How to read a paper 01

so now in this video we're gonna go over how to read a paper that was a comment a bunch of students had asked they wanted a good organized way of how to do it the way I used to do it was I would start at the beginning and then I would just flip through the next page so the next page to the next page till I got to the end and you know by the time I got halfway through I was so mentally exhausted I had no idea what I was reading so obviously since I'm making a video about this there's got to be a different way a better way and of course there is so in short what you then do is go to first read the abstract and then the next thing that I do is I read every single table and try to make sense of what it's saying and then go back and read the methods and the results this will make a lot more sense now that you have looked at and understood the tables because they're especially the results they basically are explaining what they found and it's really hard to read a description of numbers unless you kind of know what the numbers already say and now that you've read the method and results and you can decide what conclusions you would draw from this and then after that you can read what they what conclusions they drew by reading their results in their discussion section so let's go through this process together and we're gonna use this paper yes it's an old ones from 1981 but it's one that is commonly taught in DBM courses as you'll see why and it's the the title of it is coffee and cancer of the pancreas I know what a nightmare that would be coffee causes cancer the pancreas because coffee is so delicious and the cancer the pancreas is so terrible but let's read on so let's start with the abstract okay and the abstract says they question 369 patients with histologically proven cancer the pancreas of 640 for control patients about the use of tobacco alcohol tea and coffee so if you couldn't tell before I'm a very visual person so I like a draw out what's going on so we had 369 patients with pancreatic cancer and 644 who didn't have any cancer they asked both groups about the use of cigarettes alcohol coffee and tea so you may immediately recognize that this is a case control study design those who have the pancreatic cancer are the cases and those who don't are the controls alright so now let's get back to the paper so they say there was a weak positive association between pancreatic cancer and cigarette smoking but they found no association with cigars none with pipe tobacco none with alcohol or none with tea however they did find a strong association between coffee consumption and pancreatic cancer in both sexes the Association was not affected by controlling for cigarette use now you'd want to control right because people at least maybe in the 80s when they drank the coffee they often smoked cigarettes too so you want to control for the cigarette so you so that you can't blame the cancer on the cigarettes instead of the coffee for the sexes combined there was significant dose-response relation after adjustment for cigarette smoking and so that means the more coffee you drank the more cancer you're gonna more likely gonna have to have cancer the relative risk associated with the drinking up to 2 cups of coffee per day was 1.8 i with a confidence live over the conference interval of 1 to 3 point oh ok now we already know that a confidence interval that includes 1 is not statistically significant if we're looking at risk ratios but and that with 3 or more cups of coffee was 2.7 the association should be evaluated with other data it reflects a causal relationship between coffee and drinking and pancreatic cancer it does but ok coffee use might account for a substantial proportion of the cases of this disease in the United States so in fact I believe these authors who are GI doctors had everybody in their clinic stop drinking coffee and and only drink tea so let's think about what they were saying again they're saying that of the people who had pancreatic cancer more of them drank coffee than the ones who didn't have pancreatic cancer and they're calling that a a causal relationship meaning that coffee causes pancreatic cancer your EBM spidey-sense should be tingling now because you would say this is a case control study we don't make causal relationships from a case control study we talk about associations correlations but not call it cause and effect okay we read the abstract step two let's go look at the tables and so here we are at table 1 at table 1 in most papers is going to give you the demographics of each group of the cases in the CRO controls how many men there were how many people had comorbidities whatever information that might be of relevance in this case it's a little bit different and we're looking at the distribution of cases and controls according to cigarette smoking habits and estimates of risk ratios okay and so we got they split them up by men and women and the cases that means the people who have pancreatic cancer and the controls nose who have no pancreatic cancer in both men and women all right so then across the top we have those who never smoked those who are ex-smokers those who smoked less than a pack a day and more than a pack a day all right and then they can compare the two groups and they calculated the relative risk ratios and you can see these are here so with smokers you could see the ones the denominator is going to be the never smoked group so that's why there's no that's what this will always be one here because you're comparing it to itself and then we would be comparing X smokers to the never smokers and current smokers - they're never smokers alright and so the risk ratio for the people who