EBM & screening

Question 1

what is screening?

Answer 1

process of identifying apparently healthy people who may be at increased risk of a disease or condition

Question 2

what are some examples of screening?

Answer 2

• bowel cancer, from age 50, every 2 years
• breast cancer, from age 50, every 3 years
• cervical cancer, every 3 years from 25-49 then every 5 years
• lots in pregnancy
  -lots in newborns

Question 3

how do you decide what to screen for?

Answer 3

Wilson Jungner criteria
= important public health problem
= need to understand natural history
= recognizable latent or early symptomatic screening

Question 4

what is the wilson jungner criteria?

Answer 4

• simple, safe, precise & validated
• acceptable to public (clinically,socially & ethically acceptable)
• distribution of test results know & cut-off defined
• agreed policy on further diagnostic investigations

treatment = effective & available

Question 5

what is difference between diagnostic & screening testing?

Answer 5

not the same!

positive screening result doesn’t necessarily mean disease-free

Question 6

what is sensitivity?

Answer 6

= how well the test detects having the disease when it is “really” present
= true positives (can also then work out false negatives)

Question 7

how do you calculate sensitivity?

Answer 7

n.o results where disease is detected in people with the disease DIVIDED BY n.o of people with disease x100

Question 8

what is specificity?

Answer 8

= how well the test detects not having the disease
= true negatives (means you can also find false +ve)

Question 9

how do you calculate specificity?

Answer 9

number of normal or negative results where disease is NOT detected in people without the disease DIVIDED BY number of people without the disease x100

Question 10

what does it mean if
(A) high sensitivity?
(B) high specificity?

Answer 10

(A) picks up most of disease, very few false -ve
(B) correctly detects no disease when not present, very few falso positives

Question 11

what is positive predictive value?

Answer 11

= how reliable the test result is when it shows the disease is present

Question 12

what is negative predictive value?

Answer 12

= how reliable the test result is when it shows disease is not present

Question 13

how do you calculate positive predictive value?

Answer 13

n.o people with positive test result & have disease/ n.o of people with positive test result x100

Question 14

how do you calculate negative predictive value?

Answer 14

n.o people who have negative test result & do not have disease/ n.o people with negative test result x100

Question 15

what effect does prevalence have on PPV and NPV?

Answer 15

changes in prevalence = changes in PPV & NPV

Question 16

what effect does prevalence have on sensitivity & specificity?

Answer 16

changes in prevalence = no effect

Question 17

why do screening?

Answer 17

• reduce disease incidence
• reduce disease mortality
• earlier protection - treat earlier →less radical treatment (extreme)
• more cost effective as later detected often means more extreme treatment

Question 18

what are possible disadvantages of screening?

Answer 18

• false reassurance = negative test doesn’t always mean negative disease
• over investigation and treatment if false positive
• anxiety increase in participants
• longer period of morbidity with unaltered prognosis (if treatment doesn’t have much effect on prognosis)
• potential harm if screening test like X-ray etc
• opportunity cost - if you spend money on screening programme you’ve lost opportunity to spend money on something else

Question 19

how do you determine screening frequency?

Answer 19

it’s tricky - a trade off with acceptability:

= short enough to not miss new cases but not too long so it’s annoying and still cost effective

Question 20

what is lead time bias?

Answer 20

potential issue with introduction of screening - survival time appears to be increased due to earlier detection which may occur before symptoms

Question 21

how can lead time bias be adjusted for in RCT?

Answer 21

by taking start time as point of randomisation so not the point of diagnosis
or
comparing number of deaths occuring in fixed period of time instead or as well as number of people surviving

Question 22

is screening mandatory?

Answer 22

no - means you need informed consent! so patients need lots of info and support to know what they’re doing (from leaflets, social media, education etc)