WEBVTT 1 00:00:01.470 --> 00:00:04.320 Hello, in this presentation I'll introduce you 2 00:00:04.320 --> 00:00:06.540 to Hill's Causal Criteria 3 00:00:06.540 --> 00:00:09.270 before walking through a case study. 4 00:00:09.270 --> 00:00:10.890 Hopefully the case study will help you 5 00:00:10.890 --> 00:00:14.550 to think about your causality assignment for this week. 6 00:00:14.550 --> 00:00:16.770 I strongly recommend that you watch the lecture 7 00:00:16.770 --> 00:00:19.530 on the different types of epidemiological studies 8 00:00:19.530 --> 00:00:22.500 before going through this lecture because an understanding 9 00:00:22.500 --> 00:00:24.750 of study types will help you to follow along. 10 00:00:28.200 --> 00:00:30.210 In epidemiology, we are often interested 11 00:00:30.210 --> 00:00:34.140 in whether a given exposure causes a specific disease. 12 00:00:34.140 --> 00:00:37.530 If it does, then prevention is possible. 13 00:00:37.530 --> 00:00:38.700 We may also be interested in 14 00:00:38.700 --> 00:00:42.330 whether a given exposure protects against disease. 15 00:00:42.330 --> 00:00:46.170 If yes, then health promotion is possible. 16 00:00:46.170 --> 00:00:48.480 But how do we determine whether an association 17 00:00:48.480 --> 00:00:51.753 between an exposure and a health outcome is causal? 18 00:00:54.720 --> 00:00:56.160 First, it's important to note 19 00:00:56.160 --> 00:01:00.120 that correlation does not equal causation. 20 00:01:00.120 --> 00:01:01.950 In this graph here, it looks like 21 00:01:01.950 --> 00:01:05.037 per capita cheese consumption is correlated with 22 00:01:05.037 --> 00:01:09.240 the number of people who die tangled in their bedsheets. 23 00:01:09.240 --> 00:01:10.620 However, that doesn't mean 24 00:01:10.620 --> 00:01:14.523 that eating more cheese causes death by bedsheet tangle. 25 00:01:18.570 --> 00:01:22.020 Here's another example, just by looking at the graph, 26 00:01:22.020 --> 00:01:24.270 you can see the ice cream sales 27 00:01:24.270 --> 00:01:26.335 and shark attacks both go up in the summer 28 00:01:26.335 --> 00:01:30.570 but that doesn't mean that one causes the other. 29 00:01:30.570 --> 00:01:32.610 These may seem like silly examples 30 00:01:32.610 --> 00:01:34.020 but it can be an easy mistake 31 00:01:34.020 --> 00:01:37.083 to make an epidemiology so you really need to be aware. 32 00:01:39.870 --> 00:01:44.130 So how do we determine if an association is causal? 33 00:01:44.130 --> 00:01:46.020 A common way to assess causality 34 00:01:46.020 --> 00:01:49.560 is through the use of Hill's Causal Criteria. 35 00:01:49.560 --> 00:01:51.630 However, Hill's Causal Criteria should be thought 36 00:01:51.630 --> 00:01:54.723 of as guidelines because they're not foolproof. 37 00:01:57.060 --> 00:02:00.000 So first is strength of association. 38 00:02:00.000 --> 00:02:02.577 Strength of association between the exposure of interest 39 00:02:02.577 --> 00:02:06.030 and the outcome is often measured via risk ratios 40 00:02:06.030 --> 00:02:09.360 rate ratios, or odds ratios. 41 00:02:09.360 --> 00:02:11.790 Hill thought that causal relationships were more 42 00:02:11.790 --> 00:02:16.790 likely to show strong associations than non causal agents. 43 00:02:17.850 --> 00:02:20.400 However, a weak association doesn't necessarily 44 00:02:20.400 --> 00:02:23.760 indicate non causality, and this is particularly 45 00:02:23.760 --> 00:02:26.673 true when the outcome of interest is relatively common. 46 00:02:27.600 --> 00:02:31.500 Additionally, a strong association alone doesn't indicate 47 00:02:31.500 --> 00:02:34.623 causality and could be influenced by confounding. 48 00:02:37.705 --> 00:02:41.670 Consistency refers to the reproducibility of results. 49 00:02:41.670 --> 00:02:42.810 Do you see the same results 50 00:02:42.810 --> 00:02:45.990 across various populations and situations? 