Future Directions

2. Screening test

• Understanding that screening tests may have benefits AND harms.
• Understanding that screening tests can make two errors: false positive AND false negative.
• Understanding how to translate specifics, sensitivities, and other conditional probabilities into natural frequencies.
• Understanding that the goal of screening is not simply the early detection of disease; it is mortality reduction or improvement of quality of life.

 

3. Treatment

• Understanding that treatments typically have benefits AND harms.
• Understanding the size of the benefit (and that it is generally over-estimated) AND harm (and that it is always under-estimated)

 

4. Questions about the science behind the numbers

• Understanding the difference between evidence based on randomized control trials, well-designed cohort or case-control studies and opinions.
• Recognizing potential conflicts of interest.

 

The authors also insightfully discuss the causes and potential remedies to statistical illiteracy and conclude with four simple recommendations that could immediately ameliorate, albeit not eliminate, the problem:

1. Use frequency statements (e.g., out of every 10 patients who take drug x, 3-5 will develop a serious rash) rather than single-event probabilities (e.g., If you take drug x, the probability you will get a rash is between 30 and 50%).

2. Use absolute risks (e.g., mammogram screening reduces the risk of dying from breast cancer by about 1 in 1,000 from about 5 in 1,000 to about 4 in 1,000) rather than relative risk (e.g., mammogram screening reduces the risk of dying from breast cancer by about 20%.)

3. Use mortality rates (e.g., there are 26 prostate cancer deaths per 100,000 American men vs. 27 per 100,000 men in Britain) rather than survival rates (e.g., the five year survival rate for people with prostate cancer is 98% in the USA vs. 71% in Britain)

4. Use natural frequencies (e.g., the probability that a woman who tests positive on a mammogram actually has breast cancer is about 10% since for every 1,000 women 9 with breast cancer and 89 without breast cancer will test positive) rather than

 

5. Conditional probabilities (e.g., the probability that a woman who tests positive on mammogram is about 10% because the probability she tests positive given she has breast cancer is 90% and the probability she tests positive given she doesn't have breast cancer is 9% and the overall probability of having breast cancer is 1%)."

 

6. Understand the difference between sensitivity and specificity

Sensitivity: The probability that an individual who has the disease of interest will have a positive screening test result. A test with high sensitivity has few false negatives, so you do not miss people with the disease.

Specificity: The probability that an individual who does not have the disease of interest will have a negative screening test result. A test with high specificity has few false positives so you do not identify people with no disease are rarely missed or told they don't have the disease.

• Diseased individuals with a positive screening test (true positives)
• Non diseased individuals with a positive screen (false positive)
• Diseased individuals with a negative screen ( false negative)
• Nondiseased individuals with a negative screen (true negative)

		We hope this information helps you understand a very difficult topic!!!