Screening!

Question 1

Sensitivity calculation?

Answer 1

TPR (TP/TP+FN)

Question 2

Specificity calculation?

Answer 2

TNR (TN/TN+FP)

Question 3

Reasons why may have low cancer survival in UK?

Answer 3

Late diagnosis, unhealthy population, less healthcare spending per capita, slower to adapt to innovation, underuse of successful treatments, poor access to treatment for some (particularly the elderly), air pollution, screening uptake not ideal (for example smear uptake reduced). Could also make the point that survival is not the be-all and end-all.

Question 4

Wilson and Jungner?

Answer 4

1. Important health problem.
2. There should be a suitable test
3. Treatment must be acceptable and more effective if started early.
4. Must be recognisable latent phase.
5. Natural history must be well understood.
6. Must be clear policy on who to treat.
7. Must be cost effective (in total) inc. opportunity costs.
8. Case finding should be a continuous process
9. Test should be acceptable to population
10. Facilities for diagnosis and treatment should be available.

Question 5

Way that technology could be used to improve screening?

Answer 5

Increased use of genomics challenge classic criteria. Could go from preclinical to prepathological stage.

Question 6

Screening high risk groups?

Answer 6

Get increased cost-effectiveness as get higher yield per scan (considered re cytology or chest imaging in smokers). Problem is that there are very few ways, other than age, to stratify the population into groups that are large enough for mortality to be improved by screening.

Question 7

Why is lung cancer not screened?

Answer 7

Sputum cytology and CXR lack sensitivity; consistent evidence that has no effect on mortality; often incurable when detected; follow up investigations are radiation-intensive, invasive (bronchoscopy) and expensive if false positive. Obviously is good choice because is leading cause of cancer mortality for both sexes.

Question 8

Lead and length time bias?

Answer 8

Lead time bias refers to artificially inflated survival because longer time from diagnosis till death even though disease has not been modified. Length time bias refers to screening detecting lots of slow-growing indolent cancers which are then used to demonstrate that screening improves survival.

Question 9

Estimating sensitivity?

Answer 9

Use proportional incidence (if predicted cancers in screened group is 100, and in the interval period only 20 are detected, then screening can be said to have prevented/detected 80).

Question 10

Estimating specificity?

Answer 10

Use FPR based on gold-standard investigation. The importance of this depends on whether or not this procedure is expensive/invasive

Question 11

What % of FOB are positive?

Answer 11

2%

Question 12

New FOB test?

Answer 12

FIT (faecal immunochemical test): only do once, more sensitive to lower blood levels. Problem is limited colonoscopy workspace which could be overwhelmed

Question 13

Comparisons of breast and CRC screening?

Answer 13

CRC is lower recall; 1/10 of those investigated have cancer vs 1/5 of those recalled in breast. More frequent, older age, lower uptake, test has no AEs, no unequivocal evidence of mortality reduction. Can affect incidence in way that breast cancer cannot.

Question 14

When was NHSBSP introduced?

Answer 14

1988

Question 15

Recall rates for breast?

Answer 15

1/25 recalled; 1/5 of these will have cancer

Question 16

High grade DCIS progression?

Answer 16

40%

Question 17

Breast screening QALYs?

Answer 17

£20,000 per QALY saved

Question 18

Extra cancer with radiation for breast screening?

Answer 18

3-6 extra cancers for 10,000 women screened for 20 years. Marmot review.

Question 19

Differences with screened breast cancers?

Answer 19

Less likely to need chemo/radiotherapy, more likely to be of special type, early and well-differentiated

Question 20

Checking how individual units perform in breast screening?

Answer 20

Use standardised detection ratios

Question 21

Considerations in designing programme?

Answer 21

Age, how to word it, how often, percentage uptake needed for mortality benefit, cost (not just of screening but of test, follow up, counselling [note FIC]; balance vs often cheaper to treat disease early rather than complex surgery or prolonged chemotherapy; think about sensitivity and specificity when using ROC curves; frequency; also ethics. Cost vs yield, multiphase screening, % that progresses, cut-offs, disease modifiying, % that progresses (COMET)

Question 22

Selection bias arising from screening?

Answer 22

Non-attendees may be more likely to present late when DO have symptoms as well (same characteristics) and therefore artificially increased mortality in unscreened population

Question 23

How can screening improve QOL in breast cancer?

Answer 23

Screened cancer more likely to early stage, of special type, node negative and therefore more conducive to BCS which has psychological benefits

Question 24

Why is flexi sig different to other screening methods?

Answer 24

Because “intervene and screen”; this affects incidence in screened so hard to measure efficacy

Question 25

1 million study?

Answer 25

2% FOB +ve; 1/10 of those cancer. 70% Dukes A-B1

Question 26

Why is BRCA more likely to be missed by screening?

Answer 26

High grade, rapidly growing, no microcalcifications

Question 27

Overdiagnosis in breast?

Answer 27

Marmot review found that 1/4 of those diagnosed are overdiagnosed. For every 1,000 women screened, 17 overdiagnosed and 5 lives saved

Question 28

Significant benefit of CRC for screening?

Answer 28

Can treat incidence more easily because polyps are easier to remove “as you go”; halt progression before become malignant. No such equivalent in breast.