51 00:02:45.990 --> 00:02:48.750 Consistency can be used to rule out other explanations 52 00:02:48.750 --> 00:02:50.610 for the outcome of interest. 53 00:02:50.610 --> 00:02:51.690 Generally, the greater 54 00:02:51.690 --> 00:02:55.480 the consistency the more likely the causal association 55 00:02:57.870 --> 00:03:00.330 Specificity a tricky criterion 56 00:03:00.330 --> 00:03:03.930 and can be invalid in a number of instances. 57 00:03:03.930 --> 00:03:06.810 The idea was that a specific exposure leads solely 58 00:03:06.810 --> 00:03:08.850 to one specific outcome. 59 00:03:08.850 --> 00:03:10.950 However, in the case of smoking, we know 60 00:03:10.950 --> 00:03:14.310 that smoking can lead to lung cancer as well as a number 61 00:03:14.310 --> 00:03:18.120 of other clinical disorders, although it should be noted 62 00:03:18.120 --> 00:03:20.070 that cigarettes do contain a large mix 63 00:03:20.070 --> 00:03:22.200 of different substances. 64 00:03:22.200 --> 00:03:26.610 However, a one-to-one relationship does exist in some cases. 65 00:03:26.610 --> 00:03:29.790 Take for example, specific pathogens 66 00:03:29.790 --> 00:03:32.940 which are necessary to produce a specific disease. 67 00:03:32.940 --> 00:03:37.940 So the virus SARS-COV-2 is necessary to produce COVID-19. 68 00:03:42.344 --> 00:03:45.000 Temporality is an essential criteria though. 69 00:03:45.000 --> 00:03:49.470 For an exposure to be causal, its presence must precede 70 00:03:49.470 --> 00:03:51.183 the development of the outcome. 71 00:03:54.570 --> 00:03:57.150 The presence of a dose response relationship 72 00:03:57.150 --> 00:04:00.600 between an exposure and outcome can provide good evidence 73 00:04:00.600 --> 00:04:02.370 for a causal relationship. 74 00:04:02.370 --> 00:04:05.700 But lack of a dose response relationship doesn't rule out 75 00:04:05.700 --> 00:04:07.770 a causal relationship because some health 76 00:04:07.770 --> 00:04:11.730 outcomes don't display a typical dose response relationship 77 00:04:11.730 --> 00:04:13.470 with an exposure. 78 00:04:13.470 --> 00:04:17.220 For example, there may be a threshold association instead 79 00:04:17.220 --> 00:04:19.200 and this is something we'll get into 80 00:04:19.200 --> 00:04:20.970 in the next module when we start talking 81 00:04:20.970 --> 00:04:22.083 about toxicology. 82 00:04:26.945 --> 00:04:29.580 Information of biological plausibility is often garnered 83 00:04:29.580 --> 00:04:31.230 through laboratory data 84 00:04:31.230 --> 00:04:33.813 and can further support a causal relationship. 85 00:04:35.790 --> 00:04:39.270 Coherence represents the idea that for a causal association 86 00:04:39.270 --> 00:04:42.600 to be supported, any new data shouldn't provide 87 00:04:42.600 --> 00:04:45.090 evidence against causality. 88 00:04:45.090 --> 00:04:47.490 However, you have to be cautious about making conclusions 89 00:04:47.490 --> 00:04:50.520 about causality based on lack of coherence 90 00:04:50.520 --> 00:04:53.193 because information can be biased or incorrect. 91 00:04:55.290 --> 00:04:57.600 Experiment evidence can come 92 00:04:57.600 --> 00:05:00.900 from laboratory studies, epidemiological studies 93 00:05:00.900 --> 00:05:03.900 clinical trials, or other research. 94 00:05:03.900 --> 00:05:06.570 Ideally, the research would involve a well controlled study 95 00:05:06.570 --> 00:05:10.650 with randomized control trials being a gold standard. 96 00:05:10.650 --> 00:05:12.960 These types of studies can support causality 97 00:05:12.960 --> 00:05:16.683 by showing that altering the cause alters the effect. 98 00:05:17.910 --> 00:05:21.360 And then finally, analogy is one of the weaker criterion 99 00:05:21.360 --> 00:05:26.360 because it is speculative and depends on subjective opinion. 