Composer: Visual Cohort Analysis of Patient Outcomes Jennifer - - PowerPoint PPT Presentation

▶
composer visual cohort analysis of patient outcomes
SMART_READER_LITE
LIVE PREVIEW

Composer: Visual Cohort Analysis of Patient Outcomes Jennifer - - PowerPoint PPT Presentation

Mar 04, 2024 176 likes •826 views

Composer: Visual Cohort Analysis of Patient Outcomes Jennifer Rogers, Nicholas Spina, Ashley Neese, Rachel Hess, Darrel Brodke, Alexander Lex 1 Lower Back Pain is a Significant Health Burden 2.6 Million Emergency Room visits Treatment exceeding

slide-1
SLIDE 1

Composer: Visual Cohort Analysis

  • f Patient Outcomes

Jennifer Rogers, Nicholas Spina, Ashley Neese, Rachel Hess, Darrel Brodke, Alexander Lex

1

slide-2
SLIDE 2

Lower Back Pain is a Significant Health Burden

2.6 Million Emergency Room visits Treatment exceeding $100 Billion

2

slide-3
SLIDE 3

This is Frank. He has a herniated disc.

3

slide-4
SLIDE 4

4

Intervertebral herniated disc lower back pain weakness in legs bladder and bowel problems

Dakota Harr

slide-5
SLIDE 5

Three treatment options to consider with his doctor.

5

slide-6
SLIDE 6

6

Surgery mostly effective for persistent symptoms 12% will need another one within 4 years. 43% of these will need fusion Risk involved, takes time to recover

slide-7
SLIDE 7

Frank has some pre-existing conditions.

Age 55 BMI: 29 Diabetic Tried physical therapy

7

slide-8
SLIDE 8

Takes these into account along with past experience and clinical guidelines.

8

slide-9
SLIDE 9

general population may not provide an accurate reflection

  • f potential outcomes for patients with pre-existing

conditions.

9

BMI: 30 diabetic

slide-10
SLIDE 10

10

EHR for evidence based comparisons Identify factors that can influence recovery and more accurately predict outcomes

slide-11
SLIDE 11

11

Dataset of Prior Cases Outcome Measures at Many Timepoints Accurate Cohort Definition Prognosis Under Different Treatment Options

slide-12
SLIDE 12

Cohorts: subset of the general population shares defining characteristics

12

slide-13
SLIDE 13

Effective for identifying influential factors.

13

slide-14
SLIDE 14

14

Investigating Patient Reported Outcomes as measure of well-being

slide-15
SLIDE 15

PROMIS

Patient Reported Outcome Measurement Information System.

15

Evaluate and monitor physical, mental, social health.

slide-16
SLIDE 16

Focus on PROMIS physical function scores.

16

slide-17
SLIDE 17

Way to quantify the physical ability

32: Can stand for short time. 55: Can go on a short hike. 72: Can run 10 miles.

17

slide-18
SLIDE 18

Collected over time Track patient progression

32: Can stand for short time. 72: Can run 10 miles.

18

34: 2 weeks after surgery 55: 1 month after surgery. 65: 2 months after surgery.

slide-19
SLIDE 19

use PROMIS PF to more accurately evaluate progression To compare outcomes

BMI: 28 Age: 55 CCI: 1 Age: 70 Smokes Age: 45 Diabetic Age: 35

19

slide-20
SLIDE 20

Lack tools that use PROMIS PF trajectories

BMI: 28 Age: 55 CCI: 1 Age: 70 Smokes Age: 45 Diabetic Age: 35

20

slide-21
SLIDE 21

21

PROMIS PF scores for 6071 patients beginning in 2013 Range of 1 to more than 20 scores ICD/CPT codes, demographic data, comorbidities Dataset

slide-22
SLIDE 22

22

3 requirements for functionality

  • 1. Define meaningful cohorts of patients
  • 2. Compare outcomes of different cohorts
  • 3. Compare outcomes of different treatments
slide-23
SLIDE 23
  • 1. Define meaningful cohorts of patients.

23

Domain Requirements

slide-24
SLIDE 24
  • 1. Define meaningful cohorts of patients.

24

Domain Requirements

slide-25
SLIDE 25
  • 2. Compare outcomes of different cohorts.

