Interpreting Cancer Statistics

1
Interpreting Cancer Statistics

2
Incidence Rate

• Number of new cases
• per time period
• per 100,000 population

3
Mortality Rate

• Number of deaths
• per time period
• per 100,000 population

4
Time Period

• Generally 5-year interval

5
Per 100,000 at risk

• Most rates are expressed per 100,000
• Even rates from small units
• like counties

6
Always Use Age-Adjusted Rates

• Reference population
• 2000 US Census Population

7
Age-Adjusted

• As if each population had
• same sex and age structure
• as reference population

8

• The number of cases expected
• if county had population of 100,000
• with the same sex and age distribution
• as the US in 2000

9
Always UseConfidence Intervals

• All incidence and mortality rates
• have Confidence Intervals
• By convention 95

10
95 Confidence Interval

• All rates have margin of error.
• The smaller the number of cases
• and
• the smaller the population,
• the bigger the error.

11
Confidence Interval isMargin of Error

• We are 95 sure the true rate
• lies within the Confidence Interval

12
Comparing Confidence Intervals

• If Confidence Intervals overlap,
• rates are not statistically different.
• If rates are not statistically different,
• they are not different, period.

13
Example

• Colorectal Cancer per 100,000
• Incidence rate 95 CI
• Montana 45.7 43.9 - 47.6
• Carter 102.4 38.5 - 166.4

14
Example

• Colorectal Cancer per 100,000
• Montana 45.7 43.9 - 47.6
• Powell 25.8 10.5 - 41.2

15
Example

• Colorectal Cancer per 100,000
• Montana 2002-06 45.7 43.9 - 47.6
• Montana 2000-04 49.1 47.1 - 51.0

16
Example

• Colorectal Cancer per 100,000
• Montana 2002-06 45.7 43.9 - 47.6
• Montana 2000-04 49.1 47.1 - 51.0
• Powell 2002-06 25.8 10.5 - 41.2
• Powell 2000-04 44.5 23.7 - 65.3

17
Example

• Brain and CNS Cancers per 100,000
• Montana 7.5 6.8 - 8.3
• Meagher 44.1 0.0 - 89.8

18

• Never compare
• number of cases

19

• Never compare
• rates

20

• Always compare
• Confidence Intervals

21
Montana Counties are Small

• Incidence and mortality rates
• for counties are unstable
• and have large margins of error

22
Unstable

• Unstable
• fluctuate a lot from year to year

23
Age-Adjusted to 2000 Population

• Can compare state to US
• County to state
• Do not compare county to county

24

• Questions?