100 00:05:26.580 --> 00:05:30.270 But taken together, you would look at all of these criteria 101 00:05:30.270 --> 00:05:33.843 and then determine, make a determination of causality. 102 00:05:36.480 --> 00:05:39.510 So now let's walk through a case study. 103 00:05:39.510 --> 00:05:43.620 Repetitive head impacts defined as the cumulative exposure 104 00:05:43.620 --> 00:05:48.150 to recurrent concussive and sub-concussive events have long 105 00:05:48.150 --> 00:05:50.850 been associated with the neurodegenerative disease 106 00:05:50.850 --> 00:05:55.850 chronic traumatic encephalopathy, or CTE. 107 00:05:56.340 --> 00:05:58.830 However, in recent years, some people have questioned 108 00:05:58.830 --> 00:06:01.800 whether the relationship between the repetitive head impacts 109 00:06:01.800 --> 00:06:03.753 and CTE is causal. 110 00:06:05.160 --> 00:06:10.140 This article shown on the slide applies Hill's 111 00:06:10.140 --> 00:06:12.153 Causal Criteria to this question. 112 00:06:17.265 --> 00:06:21.990 Really quickly, you might be wondering what exactly is CTE? 113 00:06:21.990 --> 00:06:24.390 So the graphic on this page shows 114 00:06:24.390 --> 00:06:29.390 the four stages of CTE starting with no symptoms and with 115 00:06:29.490 --> 00:06:32.640 symptoms progressively worsening until reaching stage four, 116 00:06:32.640 --> 00:06:33.873 advanced dementia. 117 00:06:40.410 --> 00:06:44.100 Strength of association is the first criterion. 118 00:06:44.100 --> 00:06:45.930 An analysis of strength 119 00:06:45.930 --> 00:06:48.480 of association starts from the premise 120 00:06:48.480 --> 00:06:52.080 that for an association to be statistically significant, 121 00:06:52.080 --> 00:06:53.820 the lower confidence limit 122 00:06:53.820 --> 00:06:57.565 on the odds ratio or relative risk must be greater than one, 123 00:06:57.565 --> 00:07:00.385 that is very unlikely 124 00:07:00.385 --> 00:07:02.733 to be caused by random chance. 125 00:07:03.840 --> 00:07:07.290 Other things being equal, the larger the central estimate 126 00:07:07.290 --> 00:07:09.720 of the odds ratio or relative risk, 127 00:07:09.720 --> 00:07:12.720 the less likely that the significant association 128 00:07:12.720 --> 00:07:16.683 is an artifact of confounding bias or misclassification. 129 00:07:18.240 --> 00:07:22.170 An odds ratio or relative risk that is equal 130 00:07:22.170 --> 00:07:27.170 to 2.0 or greater is generally regarded as strong. 131 00:07:27.180 --> 00:07:29.610 This numerical choice is subjective, but 132 00:07:29.610 --> 00:07:32.790 the appeal of using 2.0 as a benchmark 133 00:07:32.790 --> 00:07:36.270 is that above this ratio is more likely 134 00:07:36.270 --> 00:07:40.020 than not that any individual person with the exposure 135 00:07:40.020 --> 00:07:43.260 and the disease will have contracted the disease 136 00:07:43.260 --> 00:07:46.923 because of the exposure than for any other reason. 137 00:07:48.240 --> 00:07:50.220 So cohort studies are a common method 138 00:07:50.220 --> 00:07:53.340 by which to identify odds ratios, but 139 00:07:53.340 --> 00:07:55.920 in this case there were no cohort studies available 140 00:07:55.920 --> 00:08:00.150 that followed both symptomatic and asymptomatic individuals 141 00:08:00.150 --> 00:08:04.080 through to autopsy to confirm the presence or absence 142 00:08:04.080 --> 00:08:09.000 of CTE because it has to be confirmed during an autopsy. 143 00:08:09.000 --> 00:08:10.590 So in this case 144 00:08:10.590 --> 00:08:14.250 the authors identified six case control studies 145 00:08:14.250 --> 00:08:17.040 where the researchers made a reasonable attempt 146 00:08:17.040 --> 00:08:20.460 to identify RHI history 147 00:08:20.460 --> 00:08:24.030 and had more than 50 subjects to be sufficiently powered 148 00:08:24.030 --> 00:08:26.373 for statistical significance. 