SlideShare a Scribd company logo
1 of 49
Download to read offline
Statistical Analysis
for Recurrent Events Data
with Multiple Competing Risks
研 究 生 : 倪佳蓉
指導教授 : 王維菁 教授
Motivating Example (1)
• The dialysis shunt thrombosis (透析瘻管栓塞) data
• Provided by 臺大新竹分院, 腎臟科, 楊忠煒主任
• Patients with shunt thromboses (瘻管栓塞)
• Features:
• Recurrence
• Competing Risks
• Acute: 需動手術處理
• Non-acute : 可簡單排除
• Single-type Endpoints
• Independent censoring
2
Dialysis Shunt Thrombosis Data
• Sample size = 2886
• Figure 1. Selected sample path for shunt thrombosis data
3
Motivating Example (2)
• The peritoneal dialysis (腹膜透析) data
• Provided by 臺中榮總
• Patients with peritonitis infection (腹膜炎感染)
• Features:
• Recurrence
• Competing Risks
• Gram-positive
(革蘭氏陽性菌)
• Non-Gram-positive
(革蘭氏陰性菌)
• Two-type Endpoints
• Independent censoring
• Dependent censoring
4
Peritoneal Dialysis Data
• Sample size = 575
• Figure 2. Selected sample path for peritonitis infection data
5
Gram-positive
Non-gram-positive
Our Interests
• Literature developments
• How to deal with competing risks?
• How to handle association?
• How to generate such data?
6
Literature Developments
7
Examinations of Data Structures
Recurrence Competing risks
Semi-
competing risks
Typical univariate survival
data × × ×
Competing risks data
without recurrences × ∨ ×
Semi-competing risks
survival data × × ∨
Recurrent events data with
multiple competing risks ∨ ∨ ×
8
Data structure form of our two motivating examples
Possible Sample Paths for Survival Data
9
Typical Univariate Survival Data
• Observed variables:
𝑋 = 𝐷 ∧ 𝐶
𝛿 = 𝐼 𝐷 ≤ 𝐶
Initialstate
death
censored
10
D
C
Competing Risks Data without Recurrences
• Observed variables:
෨𝑌 = Y ∧ 𝐶 ∧ 𝐷
෨∆= 𝐼 𝑌 ≤ 𝐶 ∆
Initialstate
death
censored
Type-1 event
Type-2 event
11
C
Y
D
Semi-Competing Risks Survival Data
• Observed variables:
𝑌 ∧ 𝐷 ∧ 𝐶
𝐼 𝑌 ≤ 𝐷 ∧ 𝐶
𝐷 ∧ 𝐶
𝐼 𝐷 ≤ 𝐶
• Semi-competing risks: 𝐷 is a competing risks for 𝑌 but not vice versa
Initialstate
death
rejection
Y
D
12
• Observed variables:
෩𝑌𝑗 = 𝑌𝑗 ∧ 𝐶 ∧ 𝐷
෩𝑇𝑗 = ෩Y𝑗 − ෩Y𝑗−1
෩∆𝑗= ∆𝑗 × 𝐼(𝑌𝑗 < 𝐶 ∧ 𝐷)
෩𝑇𝑗 = 𝑇𝑗 , ∀ 𝑗 = 1, ⋯ , 𝑀 − 1 ; but ෨𝑇 𝑀 < 𝑇 𝑀
Initialstate
Stage 0
death
Type-2 event
censored
Type-1 event
death
Type-2 event
censored
Type-1 event
Stage 1
Stage 2
death
censored
Stage M
13
Recurrent Events Data
with Multiple Competing Risks
Deal with Competing Risks
14
Competing Risks
- Classical Definition
• Every human is continuously exposed to many risks of death
• Cancer
• Heart diseases
• Pneumonia
• …
• Because death is not a repetitive event and is usually attributed to a
single cause, these risks compete with one another for the life of a
person.
• What if we are interested in a particular cause of death?
• Identifiability issue: Tsiatis (1975)
• Crude approach
• Net approach
15
Competing Risks
- Crude Approach
• Assumption:
• Recognize the presence of all competing risks
• 𝑌 = 𝑌1 ∧ ⋯ ∧ 𝑌𝐾
• ∆ = cause of the event
• Quantity of interest:
• Cumulative incidence function (abbr. CIF)
• CIF: 𝐹𝑘 𝑡 = 𝑃𝑟 𝑌 ≤ 𝑡, ∆= 𝑘
• Note that: lim
𝑡→∞
𝐹𝑘 𝑡 < 1
• 𝐹𝑘 𝑡 is an improper function
• Remark: This approach is constructed on a real world
16
Competing Risks
- Net Approach
• Assumptions:
• Other competing risks can be removed
• 𝑌 = 𝑌1 ∧ ⋯ ∧ 𝑌𝐾
• ∆ = cause of the event
• Quantity of interest
• Survival function: 𝑆 𝑘 𝑡 = 𝑃𝑟 𝑌𝑘 > 𝑡
• Note that: lim
𝑡→∞
𝑆 𝑘 𝑡 = 0
• 𝑆 𝑘 𝑡 is a proper survival function
• Remark: This approach is constructed on a hypothetical world
17
Semi-Competing Risks Data
- Revisited
What we observed Exist but censored
18
• The implicit assumption is that the censoring events & the terminal
event can both be removed.
• Although we cannot observe 𝑌 for 𝑌 > 𝐷, we still assume that there is
a hypothetical world that such Y exists.
?
?
?
Handel Associations
19
Modeling Associations
• There are three possible types of association
• Between gap-times
• Between competing risks
• Between recurrences and death
• Association models
• Frailty (random effect) modeling
• Copula modeling
• Frailty model and Archimedean Copula model
• Equivalence under some conditions in theory
• Data generation algorithms are different
20
Modeling Associations
- Frailty and Copula Approaches
• There are three possible types of association
• Between gap-times
• Between competing risks
• Between recurrences and death
• 𝑊 : frailty (unobservable)
• How 𝑊 affects 𝑇𝑗  usually Cox PH model
• Conditional independence given W s.t. 𝑇𝑗 ⊥ 𝑇𝑗′ | 𝑊
• The frailty distribution is specified  usually Gamma dist.
• For the frailty control the association between competing risks,
• We use the copula package in R
• The built-in function rcopula()
21
Frailty Modeling
- before specifying frailty distribution
• Assume the frailty variable 𝑊 has a PH effect on the hazard of 𝑇𝑗
• 𝜆𝑗 𝑡 𝑤 = 𝑤𝜆0𝑗 𝑡 exp 𝑍 𝑇 𝛽
𝑠𝑖𝑚𝑝𝑙𝑖𝑓𝑦
𝜆𝑗 𝑡 𝑤 = 𝑤𝜆0𝑗 𝑡
• Λ𝑗 𝑡 𝑤 = ‫׬‬0
𝑡
𝜆𝑗 𝑢 𝑤 𝑑𝑢 = 𝑤Λ0𝑗 𝑡
• 𝑆𝑗 𝑡 𝑤 = exp −Λ𝑗 𝑡 𝑤 = exp −Λ0𝑗 𝑡
𝑤
= 𝐵𝑗(𝑡)
𝑤
where 𝐵𝑗(𝑡) is a continuous baseline survivor function of 𝑇𝑗
• Derive the unconditional marginal survivor function of 𝑇𝑗
• 𝑆𝑗 𝑡 = ‫׬‬𝑤
𝑆𝑗 𝑡 𝑤 𝑑𝐹 𝑊(𝑤) = ‫׬‬𝑤
𝐵𝑗(𝑡)
𝑤
𝑑𝐹 𝑊(𝑤)
= න
𝑤
𝑒 𝑤 log 𝐵 𝑗(𝑡)
𝑑𝐹 𝑊(𝑤) = 𝑝 − log 𝐵𝑗(𝑡)
, where 𝑝 𝑢 = 𝐸 𝑒−𝑢𝑊 is the Laplace transform of 𝑊
22
𝑆𝑗 𝑡 𝑤
𝑆𝑗(𝑡)
𝑝 𝑢 𝑝−1 𝑢
Frailty Modeling
- after specifying frailty distribution
• Specify the form of 𝐹 𝑊 𝑤 as W~𝑔𝑎𝑚𝑚𝑎 𝛼, 𝛽
• Since its Laplace transform has a closed form
• Derive the Laplace transform as 𝑝 𝑢 = 𝐸 𝑒−𝑢𝑊 = 1 + 𝛽𝑢 −𝛼
