YouTube2Text

Summary Week 11 — Transcript

by Tessa Crume · 1,409 words · 245 segments · language en · Watch on YouTube

Full transcript

  1. 0:02this week we learned about
  2. 0:05standardization as a technique to adjust
  3. 0:10for
  4. 0:11confounding when we're comparing two
  5. 0:14different
  6. 0:15populations that have
  7. 0:18different uh composition differences so
  8. 0:21populations that have differences in age
  9. 0:24distribution socioeconomic distribution
  10. 0:27different demographic differences across
  11. 0:29po
  12. 0:30calulations when we look at results from
  13. 0:33public health
  14. 0:35agencies or descriptive studies we are
  15. 0:39often looking at rates that have been
  16. 0:42standardized in order to be able to
  17. 0:45compare and draw conclusions about any
  18. 0:49excess categories or any excess
  19. 0:53demographic groups that may have higher
  20. 0:57rates than
  21. 0:58another standardization helps to remove
  22. 1:02differences in the composition between
  23. 1:05po populations that are being compared
  24. 1:09by mathematical procedures so it's a
  25. 1:13technique that allows us to anal
  26. 1:16analytically adjust for a biasing factor
  27. 1:21on the basis of differences in
  28. 1:24population
  29. 1:27composition we uh learned about
  30. 1:30standardization as involving data from
  31. 1:33two
  32. 1:34populations are study populations or
  33. 1:37populations of Interest which are the
  34. 1:40ones we are going to
  35. 1:41standardize and then a standard or
  36. 1:45reference population that we will
  37. 1:47standardize
  38. 1:49against so I taught you about two forms
  39. 1:52of standardization direct
  40. 1:54standardization and indirect
  41. 1:58standardization Direct standardization
  42. 2:00is always what we would prefer to do
  43. 2:03because it allows us to then compare
  44. 2:07rates between study populations that
  45. 2:09were adjusted to the same standard
  46. 2:12population so for example if I wanted to
  47. 2:16compare rates in covid infections
  48. 2:20between different states in the United
  49. 2:23States I could standardize them all to
  50. 2:25the US population as my standard
  51. 2:28population and then I could compare
  52. 2:32those rates between states even though
  53. 2:34they all might have differences and
  54. 2:36they're underlying population
  55. 2:39distributions or if I wanted to look at
  56. 2:43rates
  57. 2:45of suicide
  58. 2:47mortality or mortality from different
  59. 2:52cancers I could do that by adjusting
  60. 2:55them to the same standard
  61. 2:57population the choice of a standard
  62. 2:59population is somewhat arbitrary there
  63. 3:02are different situations that inform our
  64. 3:05choice of a standard population but most
  65. 3:09commonly we're going to choose it as a
  66. 3:11larger population so for example if I'm
  67. 3:14standardizing rates across Colorado
  68. 3:17counties I would choose Colorado as a
  69. 3:20whole as the standard population or if I
  70. 3:24wanted to compare rates across States I
  71. 3:26would choose the us as a whole as the
  72. 3:29state standard
  73. 3:31population but in order to do direct
  74. 3:34standardization we have to have Stratus
  75. 3:37specific rates from our population of
  76. 3:40Interest the one that we are
  77. 3:43standardizing so if I wanted to look at
  78. 3:46rates in uh covid infection and I'm
  79. 3:50concerned about age different ages of
  80. 3:53the population as being a potential
  81. 3:56confounding Factor then I would need
  82. 4:00Stratus specific rates of covid
  83. 4:02infection for the different age
  84. 4:06categories and for a variety of reasons
  85. 4:09sometimes we don't have these so that's
  86. 4:12why we have to turn to indirect
  87. 4:15standardization but direct
  88. 4:17standardization is always preferable
  89. 4:20because it allows us to answer the
  90. 4:22question how many events would occur
  91. 4:25among the standard population if they
  92. 4:27had the same rates as the study
  93. 4:30population and once we calculate our
  94. 4:33adjusted rates we can compare them to
  95. 4:35each other if they were adjusted to the
  96. 4:38same standard
  97. 4:40population the situation in which we
  98. 4:43cannot do direct standardization is if
  99. 4:46we don't have the stratus specific rates
  100. 4:49so maybe I don't know what the age
  101. 4:52specific rates uh covid infection are
  102. 4:56but I'm still worried about differences
  103. 4:58in population
  104. 5:01composition and um then I would need to
  105. 5:04turn to indirect standardization so we
  106. 5:07use this when we know the total number
  107. 5:09of cases in our study population but we
  108. 5:12don't have Stratus specific rates some
  109. 5:16of the reasons that we may not have
  110. 5:18these rates is because we just don't
  111. 5:20have that data available to us we're
  112. 5:22turning to a surveillance source that
  113. 5:25doesn't provide Stratus specific rates
  114. 5:28but most commonly the reason we don't
  115. 5:31have these strus specific rates is
  116. 5:33because they're based on small numbers
  117. 5:35and when we calculate rates based on
  118. 5:37small numbers they're very unstable they
  119. 5:40lack
  120. 5:41precision and the difference between
  121. 5:44adding one or two more cases can result
  122. 5:47in a huge jump in that rate which we
  123. 5:50don't think is really reflective of
  124. 5:53changes in the population but it's a
  125. 5:56problem with Precision because we're
  126. 5:58dealing with really small number
  127. 6:00numbers so we turn to indirect