149 00:08:27.390 --> 00:08:30.510 So all six studies reported large 150 00:08:30.510 --> 00:08:35.490 and significant odds ratios of CTE cases in subjects 151 00:08:35.490 --> 00:08:40.490 with substantial exposure to RHI compared to the controls. 152 00:08:40.650 --> 00:08:42.040 So the clear finding was 153 00:08:43.015 --> 00:08:47.340 that the associations between RHI and CTE cases are strong 154 00:08:47.340 --> 00:08:51.063 and they're consistently above the 2.0 benchmark. 155 00:08:54.300 --> 00:08:57.270 The next criterion is consistency. 156 00:08:57.270 --> 00:08:58.950 So is the association 157 00:08:58.950 --> 00:09:02.280 between repetitive head injuries or impacts 158 00:09:02.280 --> 00:09:07.280 and CTE demonstrated across multiple studies conducted 159 00:09:07.290 --> 00:09:09.780 by independent groups investigating different 160 00:09:09.780 --> 00:09:13.830 patient populations and using different approaches? 161 00:09:13.830 --> 00:09:15.780 The short answer is yes. 162 00:09:15.780 --> 00:09:19.980 Investigations around the world have described CTE pathology 163 00:09:19.980 --> 00:09:21.600 in former athletes and 164 00:09:21.600 --> 00:09:25.290 in military veterans, all individuals who are subject 165 00:09:25.290 --> 00:09:28.293 to repetitive head impact exposure. 166 00:09:29.975 --> 00:09:31.440 Overall the current literature 167 00:09:31.440 --> 00:09:34.920 from independent researchers and from brain banks, 168 00:09:34.920 --> 00:09:39.360 brain banks being research institutes that accept donations 169 00:09:39.360 --> 00:09:44.360 of brains so that they can examine them postmortem, 170 00:09:46.380 --> 00:09:48.420 all of these provide strong support 171 00:09:48.420 --> 00:09:50.733 for the consistency criterion. 172 00:09:54.300 --> 00:09:57.660 Specificity is the next criterion 173 00:09:57.660 --> 00:10:01.830 and I've included a brief quote from the paper. 174 00:10:01.830 --> 00:10:04.295 Specificity can be observed in a number of ways 175 00:10:04.295 --> 00:10:07.470 but perhaps most significantly by the fact 176 00:10:07.470 --> 00:10:12.470 that prior exposure to RHI is the only known unifying factor 177 00:10:12.480 --> 00:10:16.410 among CTE cases reported in the literature to date. 178 00:10:16.410 --> 00:10:18.813 So it clearly meets this criterion. 179 00:10:20.220 --> 00:10:22.503 The next criterion is temporality. 180 00:10:23.610 --> 00:10:27.810 RHI exposure clearly proceeds CTE pathology, 181 00:10:27.810 --> 00:10:30.090 typically years or decades earlier. 182 00:10:30.090 --> 00:10:33.240 So often we don't know that someone had CTE 183 00:10:33.240 --> 00:10:36.210 until the autopsy occurs, 184 00:10:36.210 --> 00:10:40.800 and the head injury or head impact exposure often happened 185 00:10:40.800 --> 00:10:42.570 many years beforehand. 186 00:10:42.570 --> 00:10:45.300 So that clearly fulfills this criterion. 187 00:10:45.300 --> 00:10:47.220 However, because we can only 188 00:10:47.220 --> 00:10:50.610 confirm CTE pathology postmortem 189 00:10:50.610 --> 00:10:54.150 it's unknown when the pathological process actually 190 00:10:54.150 --> 00:10:56.490 manifests for this disease. 191 00:10:56.490 --> 00:10:59.430 So it's important to ask could the pathology develop 192 00:10:59.430 --> 00:11:02.970 before the RHI exposure actually happened? 193 00:11:02.970 --> 00:11:05.470 There are several factors that make this unlikely. 194 00:11:06.600 --> 00:11:08.340 I won't go into it now, but if you're interested 195 00:11:08.340 --> 00:11:11.850 you can read through the entire article. 