• 𝑊 : frailty (unobservable)
• To ensure the identifiability of model  𝐸 𝑊 = 1
• To allow within-cluster dependence  𝑉𝑎𝑟 𝑊 = 𝜃1 ≥ 0
• To satisfy the above constraints, assume
• W~𝑔𝑎𝑚𝑚𝑎 𝛼 =
1
𝜃1
, 𝛽 = 𝜃1
• so the Laplace transform is
• 𝑝 𝑢 = 𝐸 𝑒−𝑢𝑊 = 1 + 𝛽𝑢 −𝛼 = 1 + 𝜃1 𝑢
−
1
𝜃1
23
𝑆𝑗 𝑡 𝑤
𝑆𝑗(𝑡)
𝑝 𝑢 𝑝−1 𝑢
Proposal for
Data Generation Algorithms
24
Sketch of Simulation Study
• Hypothetical world – Net approach
• No terminal event or with terminal event
• Steps:
• Generate multivariate failure times
• Take minimum as observed variables
• Real world – Crude approach
• No terminal event or with terminal event
• Steps:
• Generate data from improper distribution
• Ref: Cheng & Fine (2012)
• We propose to use different frailty variables to account for
different associations.
• Use generated data to evaluate the performance of CIFs
Hypothetical
𝐶
𝐶, 𝐷
Real world
𝐶
𝐶, 𝐷
25
• Under the hypothetical world assumption
• Remove the influence of a terminal event
by taking 𝐷 = ∞
Cases 1.
Data Generation Algorithms
𝑇𝑗(1) 𝑇𝑗(2)
• 𝑇𝑗 = min{ 𝑇𝑗(1), 𝑇𝑗(2)}
• ∆𝑗 = ቊ
1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(1)
2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(2)
Compete
𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3
⋯
Frailty H
Frailty W
Hypothetical
𝐶
𝐶, 𝐷
Real world
𝐶
𝐶, 𝐷
26
Cases 1.
Data Generation Algorithms (cont’d)
• Step 1:
For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎
1
𝜃1
, 𝜃1 and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾
• Step 2:
For each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use 𝑈𝑗(𝑘) to
generate 𝑇𝑗(𝑘) as follows:
𝜆𝑗 𝑡|𝑤 = 𝑤𝜆0𝑗 𝑡 = 𝑤𝜉𝑗
 S(𝑗) 𝑡|𝑤 = 𝑒𝑥𝑝 −Λ𝑗 𝑡 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑤
 𝑈𝑗(𝑘) = 𝑒𝑥𝑝 − 𝜉𝑗 𝑇𝑗(𝑘)
𝑤
 𝑇𝑗(𝑘) = −
1
𝑤𝜉 𝑗
𝑙𝑜𝑔𝑈𝑗 𝑘
So, define 𝑇𝑗 = min{𝑇𝑗 1 , 𝑇𝑗(2)} and ൝
Δ𝑗 = 1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 1
Δ𝑗 = 2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 2
.
Then, set 𝑌𝑗 = σ 𝓂=1
𝑗
𝑇 𝑚.
27
Cases 1.
Data Generation Algorithms (cont’d)
• Step 3:
If 𝑌𝑗 ≤ 𝐶, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗.
Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤ 𝐶 < 𝑌 𝑀.
Finally, we set ෪𝑌 𝑀 = 𝐶, ෪𝑇 𝑀 = 𝐶 − 𝑌 𝑀−1, and ෪∆ 𝑀= 0.
28
• Under the hypothetical world assumption
• Consider the influence of a terminal event
by taking 𝐷 < ∞
Cases 2.
Data Generation Algorithms
𝑇𝑗(1) 𝑇𝑗(2)
• 𝑇𝑗 = min{ 𝑇𝑗(1), 𝑇𝑗(2)}
• ∆𝑗 = ቊ
1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(1)
2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(2)
Compete
𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3
⋯
𝐷
Frailty W
Frailty 𝐻
Frailty V
Hypothetical
𝐶
𝐶, 𝐷
Real world
𝐶
𝐶, 𝐷
29
Cases 2.
Data Generation Algorithms (cont’d)
• Step 1:
For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎
1
𝜃1
, 𝜃1 , 𝑉~𝑔𝑎𝑚𝑚𝑎
1
𝜃2
, 𝜃2
and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾
• Step 2:
For each stage, generate 𝑈D~unif 0,1 , then use 𝑈D to generate 𝐷 by
𝜆 𝐷 𝑡|𝑣 = 𝑣𝜆0𝐷 𝑡 = 𝑣𝜂
 𝑆 𝐷 𝑡|𝑣 = 𝑒𝑥𝑝 −Λ 𝐷 𝑡 = 𝑒𝑥𝑝 − 𝜂𝑡 𝑣
 𝑈 𝐷 = 𝑒𝑥𝑝 − 𝜂𝐷 𝑣
 𝐷 = −
1
𝑣𝜂
𝑙𝑜𝑔𝑈 𝐷
Moreover, for each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use
𝑈𝑗(𝑘) to generate 𝑇𝑗(𝑘) as follows:
𝜆𝑗 𝑡|𝑣, 𝑤 = 𝑣𝑤𝜆0𝑗 𝑡 = 𝑣𝑤𝜉𝑗
 S(𝑗) 𝑡|𝑣, 𝑤 = 𝑒𝑥𝑝 −Λ𝑗 𝑡 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑣𝑤
30
Cases 2.
Data Generation Algorithms (cont’d)
 𝑈𝑗(𝑘) = 𝑒𝑥𝑝 − 𝜉𝑗 𝑇𝑗(𝑘)
𝑣𝑤
 𝑇𝑗(𝑘) = −
1
𝑣𝑤𝜉 𝑗
𝑙𝑜𝑔𝑈𝑗 𝑘
So, define 𝑇𝑗 = min{𝑇𝑗 1 , 𝑇𝑗(2)} and ൝
Δ𝑗 = 1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 1
Δ𝑗 = 2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 2
.
Then, set 𝑌𝑗 = σ 𝓂=1
𝑗
𝑇 𝑚.
• Step 3:
If 𝑌𝑗 ≤ min{𝐶, 𝐷}, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗.
Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤
min{C, D} < 𝑌 𝑀.
Finally, we set ෪𝑌 𝑀 = min{𝐶, 𝐷}, ෪𝑇 𝑀 = min{𝐶, 𝐷} − 𝑌 𝑀−1, and
෪∆ 𝑀= ቊ
0, 𝑖𝑓 𝐶 ≤ 𝐷
3, 𝑖𝑓 𝐶 > 𝐷
.
31
• Under the real world assumption
• Remove the influence of a terminal
event by taking 𝐷 = ∞
Cases 3.
Data Generation Algorithms
𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3
⋯
Frailty W
𝜂1 ≜ 𝐹 𝑗 ∞ 𝑤
𝑡
1
𝜂1
𝐹 𝑗 (𝑡|𝑤)
Frailty 𝐻
• 𝜂1 ≜ 𝐹1
(𝑗)
∞ 𝑤
• 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0
• 𝐼𝑓 𝑈𝑗(1) ≤ 𝜂1, 𝑡ℎ𝑒𝑛 𝑢𝑠𝑒 𝑈𝑗(1) 𝑡𝑜 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒 𝑇𝑗 = 𝑇𝑗(1) 𝑎𝑛𝑑 ∆𝑗= 1; 𝑜𝑟 𝑣𝑖𝑐𝑒 𝑣𝑒𝑟𝑠𝑎.
Hypothetical
𝐶
𝐶, 𝐷
Real world
𝐶
𝐶, 𝐷
32
Cases 3.
Data Generation Algorithms (cont’d)
• Step 1:
For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎
1
𝜃1
, 𝜃1 and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾
• Step 2:
For each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use 𝑈𝑗(𝑘) to
generate 𝑇𝑗 as follows:
S(𝑗)
𝑡|𝑤 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑤
 S(𝑗)
𝑡 = ‫׬‬𝑤
S(𝑗)
𝑡|𝑤 𝑑𝐹 𝑊(𝑤) = 1 + 𝜃1 𝜉𝑗 𝑡
−
1
𝜃1
 F1
𝑗
𝑡 + F2
𝑗
𝑡 = 1 − S 𝑗
𝑡 = 1 − 1 + 𝜃1 𝜉𝑗 𝑡
−
1
𝜃1
 ൞
F1
𝑗
𝑡 = 𝜓𝑗 1 − 1 + 𝜃1 𝜉𝑗 𝑡
−
1
𝜃1
F2
𝑗
𝑡 = 1 − 𝜓𝑗 1 − 1 + 𝜃1 𝜉𝑗 𝑡
−
1
𝜃1
 F 𝑘
𝑗
𝑡|𝑤 = 𝐵𝑘
𝑗
(𝑡)
𝑤
= 𝑒𝑥𝑝 −𝑞 F 𝑘
𝑗
𝑡
𝑤
33
Cases 3.
Data Generation Algorithms (cont’d)
 Have explicit forms for F1
𝑗
𝑡|𝑤 & F2
𝑗
𝑡|𝑤 .
 Set lim
𝑡→∞
F 𝑘
𝑗
𝑡|𝑤 = 𝜂1
- If 𝑈𝑗(1) ≤ 𝜂1, generate Δ𝑗 = 1 and 𝑇𝑗 = 𝑇𝑗 1 = F1
𝑗 −1
𝑈𝑗(1)|𝑊 = 𝑤
- If 𝑈𝑗(1) > 𝜂1, generate Δ𝑗 = 2 and 𝑇𝑗 = 𝑇𝑗 2 , where 𝑇𝑗 2 is generated
from 𝑃𝑟 𝑇𝑗 ≤ 𝑡|Δ𝑗 = 2 =
F2
𝑗
𝑡
F2
𝑗
∞
Then, set 𝑌𝑗 = σ 𝓂=1
𝑗
𝑇 𝑚.
• Step 3:
If 𝑌𝑗 ≤ 𝐶, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗.
Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤ 𝐶 < 𝑌 𝑀.
Finally, we set ෪𝑌 𝑀 = 𝐶, ෪𝑇 𝑀 = 𝐶 − 𝑌 𝑀−1, and ෪∆ 𝑀= 0.
34
