# Need help with a simple formula i think?

#### not so clever doc

##### New Member
Anyone who can help, would be great please.... essentially, I have a very simple question that can be answered by most clever stats guru's, of which I am not one, so ever grateful and hoping I get a reply:
I want to test/screen a population for lets say a an illness (any random illness will do)...and we are looking at 1 illness only
I want a formula that will take into account the following:
Population size
'Level of screening' ie how many people? do I screen how often? to get say …… 10% or 20% or 25% etc to say 100 % of the population in say 1 month 4 months, 6 months and 12 months...…… I hope that makes sense...ie I want to is screen people of a population and understand how many people I screen and how often I screen to get these levels of 'coverage' for the population

#### staassis

##### Active Member
The problem is not clear. What do you want to screen the population for? What do you want to detect? What is your criterion?... What information is available to you for each member of the population?

#### not so clever doc

##### New Member
Hi there, thank you for your reply.....
Hope this helps.... I want to screen a population for a sexually transmitted disease (let’s say gonorrhoea) .... I want to know how many people (of a population say 500, 1000, 2000 people etc) I have to screen how often ( monthly, quarterly, annually) to give me a percentage of screening say 10%, 20 % etc ie how many needs testing for a test that is either positive or negative for how often to give me a 10% level of screening threshold ....
Or otherwise how often should I screen to make sure I cover the whole population in a timescale of monthly, every other month etc

#### staassis

##### Active Member
You said that already: "screen a population for a... disease"... So I can only reiterate: "What do you want to detect? What is your criterion?... What information is available to you for each member of the population?"

#### not so clever doc

##### New Member
You said that already: "screen a population for a... disease"... So I can only reiterate: "What do you want to detect? What is your criterion?... What information is available to you for each member of the population?"
Thank you Staassis, and my apologies for the frustration caused.... I am no statistician....
Essentially I want to screen for the incidence (new cases) of a disease in a fixed number of people over time (my population) over set periods be it monthly etc. And I want to have a formula to work out how many people of my pop. I have to screen to detect say, a 10%, 20% etc of the cases ie how many people to I screen to find this our ie clearly if I want it be 100 % certain of screening for all possible cases in a month then I simply screen all in the population.....sorry , maybe I am not esplaining clearly..... ps, I am sorry I don’t understand or may use the wrong terminology.... thanks you for your patience....

#### staassis

##### Active Member
This statement of yours does not make sense even from the point of view of standard English grammar: "I have to screen to detect say, a 10%, 20% etc of the cases"... I have to ask again: what are you trying to find out? What is your criterion?... What information is available to you for each member of the population?.. Step by step, second by second: what are you doing?

#### not so clever doc

##### New Member
Hmmmmnnnnn, not taking into account grammar, even though English is not my first language I will try a step by step approach. I am not sure which criterion(or is it criteria) you refer to, or whether this has a special meaning in statistics, but here goes .... step by step...
1. We are offering a heath screening test to an employer who employs a set number of people.
2. We are offering ‘levels’ of screening on a pre-agreed plan that will be charged at different rates to the employer
3. The test for our screen is either positive or negative ..... there are no inbetweens or maybe?
4.So, nurses will go out and randomly screen the population with urine screens
5. I want to be able to work out how many people I have to screen, how often to offer a plan with pre-agreed ‘level of screening’ ..
6. In other words, how many people do I screen in my population so I know I’ve picked up say 25, 50, 75 or 100% of the cases.
7. Another factor of course will be how often do I screen to achieve a level of screening...
I hope this helps .... please feel free to ask more questions ...
In short I would like to be able to say to an employer ....
I order to pick up ‘25% of cases’ in your workforce of ‘1000 people’, we have to screen ‘200’ employees ‘monthly’..... these numbers are for illustration purposes only, but they are the variables an I am hopeful that this is what you refer to when you say ‘criterion’..... thank you for your help staassis.....

#### staassis

##### Active Member
1. To know "how often do you screen", you need to develop a stochastic model for evolution of the decease. Look at the relevant epidemiology literature. Formulas there are far from being trivial.

2. Contracting a decease is random. Therefore, you will never know that you have "picked up ...100% of the cases" unless you have tested 100% of people. Typically, in such settings researchers talk in the language of estimating probabilities for decease incidence (fraction of people affected) and building confidence intervals for those.

When the underlying world is random, you should stay away from words "I've picked up" and "made sure". Instead you should be talking in the language of "estimates" and "averages".

Last edited:

#### not so clever doc

##### New Member
1. To know "how often do you screen", you need to develop a stochastic model for evolution of the decease. Look at the relevant epidemiology literature. Formulas there are far from being trivial.

2. Contracting a decease is random. Therefore, you will never know that you have "picked up ...100% of the cases" unless you have tested 100% of people. Typically, in such settings researchers talk in the language of estimating probabilities for decease incidence (fraction of people affected) and building confidence intervals for those.

When the underlying world is random, you should stay away from words "I've picked up" and "made sure". Instead you should be talking in the language of "estimates" and "averages".

#### not so clever doc

##### New Member
Ok, I think we’re getting there..... let’s see....
I hope we are on track, otherwise I might be on the wrong forum perhaps ....:
Not convinced we need stochastic modelling (I am sure these formulae are not trivial) with epidemiological data or disease evolution as such: let’s assume, hypothetically, that we know the disease is prevalent in our population.... no ifs, ands or buts we KNOW it’s present although we don’t know the prevalence or incidence. Let’s assume also that our test is clear.... you either have the disease or you don’t.....

Can we now answer this question? Ie how many do I screen of the population , how often to ‘estimate’ that I will be ‘likely or probable’ to have detected a preset percentage of the populations that is likely to be positive....