NUR 7630 Module 6 Practice Decision Paper: Absolute Risk and Natural Frequencies With a Patient Example

Twenty-Six Percent or Eight in Ten Thousand: Turning Relative Risks Into Natural Frequencies for a 53-Year-Old Deciding About Hormone Therapy

Student Name

Doctor of Nursing Practice, Thomas Edison State University

NUR-7630: Biostatistics and Clinical Epidemiology

Mentor Name

Month Day, Year

Twenty-Six Percent or Eight in Ten Thousand: Turning Relative Risks Into Natural Frequencies for a 53-Year-Old Deciding About Hormone Therapy

A Frightening Number

Ms. P., an invented patient, is 53 and teaches fourth grade in Cherry Hill. Since her periods stopped a year ago, she has had up to fifteen hot flashes a day and wakes four or five times a night drenched in sweat. Her concentration in the classroom is suffering. Her nurse practitioner suggested a short course of hormone therapy, but Ms. P. had read online that it "raises breast cancer risk by 26 percent" and left the visit in tears. She has no history of breast cancer, clots or stroke, and her mother lived to 90. As the clinic's DNP-prepared quality director, I was asked to help create a way to present the risks honestly. This paper explains why the number frightened her and how to present it in a form she can use.

Relative and Absolute Risk

A relative risk compares the risk in one group with the risk in another as a ratio. The 26 percent figure comes from the Women's Health Initiative, which reported a hazard ratio of 1.26 for invasive breast cancer with combined estrogen and progestin (Writing Group for the Women's Health Initiative Investigators, 2002). That is a relative increase: whatever a woman's baseline risk, it is about a quarter higher.

An absolute risk is the actual probability of the event over a period of time. The trial report also counted the effect in women: among 10,000 taking combined hormones for a year, about eight additional breast cancers, seven additional coronary events and eight each of additional strokes and lung clots, set against about six colon cancers and five hip fractures avoided. A 26 percent relative increase and 8 extra cases per 10,000 women a year describe the same finding.

Why the Format Matters

Gigerenzer et al. (2007) reviewed evidence that many physicians, patients and journalists misunderstand health statistics. Relative risks make effects seem larger than absolute risks do, and conditional probabilities, such as the chance of disease given a positive test, confuse even experienced clinicians. They recommend natural frequencies, statements of the form "out of 1,000 women like you, this many," because they make the reference group explicit and allow simple counting. Our minds handle counts of people more easily than percentages of percentages.

The authors also describe how the same information framed as a relative risk reduction persuades more people to accept a treatment than when it is framed as an absolute reduction, even though the facts are identical. That finding cuts both ways: a relative increase in harm can frighten patients away from a treatment that would help them, which is what happened to Ms. P.

Converting for Ms. P.

Over five years, the Women's Health Initiative's figures translate roughly into 4 extra breast cancers among 1,000 women taking combined hormones, along with similar small increases in stroke and clots and small decreases in fractures and colorectal cancer. Ms. P. is younger and closer to menopause than the average trial participant, and later analyses suggest that absolute risks of heart disease and stroke are lower in women her age, so these figures are, if anything, conservative for her. Her nurse practitioner would also consider whether she has a uterus, since estrogen alone carries different risks.

One Benefit, One Harm, in Counts

Laupacis et al. (1988) proposed expressing treatment effects as the number of patients who must be treated for one to benefit. For harms, the same idea gives a number needed to harm. With about 8 extra breast cancers per 10,000 women a year, roughly 1,250 women would need to take combined hormones for a year for one additional breast cancer to occur, or about 250 women for five years. For symptom relief, the balance is very different: in women with severe hot flashes, most experience substantial relief within weeks, so the number needed to treat for symptom benefit is small. Placing a large number needed to harm beside a small number needed to treat for her main concern helps her see the trade-off.

The comparison must still be honest about uncertainty. The trial's estimates have confidence intervals, her personal risk could be higher or lower than average, and the harms are serious when they occur. The goal is not to talk her into treatment but to replace a number that misled her with numbers she can weigh.

The One-Page Summary

The summary for Ms. P. uses natural frequencies and pictures. It shows 1,000 small figures of women, with the handful affected by each outcome shaded, and states plainly: "If 1,000 women like you take this treatment for five years, about 4 more will develop breast cancer than if they did not take it, and most of those taking it will have far fewer hot flashes." It lists the alternatives, such as nonhormonal medications and cognitive behavioral therapy for hot flashes, with their own benefits and side effects, and it invites her to decide with her nurse practitioner after reading it at home.

A Shared Decision

At her next visit, Ms. P. reads the summary and asks how long she would need the treatment. Her nurse practitioner suggests starting at a low dose and agreeing now to revisit the question every twelve months. She decides to try it for a year and then reassess. She says the number that frightened her now seems manageable because she can picture it.

Her decision is recorded in her chart with the summary attached, so that at next year's review she and her nurse practitioner can return to the same numbers, updated if new evidence has emerged, rather than starting the conversation again from fear.

Conclusion

A 26 percent relative increase in breast cancer and 4 extra cases per 1,000 women over five years are the same finding expressed in two ways, and only the second lets a patient weigh it against her symptoms. Converting relative risks to absolute risks and natural frequencies, adding numbers needed to treat and harm, and presenting them simply allowed Ms. P. to make an informed choice. That conversion is the practical end of every calculation in this course.

References

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. https://doi.org/10.1111/j.1539-6053.2008.00033.x

Laupacis, A., Sackett, D. L., & Roberts, R. S. (1988). An assessment of clinically useful measures of the consequences of treatment. New England Journal of Medicine, 318(26), 1728-1733. https://doi.org/10.1056/NEJM198806303182605

Writing Group for the Women's Health Initiative Investigators. (2002). Risks and benefits of estrogen plus progestin in healthy postmenopausal women: Principal results from the Women's Health Initiative randomized controlled trial. JAMA, 288(3), 321-333. https://doi.org/10.1001/jama.288.3.321