• Under the real world assumption
• Consider the influence of a
terminal event by taking 𝐷 < ∞
Cases 4.
Data Generation Algorithms
𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3
⋯
𝐷
Frailty W
Frailty V
𝜂1 ≜ 𝐹 𝑗 ∞ 𝑤
𝑡
1
𝜂1
𝐹 𝑗 (𝑡|𝑤)
Frailty 𝐻
• 𝜂1 ≜ 𝐹1
(𝑗)
∞ 𝑤
• 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0
• 𝐼𝑓 𝑈𝑗(1) ≤ 𝜂1, 𝑡ℎ𝑒𝑛 𝑢𝑠𝑒 𝑈𝑗(1) 𝑡𝑜 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒 𝑇𝑗 = 𝑇𝑗(1) 𝑎𝑛𝑑 ∆𝑗= 1; 𝑜𝑟 𝑣𝑖𝑐𝑒 𝑣𝑒𝑟𝑠𝑎.
Hypothetical
𝐶
𝐶, 𝐷
Real world
𝐶
𝐶, 𝐷
35
Cases 4.
Data Generation Algorithms (cont’d)
• Step 1:
For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎
1
𝜃1
, 𝜃1 , 𝑉~𝑔𝑎𝑚𝑚𝑎
1
𝜃2
, 𝜃2
and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾
• Step 2:
For each stage, generate 𝑈D~unif 0,1 , then use 𝑈D to generate 𝐷 by
𝜆 𝐷 𝑡|𝑣 = 𝑣𝜆0𝐷 𝑡 = 𝑣𝜂
 𝑆 𝐷 𝑡|𝑣 = 𝑒𝑥𝑝 −Λ 𝐷 𝑡 = 𝑒𝑥𝑝 − 𝜂𝑡 𝑣
 𝑈 𝐷 = 𝑒𝑥𝑝 − 𝜂𝐷 𝑣
 𝐷 = −
1
𝑣𝜂
𝑙𝑜𝑔𝑈 𝐷
36
Cases 4.
Data Generation Algorithms (cont’d)
Moreover, for each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 ,
then use 𝑈𝑗(𝑘) to generate 𝑇𝑗 as follows:
𝜆𝑗 𝑡|𝑣, 𝑤 = 𝑣𝑤𝜆0𝑗 𝑡 = 𝑣𝑤𝜉𝑗
 S(𝑗) 𝑡|𝑣, 𝑤 = 𝑒𝑥𝑝 −Λ𝑗 𝑡 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑣𝑤
 F1
𝑗
𝑡|𝑣, 𝑤 + F2
𝑗
𝑡|𝑣, 𝑤 = 1 − S(𝑗) 𝑡|𝑣, 𝑤
= 1 − 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑣𝑤
 ൞
F1
𝑗
𝑡|𝑤 = 𝜓𝑗 1 − 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑣𝑤
F2
𝑗
𝑡|𝑤 = 1 − 𝜓𝑗 1 − 𝑒𝑥𝑝 − 𝜉𝑗 𝑡
𝑣𝑤
 Have explicit forms for F1
𝑗
𝑡|𝑣, 𝑤 & F2
𝑗
𝑡|𝑣, 𝑤 .
 Set lim
𝑡→∞
F 𝑘
𝑗
𝑡|𝑤 = 𝜂1
37
Cases 4.
Data Generation Algorithms (cont’d)
- If 𝑈𝑗(1) ≤ 𝜂1, set Δ𝑗 = 1 and 𝑇𝑗 = 𝑇𝑗 1 = F1
𝑗
−1
𝑈𝑗(1)|𝑉 = 𝑣, 𝑊 = 𝑤
- If 𝑈𝑗(1) > 𝜂1, set Δ𝑗 = 2 and 𝑇𝑗 = 𝑇𝑗 2 , where 𝑇𝑗 2 is generated from
𝑃𝑟 𝑇𝑗 ≤ 𝑡|Δ𝑗 = 2 =
F2
𝑗
𝑡
F2
𝑗
∞
Then, set 𝑌𝑗 = σ 𝓂=1
𝑗
𝑇 𝑚.
• Step 3:
If 𝑌𝑗 ≤ min{𝐶, 𝐷}, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗.
Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤
min{C, D} < 𝑌 𝑀.
Finally, we set ෪𝑌 𝑀 = min{𝐶, 𝐷}, ෪𝑇 𝑀 = min{𝐶, 𝐷} − 𝑌 𝑀−1, and
෪∆ 𝑀= ቊ
0, 𝑖𝑓 𝐶 ≤ 𝐷
3, 𝑖𝑓 𝐶 > 𝐷
.
38
Simulations
39
Simulation Study
Hypothetical
𝐶
𝐶, 𝐷
Real world
𝐶
𝐶, 𝐷
40
Case (i) Case (ii) Case (iii) Case (iv)
Between competing risks
𝜃0
indep.
(1)
dep.
(1.25)
indep.
(1)
dep.
(1.25)
Between gap-times
𝜃1
indep.
(1)
indep.
(1)
dep.
(1.25)
dep.
(1.25)
• Ref: Ph.D. thesis of Bowen Li (2016)
• IPCW (inverse probability
censoring weighting) to adjust bias
• Parameter settings:
• 𝜓𝑗 = 0.25
• 𝜉𝑗 = 0.25
• 𝐶~𝑢𝑛𝑖𝑓 0, 500
• 𝑁 = 1000
• 𝐵 = 500
• Assume no correlation,
so frailties parameters:
• 𝜃0 = 1
• 𝜃1 = 1
Simulation Study – Case (i)
41
• Parameter settings:
• 𝜓𝑗 = 0.25
• 𝜉𝑗 = 0.25
• 𝐶~𝑢𝑛𝑖𝑓 0, 500
• 𝑁 = 1000
• 𝐵 = 500
• Assume positive correlation,
so frailties parameters:
• 𝜃0 = 1.25
• 𝜃1 = 1
Simulation Study – Case (ii)
42
• Parameter settings:
• 𝜓𝑗 = 0.25
• 𝜉𝑗 = 0.25
• 𝐶~𝑢𝑛𝑖𝑓 0, 500
• 𝑁 = 1000
• 𝐵 = 500
• Assume positive correlation,
so frailties parameters:
• 𝜃0 = 1
• 𝜃1 = 1.25
Simulation Study – Case (iii)
43
• Parameter settings:
• 𝜓𝑗 = 0.25
• 𝜉𝑗 = 0.25
• 𝐶~𝑢𝑛𝑖𝑓 0, 500
• 𝑁 = 1000
• 𝐵 = 500
• Assume positive correlation,
so frailties parameters:
• 𝜃0 = 1.25
• 𝜃1 = 1.25
Simulation Study – Case (iv)
44
Conclusions
• Two real-world examples
• Analyze recurrent events data with competing risks
• Literature development
• Review survival data structure
• Deal with competing risks
• Account for Identifiability issue
• “Net” vs. “Crude” approach
• Handle association by frailty approach
• Between gap-times
• Between competing risks
• Between recurrence processes and death
• Simulation
• Propose data generation algorithms
• Conduct simulation analysis to examine the validity
45
References (1)
• Chen, C. M., Chuang, Y. W., & Shen, P. S. (2015). Two‐stage
estimation for multivariate recurrent event data with a dependent
terminal event. Biometrical Journal, 57(2), 215-233.
• Chang, W. H., & Wang, W. (2009). Regression analysis for
cumulative incidence probability under competing risks. Statistica
Sinica, 391-408
• Cheng, Y., & Fine, J. P. (2012). Cumulative incidence association
models for bivariate competing risks data. Journal of the Royal
Statistical Society: Series B (Statistical Methodology), 74(2), 183-
202.
• Goethals, K., Janssen, P., & Duchateau, L. (2008). Frailty models
and copulas: similarities and differences. Journal of Applied
Statistics, 35(9), 1071-1079.
46
References (2)
• Oakes, D. (1982). A model for association in bivariate survival
data. Journal of the Royal Statistical Society. Series B
(Methodological), 414-422.
• Oakes, D. (1989). Bivariate survival models induced by
frailties. Journal of the American Statistical Association, 84(406),
487-493.
• Tsiatis, A. (1975). A nonidentifiability aspect of the problem of
competing risks. Proceedings of the National Academy of
Sciences, 72(1), 20-22.
• Wang, W., & Wells, M. T. (2000). Estimation of Kendall's tau under
censoring. Statistica Sinica, 1199-1215.
47
Thank you for your attention!
48
Appendix
- CIF Estimation
• Ref: Ph.D. thesis of Bowen Li (2016)
• IPCW (inverse probability censoring weighting) to adjust bias
 is due to induced dependent censoring
• Goal: Estimate 𝐹𝑘
𝑗
𝑡 = 𝐸 𝐼 𝑇𝑗 ≤ 𝑡, ∆𝑗= 𝑘 for 𝑗 ≥ 2
• Empirical proxy:
𝐼 ෩𝑇𝑗 ≤ 𝑡, ෩∆𝑗= 𝑘 = 𝐼 𝑇𝑗 ≤ 𝑡, ∆𝑗= 𝑘, 𝐶 > 𝑌𝑗
• By conditional expectation,
𝐹𝑘
𝑗
𝑡 = 𝐸
𝐸 𝐼 ෩𝑇𝑗 ≤ 𝑡, ෩∆𝑗= 𝑘 |𝑌𝑗, 𝑌𝑗−1, ∆𝑗
𝐺(෩𝑌𝑗)
, where 𝐺 𝑡 = 𝑃𝑟 𝐶 > 𝑡 is the survival function for 𝐶
• CIFs estimators based on observed data can be expressed as
෠𝐹𝑘
(𝑗)
= ෍
𝑖=1
𝑛 𝐼 ෪𝑇𝒊𝑗 ≤ 𝑡, ෪∆𝒊𝑗= 𝑘
𝑛 ෠𝐺(෪𝑌𝒊𝑗)
49