are Exum of course is 1.4 however the confidence interval here you see includes one right so this is not statistically significant so whatever we don't care about that this one's 1.1 confidence interval again includes one we don't care about that 1.4 this confidence interval includes 11.4 confidence interval includes one we don't care about that what about with women do we have anything else look this this one includes one this one includes one this one includes one this one includes one and so we can't say that we found some any stats Stickley significant association between cigarette use and pancreatic cancer according to this table that's how I would interpret it now let's go down to table two and this asks a very similar thing except now we're looking at alcohol drinking habits okay and again we got our men and our women our cases and controls and we got those who never drank occasional drinkers and regular drinkers okay and so we do the same thing again right this is going to be our reference group here the ones who never drank those are the ones that we're going to compare everyone else to those are the denominator so the occasional drinkers 1.3 but confidence interval here it includes one so we don't care about that this includes one that includes one so we don't really find here any association that is statistically significant between alcohol use and pancreatic cancer all right let's go to the next table all right so this one here is looking at tea drinking habits okay and the risk ratios and so same same cases and control men and women and then we got people who drink 0 1 2 2 and greater than 3 cups of tea a day ok this is our reference group this is one that we're comparing them to and again let's look at our confidence intervals here and these all include one but if you you know sometimes people will say hey but it's trending towards better so it looks like 0.7 you're less likely to have pancreatic cancer than if you didn't drink if you drank tea then didn't drink tea but it's not statistically significant you say well it's trending towards that but I don't think you didn't really say that so nothing there and how about with the women same thing so we can't really find an association here with tea between tea drinking and pancreatic cancer okay now we're gonna look at coffee drinking right and it's the same thing again cases controls men and women this is our reference group the one who doesn't drink any coffee and in the across the top we have one to two to three to four and greater than five cups a day all right again now we look at our numbers here these are bigger numbers and we've seen before here and accompany this conference interval does not include one so this might be something of significance and then this would also it it hits one so I'm not gonna count that one you know what about this one this one also doesn't include one and if you take them all together look this confidence interval does not include one and let's look for the women this one comfortable includes one don't care ooh this one doesn't and this one doesn't and this one doesn't alright so now we're starting to see something here the more coffee you drink and the more likely you are to have pancreatic cancer and so that's what they said but you know what remember they said it often times you smoke while you drink your coffee so maybe we need to adjust for that so now we move to this table over here and this looks at the relative risk of cancer of the pancreas associated with the use of coffee and cigarettes and so this table is a little bit different than the ones that we were looking at along this side we have our cigarette smoking right so we have our never smokers X smokers and current smokers then on this side we have our coffee drinkers and we have our zero cup 1 to 2 cup in the greater than 3 cup drinkers and so our reference group is of course gonna be the one who doesn't do anything and that's this and so that's who then compared what else to and so let's look at our people who never who don't smoke and drink coffee and look they have risk ratios that are pretty high right now they don't give us confidence intervals for these but they do further totals so for the ones who this is gonna be a referent group if we compare across this row here is gonna be this one here right and so then for the never smokers for the ones who are ex-smokers regardless you know of how much coffee they drink doesn't make a difference right because our confidence interval includes one and for the current smokers regardless of the coffee they drink again it doesn't make a difference because our confidence interval here includes one now let's look at our coffee drinking okay and so here we got though this is our gun our referent group this is the people who drink one to two cups of coffee a day and regardless of how much they smoke the conference interval includes one so we don't care about this one but here we do look at that this conference interval does not include one and so regardless of how much whether you smoke or not if you just drink three or more cups of coffee a day you got an increased risk of pancreatic cancer okay so we're halfway through what we need to do and near the end of this video but what we've learned so far is that this is a case control trial and we found that there's a strong association with coffee consumption and even if you include the cigarettes and it still stays there but not for tea or alcoholic beverage alcoholic beverages so now let's go through the rest of the paper in the next video

Channel: Rahul Patwari

Views: 22,569

Rating: 4.9175258 out of 5

Keywords:

Id: SOIF7YIIlFg

Length: 11min 51sec (711 seconds)

Published: Tue Dec 29 2015