You Are Not a Number: Explaining the Statistics of a Cancer Diagnosis

A cancer diagnosis comes with a lot of questions. Our worrying minds may crave hard facts in the face of uncertainty. Sometimes, we try to answer these questions with statistics from research studies. But when taken out of context, these statistics can give us a false understanding of our own circumstances. 

Here, we demystify some common misconceptions about what cancer statistics mean — and what they don’t mean. This way, you can make better decisions about your care and your future. 

1) Statistics apply to populations, not individuals 

It’s important to remember that statistics are designed to describe populations. So when we apply those numbers to individuals, it leads to assumptions that aren’t always true. 

Anytime you read a cancer statistic (like survival, recurrence rate, or mortality), there are two key concepts to keep in mind:

1. You’re not always the same as the group that best characterizes you. 

2. Many times, statistics come from a group that doesn’t actually represent you at all.

Let’s explain these concepts in a little more detail. 

In a study, researchers will collect a sample of people from the group they’re trying to understand. Let’s say a researcher wants to know if there’s a link between someone’s height and how much pizza they eat. And they find that tall people eat more pizza than short people. Now say we want to use this data to compare two people. Does this mean the taller person is going to eat more pizza than the shorter person? Not necessarily. Because we can't assume an individual is going to behave the same way as their corresponding group.   

What’s more, a lot of times we read a statistic without realizing the study population is actually quite different from us. So if the population for our pizza study was only people from 20 to 30 years old, we can’t really apply those results to someone who is 50 years old. 

This scenario can be particularly common in cancer research. Clinical trials often do not include a variety of different people. For example, the following groups are commonly underrepresented:

• Underserved populations, like Black and Hispanic people

• People who are older

• People who live in rural areas

• People who are poor

Experts are increasingly aware of this problem, and they’re trying to be more inclusive in their clinical trials. But this change is slow.

2) Cancer research is fast, but study timelines are slow

Cancer is one of the most quickly evolving research fields in medicine. It seems like there’s a new study or treatment available every day. But it takes a long time to develop a study, collect all the data, then publish the results. And so today’s published statistics reflect yesterday’s medical technology. 

Here is a good example. This study asked the question, “Do mammograms improve breast cancer mortality?” The study was published in 2014. But data collection began in 1980. So if you use this study to decide whether to get a mammogram, you’re basing your decision on technology from 20 years ago. (And mammograms have improved a lot since then!) 

3) Measured outcomes are not necessarily the ones that matter most to you 

It can be hard to predict which treatment option will be the best for you. This is because many studies look at outcomes that aren’t meaningful to the patient. Researchers often study how a treatment affects the cancer — but not the individual. 

Researchers use “outcomes” to study a treatment. Outcomes are the end results that determine if a treatment is effective. Let's say you wanted to study a new cancer drug. You could choose several different outcomes to measure:

• Change in tumor size

• Blood tests (“tumor markers”) that show how many cancer cells remain in the body

• Mortality rate in people receiving the drug

But the measured outcome may not be the one you actually care about. What good is a cancer drug that shrinks a tumor but doesn’t improve survival? Especially if that drug makes people sicker than they would be with the current treatment.

If you’re reading about treatment outcomes, there are a few things to look out for:

• Researchers choose outcomes that are quick and easy to measure. And something like tumor size is easier to measure than cure rate. Even though the cure rate is more meaningful to a patient.

• Statistics can tell us whether a treatment makes a difference. But it doesn't always tell us if that difference is important. Let's say a cancer drug “significantly reduced tumor size” in the treatment group. But that reduction may be only a small fraction of a centimeter. And it may have no impact on survival.

• Many studies do not consider how well a patient can tolerate a treatment. So a treatment may be effective. But it also may come with worse side effects. 

So outcome measures do not always align with a person’s priorities.

4) Survival statistics aren’t the whole picture 

After someone receives a cancer diagnosis, they want to understand their prognosis. This often brings people to reports of survival and mortality rates. But these numbers can often be misleading. And survival data can bring its own specific pitfalls.

Problems with mortality rates

Mortality rates often include deaths from cancer — plus death from any other cause. And a one size-fits-all approach isn’t helpful to someone who is trying to understand their risk of death. Mortality rates can ignore other important factors that contribute to death: 

• Age at the time of diagnosis

• Cancer stage at the time of diagnosis

• Other medical conditions (like heart disease, diabetes, and the like)

• Toxicity or side effects due to the cancer treatment itself

Researchers will sometimes attempt to separate cancer deaths from non-cancer deaths. But it isn’t so easy to determine the cause of death based on medical records. And someone with cancer often has cancer listed on their death certificate, even when cancer isn’t the direct cause. 

There are even more flaws to mortality rates when you consider someone’s age at the time of diagnosis. Let’s say a study is reporting death caused by prostate cancer. An older person is more likely to die from some other condition than prostate cancer. A young person, with no other medical conditions, is more likely to die from prostate cancer than something else. But they’re still less likely than an older person to die from prostate cancer. So this study may make it look like prostate cancer mortality is higher in a young person. But in reality, prognosis is better for a younger person. 

Problems with survival rates

People with cancer often want to get a sense of their long-term life expectancy. But survival statistics don’t always paint an accurate picture of this. There are a couple reasons for this:

• It takes a lot of time to study survival. So many studies will report short-term survival instead (like 1 or 3 years). This does not tell someone if they will be alive and cancer-free 10 years later.

• To study long-term survival, researchers need to study people diagnosed with cancer a long time ago! And this group isn’t an accurate representation of people diagnosed today. People diagnosed today likely have safer and more effective treatments.

What’s more, someone’s prognosis changes (and often improves) every extra year they survive. This is especially true if they have survived successful treatment. So most survival rates are only relevant to someone at the beginning of their diagnosis. Once they start treatment, or live for another year, their own survival rate improves. And it’s likely better than the one reported in the study. 

Why trust our experts?

Written by:

Katie E. Golden, MD

Katie E. Golden, MD, is a board-certified emergency medicine physician and a medical editor at GoodRx.

Reviewed by:

Sophie Vergnaud, MD

Sophie Vergnaud, MD, is the Senior Medical Director for GoodRx Health. An experienced and dedicated pulmonologist and hospitalist, she spent a decade practicing and teaching clinical medicine at academic hospitals throughout London before transitioning to a career in health education and health technology.

Our editorial standards

Meet our experts