More Related Content

What's hot

SURVIVAL ANALYSIS.ppt
SURVIVAL ANALYSIS.pptSURVIVAL ANALYSIS.ppt
SURVIVAL ANALYSIS.pptmbang ernest
 
Mantel Haenszel methods in epidemiology (Stratification)
Mantel Haenszel methods in epidemiology (Stratification) Mantel Haenszel methods in epidemiology (Stratification)
Mantel Haenszel methods in epidemiology (Stratification) Rizwan S A
 
Epidemiolgy and biostatistics notes
Epidemiolgy and biostatistics notesEpidemiolgy and biostatistics notes
Epidemiolgy and biostatistics notesCharles Ntwale
 
Inferential statistics (2)
Inferential statistics (2)Inferential statistics (2)
Inferential statistics (2)rajnulada
 
Data analysis using spss
Data analysis using spssData analysis using spss
Data analysis using spssSyed Faisal
 
4.4. effect modification
4.4. effect modification4.4. effect modification
4.4. effect modificationA M
 
Epidemiological Studies
Epidemiological StudiesEpidemiological Studies
Epidemiological StudiesINAAMUL HAQ
 
Relative and Attributable Risk For Graduate and Postgraduate Students
Relative and Attributable Risk For Graduate and Postgraduate StudentsRelative and Attributable Risk For Graduate and Postgraduate Students
Relative and Attributable Risk For Graduate and Postgraduate StudentsTauseef Jawaid
 
Survival Analysis Using SPSS
Survival Analysis Using SPSSSurvival Analysis Using SPSS
Survival Analysis Using SPSSNermin Osman
 
unmatched case control studies
unmatched case control studiesunmatched case control studies
unmatched case control studiesMrinmoy Bharadwaz
 
Probability sampling
Probability samplingProbability sampling
Probability samplingBhanu Teja
 
Hypothesis testing
Hypothesis testingHypothesis testing
Hypothesis testingArnab Sadhu
 
Workshop on SPSS: Basic to Intermediate Level
Workshop on SPSS: Basic to Intermediate LevelWorkshop on SPSS: Basic to Intermediate Level
Workshop on SPSS: Basic to Intermediate LevelHiram Ting
 
Intro biostat1&2
Intro biostat1&2Intro biostat1&2
Intro biostat1&2Lucidante1
 
Sample size determination
Sample size determinationSample size determination
Sample size determinationGopal Kumar
 
Univariate Analysis
 Univariate Analysis Univariate Analysis
Univariate AnalysisSoumya Sahoo
 

What's hot (20)

SURVIVAL ANALYSIS.ppt
SURVIVAL ANALYSIS.pptSURVIVAL ANALYSIS.ppt
SURVIVAL ANALYSIS.ppt
 
Mantel Haenszel methods in epidemiology (Stratification)
Mantel Haenszel methods in epidemiology (Stratification) Mantel Haenszel methods in epidemiology (Stratification)
Mantel Haenszel methods in epidemiology (Stratification)
 
Epidemiolgy and biostatistics notes
Epidemiolgy and biostatistics notesEpidemiolgy and biostatistics notes
Epidemiolgy and biostatistics notes
 
Inferential statistics (2)
Inferential statistics (2)Inferential statistics (2)
Inferential statistics (2)
 
Data analysis using spss
Data analysis using spssData analysis using spss
Data analysis using spss
 
4.4. effect modification
4.4. effect modification4.4. effect modification
4.4. effect modification
 
Epidemiological Studies
Epidemiological StudiesEpidemiological Studies
Epidemiological Studies
 
Relative and Attributable Risk For Graduate and Postgraduate Students
Relative and Attributable Risk For Graduate and Postgraduate StudentsRelative and Attributable Risk For Graduate and Postgraduate Students
Relative and Attributable Risk For Graduate and Postgraduate Students
 
Survival Analysis Using SPSS
Survival Analysis Using SPSSSurvival Analysis Using SPSS
Survival Analysis Using SPSS
 
Survival analysis
Survival  analysisSurvival  analysis
Survival analysis
 
unmatched case control studies
unmatched case control studiesunmatched case control studies
unmatched case control studies
 
Parametric test
Parametric testParametric test
Parametric test
 
Probability sampling
Probability samplingProbability sampling
Probability sampling
 
Hypothesis testing
Hypothesis testingHypothesis testing
Hypothesis testing
 
Sample size calculation
Sample size calculationSample size calculation
Sample size calculation
 
Workshop on SPSS: Basic to Intermediate Level
Workshop on SPSS: Basic to Intermediate LevelWorkshop on SPSS: Basic to Intermediate Level
Workshop on SPSS: Basic to Intermediate Level
 
Intro biostat1&2
Intro biostat1&2Intro biostat1&2
Intro biostat1&2
 
Sample size determination
Sample size determinationSample size determination
Sample size determination
 
Univariate Analysis
 Univariate Analysis Univariate Analysis
Univariate Analysis
 
Estimating a Population Proportion
Estimating a Population Proportion  Estimating a Population Proportion
Estimating a Population Proportion
 

Similar to Statistical analysis for recurrent events data with multiple competing risks

Intro to Quant Trading Strategies (Lecture 2 of 10)
Intro to Quant Trading Strategies (Lecture 2 of 10)Intro to Quant Trading Strategies (Lecture 2 of 10)
Intro to Quant Trading Strategies (Lecture 2 of 10)Adrian Aley
 
Deep learning study 2
Deep learning study 2Deep learning study 2
Deep learning study 2San Kim
 
Efficient anomaly detection via matrix sketching
Efficient anomaly detection via matrix sketchingEfficient anomaly detection via matrix sketching
Efficient anomaly detection via matrix sketchingHsing-chuan Hsieh
 
Randomized Algorithm- Advanced Algorithm
Randomized Algorithm- Advanced AlgorithmRandomized Algorithm- Advanced Algorithm
Randomized Algorithm- Advanced AlgorithmMahbubur Rahman
 
Linear regression, costs & gradient descent
Linear regression, costs & gradient descentLinear regression, costs & gradient descent
Linear regression, costs & gradient descentRevanth Kumar
 
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...irjes
 
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...IJRES Journal
 
Direct solution of sparse network equations by optimally ordered triangular f...
Direct solution of sparse network equations by optimally ordered triangular f...Direct solution of sparse network equations by optimally ordered triangular f...
Direct solution of sparse network equations by optimally ordered triangular f...Dimas Ruliandi
 
Basic Concepts of Standard Experimental Designs ( Statistics )
Basic Concepts of Standard Experimental Designs ( Statistics )Basic Concepts of Standard Experimental Designs ( Statistics )
Basic Concepts of Standard Experimental Designs ( Statistics )Hasnat Israq
 
Lecture 3: Stochastic Hydrology
Lecture 3: Stochastic HydrologyLecture 3: Stochastic Hydrology
Lecture 3: Stochastic HydrologyAmro Elfeki
 

Similar to Statistical analysis for recurrent events data with multiple competing risks (20)

Intro to Quant Trading Strategies (Lecture 2 of 10)
Intro to Quant Trading Strategies (Lecture 2 of 10)Intro to Quant Trading Strategies (Lecture 2 of 10)
Intro to Quant Trading Strategies (Lecture 2 of 10)
 
ERF Training Workshop Panel Data 5
ERF Training WorkshopPanel Data 5ERF Training WorkshopPanel Data 5
ERF Training Workshop Panel Data 5
 
Deep learning study 2
Deep learning study 2Deep learning study 2
Deep learning study 2
 
Stochastic Optimization
Stochastic OptimizationStochastic Optimization
Stochastic Optimization
 
Efficient anomaly detection via matrix sketching
Efficient anomaly detection via matrix sketchingEfficient anomaly detection via matrix sketching
Efficient anomaly detection via matrix sketching
 
Randomized Algorithm- Advanced Algorithm
Randomized Algorithm- Advanced AlgorithmRandomized Algorithm- Advanced Algorithm
Randomized Algorithm- Advanced Algorithm
 
Seminar9
Seminar9Seminar9
Seminar9
 
Av 738- Adaptive Filtering - Wiener Filters[wk 3]
Av 738- Adaptive Filtering - Wiener Filters[wk 3]Av 738- Adaptive Filtering - Wiener Filters[wk 3]
Av 738- Adaptive Filtering - Wiener Filters[wk 3]
 
Formulario estadistica
Formulario   estadisticaFormulario   estadistica
Formulario estadistica
 
Lec05.pptx
Lec05.pptxLec05.pptx
Lec05.pptx
 
Linear regression, costs & gradient descent
Linear regression, costs & gradient descentLinear regression, costs & gradient descent
Linear regression, costs & gradient descent
 
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
 
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
A Mathematical Model for the Hormonal Responses During Neurally Mediated Sync...
 
