Deterministic sensitivity analysis

CJ
by Clarieke Jonker4 views·Updated Oct 9, 2020
Create your own

Audio TranscriptAuto-generated

  1. hi, everyone.
  2. In a previous knowledge clip, you learned about what the
  3. deterministic sensitivity analysis face today.
  4. I will tell you more about how to present ourselves
  5. off the DNA on behalf of crude to, as you
  6. already know, that these days away to address uncertainty.
  7. And it's about analyzing the sensitivity off one or more
  8. para meters on the outcome.
  9. Carrie Wells.
  10. There are different types off ESAs in this knowledge ship.
  11. We will give you an example off one way, two
  12. way and three way the ESA and how to present
  13. them. We will also provide examples to make it easier
  14. to understand in a one radius A.
  15. The one refers to the affected only one para meter
  16. exchanged at the time.
  17. It is important, however, to remember that one way that
  18. he is a is often done various times.
  19. This is because researchers want to know which para meter.
  20. That's the biggest influence on the out home.
  21. The para meters are not changed at the same time,
  22. but separately.
  23. Then they are put together in a tornado diagram or
  24. a table.
  25. Through this tornado diagram or table, it becomes clear toe
  26. what extent a surgeon change off the perimeter changes the
  27. outcome variable often the Isar.
  28. So let's say we are doing an economic evaluation on
  29. the vaccination for her peace among the elderly.
  30. Been looking at cost and effects.
  31. We obtained a Nizer off 2000 364 correct $23,646 for
  32. quality gains.
  33. Now we want to conduct a one way DSA.
  34. Some of the parameters we take into account our vaccination
  35. rates in discounting rates, the vaccine price and the duration
  36. off the vaccine.
  37. We start with changing the vaccination rates.
  38. At the initial analysis, our Isar waas 23,646 here we
  39. assumed the vaccination rate 20% off the entire population.
  40. In our one way DSA we're going to see what
  41. the Isar is if we assume a higher and lower
  42. vaccination rates often when doing a DSA a fix change
  43. is used 10% higher or lower for all para meters.
  44. After this, we repeat this for the other para meters.
  45. It is important to choose a fella's amount of change
  46. in the perimeter based on literature and not on arbitrary
  47. one. When displaying our tornado diagram, we often put the
  48. para meter with the highest effects change on the ice
  49. air at the top, the second we put seconds.
  50. And Joel this Verceles any form off a tornado and
  51. is why the tornado diagram is called that way.
  52. In this case, the tornado diagram could be the following
  53. the middle line.
  54. They were percent SharePoint estimate the ice are 24,646 on
  55. the why exists, the bearer meters are presented and they
  56. change off.
  57. Your outcome is displayed on the X axis as we
  58. can see changing the vaccination rates.
  59. Brussels in the highest change in nicer discounting rates as
  60. well, but affects in price and duration off the vaccine
  61. Not so much.
  62. What does this mean for your cost effectiveness analysis?
  63. It means if, if the vaccination rate is higher than
  64. you initially assumed, your eyes, er could change substantially.
  65. Your estimate off course and effects are less progressed.
  66. This could affect your decision on whether or not to
  67. implement on intervention.
  68. Based on your initial Isar off 24,646.
  69. You might decide to implement the herpes vaccination.
  70. However, are you certain that you assumed the vaccination rate
  71. off 10% is realistic if the vaccination rate turns out
  72. to be way different, it affects your eyes, er, in
  73. this case, and therefore your decision to implement the intervention
  74. or not.
  75. If your model is very susceptible to change yourself para
  76. meters, it is more insurgent.
  77. If you're if your cost and effectiveness estimates are correct
  78. and you will need to decide if you still want
  79. to implement the intervention, is it susceptible to a lot
  80. of para meters, or to only some?
  81. What are these bearer meters?
  82. The more para meter shell model is susceptible toe, the
  83. less regressed it ISS and therefore it is more uncertain.
  84. In this case, we might say that the model is
  85. susceptible to some pair meters but can still be considered
  86. to be quite rivest.
  87. The result off several one way visas also can be
  88. presented in a table or can easily be displayed.
  89. Food text.
  90. The outcome ranged from X to Y when para Meter
  91. said waas for right from A to B.
  92. Next up is the multiple a sensitivity analysis.
  93. As you can probably imagine, it is possible that multiple
  94. parameters can have an impact on a model at the
  95. same time.
  96. In this case, a two way sensitivity analysis or even
  97. a three way sensitivity analysis is used there are used
  98. decent are used in to our mirror.
  99. More para meters changed simultaneously and therefore have an impact
  100. on the model.
  101. So let's say we want to know how the price
  102. often intervention together with the effectiveness off the same intervention
  103. influence says the costs.
  104. To do so, we could make a table with the
  105. effectiveness on the Y axis and the price of the
  106. intervention on the X excess, like in the following picture.
  107. In this figure, the price of the intervention uses steps
  108. off $100 and the effectiveness steps off 5% in the
  109. stable, it is easy to read what the total cost
  110. will be.
  111. If the intervention cost $100 and as an effectiveness off
  112. 30% for example, in this case, you can see that
  113. the total cost will end up at $2048.
  114. Suppose the intervention is $200 more expensive and therefore cost
  115. a total off $300.
  116. In this case, you are still looking at the 30%
  117. effectiveness. Only now in the $300 column, you can see
  118. that the total cost skyrockets to $6349.
  119. Of course, it is also possible to set other parameters
  120. against each other.
  121. So a two sided sensitivity analysis is an easy way
  122. to quickly get an overview of what happens to your
  123. desired outcome.
  124. When two para meters change at the same time, then
  125. a three way sensitivity analysis often leads to official overload
  126. and confusion and will therefore only work if the first
  127. para meter can be represented as a discrete value when
  128. reporting a disease a enforcing three or more comparator xcx
  129. This kind of analysis is very usual.
  130. The relevant question you want toe answer in a three
  131. way sensitivity analysis is which alternative is cost effective at
  132. the threshold of X to portray the answer for two
  133. para meters partition that quadrant to show which alternative is
  134. cost effective at various isil thresholds and for different combination
  135. off para meters results for different decision criteria are best
  136. presented in separate panels off a graphical dismayed we've come
  137. to the end of the knowledge ship.
  138. Hopefully, you now know everything about how to present themselves
  139. over one way analysis and multi way analysis.
  140. I want to thank you for listening