Difference between revisions of "Impact numbers"
Bosmana fem (talk | contribs) (Created page with "Impact numbers are a relatively new concept. They are measures derived from case-control and cohort studies that are intended to be simple to understand and to help compare th...") |
Bosmana fem (talk | contribs) m |
||
| (3 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
| − | Impact numbers are a relatively new concept. They are measures derived from case-control and cohort studies that are intended to be simple to understand and to help compare the population impact of different interventions | + | Impact numbers are a relatively new concept. They are measures derived from case-control and cohort studies that are intended to be simple to understand and to help compare the population impact of different interventions <ref>Heller RF, Dobson AJ. Disease impact number and population impact number: population perspectives to measures of risk and benefit. Br Med J 2000;321:905-2.</ref> <ref>Heller RF, Dobson AJ, Attia J, Page J. Impact numbers: measures of risk factor impact on the whole population from case-control and cohort studies. J Epidemiol Community Health 2002;56:606-10.</ref>. |
They are analogous to the concept of the number needed to treat (NNT) used in clinical medicine. This is the number of persons with a disease that, on average, must be treated in order to achieve one beneficial outcome (e.g. cure) or to prevent one adverse outcome (e.g. relapse). | They are analogous to the concept of the number needed to treat (NNT) used in clinical medicine. This is the number of persons with a disease that, on average, must be treated in order to achieve one beneficial outcome (e.g. cure) or to prevent one adverse outcome (e.g. relapse). | ||
| Line 74: | Line 74: | ||
|} | |} | ||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
==Population impact number (PIN)== | ==Population impact number (PIN)== | ||
This is the number in the whole population among whom one case is attributable to the exposure or risk factor. It can also signify, for a protective factor, the number in the whole population among whom one case will be prevented by the exposure or intervention. | This is the number in the whole population among whom one case is attributable to the exposure or risk factor. It can also signify, for a protective factor, the number in the whole population among whom one case will be prevented by the exposure or intervention. | ||
| − | It is equivalent to the reciprocal of the population attributable risk (ARpop). | + | It is equivalent to the reciprocal of the population-attributable risk (ARpop). |
| + | [[File:5611.PIN.gif-550x0.png|400px|frameless|left]] | ||
| + | |||
| − | + | In the example of drunk driving and driving-related deaths (table - death from drunk driving), there were 4 deaths per 1,000 drivers in one year attributable to drunk driving. We thus have: | |
| − | + | [[File:5611.PIN-2.gif-550x0.png|400px|frameless|left]] | |
| − | |||
| − | |||
| − | + | This means that, for every 250 people in Anystate, there is one driving-related death attributable to drunk driving on average per year. | |
==Case impact number (CIN)== | ==Case impact number (CIN)== | ||
| Line 128: | Line 97: | ||
In the example of drunk driving and driving related deaths (table - death from drunk driving), 22% of driving related deaths in the population could be attributed to drunk driving. We thus have: | In the example of drunk driving and driving related deaths (table - death from drunk driving), 22% of driving related deaths in the population could be attributed to drunk driving. We thus have: | ||
| − | + | [[File:7065.CIN.gif-550x0.png|400px|frameless|left]] | |
This means that, for every 4.5 driving related deaths, one is attributable to drunk driving on average. | This means that, for every 4.5 driving related deaths, one is attributable to drunk driving on average. | ||
| Line 141: | Line 110: | ||
In the example of drunk driving and driving related deaths (table - death from speeding or drunk driving), there were 136 deaths per 1,000 drunk drivers in one year attributable to drunk driving. We thus have: | In the example of drunk driving and driving related deaths (table - death from speeding or drunk driving), there were 136 deaths per 1,000 drunk drivers in one year attributable to drunk driving. We thus have: | ||