Statistics (recap)
Statistics (recap)Statistics (recap)
Statistics (recap)
 
Direct solution of sparse network equations by optimally ordered triangular f...
Direct solution of sparse network equations by optimally ordered triangular f...Direct solution of sparse network equations by optimally ordered triangular f...
Direct solution of sparse network equations by optimally ordered triangular f...
 
Basic Concepts of Standard Experimental Designs ( Statistics )
Basic Concepts of Standard Experimental Designs ( Statistics )Basic Concepts of Standard Experimental Designs ( Statistics )
Basic Concepts of Standard Experimental Designs ( Statistics )
 
Stat 3203 -pps sampling
Stat 3203 -pps samplingStat 3203 -pps sampling
Stat 3203 -pps sampling
 
MLU_DTE_Lecture_2.pptx
MLU_DTE_Lecture_2.pptxMLU_DTE_Lecture_2.pptx
MLU_DTE_Lecture_2.pptx
 
Lecture 3: Stochastic Hydrology
Lecture 3: Stochastic HydrologyLecture 3: Stochastic Hydrology
Lecture 3: Stochastic Hydrology
 
GDRR Opening Workshop - Variance Reduction for Reliability Assessment with St...
GDRR Opening Workshop - Variance Reduction for Reliability Assessment with St...GDRR Opening Workshop - Variance Reduction for Reliability Assessment with St...
GDRR Opening Workshop - Variance Reduction for Reliability Assessment with St...
 

Recently uploaded

04242024_CCC TUG_Joins and Relationships
04242024_CCC TUG_Joins and Relationships04242024_CCC TUG_Joins and Relationships
04242024_CCC TUG_Joins and Relationshipsccctableauusergroup
 
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一fhwihughh
 
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...dajasot375
 
Kantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdf
Kantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdfKantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdf
Kantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdfSocial Samosa
 
INTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTDINTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTDRafezzaman
 
VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...
VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...
VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...Suhani Kapoor
 
Industrialised data - the key to AI success.pdf
Industrialised data - the key to AI success.pdfIndustrialised data - the key to AI success.pdf
Industrialised data - the key to AI success.pdfLars Albertsson
 
Amazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptx
Amazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptxAmazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptx
Amazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptxAbdelrhman abooda
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]📊 Markus Baersch
 
Data Science Jobs and Salaries Analysis.pptx
Data Science Jobs and Salaries Analysis.pptxData Science Jobs and Salaries Analysis.pptx
Data Science Jobs and Salaries Analysis.pptxFurkanTasci3
 
Call Girls In Mahipalpur O9654467111 Escorts Service
Call Girls In Mahipalpur O9654467111  Escorts ServiceCall Girls In Mahipalpur O9654467111  Escorts Service
Call Girls In Mahipalpur O9654467111 Escorts ServiceSapana Sha
 
How we prevented account sharing with MFA
How we prevented account sharing with MFAHow we prevented account sharing with MFA
How we prevented account sharing with MFAAndrei Kaleshka
 
Customer Service Analytics - Make Sense of All Your Data.pptx
Customer Service Analytics - Make Sense of All Your Data.pptxCustomer Service Analytics - Make Sense of All Your Data.pptx
Customer Service Analytics - Make Sense of All Your Data.pptxEmmanuel Dauda
 
꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call
꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call
꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Callshivangimorya083
 
VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130
VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130
VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130Suhani Kapoor
 
Schema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdfSchema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdfLars Albertsson
 
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptxEMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptxthyngster
 
Brighton SEO | April 2024 | Data Storytelling
Brighton SEO | April 2024 | Data StorytellingBrighton SEO | April 2024 | Data Storytelling
Brighton SEO | April 2024 | Data StorytellingNeil Barnes
 
High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...
High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...
High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...soniya singh
 
(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service
(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service
(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Serviceranjana rawat
 

Recently uploaded (20)

04242024_CCC TUG_Joins and Relationships
04242024_CCC TUG_Joins and Relationships04242024_CCC TUG_Joins and Relationships
04242024_CCC TUG_Joins and Relationships
 
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
办理学位证纽约大学毕业证(NYU毕业证书)原版一比一
 
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
Indian Call Girls in Abu Dhabi O5286O24O8 Call Girls in Abu Dhabi By Independ...
 
Kantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdf
Kantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdfKantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdf
Kantar AI Summit- Under Embargo till Wednesday, 24th April 2024, 4 PM, IST.pdf
 
INTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTDINTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTD
 
VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...
VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...
VIP High Class Call Girls Jamshedpur Anushka 8250192130 Independent Escort Se...
 
Industrialised data - the key to AI success.pdf
Industrialised data - the key to AI success.pdfIndustrialised data - the key to AI success.pdf
Industrialised data - the key to AI success.pdf
 
Amazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptx
Amazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptxAmazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptx
Amazon TQM (2) Amazon TQM (2)Amazon TQM (2).pptx
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]
 
Data Science Jobs and Salaries Analysis.pptx
Data Science Jobs and Salaries Analysis.pptxData Science Jobs and Salaries Analysis.pptx
Data Science Jobs and Salaries Analysis.pptx
 
Call Girls In Mahipalpur O9654467111 Escorts Service
Call Girls In Mahipalpur O9654467111  Escorts ServiceCall Girls In Mahipalpur O9654467111  Escorts Service
Call Girls In Mahipalpur O9654467111 Escorts Service
 
How we prevented account sharing with MFA
How we prevented account sharing with MFAHow we prevented account sharing with MFA
How we prevented account sharing with MFA
 
Customer Service Analytics - Make Sense of All Your Data.pptx
Customer Service Analytics - Make Sense of All Your Data.pptxCustomer Service Analytics - Make Sense of All Your Data.pptx
Customer Service Analytics - Make Sense of All Your Data.pptx
 
꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call
꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call
꧁❤ Greater Noida Call Girls Delhi ❤꧂ 9711199171 ☎️ Hard And Sexy Vip Call
 
VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130
VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130
VIP Call Girls Service Miyapur Hyderabad Call +91-8250192130
 
Schema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdfSchema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdf
 
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptxEMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptx
 
Brighton SEO | April 2024 | Data Storytelling
Brighton SEO | April 2024 | Data StorytellingBrighton SEO | April 2024 | Data Storytelling
Brighton SEO | April 2024 | Data Storytelling
 
High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...
High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...
High Class Call Girls Noida Sector 39 Aarushi 🔝8264348440🔝 Independent Escort...
 
(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service
(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service
(PARI) Call Girls Wanowrie ( 7001035870 ) HI-Fi Pune Escorts Service
 