  128. 6:03standardization which uses strata
  129. 6:06specific rates from a standard
  130. 6:08population a larger population in which
  131. 6:12the stratus specific rates are stable
  132. 6:14because they're based on larger numbers
  133. 6:17and we use the total number of observed
  134. 6:21cases from our study
  135. 6:23population this is very common in a lot
  136. 6:27of uh looking at hand
  137. 6:30rates where we might not we might have a
  138. 6:33small number of total cases thank
  139. 6:35goodness but um we don't have large
  140. 6:39enough numbers to calculate Stratus
  141. 6:42specific rates or those rates are just
  142. 6:44unstable because they're based on very
  143. 6:46few cases so when we use indirect
  144. 6:49standardization what we calculate is
  145. 6:52called a standardized morbility sorry
  146. 6:56mortality or incidence ratio so an SMR
  147. 7:00or an
  148. 7:02S and this is calculated as the ratio of
  149. 7:06observed versus expected number of cases
  150. 7:10The observed number of cases come from
  151. 7:13our study population and the expected
  152. 7:16number of cases come from the rates that
  153. 7:19we're able to get from our standard
  154. 7:22population applied to the stratus
  155. 7:25specific numbers for uh demographics in
  156. 7:29our study
  157. 7:30population it answers a slightly
  158. 7:33different question than we can answer
  159. 7:35with direct
  160. 7:36standardization and it answers a more
  161. 7:38limited question so indirect
  162. 7:42standardization where we calculate smrs
  163. 7:45or S IRS answers the question what is
  164. 7:49the excess mortality in the study
  165. 7:51population compared to the reference
  166. 7:55population so for example if I know the
  167. 7:58total number number of cases of
  168. 8:01pancreatic cancer in Colorado but I
  169. 8:04don't know what those rates are
  170. 8:07according to age group and I think that
  171. 8:09age could be an important confounding
  172. 8:12factor I could use the US population as
  173. 8:16the standard population in which there
  174. 8:18are enough cases to generate age
  175. 8:22specific rates and I apply that to the
  176. 8:26demographic composition of my study
  177. 8:28population
  178. 8:30Colorado according to the age groups and
  179. 8:33I calculate the expected number of cases
  180. 8:36for each age group that would occur if
  181. 8:40Colorado had the same age specific rates
  182. 8:43as my standard population of the
  183. 8:46US and then I calculate my standardized
  184. 8:51incidence
  185. 8:52ratio and I can interpret it as the
  186. 8:57amount of mortality in my study
  187. 9:00population relative to my standard
  188. 9:05population one important limitation of
  189. 9:08indirect standardization where we
  190. 9:10calculate smrs and SS is that we cannot
  191. 9:14compare them to each other even if
  192. 9:17they're adjusted to the same standard
  193. 9:20population the reason is because we're
  194. 9:24using the rates from the standard
  195. 9:27population in indirect Ste
  196. 9:29ization so they are not coming from our
  197. 9:34study population so if I calculate an
  198. 9:38smmr for
  199. 9:40Colorado using the us as the standard
  200. 9:43population and I calculate another one
  201. 9:46for New Mexico again using the us as the
  202. 9:49study population with indirect
  203. 9:52standardization I could not compare that
  204. 9:55SMR from Colorado to that of New Mexico
  205. 9:59C because indirect standardization has a
  206. 10:03more limited interpretation the only
  207. 10:06interpretation is the excess mortality
  208. 10:10excess or deficit mortality in the study
  209. 10:13population compared to the standard
  210. 10:17population so this week I walk through
  211. 10:21these calculations in the lecture make
  212. 10:23sure that you look at the flow of the
  213. 10:26calculation it can be a little confusing
  214. 10:28to think about these concepts of a
  215. 10:30standard population and a study
  216. 10:34population and then also thinking about
  217. 10:36our Stratus specific rates from these
  218. 10:40populations and our Stratus specific
  219. 10:43demographic composition so the challenge
  220. 10:47in mastering these
  221. 10:48calculations is understanding the inputs
  222. 10:52necessary for direct and indirect
  223. 10:55standardization and the appropriate
  224. 10:58interpretations to Aid in helping you
  225. 11:01master the flow of these calculations we
  226. 11:04have an applied EP exercise this week
  227. 11:07please take time to work through that
  228. 11:09exercise you need practice applying
  229. 11:11these
  230. 11:12calculations and then I will host a a
  231. 11:17virtual live recitation section next
  232. 11:21week to walk through these calculations
  233. 11:24as a group remember your applied EP
  234. 11:26exercises are graded on compl completion
  235. 11:30not correctness so don't worry if you're
  236. 11:34not able to work through all these
  237. 11:36calculations I want you to attempt them
  238. 11:39and turn in your answers and then we'll
  239. 11:41review them together as a group to make
  240. 11:44sure that you understand the flow of
  241. 11:46these
  242. 11:47calculations as always I'm available in
  243. 11:50office hours on Thursdays at noon and
  244. 11:53I'm happy to work through
  245. 11:55additional um examples at that time

About this transcript

This page contains the full transcript of Summary Week 11 by Tessa Crume, generated from the public captions YouTube serves with the video. The transcript has 1,409 words across 245 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.