Print Person Time with airSlate SignNow

Add person calculated

welcome back we're continuing on with measuring outcomes and are still looking at incidence and I drew a similar graph as we had on the previous ones we have a study here which extends over ten years and the time frame of the study I drew is this pink box and you can see here this is the time frame over ten years and if you remember this represents the length of disease of that a patient had this was their onset and then this is when they died of that disease and here's the onset and when they died of the disease again for these people I added here this y-axis so we can kind of keep track of people that are in here okay so let's add a few more cases in here to make things a little more complicated so we got a few patients in here who never got cancer so they didn't have an onset so there's no little circle to draw and if they never got the cancer they never died from the cancer either and we had one patient in our study who never got cancer but then died of a car crash over here and this lucky patient decided to leave and move to Paris halfway through the study this unfortunate guy died of a gunshot before the end of the study and this poor soul died of a heart attack before the end of the study so you can see there's a lot of things going on here this is life right people are going to move away people are going to die from other reasons not just from the cancer so let's say that we're doing a study this over this ten years and we want to count the incidence of death from cancer so I put the definition again of incidence here's the number of new cases over the number of people at risk in a given time frame and so since we sent our case is going to be death in the time frame we see good there's one there's two and there's three and so we've got three new cases now the next thing that we need to do is figure out what to put in the denominator who are the people that are at risk in here and we know that this person is this person is this person is at risk this person well it went they were at risk and then they died and so now from a car crash so they're no longer at risk for this portion similarly this guy had the heart attack was at risk here of getting cancer and dying but then he had heart attack and died and so he was no longer at risk here this person moved away so who knows what happened right so they're no longer in our studies so they may have gotten cancer while they were in Paris and died and we just don't know and same thing with the gunshot here and so this is real life again that we're during our study we're losing some of our patients to other causes of death nor losing them to follow up as they move away so while people are dying it's a different cause it's a different risk so we cause these we call these other risks competing risks because these are not the ones that we were looking for which is death by cancer so then how do we banish this how do we know which patients were going to include in the denominator well this is where something called incidence rate comes into play it's also called person time and what this does is it allows us to chop this up into pieces such that we can count the the portion of time that this person was at risk while excluding the portion where they were gone and no longer at risk and so what I've done here is I've tried to make these bigger pink stripes and each one represents a year and the lighter pink is a year so each one of these little intervals represents a year so now we've broken up each one of these timelines into these separate areas that we can count how many years is each person at risk and so that's where we get the term person time each person for the unit of time we pick is a year so each person year so let's let's do this so for the first person we have one year here a second year three four five six seven eight and then they died in the eighth year so not for nine and not for ten so this person has eight person years that they are at risk all right what about the next person so they got one two three four five six seven eight nine ten ten percent years at risk and this next person has one two three four five they also have ten and how about this next one I want to died in the car crash one two three four five six and then I died in the 60s only six years six percent years so I'm not going to make you go make you watch me go through all this but why don't you pause the video now and count how many you think there are and then when you're unpause I will fill the rest in okay all right so I'm going to fill the rest in now okay so I filled in every year that each one of these people were at risk and so they're at risk anytime that they don't have cancer and they're alive so as soon as they die and they're no longer at risk so the sky died of cancer so no longer is clear this person died of a gunshot so no longer was clear and this person moved away so they're he's no longer a risk here so we have one two three four five six seven eight nine ten eleven people total and the sum of this so the sum of all this is a tea person years so we could put that over here and say a tea person years and then you get zero point zero three seven five cases over person years and so this is called incidence rate or person time and man this got messy but I hope you understood what we were doing here and if you didn't watch the video a couple times or send me an email or put some put a note in the comments so I can explain it here further but what this allows you to do as well is to have a rolling study so you could actually extend this study past ten years and just have it keep going and having rolling enrollment and people will just come in and out of the study as they die or move away or whatever and you only have to count the amount of time that they are at risk in person time and so this is one way to to take into account that fact that people are going to be in here for different lengths of time so before we go I just wanted to talk about two more things about incidents so you'll remember our definition again number of new cases over the number of people at risk and we actually had two ways of measuring these number of people at risk normal incidence which sometimes called incident proportion counts the whole time that the patient is in the study in the denominator here and for incidence rate also called person time it only counts pieces of time that the patient is in the study as you know while they're at risks or not the whole time that's called person time now since the incidence proportion is is a proportionate user basically saying we only have four people total which ones had the disease or the case that we're looking for it's only going to be a proportion of that so because it is a proportion you know this value is going to be between zero and one now an incident rate is not does not have that same limit it can actually be between zero and infinity and the reason is it's because how you define the time so let's take our pry our example which was a 0.037 five cases over person years but let's say we decided not to measure to slice things up in the years but to slice it up into centuries and so this would instead be three cases over not 80 person years but 0.8 person centuries which comes out to 3.75 so that's obviously over 1 and we could put some bigger thing we could have picked millennia in which case it would be 375 so or 37.5 and so you can see that this number is not limited to being between 0 & 1 but it's really between 0 and infinity it all depends on what time frame is now we could have made it really really small instead of using person years we could have used person months or person days or person minutes in which case this number would get smaller and smaller now the other thing we should talk about is something called the waiting time so let's take our incidence again and you could say technically that a case really is a person right so these two units can actually cancel out and you're just left with years in the denominator so the units of this thing would be 1 over years so what if we took the inverse of that we would get 1 over 0 point 0 3 7 5 1 over years which the inverse of that equals 26.67 years and what this means is you have to wait on average 26.67 years to see one case of whatever it is that you are looking for so in our previous one was a death by cancer would have to wait every 26 years you'd see someone die from cancer and this is also called the expectation of life or the survival time so I know this was a bit of a complicated topic but watch this a few times and see if you can get the hang of it and if you can't please feel free to contact me and I'll see if we can clarify any details alright we'll see you next time