'How Long Do I Have?' New Online Tool for Patients With Cancer

Roxanne Nelson, RN, BSN

February 03, 2020

Patients with cancer often want to know 'How long do I have?' Now they can take a look themselves at a newly launched and free online prognosis tool, CancerSurvivalRates.com.

It provides patients with a simple tool that can be customized according to their cancer type and specific characteristics such as grade and stage, as well as individual characteristics such as their age and gender.

Stephen Buck, CEO and cofounder of Courage Health, the company behind CancerSurvivalRates.com, emphasized that the goal of this tool is not only to provide information about prognosis, but to encourage patients to speak with their physician.

"We want the patient to do the research for the purpose of having a conversation the doctor," he told Medscape Medical News. "These are just statistics, based on their profile, but it provides the basis for further discussion."

Buck explained that it was developed as an aid for patients. "Once you hear something shocking like a cancer diagnosis, you can really be overwhelmed," he said. "And a lot of studies have shown that patients often don't understand what they are told about a prognosis. The physician, for example, may use language that may be hard for patients to understand."

Courage Health uses the same data as the American Cancer Society, the American Society of Clinical Oncology, and the National Cancer Institute (NCI), he said.

"The NCI does provide information on their website through the SEER [Surveillance, Epidemiology, and End Results] Explorer tool, but it's tricky for an inexperienced user," said Buck. "In this regard, our tool — using the same underlying data — is very similar. We just have an easier-to-use interface plus questions to ask your doctor."

In addition, this tool is a little more customizable, Buck added. "Someone with colon cancer, for example, may have a 5-year survival rate of 75% according to some data, but that might include all ages and grades. We try to tailor it a little to make it more exact."

So What Do Oncologists Think?

Medscape reached out to oncologists to gauge their reactions to the new online tool.

"I believe if you ask 50 oncologists and 50 lay people about the utility of the tool, you'll get a variety of different responses," said Ramy Sedhom, MD, an oncology fellow at Johns Hopkins Medicine in Baltimore, Maryland. He emphasized that no algorithm, or a clinician's best judgement, will be correct all of the time.  

"This is especially true considering that, as a society, we have not done our part to account for the variables that influence patient outcomes, including a lack of equity," he said. "However, I believe the training of oncologists is changing and allowing flexibility when it comes to how patient information is shared. And with the explosion of social media, Twitter, and patient advocates, there needs to be an openness to utilizing novel ways to share important information."

Sedhom said the new website provides information that is more user-friendly than other sites, as it allows for considering different variables, such as hormone receptor positivity. "Of course, the average person does not know that these statistics are meant for a population level and are not applicable for individual patients," he noted.

"With that being said, the tool is an excellent framework for a discussion about potential outcomes to allow a bigger conversation between patient and oncologist about prognosis, goals of therapy, goals of the individual patient, and potential outcomes," he said. "What I especially appreciate is the nod to palliative care, which is often taboo."

The American Cancer Society pointed out that it offers prognosis information on its own website. "The intention of the survival rate data we offer on our website and in our annual publications is to provide information about cancer survival at the population level, rather than provide data for the individual cancer patient, whose prognosis involves not only stage, grade, and age, but many other factors as well," the ACS said. "It is important to note that we also publish research studies that include more detailed survival analyses with additional prognostic factors."

While cancer statistics are available in many other places, this new website "represents the best, simplest tool that can be used by patients and providers alike," says Andrew Bruggeman, MD, an assistant professor of radiation oncology and section chief of the Radiation Medicine and Applied Sciences Palliative Care Service at the University of California, San Diego.

"This site has a simple, easy to use interface and has data for several different types of cancer," said Bruggeman, who served as an adviser on the CancerSurvivalRates.com project. "The data is highly reliable when looking at large numbers of patients with a certain type of cancer."

"If there are providers who are leery of patients using this website out of concern for the reliability of the data source, they should rest assured that the data comes from SEER data, so the models are based upon reliable, real-world data," he told Medscape Medical News.  

Having said that, he emphasized that determining an individual patient's prognosis is much more nuanced than what can be captured in a single website. "There are no perfect tools for determining an individual patient's prognosis and that is a limitation of this and any other website or data source," he added.

[It] will be useful to both patients and clinicians. Dr S. Vincent Rajkumar, professor of medicine, Mayo Clinic

S. Vincent Rajkumar, MD, professor of medicine at the Mayo Clinic in Rochester, Minnesota, agreed that this website is very easy to navigate and is convenient for obtaining prognostic information. "It is well referenced and will be useful to both patients and clinicians," he said.  

Elsewhere on the Internet, the "rates provided can vary widely…I checked out the myeloma rates with various parameters, and they are accurate, said Rajkumar, a myeloma specialist.

The site is intended for both for patients and physicians, and Rajkumar thinks that oncologists will find it useful as well. "And as they use it more, they may be more receptive," he said.

However, the problem with survival rates is interpretation, and he agrees that patients may not realize that these estimates are for populations. "No one can predict that for an individual patient," he said. "But it gives an idea, and that is useful for reassurance and for planning. Another factor is that the numbers depend on data that is a few years old. Outcomes with current therapy maybe better, and that should offer hope."

May Help Toward Having a "Better Conversation"

The simplicity of this tool may help oncologists put prognosis into better perspective when having that discussion with patients, commented Alex R. Menter, MD, an oncologist with Kaiser Permanente in Lone Tree, Colorado. He added that he has "played" with the NCI calculator in the past, but found it cumbersome to use.

"The version I used previously looked at overall mortality based on comorbidities, but did not look at cancer-specific mortality," he said. "I love this calculator because it allows me to share with my patients some real-world numbers and give them a rough idea of what they can expect with their cancer."