| − | + | [[File:1070.EIN.gif-550x0.png|400px|frameless|left]] | |
This means that, for every 7.4 drunk drivers, there is one driving related death attributable to drunk driving on average per year. | This means that, for every 7.4 drunk drivers, there is one driving related death attributable to drunk driving on average per year. | ||
| Line 153: | Line 122: | ||
In the example of drunk driving and driving related deaths (table - death from speeding or drunk driving), 91% of driving related deaths among drunk drivers could be attributed to drunk driving. We thus have: | In the example of drunk driving and driving related deaths (table - death from speeding or drunk driving), 91% of driving related deaths among drunk drivers could be attributed to drunk driving. We thus have: | ||
| − | + | [[File:0081.ECIN.gif-550x0.png|400px|frameless|left]] | |
| Line 161: | Line 130: | ||
==Population impact number of eliminating a risk factor (PIN-ER-t)== | ==Population impact number of eliminating a risk factor (PIN-ER-t)== | ||
| − | This is derived from the population attributable fraction. It is calculated by multiplying the population size (n) by the risk of an event in the next t years (Ipop) and by the population attributable fraction (AFpop) | + | This is derived from the population attributable fraction. It is calculated by multiplying the population size (n) by the risk of an event in the next t years (Ipop) and by the population attributable fraction (AFpop) <ref>Heller RF, Buchan I, Edwards R, Lyratzopulos G, McElduff P, St Leger S. Communicating risks at the population level: application of population impact numbers. Br Med J 2003;327:1162-5.</ref>. |
| − | |||
| − | |||
| − | + | [[File:0081.ECIN.gif-550x0.png|400px|frameless|left]] | |
| − | |||
| − | AFpop = attributable fraction in population | + | * n = population size |
| + | * Ipop = incidence in population (over t years) | ||
| + | * AFpop = attributable fraction in population | ||
In the example of drunk driving and driving related deaths (table - death from drunk driving), there were 10,000 drivers in the study (n), 18 deaths per 1,000 drivers in one year (Ipop), and 22% of driving related deaths in the population could be attributed to drunk driving (AFpop). We thus have: | In the example of drunk driving and driving related deaths (table - death from drunk driving), there were 10,000 drivers in the study (n), 18 deaths per 1,000 drivers in one year (Ipop), and 22% of driving related deaths in the population could be attributed to drunk driving (AFpop). We thus have: | ||
| − | + | [[File:2251.PINERt-2.gif-550x0.png|400px|frameless|left]] | |
| Line 178: | Line 146: | ||
=References= | =References= | ||
| − | + | <references/> | |
| − | + | ||
| − | + | ||
| + | [[Category:Measures of Disease Impact]] | ||
Latest revision as of 21:03, 22 March 2023
Impact numbers are a relatively new concept. They are measures derived from case-control and cohort studies that are intended to be simple to understand and to help compare the population impact of different interventions [1] [2].
They are analogous to the concept of the number needed to treat (NNT) used in clinical medicine. This is the number of persons with a disease that, on average, must be treated in order to achieve one beneficial outcome (e.g. cure) or to prevent one adverse outcome (e.g. relapse).
The impact number reflects the number of people in each population (the whole population, the cases, all those exposed, and the exposed cases) among whom one case is attributable to the particular exposure or risk factor.
The following types of impact number are described below:
- population impact number (PIN)
- case impact number (CIN)
- exposure impact number (EIN)
- exposed cases impact number (ECIN)
- population impact number of eliminating a risk factor (PIN-ER-t).
The table below shows these impact numbers calculated from the example of the study of drunk driving and automobile related deaths.
