Why Median Survival Times Aren't (That) Helpful

The median is one of the most common statistical tools used in clinical research reports in oncology. They're also, from the patient's perspective, not as helpful as they seem at first glance. It's very important to be realistic about what statistics mean, especially when our lives are on the line. This doesn't always mean being pessimistic about the data. Rather, it means understanding what the data tells us -- and, equally importantly, what it doesn't tell us.

A median is, briefly, one type of average. There are three types of averages: mean (the most common type outside of clinical research), mode, and median. In short, the median is the point at which 50% of a sample have a particular attribute, or have accomplished some particular task. For instance, a country's median income is a dollar figure which half of the people earn less than, and half of the people earn more than. In cancer trials, median survival means the point at which half the patients recruited into the trial have died. And that's all it means -- it doesn't say when the 50% who died did so, or how long the other 50% are likely to live.

95% of Statistics are Bogus (Including This One)

My own Statistics page still languishes in its unstarted state, but a new article in Science News manages to present, more or less in layman's terms, precisely the sorts of problems that I wanted to highlight for readers here. So in the meantime, it's a useful starting point.

Tom Siegfried explains very well that the scientific method and statistical analysis are not necessarily natural bed-partners, and that even the games we play to make statistics seem "reliable" can end up, in certain cases, actually making them more misleading. Most of this will be old news to people with stats training, but it's still a good refresher. There's also a section specifically on the problems inherent in clinical trials:

Statistical problems also afflict the “gold standard” for medical research, the randomized, controlled clinical trials that test drugs for their ability to cure or their power to harm. Such trials assign patients at random to receive either the substance being tested or a placebo, typically a sugar pill; random selection supposedly guarantees that patients’ personal characteristics won’t bias the choice of who gets the actual treatment. But in practice, selection biases may still occur, Vance Berger and Sherri Weinstein noted in 2004 in Controlled Clinical Trials. “Some of the benefits ascribed to randomization, for example that it eliminates all selection bias, can better be described as fantasy than reality,” they wrote.

Better off in BC? New Cancer Care Comparisons

The Lancet is publishing a new survey of cancer care outcomes in six countries -- Canada, Australia, Denmark, Norway, Sweden, and Britain -- between 1995 and 2007 (report here; registration is required, but free). It's not of direct relevance to Northoma, because the statistics in question don't include lymphoma. However, it's always helpful to understand how Canadian healthcare compares to systems in other countries. Frustratingly, the United States was not part of this data set. I'll write more about that particular comparison later.

In the meantime, how did Canada measure up?

Are Survival Statistics for Follicular Lymphoma Changing?

As I have warned before, survival statistics for cancer, especially for indolent cancers like follicular lymphoma, can be highly misleading. Median figures say little about where most people actually fall on a curve. And it is still a curve -- there's no way to say where on that curve your particular data point sits. Plus, that curve applies only to historical patients, not (necessarily) to current ones. The 5-year survival curve published today applies only to patients diagnosed 5 years ago.

One historical problem is clear in follicular lymphoma, though: it's easy to induce a partial remission with chemotherapy, and fiendishly difficult to actually extend lifespan with effective treatment. This is one of the basic realizations that informs the "watch & wait" strategy of today: hold off on the hard drugs, because early treatment doesn't really help much. Alkalating agent-based protocols, like CHOP, had only modest effects on overall survival.

Then rituximab came along. Actually, that particular anti-CD20 monoclonal antibody was simply the first in a series of revolutionary new lymphoma drugs, but it has resulted in massive change in the statistical picture. The question now is how much overall survival has changed in the past decade, and whether that means altering our understanding of how follicular lymphoma should be treated. Which is why studies like the recent immediate-maintenance-rituximab trial get press despite their obvious flaws. So how far have we come, anyways?

Chances of Getting Lymphoma – Plus, My Experience

My own sluggish path towards a diagnosis continues. I had an ultrasound yesterday which will apparently have results available within five days (although my clinic's office said 14 days, so who knows?). Judging the size and shape of the one node I did glimpse by craning my neck towards the screen (2.0 x 0.6 cm), it's not exactly great news. On the other hand, it's always possible I'll luck out.

In the meantime, I know that one of the most worrying periods in any person's life is the one in which they're not quite sure (yet) whether they're really sick. In the vast majority of cases, we actually aren't – or at least, not seriously. So, on the assumption that some such nervous readers will eventually find their way here, I thought I'd share some potentially comforting statistics.