196 00:11:11.850 --> 00:11:13.620 There's additional temporal evidence 197 00:11:13.620 --> 00:11:15.750 that comes from the clinical manifestation 198 00:11:15.750 --> 00:11:20.463 of CTE which is traumatic encephalopathy syndrome or TES. 199 00:11:21.390 --> 00:11:24.810 So all criteria for TES require a history 200 00:11:24.810 --> 00:11:26.790 of exposure to head injuries 201 00:11:26.790 --> 00:11:28.710 and the symptoms closely associated 202 00:11:28.710 --> 00:11:30.720 with TES appear to manifest 203 00:11:30.720 --> 00:11:32.790 in the years following the exposure. 204 00:11:32.790 --> 00:11:35.103 So that helps to confirm this criterion. 205 00:11:37.770 --> 00:11:40.473 The next criterion is dose response. 206 00:11:42.885 --> 00:11:46.200 CTE has been recognized by multiple studies over the course 207 00:11:46.200 --> 00:11:49.470 of a century as having a strong dose response relationship 208 00:11:49.470 --> 00:11:53.010 with RHI in several different sports. 209 00:11:53.010 --> 00:11:56.643 Examples include boxing and American football. 210 00:11:57.930 --> 00:11:59.880 So ringside observations 211 00:11:59.880 --> 00:12:03.810 of boxers as far back as 1928 included 212 00:12:03.810 --> 00:12:06.270 the recognition of a dose response relationship 213 00:12:06.270 --> 00:12:09.483 between RHI and possible CTE symptoms. 214 00:12:10.320 --> 00:12:13.260 And the paper states that boxers who had the greatest number 215 00:12:13.260 --> 00:12:15.930 of fights, who had a long duration 216 00:12:15.930 --> 00:12:19.260 of fighting, who were sluggers, meaning they took a lot 217 00:12:19.260 --> 00:12:23.490 of blows, or who sparred extensively were most likely 218 00:12:23.490 --> 00:12:27.780 to show a clinical triad of mental confusion 219 00:12:27.780 --> 00:12:31.440 and recent memory impairment, emotional liability 220 00:12:31.440 --> 00:12:35.370 with mood swings and diminished ability to control temper, 221 00:12:35.370 --> 00:12:38.545 and Parkinsonian signs including tremor, 222 00:12:38.545 --> 00:12:39.873 rigidity and bradykinesia. 223 00:12:42.993 --> 00:12:44.878 With regards to American football, 224 00:12:44.878 --> 00:12:48.270 the Boston University CTE Center researchers showed 225 00:12:48.270 --> 00:12:53.270 that each additional year of football play was associated 226 00:12:53.490 --> 00:12:55.180 with an increased odds ratio 227 00:12:56.465 --> 00:13:00.893 of 1.3 with a 95% confidence interval of 1.19 to 1.41 228 00:13:03.930 --> 00:13:06.670 effectively doubling the odds of disease 229 00:13:07.662 --> 00:13:09.063 every 2.6 years played. 230 00:13:09.990 --> 00:13:13.440 The study also found that each additional year 231 00:13:13.440 --> 00:13:17.760 of football play was associated with an increased odds ratio 232 00:13:17.760 --> 00:13:22.760 of 1.14 with a 95% confidence interval 233 00:13:22.890 --> 00:13:27.890 of 1.07 to 1.22 for having more severe CTE 234 00:13:29.280 --> 00:13:32.070 so effectively doubling the odds of having severe 235 00:13:32.070 --> 00:13:37.070 versus mild CTE every additional 5.3 years of play. 236 00:13:42.810 --> 00:13:46.290 Next, what about biological plausibility? 237 00:13:46.290 --> 00:13:48.960 There's still a lot that we don't know about CTE 238 00:13:48.960 --> 00:13:50.640 but the current literature supports 239 00:13:50.640 --> 00:13:53.853 the plausibility of RHI causing CTE. 240 00:13:55.020 --> 00:13:59.190 The evidence comes from mouse models, computational models, 241 00:13:59.190 --> 00:14:01.320 and studies of wildlife. 242 00:14:01.320 --> 00:14:05.520 So interestingly, researchers have found CTE lesions 243 00:14:05.520 --> 00:14:09.960 in the brains of musk oxen and bighorn sheep, 244 00:14:09.960 --> 00:14:12.090 animals that share certain brain similarities 245 00:14:12.090 --> 00:14:14.520 with humans that are exposed 246 00:14:14.520 --> 00:14:18.483 to RHI through combative headbutting in the wild. 247 00:14:23.160 --> 00:14:24.720 So coherence requires 248 00:14:24.720 --> 00:14:28.710 that the interpretation of the data not seriously conflict 249 00:14:28.710 --> 00:14:30.120 with what is already known 250 00:14:30.120 --> 00:14:32.493 about the disease or the exposure. 