Table. Impact numbers estimated from a cohort study (n=10,000) of drunk driving and automobile related deaths, Anystate, 2010
| Impact measures | Abbrev. | Rate or number |
|---|---|---|
| Death rate in drunk drivers | Ie | 150/1,000 |
| Death rate in non-drunk drivers | Iu | 14/1,000 |
| Death rate in all drivers | Ipop | 18/1,000 |
| Attributable risk among exposed | ARe | 136/1,000 |
| Attributable fraction among exposed | AFe | 0.91 |
| Population attributable risk | ARpop | 4/1,000 |
| Population attributable fraction | AFpop | 0.22 |
| Population impact number | PIN | 250 |
| Case impact number | CIN | 4.5 |
| Exposure impact number | EIN | 7.4 |
| Exposed case impact number | ECIN | 1.1 |
| Population impact number of eliminating a risk factor | PIN-ER-t | 40 |
Contents
Population impact number (PIN)
This is the number in the whole population among whom one case is attributable to the exposure or risk factor. It can also signify, for a protective factor, the number in the whole population among whom one case will be prevented by the exposure or intervention.
It is equivalent to the reciprocal of the population-attributable risk (ARpop).
In the example of drunk driving and driving-related deaths (table - death from drunk driving), there were 4 deaths per 1,000 drivers in one year attributable to drunk driving. We thus have:
This means that, for every 250 people in Anystate, there is one driving-related death attributable to drunk driving on average per year.
Case impact number (CIN)
This is the number of people with the disease or outcome among whom one case is attributable to the exposure or risk factor. It can also signify, for a protective factor, the number of people with the disease among whom one case will be prevented by the exposure or intervention.
It is equivalent to the reciprocal of the population attributable fraction (AFpop).
In the example of drunk driving and driving related deaths (table - death from drunk driving), 22% of driving related deaths in the population could be attributed to drunk driving. We thus have:
This means that, for every 4.5 driving related deaths, one is attributable to drunk driving on average.
Exposure impact number (EIN)
This is the number of people with the exposure among whom one excess case is attributable to the exposure.
It is equivalent to the reciprocal of the attributable risk in the exposed (ARe).
In the example of drunk driving and driving related deaths (table - death from speeding or drunk driving), there were 136 deaths per 1,000 drunk drivers in one year attributable to drunk driving. We thus have:
This means that, for every 7.4 drunk drivers, there is one driving related death attributable to drunk driving on average per year.
Exposed cases impact number (ECIN)
This is the number of exposed cases among whom one case is attributable to the exposure.
It is equivalent to the reciprocal of the attributable fraction in the exposed (AFe).
In the example of drunk driving and driving related deaths (table - death from speeding or drunk driving), 91% of driving related deaths among drunk drivers could be attributed to drunk driving. We thus have:
This means that, for every 1.1 drunk drivers with a driving related death, one driving related death is attributable to drunk driving on average per year.
Population impact number of eliminating a risk factor (PIN-ER-t)
This is derived from the population attributable fraction. It is calculated by multiplying the population size (n) by the risk of an event in the next t years (Ipop) and by the population attributable fraction (AFpop) [3].
- n = population size
- Ipop = incidence in population (over t years)
- AFpop = attributable fraction in population
In the example of drunk driving and driving related deaths (table - death from drunk driving), there were 10,000 drivers in the study (n), 18 deaths per 1,000 drivers in one year (Ipop), and 22% of driving related deaths in the population could be attributed to drunk driving (AFpop). We thus have:
This means that up to 40 (of 180) driving related deaths per year in Anystate could potentially be prevented by eliminating drunk driving.
References
- ↑ Heller RF, Dobson AJ. Disease impact number and population impact number: population perspectives to measures of risk and benefit. Br Med J 2000;321:905-2.
- ↑ Heller RF, Dobson AJ, Attia J, Page J. Impact numbers: measures of risk factor impact on the whole population from case-control and cohort studies. J Epidemiol Community Health 2002;56:606-10.
- ↑ Heller RF, Buchan I, Edwards R, Lyratzopulos G, McElduff P, St Leger S. Communicating risks at the population level: application of population impact numbers. Br Med J 2003;327:1162-5.
Root > Assessing the burden of disease and risk assessment > Field Epidemiology > Measurement in Field Epidemiology > Measures of Disease Impact