Statistical analysis for recurrent events data with multiple competing risks

  • 1. Statistical Analysis for Recurrent Events Data with Multiple Competing Risks 研 究 生 : 倪佳蓉 指導教授 : 王維菁 教授
  • 2. Motivating Example (1) • The dialysis shunt thrombosis (透析瘻管栓塞) data • Provided by 臺大新竹分院, 腎臟科, 楊忠煒主任 • Patients with shunt thromboses (瘻管栓塞) • Features: • Recurrence • Competing Risks • Acute: 需動手術處理 • Non-acute : 可簡單排除 • Single-type Endpoints • Independent censoring 2
  • 3. Dialysis Shunt Thrombosis Data • Sample size = 2886 • Figure 1. Selected sample path for shunt thrombosis data 3
  • 4. Motivating Example (2) • The peritoneal dialysis (腹膜透析) data • Provided by 臺中榮總 • Patients with peritonitis infection (腹膜炎感染) • Features: • Recurrence • Competing Risks • Gram-positive (革蘭氏陽性菌) • Non-Gram-positive (革蘭氏陰性菌) • Two-type Endpoints • Independent censoring • Dependent censoring 4
  • 5. Peritoneal Dialysis Data • Sample size = 575 • Figure 2. Selected sample path for peritonitis infection data 5 Gram-positive Non-gram-positive
  • 6. Our Interests • Literature developments • How to deal with competing risks? • How to handle association? • How to generate such data? 6
  • 8. Examinations of Data Structures Recurrence Competing risks Semi- competing risks Typical univariate survival data × × × Competing risks data without recurrences × ∨ × Semi-competing risks survival data × × ∨ Recurrent events data with multiple competing risks ∨ ∨ × 8 Data structure form of our two motivating examples
  • 9. Possible Sample Paths for Survival Data 9
  • 10. Typical Univariate Survival Data • Observed variables: 𝑋 = 𝐷 ∧ 𝐶 𝛿 = 𝐼 𝐷 ≤ 𝐶 Initialstate death censored 10 D C
  • 11. Competing Risks Data without Recurrences • Observed variables: ෨𝑌 = Y ∧ 𝐶 ∧ 𝐷 ෨∆= 𝐼 𝑌 ≤ 𝐶 ∆ Initialstate death censored Type-1 event Type-2 event 11 C Y D
  • 12. Semi-Competing Risks Survival Data • Observed variables: 𝑌 ∧ 𝐷 ∧ 𝐶 𝐼 𝑌 ≤ 𝐷 ∧ 𝐶 𝐷 ∧ 𝐶 𝐼 𝐷 ≤ 𝐶 • Semi-competing risks: 𝐷 is a competing risks for 𝑌 but not vice versa Initialstate death rejection Y D 12
  • 13. • Observed variables: ෩𝑌𝑗 = 𝑌𝑗 ∧ 𝐶 ∧ 𝐷 ෩𝑇𝑗 = ෩Y𝑗 − ෩Y𝑗−1 ෩∆𝑗= ∆𝑗 × 𝐼(𝑌𝑗 < 𝐶 ∧ 𝐷) ෩𝑇𝑗 = 𝑇𝑗 , ∀ 𝑗 = 1, ⋯ , 𝑀 − 1 ; but ෨𝑇 𝑀 < 𝑇 𝑀 Initialstate Stage 0 death Type-2 event censored Type-1 event death Type-2 event censored Type-1 event Stage 1 Stage 2 death censored Stage M 13 Recurrent Events Data with Multiple Competing Risks
  • 15. Competing Risks - Classical Definition • Every human is continuously exposed to many risks of death • Cancer • Heart diseases • Pneumonia • … • Because death is not a repetitive event and is usually attributed to a single cause, these risks compete with one another for the life of a person. • What if we are interested in a particular cause of death? • Identifiability issue: Tsiatis (1975) • Crude approach • Net approach 15
  • 16. Competing Risks - Crude Approach • Assumption: • Recognize the presence of all competing risks • 𝑌 = 𝑌1 ∧ ⋯ ∧ 𝑌𝐾 • ∆ = cause of the event • Quantity of interest: • Cumulative incidence function (abbr. CIF) • CIF: 𝐹𝑘 𝑡 = 𝑃𝑟 𝑌 ≤ 𝑡, ∆= 𝑘 • Note that: lim 𝑡→∞ 𝐹𝑘 𝑡 < 1 • 𝐹𝑘 𝑡 is an improper function • Remark: This approach is constructed on a real world 16
  • 17. Competing Risks - Net Approach • Assumptions: • Other competing risks can be removed • 𝑌 = 𝑌1 ∧ ⋯ ∧ 𝑌𝐾 • ∆ = cause of the event • Quantity of interest • Survival function: 𝑆 𝑘 𝑡 = 𝑃𝑟 𝑌𝑘 > 𝑡 • Note that: lim 𝑡→∞ 𝑆 𝑘 𝑡 = 0 • 𝑆 𝑘 𝑡 is a proper survival function • Remark: This approach is constructed on a hypothetical world 17
  • 18. Semi-Competing Risks Data - Revisited What we observed Exist but censored 18 • The implicit assumption is that the censoring events & the terminal event can both be removed. • Although we cannot observe 𝑌 for 𝑌 > 𝐷, we still assume that there is a hypothetical world that such Y exists. ? ? ?
  • 20. Modeling Associations • There are three possible types of association • Between gap-times • Between competing risks • Between recurrences and death • Association models • Frailty (random effect) modeling • Copula modeling • Frailty model and Archimedean Copula model • Equivalence under some conditions in theory • Data generation algorithms are different 20
  • 21. Modeling Associations - Frailty and Copula Approaches • There are three possible types of association • Between gap-times • Between competing risks • Between recurrences and death • 𝑊 : frailty (unobservable) • How 𝑊 affects 𝑇𝑗  usually Cox PH model • Conditional independence given W s.t. 𝑇𝑗 ⊥ 𝑇𝑗′ | 𝑊 • The frailty distribution is specified  usually Gamma dist. • For the frailty control the association between competing risks, • We use the copula package in R • The built-in function rcopula() 21
  • 22. Frailty Modeling - before specifying frailty distribution • Assume the frailty variable 𝑊 has a PH effect on the hazard of 𝑇𝑗 • 𝜆𝑗 𝑡 𝑤 = 𝑤𝜆0𝑗 𝑡 exp 𝑍 𝑇 𝛽 𝑠𝑖𝑚𝑝𝑙𝑖𝑓𝑦 𝜆𝑗 𝑡 𝑤 = 𝑤𝜆0𝑗 𝑡 • Λ𝑗 𝑡 𝑤 = ‫׬‬0 𝑡 𝜆𝑗 𝑢 𝑤 𝑑𝑢 = 𝑤Λ0𝑗 𝑡 • 𝑆𝑗 𝑡 𝑤 = exp −Λ𝑗 𝑡 𝑤 = exp −Λ0𝑗 𝑡 𝑤 = 𝐵𝑗(𝑡) 𝑤 where 𝐵𝑗(𝑡) is a continuous baseline survivor function of 𝑇𝑗 • Derive the unconditional marginal survivor function of 𝑇𝑗 • 𝑆𝑗 𝑡 = ‫׬‬𝑤 𝑆𝑗 𝑡 𝑤 𝑑𝐹 𝑊(𝑤) = ‫׬‬𝑤 𝐵𝑗(𝑡) 𝑤 𝑑𝐹 𝑊(𝑤) = න 𝑤 𝑒 𝑤 log 𝐵 𝑗(𝑡) 𝑑𝐹 𝑊(𝑤) = 𝑝 − log 𝐵𝑗(𝑡) , where 𝑝 𝑢 = 𝐸 𝑒−𝑢𝑊 is the Laplace transform of 𝑊 22 𝑆𝑗 𝑡 𝑤 𝑆𝑗(𝑡) 𝑝 𝑢 𝑝−1 𝑢