251 00:14:34.110 --> 00:14:35.370 There are a couple of areas 252 00:14:35.370 --> 00:14:38.343 of coherence in the literature related to CTE. 253 00:14:39.465 --> 00:14:41.970 So for example, CTE has yet 254 00:14:41.970 --> 00:14:45.660 to be diagnosed in a female athlete worldwide 255 00:14:45.660 --> 00:14:50.250 and this may reflect differences in sports RHI exposure. 256 00:14:50.250 --> 00:14:52.680 So some sex specific rules are 257 00:14:52.680 --> 00:14:54.510 in place for sports like ice hockey 258 00:14:54.510 --> 00:14:58.740 and lacrosse where body checking is illegal for females. 259 00:14:58.740 --> 00:15:02.790 Also, historical discrimination in sports participation 260 00:15:02.790 --> 00:15:06.090 mean that there are fewer female former contact sports 261 00:15:06.090 --> 00:15:09.510 participants in the US over the age of 60. 262 00:15:09.510 --> 00:15:12.210 Bear in mind that Title IX was enacted 263 00:15:12.210 --> 00:15:14.970 in 1972 and provided females 264 00:15:14.970 --> 00:15:18.123 with equal opportunity to participate in school sports. 265 00:15:20.430 --> 00:15:22.775 Another way to look at coherence 266 00:15:22.775 --> 00:15:26.490 is to explore alternative hypotheses. 267 00:15:26.490 --> 00:15:29.130 So many variables have been highlighted 268 00:15:29.130 --> 00:15:34.130 as potential confounders for CTE including substance abuse, 269 00:15:34.560 --> 00:15:37.380 mental health disorders, genetics, 270 00:15:37.380 --> 00:15:40.770 lifestyle factors like diet and exercise. 271 00:15:40.770 --> 00:15:42.720 however, so far, none 272 00:15:42.720 --> 00:15:45.420 of these appear to explain the association 273 00:15:45.420 --> 00:15:50.420 between RHI and CTE and RHI remains the only candidate 274 00:15:51.510 --> 00:15:54.933 risk factor for CTE causation. 275 00:15:59.310 --> 00:16:01.200 Experimental evidence can be difficult 276 00:16:01.200 --> 00:16:04.800 to obtain in the context of environmental exposures. 277 00:16:04.800 --> 00:16:08.070 However, researchers can ask themselves 278 00:16:08.070 --> 00:16:10.230 if a preventive action is taken 279 00:16:10.230 --> 00:16:12.300 due to an observed association 280 00:16:12.300 --> 00:16:14.760 between an exposure and a disease, 281 00:16:14.760 --> 00:16:18.720 does that action in fact prevent the disease outcome? 282 00:16:18.720 --> 00:16:20.370 They can also ask themselves 283 00:16:20.370 --> 00:16:24.033 do other experimental studies suggest a causal relationship? 284 00:16:25.650 --> 00:16:29.250 In this case, CTE has only been widely recognized outside 285 00:16:29.250 --> 00:16:30.720 of the sport of boxing 286 00:16:30.720 --> 00:16:33.345 for about 20 years which isn't enough time for there 287 00:16:33.345 --> 00:16:38.100 to have been meaningful studies of CTE prevention efforts. 288 00:16:38.100 --> 00:16:40.500 It would also be difficult to control all variables 289 00:16:40.500 --> 00:16:44.133 of interest in such a study to draw any firm conclusions. 290 00:16:45.030 --> 00:16:46.140 However, findings 291 00:16:46.140 --> 00:16:50.310 from preclinical RHI studies that replicate elements 292 00:16:50.310 --> 00:16:54.360 of observed human pathophysiological imaging and 293 00:16:54.360 --> 00:16:58.113 epidemiological studies do suggest the causal relationship. 294 00:17:02.730 --> 00:17:04.690 Final criterion is analogy 295 00:17:05.670 --> 00:17:10.050 and the authors described two analogies in this study. 