Cohort 1 Cohort 2 Cohort 1

25

Domain Requirements

slide-26
SLIDE 26
  • 3. Compare outcomes of different treatments.

Surgery Injection

26

Domain Requirements

slide-27
SLIDE 27

Related Work

27

slide-28
SLIDE 28

Mane, K.K., Bizon, C., Schmitt, C., Owen, P., Burchett, B., Pietrobon, R. and Gersing, K., 2012. VisualDecisionLinc: A visual analytics approach for comparative effectiveness-based clinical decision support in psychiatry. Journal of Biomedical Informatics, 45(1), pp.101-106.

Patient score trajectories in the context of a similar group of patients.

28

slide-29
SLIDE 29

Bernard, Jürgen, et al. "A visual-interactive system for prostate cancer cohort analysis." IEEE computer graphics and applications 35.3 (2015): 44-55.

Iterative cohort refinement.

29

slide-30
SLIDE 30

Contributions

Comparison of treatment options measured by patient score trajectories Ability to normalize and adjust representation of trajectories Flexible definition of multiple patient cohorts for comparison

30

slide-31
SLIDE 31

31

slide-32
SLIDE 32

32

slide-33
SLIDE 33

33

slide-34
SLIDE 34

Define a cohort for Frank by filtering based on attributes.

Age 55 BMI: 29 Diabetic Tried physical therapy

34

slide-35
SLIDE 35

Filter History.

35

Patient count of cohort at each filter stage

slide-36
SLIDE 36

Filter History.

36

Remove and recalculate

slide-37
SLIDE 37

Demographic Filters.

37

slide-38
SLIDE 38

Score & CPT Filters.

38

slide-39
SLIDE 39

Filtering by attributes to define a cohort like Frank

39

slide-40
SLIDE 40

Cohort control panel.

40

Added

slide-41
SLIDE 41

Cohort control panel.

41

Branched

slide-42
SLIDE 42

Cohort control panel.

42

Remove

slide-43
SLIDE 43

How did patients like Frank progress after surgery?

43

slide-44
SLIDE 44

Realign scores to see trend after surgery

Aligned by first recorded PROMIS score Aligned by surgery

44

slide-45
SLIDE 45

Aligning by event

45

slide-46
SLIDE 46

Patient score trajectories have different baselines

46

Small change (2-8) clinically meaningful

slide-47
SLIDE 47

47

Hard to see measured change in scores

slide-48
SLIDE 48

Changing scales to relative score change

48

slide-49
SLIDE 49

This is messy.

49

slide-50
SLIDE 50

We want to see the general trend in score fluctuation

50

slide-51
SLIDE 51

Aggregation

  • f scores

51

slide-52
SLIDE 52

How did the patients with the most positive change in score progress?

52

What about the bottom quantile for score change?

slide-53
SLIDE 53

Separation of Scores by Quantiles.

Adjust the day range to calculate average score change.

53

slide-54
SLIDE 54

How did patients like Frank progress after surgery vs injection?

54

slide-55
SLIDE 55

Compare cohorts in layer view

55

slide-56
SLIDE 56

We find a patient line of interest What other events are present in their medical histories?

56

slide-57
SLIDE 57

Drill down into individual patient histories

57

slide-58
SLIDE 58

Moving Forward

Generalize to a broader clinical base Development of a shared decision-making interface

58

slide-59
SLIDE 59

Thank You

59

Learn more about our lab: http://vdl.sci.utah.edu/ Learn more on the project website: http://bit.ly/composer_paper

slide-60
SLIDE 60

60

Comparison of multiple treatment outcomes.

Franklin, L., Plaisant, C., Minhazur Rahman, K. and Shneiderman, B.,

  • 2014. TreatmentExplorer: An interactive

decision aid for medical risk communication and treatment

  • exploration. Interacting with Computers,

28(3), pp.238-252.

slide-61
SLIDE 61

Separating By quantiles

61

slide-62
SLIDE 62

Utah Health Using PROMIS scores longer than any other institution in the country. PROMIS physical function scores.

62

slide-63
SLIDE 63

Cohort control panel.

Cohorts can be added, branched and deleted.

63

slide-64
SLIDE 64

Adding, Branching, Removing Cohorts

64