  • 23. Frailty Modeling - after specifying frailty distribution • Specify the form of 𝐹 𝑊 𝑤 as W~𝑔𝑎𝑚𝑚𝑎 𝛼, 𝛽 • Since its Laplace transform has a closed form • Derive the Laplace transform as 𝑝 𝑢 = 𝐸 𝑒−𝑢𝑊 = 1 + 𝛽𝑢 −𝛼 • 𝑊 : frailty (unobservable) • To ensure the identifiability of model  𝐸 𝑊 = 1 • To allow within-cluster dependence  𝑉𝑎𝑟 𝑊 = 𝜃1 ≥ 0 • To satisfy the above constraints, assume • W~𝑔𝑎𝑚𝑚𝑎 𝛼 = 1 𝜃1 , 𝛽 = 𝜃1 • so the Laplace transform is • 𝑝 𝑢 = 𝐸 𝑒−𝑢𝑊 = 1 + 𝛽𝑢 −𝛼 = 1 + 𝜃1 𝑢 − 1 𝜃1 23 𝑆𝑗 𝑡 𝑤 𝑆𝑗(𝑡) 𝑝 𝑢 𝑝−1 𝑢
  • 25. Sketch of Simulation Study • Hypothetical world – Net approach • No terminal event or with terminal event • Steps: • Generate multivariate failure times • Take minimum as observed variables • Real world – Crude approach • No terminal event or with terminal event • Steps: • Generate data from improper distribution • Ref: Cheng & Fine (2012) • We propose to use different frailty variables to account for different associations. • Use generated data to evaluate the performance of CIFs Hypothetical 𝐶 𝐶, 𝐷 Real world 𝐶 𝐶, 𝐷 25
  • 26. • Under the hypothetical world assumption • Remove the influence of a terminal event by taking 𝐷 = ∞ Cases 1. Data Generation Algorithms 𝑇𝑗(1) 𝑇𝑗(2) • 𝑇𝑗 = min{ 𝑇𝑗(1), 𝑇𝑗(2)} • ∆𝑗 = ቊ 1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(1) 2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(2) Compete 𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3 ⋯ Frailty H Frailty W Hypothetical 𝐶 𝐶, 𝐷 Real world 𝐶 𝐶, 𝐷 26
  • 27. Cases 1. Data Generation Algorithms (cont’d) • Step 1: For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎 1 𝜃1 , 𝜃1 and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾 • Step 2: For each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use 𝑈𝑗(𝑘) to generate 𝑇𝑗(𝑘) as follows: 𝜆𝑗 𝑡|𝑤 = 𝑤𝜆0𝑗 𝑡 = 𝑤𝜉𝑗  S(𝑗) 𝑡|𝑤 = 𝑒𝑥𝑝 −Λ𝑗 𝑡 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑤  𝑈𝑗(𝑘) = 𝑒𝑥𝑝 − 𝜉𝑗 𝑇𝑗(𝑘) 𝑤  𝑇𝑗(𝑘) = − 1 𝑤𝜉 𝑗 𝑙𝑜𝑔𝑈𝑗 𝑘 So, define 𝑇𝑗 = min{𝑇𝑗 1 , 𝑇𝑗(2)} and ൝ Δ𝑗 = 1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 1 Δ𝑗 = 2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 2 . Then, set 𝑌𝑗 = σ 𝓂=1 𝑗 𝑇 𝑚. 27
  • 28. Cases 1. Data Generation Algorithms (cont’d) • Step 3: If 𝑌𝑗 ≤ 𝐶, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗. Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤ 𝐶 < 𝑌 𝑀. Finally, we set ෪𝑌 𝑀 = 𝐶, ෪𝑇 𝑀 = 𝐶 − 𝑌 𝑀−1, and ෪∆ 𝑀= 0. 28
  • 29. • Under the hypothetical world assumption • Consider the influence of a terminal event by taking 𝐷 < ∞ Cases 2. Data Generation Algorithms 𝑇𝑗(1) 𝑇𝑗(2) • 𝑇𝑗 = min{ 𝑇𝑗(1), 𝑇𝑗(2)} • ∆𝑗 = ቊ 1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(1) 2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗(2) Compete 𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3 ⋯ 𝐷 Frailty W Frailty 𝐻 Frailty V Hypothetical 𝐶 𝐶, 𝐷 Real world 𝐶 𝐶, 𝐷 29
  • 30. Cases 2. Data Generation Algorithms (cont’d) • Step 1: For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎 1 𝜃1 , 𝜃1 , 𝑉~𝑔𝑎𝑚𝑚𝑎 1 𝜃2 , 𝜃2 and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾 • Step 2: For each stage, generate 𝑈D~unif 0,1 , then use 𝑈D to generate 𝐷 by 𝜆 𝐷 𝑡|𝑣 = 𝑣𝜆0𝐷 𝑡 = 𝑣𝜂  𝑆 𝐷 𝑡|𝑣 = 𝑒𝑥𝑝 −Λ 𝐷 𝑡 = 𝑒𝑥𝑝 − 𝜂𝑡 𝑣  𝑈 𝐷 = 𝑒𝑥𝑝 − 𝜂𝐷 𝑣  𝐷 = − 1 𝑣𝜂 𝑙𝑜𝑔𝑈 𝐷 Moreover, for each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use 𝑈𝑗(𝑘) to generate 𝑇𝑗(𝑘) as follows: 𝜆𝑗 𝑡|𝑣, 𝑤 = 𝑣𝑤𝜆0𝑗 𝑡 = 𝑣𝑤𝜉𝑗  S(𝑗) 𝑡|𝑣, 𝑤 = 𝑒𝑥𝑝 −Λ𝑗 𝑡 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑣𝑤 30
  • 31. Cases 2. Data Generation Algorithms (cont’d)  𝑈𝑗(𝑘) = 𝑒𝑥𝑝 − 𝜉𝑗 𝑇𝑗(𝑘) 𝑣𝑤  𝑇𝑗(𝑘) = − 1 𝑣𝑤𝜉 𝑗 𝑙𝑜𝑔𝑈𝑗 𝑘 So, define 𝑇𝑗 = min{𝑇𝑗 1 , 𝑇𝑗(2)} and ൝ Δ𝑗 = 1, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 1 Δ𝑗 = 2, 𝑖𝑓 𝑇𝑗 = 𝑇𝑗 2 . Then, set 𝑌𝑗 = σ 𝓂=1 𝑗 𝑇 𝑚. • Step 3: If 𝑌𝑗 ≤ min{𝐶, 𝐷}, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗. Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤ min{C, D} < 𝑌 𝑀. Finally, we set ෪𝑌 𝑀 = min{𝐶, 𝐷}, ෪𝑇 𝑀 = min{𝐶, 𝐷} − 𝑌 𝑀−1, and ෪∆ 𝑀= ቊ 0, 𝑖𝑓 𝐶 ≤ 𝐷 3, 𝑖𝑓 𝐶 > 𝐷 . 31
  • 32. • Under the real world assumption • Remove the influence of a terminal event by taking 𝐷 = ∞ Cases 3. Data Generation Algorithms 𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3 ⋯ Frailty W 𝜂1 ≜ 𝐹 𝑗 ∞ 𝑤 𝑡 1 𝜂1 𝐹 𝑗 (𝑡|𝑤) Frailty 𝐻 • 𝜂1 ≜ 𝐹1 (𝑗) ∞ 𝑤 • 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 • 𝐼𝑓 𝑈𝑗(1) ≤ 𝜂1, 𝑡ℎ𝑒𝑛 𝑢𝑠𝑒 𝑈𝑗(1) 𝑡𝑜 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒 𝑇𝑗 = 𝑇𝑗(1) 𝑎𝑛𝑑 ∆𝑗= 1; 𝑜𝑟 𝑣𝑖𝑐𝑒 𝑣𝑒𝑟𝑠𝑎. Hypothetical 𝐶 𝐶, 𝐷 Real world 𝐶 𝐶, 𝐷 32
  • 33. Cases 3. Data Generation Algorithms (cont’d) • Step 1: For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎 1 𝜃1 , 𝜃1 and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾 • Step 2: For each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use 𝑈𝑗(𝑘) to generate 𝑇𝑗 as follows: S(𝑗) 𝑡|𝑤 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑤  S(𝑗) 𝑡 = ‫׬‬𝑤 S(𝑗) 𝑡|𝑤 𝑑𝐹 𝑊(𝑤) = 1 + 𝜃1 𝜉𝑗 𝑡 − 1 𝜃1  F1 𝑗 𝑡 + F2 𝑗 𝑡 = 1 − S 𝑗 𝑡 = 1 − 1 + 𝜃1 𝜉𝑗 𝑡 − 1 𝜃1  ൞ F1 𝑗 𝑡 = 𝜓𝑗 1 − 1 + 𝜃1 𝜉𝑗 𝑡 − 1 𝜃1 F2 𝑗 𝑡 = 1 − 𝜓𝑗 1 − 1 + 𝜃1 𝜉𝑗 𝑡 − 1 𝜃1  F 𝑘 𝑗 𝑡|𝑤 = 𝐵𝑘 𝑗 (𝑡) 𝑤 = 𝑒𝑥𝑝 −𝑞 F 𝑘 𝑗 𝑡 𝑤 33
  • 34. Cases 3. Data Generation Algorithms (cont’d)  Have explicit forms for F1 𝑗 𝑡|𝑤 & F2 𝑗 𝑡|𝑤 .  Set lim 𝑡→∞ F 𝑘 𝑗 𝑡|𝑤 = 𝜂1 - If 𝑈𝑗(1) ≤ 𝜂1, generate Δ𝑗 = 1 and 𝑇𝑗 = 𝑇𝑗 1 = F1 𝑗 −1 𝑈𝑗(1)|𝑊 = 𝑤 - If 𝑈𝑗(1) > 𝜂1, generate Δ𝑗 = 2 and 𝑇𝑗 = 𝑇𝑗 2 , where 𝑇𝑗 2 is generated from 𝑃𝑟 𝑇𝑗 ≤ 𝑡|Δ𝑗 = 2 = F2 𝑗 𝑡 F2 𝑗 ∞ Then, set 𝑌𝑗 = σ 𝓂=1 𝑗 𝑇 𝑚. • Step 3: If 𝑌𝑗 ≤ 𝐶, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗. Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤ 𝐶 < 𝑌 𝑀. Finally, we set ෪𝑌 𝑀 = 𝐶, ෪𝑇 𝑀 = 𝐶 − 𝑌 𝑀−1, and ෪∆ 𝑀= 0. 34
