptor positive and did not express HER2. “I think overall, you’re going to be okay.” However, given her age and the tumor size, I felt it would be reasonable to proceed with adjuvant chemotherapy and she agreed (at that point, a Recurrence Score was not yet widely utilized—yup, dating myself again).

“I want to do everything I can to never experience this again,” she said.

With that, we discussed options for treatment and based on her node-negativity, I suggested docetaxel and cyclophosphamide (TC) rather than the use of an anthracycline-based regimen. “There are data that four cycles would be as good, and I could spare you the risk of heart toxicity.”

“That sounds good to me,” she said, “but I’d like to get a second opinion.”

“Absolutely!” I said. I suggested folks in nearby comprehensive cancer centers and we made arrangements to meet after her consult.

A few weeks later she called my office. “Hi, Dr. Dizon,” she began. “I’ve decided to transfer my care. The doctor I saw in second opinion told me that no one would give me TC, and that the advice was misguided. He explained to me that doxorubicin was absolutely necessary and that if you wouldn’t give it to me, that I should go elsewhere. It really scared me and it’s my life on the line. I will be getting doxorubicin with him.”

“I am so sorry to hear this,” I said, a bit stunned. “At the end of the day, you have to feel comfortable with not only the oncologist you see, but the treatment you are getting. I think TC would be a reasonable option, but as we discussed, so is an anthracycline. Thanks for letting me know what you were told, though. If you ever need me in the future, I am here to help.”

With that, we hung up, but I was so flustered by what she told me that I reached out to the oncologist she had seen. If I was wrong about my recommendation, I needed to understand exactly why. He explained that he had not told her I was wrong but that the standard of care at his institution was to still utilize an anthracycline. We discussed the evidence supporting both options and why I had “recommended” TC. “Look,” he said, “we both know the data and we both know each is acceptable. I fall into the camp that has not adopted TC as an alternative, that’s all. But I can assure you, at no point did I say you were wrong.”

As we hung up, I had no reason to doubt his sincerity. We had known each other for many years, and we had a mutual respect for each other. It occurred to me then that what we say can be interpreted in different ways—that when suggestions diverge between a primary and secondary opinion, patients can be left trying to understand why, and in this context, what is not said can be as important as what is, and by not directly acknowledging clinical equipoise (i.e., that there is no definite right answer), a reasonable patient can make the conclusion that one of the opinions is wrong.

Following this experience, I have made it a point to acknowledge the presence of alternative approaches when they exist, and to ensure that patients walk away understanding when my approach is a recommendation based on the best data we have, and when it’s a suggestion, where reasonable clinicians may prefer to treat the same condition differently. I think I owe it to patients to be clear, but as important, I owe it to colleagues not to feel their professional opinions are less important than my own.

*Name changed to protect patient privacy.  

For more on clinical equipoise, see the earlier post, "On Patient Care and Equipoise as SWOG Meets in San Francisco."