LogicalDUAL-USE

Mortality vs. Survival Framing

What it is

Describing the same outcome as a survival rate or a mortality rate — “90 percent survive” or “1 in 10 die” — and, in screening claims, quoting five-year survival, which rises with earlier detection even when no one lives a day longer.

How it works

There are two layers. The first is a plain framing effect: McNeil, Pauker, Sox and Tversky found in 1982 that patients and physicians were far more willing to choose surgery when told 90 percent survive the operation than when told 10 percent die, though the statements are identical. The second is specific to screening and is the more consequential. Five-year survival is the share of people diagnosed who are alive five years later. Screening moves the date of diagnosis earlier (lead-time bias) and adds slow or harmless cancers that would never have caused symptoms (overdiagnosis), so five-year survival rises even when the death rate in the population is unchanged. Worked: a tumor that screening finds at 67 would otherwise be found at 70; the patient dies at 70 either way; five-year survival is 100 percent with screening and 0 percent without, and not a day has been gained. Welch, Schwartz and Woloshin showed in 2000 that rising survival across cancers bore no relation to changes in mortality. The honest measure for whether screening saves lives is mortality per 100,000 per year; the persuasion move is to quote survival because it is the number that always improves.

Real-world examples

  • •In a 2007 campaign radio advertisement Rudy Giuliani said his chance of surviving prostate cancer in the United States was 82 percent against 44 percent in England “under socialized medicine”; the figures were five-year survival, inflated in the US by widespread PSA screening, while prostate-cancer mortality in the two countries was nearly identical at about 26 per 100,000.
  • •McNeil and colleagues (1982) gave the same lung-cancer treatment outcomes to patients, students and physicians framed as survival or as mortality; the share choosing surgery over radiation shifted markedly with the frame in every group.
  • •Wegwarth and colleagues (2012) surveyed US primary-care physicians and found most judged a screening test that improved five-year survival from 68 to 99 percent as proof that it saved lives, and most were more persuaded by survival figures than by mortality figures that actually answer the question.
  • •In March 2020 the World Health Organization's statement that about 3.4 percent of reported COVID-19 cases had died was quoted to alarm, and a “99.97 percent survival rate” was quoted through 2020 and 2021 to reassure; the two described different denominators — reported cases against estimated infections — and each was chosen for its sound.

Ethical guidelines

Where the line is

Stating a survival rate is legitimate when the mortality rate stands beside it, the denominator is named and, for screening, the audience is told that survival rises with earlier diagnosis; it becomes manipulation when survival is chosen because it always improves, when a country or hospital comparison rests on different screening intensity, or when one frame is presented alone to steer a treatment choice.

  • ●For treatments, give both frames — survive and die — in the same sentence, since the evidence is that people decide differently depending on which they hear first.
  • ●For screening, report mortality rates, not five-year survival, and say why: survival rises with earlier diagnosis whether or not anyone lives longer.
  • ●State the denominator of any survival or fatality rate — diagnosed cases, reported cases, estimated infections — and do not compare rates with different denominators.
  • ●Comparing survival across countries with different screening intensity is not evidence about health systems, and presenting it as such is a known error.

How to defend against it

  • ►Flip the frame yourself: turn every survival figure into a death figure and every death figure into a survival figure before deciding; if your decision changes, the frame was doing the deciding.
  • ►For any screening claim, ask for the mortality rate — deaths per 100,000 per year, with and without screening — and treat five-year survival as uninformative on whether lives are saved.
  • ►Ask when the clock started: survival measured from diagnosis is lengthened by earlier diagnosis alone.
  • ►For fatality rates, ask what the denominator is; a rate per reported case and a rate per estimated infection can differ tenfold for the same disease.
  • ►When a politician or an advertisement compares survival rates across countries or hospitals, ask whether the places screen at the same intensity; if not, the comparison is about detection, not outcomes.

From the Defense Playbook

Ask for Absolute Numbersseconds

When someone quotes a relative change ("cuts your risk by 50 percent", "doubles the danger"), ask what the risk was before and after in plain counts out of the same number of people, because the same fact sounds enormous as a ratio and modest as a difference.

Every playbook entry states how strong its evidence is and when not to use it. Browse the full playbook.

References

  1. McNeil, B. J., Pauker, S. G., Sox, H. C., & Tversky, A. (1982). On the elicitation of preferences for alternative therapies. New England Journal of Medicine, 306(21), 1259-1262
    The survival-versus-mortality framing experiment with patients, students and physicians.
  2. Welch, H. G., Schwartz, L. M., & Woloshin, S. (2000). Are increasing 5-year survival rates evidence of success against cancer?. JAMA, 283(22), 2975-2978
    The absence of any relation between changes in five-year survival and changes in mortality across cancers.
  3. Wegwarth, O., Schwartz, L. M., Woloshin, S., Gaissmaier, W., & Gigerenzer, G. (2012). Do physicians understand cancer screening statistics? A national survey of primary care physicians in the United States. Annals of Internal Medicine, 156(5), 340-349
    Physicians misreading survival gains as evidence that screening saves lives.
  4. Gigerenzer, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L. M., & Woloshin, S. (2007). Helping doctors and patients make sense of health statistics. Psychological Science in the Public Interest, 8(2), 53-96
    The Giuliani prostate-cancer example and the explanation of lead-time and overdiagnosis bias in survival statistics.
Last reviewed
Suggest a correction

Detect Mortality vs. Survival Framing in any text

Paste any message, email, or article into our free Manipulation Detector to see if Mortality vs. Survival Framing or other techniques are being used on you.

Try the Detector FreeCreate Free Account