Time Series Analysis of Nursing Notes for Mortality Prediction via - - PowerPoint PPT Presentation

▶
time series analysis of nursing notes for mortality
SMART_READER_LITE
LIVE PREVIEW

Time Series Analysis of Nursing Notes for Mortality Prediction via - - PowerPoint PPT Presentation

Oct 04, 2022 157 likes •495 views

Time Series Analysis of Nursing Notes for Mortality Prediction via a State Transition Topic Model Date: 2016/10/18 Author: Yohan JO, Natasha Loghmanpour, Carolyn Penstein Rose Source: CIKM15 Advisor: Jia-Ling Koh Speaker: Chih-Hsuan Tzang

slide-1
SLIDE 1

Time Series Analysis of Nursing Notes for Mortality Prediction via a State Transition Topic Model

Date: 2016/10/18 Author: Yohan JO, Natasha Loghmanpour, Carolyn Penstein Rose Source: CIKM’15 Advisor: Jia-Ling Koh Speaker: Chih-Hsuan Tzang

1

slide-2
SLIDE 2

Introduction Method Experiment Conclusion

2

slide-3
SLIDE 3

Introduction Method Experiment Conclusion

3

slide-4
SLIDE 4

Introduction

  • Family support
  • Mental fitness
  • Facial expressions
  • Nurses’ intuitions & plans

4

slide-5
SLIDE 5

Introduction

Motivation:

  • Predicting a patient’s risk of mortality and taking

appropriate action are important activities in intensive care units (ICUs).

  • Nursing notes have the potential to uncover hidden

clues about a patient’s health and mental state as they change over time, such as the factors of family support and mental fitness

5

slide-6
SLIDE 6

Introduction

Goal:

  • Proposes and evaluates a model to uncover the temporal

dynamics of underlying patient states from nursing notes.

  • Evaluates the effectiveness of the identified temporal

dynamics for improving mortality prediction.

  • Offers qualitative insight into different types of textual

features regarding their roles in mortality prediction.

6

slide-7
SLIDE 7

Introduction

(STTM)

7

slide-8
SLIDE 8

Introduction Method Experiment Conclusion

8

slide-9
SLIDE 9

Method

9

slide-10
SLIDE 10

Method

  • W

e perform predictions in the short-term (one day and one week) and in the long-term (one month, six months, and one year).

  • Since some notes are too short or focus only on a single topic, all

nursing notes in the same time point are merged into one document for analysis so that the merged document can reflect the overall topics at that time point.

10

slide-11
SLIDE 11

Method

11

slide-12
SLIDE 12

Method

State T ransition Topic Model (STTM):

12

slide-13
SLIDE 13

Method

Latent Dirichlet Allocation (LDA): Example: 1. I like to eat broccoli and bananas. 2. I ate a banana and spinach smoothie for breakfast. 3. Chinchillas and kittens are cute.

  • 4. My sister adopted a kitten yesterday.

5. Look at this cute hamster munching on a piece of broccoli.

13

slide-14
SLIDE 14

Method

Latent Dirichlet Allocation (LDA):

1. I like to eat broccoli and bananas. 2. I ate a banana and spinach smoothie for breakfast. 3. Chinchillas and kittens are cute. 4. My sister adopted a kitten yesterday. 5. Look at this cute hamster munching on a piece of broccoli.

  • Sentences 1 and 2: 100% Topic A
  • Sentences 3 and 4: 100% Topic B
  • Sentences 5: 60% Topic A, 40% Topic B
  • Topic A: 30% broccoli, 15% bananas, 10%

breakfast, 10% munching

  • Topic B: 20% chinchillas, 20% kittens, 20% cute,

15% hamster,…

14

slide-15
SLIDE 15

Method

Hidden Markov Model(HMM):

15

slide-16
SLIDE 16

Method

STTM:

16

slide-17
SLIDE 17

Method

17

slide-18
SLIDE 18

Method

N-grams:

  • For each patient, unigrams and bigrams are

extracted from nursing notes.

  • use 200 and 100 top n-grams for the mortality

group and the survival group.

  • the mortality group, is much smaller than the

survival group, requires more n-grams for high recall.

  • Pointwise Mutual Information

18

slide-19
SLIDE 19

Method

Standard Topics:

  • Topic distributions learned by LDA are used.
  • Topic distributions are extracted from individual

nursing notes (not merged notes).

  • The extracted topic distributions are averaged

and aggregated into one feature vector.

  • The dimension is equal to the number of topics.

19

slide-20
SLIDE 20

Method

State-aware Topics:

  • Document-wise topic distributions learned by

STTM are used

  • STTM estimates one state per time point
  • The topic distributions of all documents are

aggregated into one feature vector as for standard topics.

20

slide-21
SLIDE 21

Method

State T ransitions:

  • a sequence of states is estimated by STTM.
  • only the latest four time points (i.e., two days) are

considered (S + S2)

  • longer sequences tend to include more states and

have lower element values.

21

slide-22
SLIDE 22

Method

22

slide-23
SLIDE 23

Method

cost-sensitive SVMs:

23

slide-24
SLIDE 24

Introduction Method Experiment Conclusion

24

slide-25
SLIDE 25

Experiment

  • Clinical data of ICU patients collected between

2001 and 2008

25

slide-26
SLIDE 26

Experiment

Task 1. State T ransitions Learned From Nursing Notes:

26

slide-27
SLIDE 27

Experiment

Task 2. Mortality Prediction with Temporal Information:

27

slide-28
SLIDE 28

Experiment

Task 2. Mortality Prediction with Temporal Information:

28

slide-29
SLIDE 29

Experiment

Task 2. Mortality Prediction with Temporal Information:

29

slide-30
SLIDE 30

Experiment

Task 3. Mortality Prediction by Individual features:

30

slide-31
SLIDE 31

Experiment

Enrichment:

31

slide-32
SLIDE 32

Introduction Method Experiment Conclusion

32

slide-33
SLIDE 33

Conclusion

  • The learned temporal informations beneficial for long- term

mortality prediction, but not much in short-term prediction.

  • STTM can be applied to any data stream.

33