  • 35. • Under the real world assumption • Consider the influence of a terminal event by taking 𝐷 < ∞ Cases 4. Data Generation Algorithms 𝑇1, Δ1 𝑇2, Δ2 𝑇3, Δ3 ⋯ 𝐷 Frailty W Frailty V 𝜂1 ≜ 𝐹 𝑗 ∞ 𝑤 𝑡 1 𝜂1 𝐹 𝑗 (𝑡|𝑤) Frailty 𝐻 • 𝜂1 ≜ 𝐹1 (𝑗) ∞ 𝑤 • 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 • 𝐼𝑓 𝑈𝑗(1) ≤ 𝜂1, 𝑡ℎ𝑒𝑛 𝑢𝑠𝑒 𝑈𝑗(1) 𝑡𝑜 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒 𝑇𝑗 = 𝑇𝑗(1) 𝑎𝑛𝑑 ∆𝑗= 1; 𝑜𝑟 𝑣𝑖𝑐𝑒 𝑣𝑒𝑟𝑠𝑎. Hypothetical 𝐶 𝐶, 𝐷 Real world 𝐶 𝐶, 𝐷 35
  • 36. Cases 4. Data Generation Algorithms (cont’d) • Step 1: For each subject, generate 𝑊~𝑔𝑎𝑚𝑚𝑎 1 𝜃1 , 𝜃1 , 𝑉~𝑔𝑎𝑚𝑚𝑎 1 𝜃2 , 𝜃2 and 𝐶~𝑢𝑛𝑖𝑓 0, 𝐾 • Step 2: For each stage, generate 𝑈D~unif 0,1 , then use 𝑈D to generate 𝐷 by 𝜆 𝐷 𝑡|𝑣 = 𝑣𝜆0𝐷 𝑡 = 𝑣𝜂  𝑆 𝐷 𝑡|𝑣 = 𝑒𝑥𝑝 −Λ 𝐷 𝑡 = 𝑒𝑥𝑝 − 𝜂𝑡 𝑣  𝑈 𝐷 = 𝑒𝑥𝑝 − 𝜂𝐷 𝑣  𝐷 = − 1 𝑣𝜂 𝑙𝑜𝑔𝑈 𝐷 36
  • 37. Cases 4. Data Generation Algorithms (cont’d) Moreover, for each stage, generate 𝑈𝑗(1), 𝑈𝑗(2) ~𝐶𝑙𝑎𝑦𝑡𝑜𝑛 𝜃0 , then use 𝑈𝑗(𝑘) to generate 𝑇𝑗 as follows: 𝜆𝑗 𝑡|𝑣, 𝑤 = 𝑣𝑤𝜆0𝑗 𝑡 = 𝑣𝑤𝜉𝑗  S(𝑗) 𝑡|𝑣, 𝑤 = 𝑒𝑥𝑝 −Λ𝑗 𝑡 = 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑣𝑤  F1 𝑗 𝑡|𝑣, 𝑤 + F2 𝑗 𝑡|𝑣, 𝑤 = 1 − S(𝑗) 𝑡|𝑣, 𝑤 = 1 − 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑣𝑤  ൞ F1 𝑗 𝑡|𝑤 = 𝜓𝑗 1 − 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑣𝑤 F2 𝑗 𝑡|𝑤 = 1 − 𝜓𝑗 1 − 𝑒𝑥𝑝 − 𝜉𝑗 𝑡 𝑣𝑤  Have explicit forms for F1 𝑗 𝑡|𝑣, 𝑤 & F2 𝑗 𝑡|𝑣, 𝑤 .  Set lim 𝑡→∞ F 𝑘 𝑗 𝑡|𝑤 = 𝜂1 37
  • 38. Cases 4. Data Generation Algorithms (cont’d) - If 𝑈𝑗(1) ≤ 𝜂1, set Δ𝑗 = 1 and 𝑇𝑗 = 𝑇𝑗 1 = F1 𝑗 −1 𝑈𝑗(1)|𝑉 = 𝑣, 𝑊 = 𝑤 - If 𝑈𝑗(1) > 𝜂1, set Δ𝑗 = 2 and 𝑇𝑗 = 𝑇𝑗 2 , where 𝑇𝑗 2 is generated from 𝑃𝑟 𝑇𝑗 ≤ 𝑡|Δ𝑗 = 2 = F2 𝑗 𝑡 F2 𝑗 ∞ Then, set 𝑌𝑗 = σ 𝓂=1 𝑗 𝑇 𝑚. • Step 3: If 𝑌𝑗 ≤ min{𝐶, 𝐷}, set ෩𝑌𝑗 = 𝑌𝑗 and ෩𝑇𝑗 = 𝑇𝑗. Repeat Step 2 for 𝑗 = 1, ⋯ , (𝑀 − 1), where 𝑀 satisfies 𝑌 𝑀−1 ≤ min{C, D} < 𝑌 𝑀. Finally, we set ෪𝑌 𝑀 = min{𝐶, 𝐷}, ෪𝑇 𝑀 = min{𝐶, 𝐷} − 𝑌 𝑀−1, and ෪∆ 𝑀= ቊ 0, 𝑖𝑓 𝐶 ≤ 𝐷 3, 𝑖𝑓 𝐶 > 𝐷 . 38
  • 40. Simulation Study Hypothetical 𝐶 𝐶, 𝐷 Real world 𝐶 𝐶, 𝐷 40 Case (i) Case (ii) Case (iii) Case (iv) Between competing risks 𝜃0 indep. (1) dep. (1.25) indep. (1) dep. (1.25) Between gap-times 𝜃1 indep. (1) indep. (1) dep. (1.25) dep. (1.25) • Ref: Ph.D. thesis of Bowen Li (2016) • IPCW (inverse probability censoring weighting) to adjust bias
  • 41. • Parameter settings: • 𝜓𝑗 = 0.25 • 𝜉𝑗 = 0.25 • 𝐶~𝑢𝑛𝑖𝑓 0, 500 • 𝑁 = 1000 • 𝐵 = 500 • Assume no correlation, so frailties parameters: • 𝜃0 = 1 • 𝜃1 = 1 Simulation Study – Case (i) 41
  • 42. • Parameter settings: • 𝜓𝑗 = 0.25 • 𝜉𝑗 = 0.25 • 𝐶~𝑢𝑛𝑖𝑓 0, 500 • 𝑁 = 1000 • 𝐵 = 500 • Assume positive correlation, so frailties parameters: • 𝜃0 = 1.25 • 𝜃1 = 1 Simulation Study – Case (ii) 42
  • 43. • Parameter settings: • 𝜓𝑗 = 0.25 • 𝜉𝑗 = 0.25 • 𝐶~𝑢𝑛𝑖𝑓 0, 500 • 𝑁 = 1000 • 𝐵 = 500 • Assume positive correlation, so frailties parameters: • 𝜃0 = 1 • 𝜃1 = 1.25 Simulation Study – Case (iii) 43
  • 44. • Parameter settings: • 𝜓𝑗 = 0.25 • 𝜉𝑗 = 0.25 • 𝐶~𝑢𝑛𝑖𝑓 0, 500 • 𝑁 = 1000 • 𝐵 = 500 • Assume positive correlation, so frailties parameters: • 𝜃0 = 1.25 • 𝜃1 = 1.25 Simulation Study – Case (iv) 44
  • 45. Conclusions • Two real-world examples • Analyze recurrent events data with competing risks • Literature development • Review survival data structure • Deal with competing risks • Account for Identifiability issue • “Net” vs. “Crude” approach • Handle association by frailty approach • Between gap-times • Between competing risks • Between recurrence processes and death • Simulation • Propose data generation algorithms • Conduct simulation analysis to examine the validity 45
  • 46. References (1) • Chen, C. M., Chuang, Y. W., & Shen, P. S. (2015). Two‐stage estimation for multivariate recurrent event data with a dependent terminal event. Biometrical Journal, 57(2), 215-233. • Chang, W. H., & Wang, W. (2009). Regression analysis for cumulative incidence probability under competing risks. Statistica Sinica, 391-408 • Cheng, Y., & Fine, J. P. (2012). Cumulative incidence association models for bivariate competing risks data. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 74(2), 183- 202. • Goethals, K., Janssen, P., & Duchateau, L. (2008). Frailty models and copulas: similarities and differences. Journal of Applied Statistics, 35(9), 1071-1079. 46
  • 47. References (2) • Oakes, D. (1982). A model for association in bivariate survival data. Journal of the Royal Statistical Society. Series B (Methodological), 414-422. • Oakes, D. (1989). Bivariate survival models induced by frailties. Journal of the American Statistical Association, 84(406), 487-493. • Tsiatis, A. (1975). A nonidentifiability aspect of the problem of competing risks. Proceedings of the National Academy of Sciences, 72(1), 20-22. • Wang, W., & Wells, M. T. (2000). Estimation of Kendall's tau under censoring. Statistica Sinica, 1199-1215. 47
  • 48. Thank you for your attention! 48
  • 49. Appendix - CIF Estimation • Ref: Ph.D. thesis of Bowen Li (2016) • IPCW (inverse probability censoring weighting) to adjust bias  is due to induced dependent censoring • Goal: Estimate 𝐹𝑘 𝑗 𝑡 = 𝐸 𝐼 𝑇𝑗 ≤ 𝑡, ∆𝑗= 𝑘 for 𝑗 ≥ 2 • Empirical proxy: 𝐼 ෩𝑇𝑗 ≤ 𝑡, ෩∆𝑗= 𝑘 = 𝐼 𝑇𝑗 ≤ 𝑡, ∆𝑗= 𝑘, 𝐶 > 𝑌𝑗 • By conditional expectation, 𝐹𝑘 𝑗 𝑡 = 𝐸 𝐸 𝐼 ෩𝑇𝑗 ≤ 𝑡, ෩∆𝑗= 𝑘 |𝑌𝑗, 𝑌𝑗−1, ∆𝑗 𝐺(෩𝑌𝑗) , where 𝐺 𝑡 = 𝑃𝑟 𝐶 > 𝑡 is the survival function for 𝐶 • CIFs estimators based on observed data can be expressed as ෠𝐹𝑘 (𝑗) = ෍ 𝑖=1 𝑛 𝐼 ෪𝑇𝒊𝑗 ≤ 𝑡, ෪∆𝒊𝑗= 𝑘 𝑛 ෠𝐺(෪𝑌𝒊𝑗) 49