Menter noted that in practice, he finds it particularly difficult to discuss prognosis with older patients, as survival rates are frequently based on estimates from clinical trials that typically enroll younger patients with better prognoses. This tool includes SEER data for actuarial outcomes for patients with different cancers, which can give patients a better real-world estimate of what could happen with their diagnosis and at their age, he said.

"I also appreciate that I can pull it up quickly on the computer in the room or on my phone if needed," he said. "This does not take into account some subsets of cancer that may have a particular targeted therapy or a much better prognosis, but I can adjust my discussion with the patient based on their tumor characteristics."

"I typically try and describe best case, worst case, and average scenarios, and these points and estimates help facilitate that discussion and understanding," he added.

Thomas LeBlanc, MD, an associate professor of medicine at Duke University School of Medicine in Durham, North Carolina, said that, although there is a plethora of information available on the Internet, it can still be incredibly difficult for patients and families to know what might apply to them. Thus, a tool that allows some personalization is helpful.   

However, the more important issue is that bulk of available data online does not account for the problem of "numeracy," he emphasized.

"Much like 'health literacy,' the concept of 'numeracy' relates to the functional understanding of numerical information and ability to apply it to one's life and decisions," LeBlanc said. "Evidence clearly shows that the average American layperson is not functionally numerate, meaning they aren't able to functionally understand, for example, what a '50% chance of remission' might actually mean for them."

Most of the data available about cancer outcomes are presented in exactly this way, and unfortunately clinicians are complicit in this problem, LeBlanc said. "We talk a lot, and we think we're doing a great job of conveying complex information, and often we're actually overwhelming patients and families, while presenting too much data and doing so in a way that isn't actually meaningfully accessible and actionable for those without high numeracy," he said.

"So, a more numeracy-friendly tool that is personalized and patient-facing is really a very welcomed advance in cancer communication," he said.

LeBlanc pointed out that multiple, high-quality studies have clearly shown that many patients with advanced incurable cancers still think their disease could be cured.

"So there's a huge gap here between what clinicians know and are trying convey, vs what patients and families hear/internalize from our encounters," he added.

"If we just continue doing what we have been doing, this gap will remain. If we figure out how to use a novel tool like this one, we might be able to more effectively help patients and families understand the likely outcomes, and have better conversations about goals, values, and preferences for treatment when they better understand the background and context for decision making."

What's My Prognosis?

The tool provides cancer survival rates for 30 types of cancer, and is available completely free of charge. The algorithm takes into account age, stage, grade, time since diagnosis, histology, and additional factors such as the Gleason score in prostate cancer. The 1-, 2-, and 5-year survival rates can then be calculated.

"Our models are Cox Proportional Hazard models, a very standard and widely accepted technique within academic literature for producing population-based survival rates," explained Buck, the Courage Health CEO. Both the modeling techniques and model output underwent expert review. "There is always some room for subjectivity in how certain things are done, but we feel strongly that our methods are in line with acceptable practices in developing survival rate models."

All statistical models use SEER data from 2004-2015 and survival rates are cause-specific, not all-cause mortality.

"This website will paint with a broad brush to help give patients an idea of 1-, 2-, and 5-year survival for 'X' type of cancer," said Bruggeman. "This is information that patients often want to know but don't always get from their providers or know how to find. It is important for them to then follow up with their individual provider to help take that information and then discuss why their individual prognosis may be similar, better, or worse."

Moving Forward

The tool is a work in progress, and is being tweaked as feedback comes in.

"Free, constructive advice for makers of this website" was offered up in a Twitter post from Vinay Prasad, MD, a practicing hematologist-oncologist and associate professor of medicine at Oregon Health and Science University, and also a Medscape contributor.  

He suggested that "instead of 5-year survival, show the 20th and 80th percentile of survival (or 15th and 85th) to provide a range of outcomes most people experience," and added that it may be useful to also "provide the stat with and without therapy."

Buck told Medscape Medical News that he thought these suggestions were "a fantastic idea to explore."

"The treatment element is certainly worth exploring in a future phase," he said. "Difficult but useful."

Paulo Nassar, MD, PhD, an intensivist at the A.C. Camargo Cancer Center in Sao Paulo, Brazil, commented that he is "always discussing prognosis with oncologists" and that he has sent the tool to all his intensive care unit colleagues, adding that he intends to use the tool for patient cases.

Nassar also had some ideas for improvements. "If I could suggest anything, it would be to include a range — maybe something like CI [confidence interval] 95% — for the survival estimates, although I'm not sure it would help patients, but certainly doctors would be helped."

Buck told Medscape Medical News that he and his team of developers have already had conversations with three major national insurers/health systems about how to integrate into patient- and physician-facing tools. "All have expressed interest in working together," he said. "And based on feedback from oncologists, we're adding additional statistical details — the percentiles for survival for the healthcare professional version only. We should have this ready soon."

Given that the tool only launched at the end of 2019, feedback from patients is still limited. Still, Buck noted, it has been interesting so far. "The number one email question/complaint we get is 'Why isn't my type of cancer included?' " he said. "These cancers usually pertain to brain, gallbladder, appendix, duodenal, and other leukemias."

There are two main reasons for not including them, at least for now. One is that the complexity and the number of individual factors involved are too numerous to provide useful comparisons. The second is the sample size. "Some of these types have less than 1000 patients a year," he said, "so building a model with multiple variables would have too much variability in survival rate estimates for subgroups."

For more from Medscape Oncology, join us on Twitter and Facebook


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.