296 00:17:10.050 --> 00:17:11.340 The first is 297 00:17:11.340 --> 00:17:13.470 that the evidence supporting traumatic brain injury 298 00:17:13.470 --> 00:17:17.040 as a causal risk factor for dementia is analogous 299 00:17:17.040 --> 00:17:20.043 to the relationship between RHI and CTE. 300 00:17:21.000 --> 00:17:22.470 The second is that the discovery 301 00:17:22.470 --> 00:17:25.860 of a link between smoking and lung cancer is analogous 302 00:17:25.860 --> 00:17:27.933 to the evidence linking RHI to CTE. 303 00:17:29.700 --> 00:17:32.315 To elaborate on this second point, 304 00:17:32.315 --> 00:17:33.945 the authors wrote, while the link 305 00:17:33.945 --> 00:17:37.770 between smoking and lung cancer is widely accepted 306 00:17:37.770 --> 00:17:41.430 key questions remain unanswered or in completely answered 307 00:17:41.430 --> 00:17:44.970 including what precisely constitutes a smoked cigarette, 308 00:17:44.970 --> 00:17:49.970 the dose, why some smokers develop cancer and others do not, 309 00:17:50.310 --> 00:17:54.420 how many cigarettes are too many or which specific cigarette 310 00:17:54.420 --> 00:17:57.990 or carcinogen sparked the lung cancer? 311 00:17:57.990 --> 00:18:00.060 The fact that these questions also remain 312 00:18:00.060 --> 00:18:03.990 for RHI and sub-concussive impacts is often raised 313 00:18:03.990 --> 00:18:06.150 in a as a reason that conclusions 314 00:18:06.150 --> 00:18:10.113 on RHI CTE causation cannot be drawn. 315 00:18:11.130 --> 00:18:14.310 These knowledge gaps have not limited the ability 316 00:18:14.310 --> 00:18:17.820 to assert a causal link between smoking and lung cancer 317 00:18:17.820 --> 00:18:21.000 and similarly should not limit the ability to determine 318 00:18:21.000 --> 00:18:24.783 the likelihood of a causal link between RHI and CTE. 319 00:18:32.445 --> 00:18:34.170 So to conclude, in the case 320 00:18:34.170 --> 00:18:36.400 of repetitive head impacts and CTE 321 00:18:37.710 --> 00:18:41.370 the researchers established the strength of the association 322 00:18:41.370 --> 00:18:45.900 confirmed consistency, specificity, temporality 323 00:18:45.900 --> 00:18:47.760 a dose response relationship, 324 00:18:47.760 --> 00:18:51.150 biological possibility, and coherence. 325 00:18:51.150 --> 00:18:53.700 They also provided the experimental evidence 326 00:18:53.700 --> 00:18:56.130 that was available and described analogies 327 00:18:56.130 --> 00:18:58.923 to other established causal relationships. 328 00:19:00.240 --> 00:19:02.760 They concluded their article by stating that 329 00:19:02.760 --> 00:19:07.410 the evidence on the link between RHI and CTE is imperfect 330 00:19:07.410 --> 00:19:08.940 and like all similar research, 331 00:19:08.940 --> 00:19:12.180 it will remain imperfect in perpetuity. 332 00:19:12.180 --> 00:19:13.860 After reviewing the medical literature 333 00:19:13.860 --> 00:19:18.270 on RHI and CTE through the Bradford Hill criteria 334 00:19:18.270 --> 00:19:19.740 we have the highest confidence 335 00:19:19.740 --> 00:19:23.490 in the conclusion that RHI causes CTE. 336 00:19:23.490 --> 00:19:25.380 We encourage the medical, scientific 337 00:19:25.380 --> 00:19:26.700 and public health communities 338 00:19:26.700 --> 00:19:29.970 to now act under the premise of a causal relationship 339 00:19:29.970 --> 00:19:34.170 and take immediate action to prevent CTE, minimize risk 340 00:19:34.170 --> 00:19:35.340 and develop therapeutics 341 00:19:35.340 --> 00:19:37.683 to slow or stop disease progression. 342 00:19:40.350 --> 00:19:42.300 If you have any additional questions 343 00:19:42.300 --> 00:19:46.320 about this study and Hill's Causal Criteria 344 00:19:46.320 --> 00:19:47.153 I would encourage you 345 00:19:47.153 --> 00:19:51.720 to read the full article at the link provided on slide six 346 00:19:51.720 --> 00:19:54.060 but hopefully this has helped to give you a good idea 347 00:19:54.060 --> 00:19:57.663 of how Hill's Causal Criteria can be applied in real life.