Skip to content

Commit 2e69022

Browse files
zachllzachll
authored
Merge pull request #184 from Display-Lab/preference-model-update
Preference model update
2 parents 360ce88 + 5fc2a35 commit 2e69022

File tree

17 files changed

+567
-389
lines changed

17 files changed

+567
-389
lines changed

‎vignettes/goal_approach_vignette/readme.md‎

Lines changed: 39 additions & 23 deletions
Original file line numberDiff line numberDiff line change
@@ -37,35 +37,51 @@ Chikondi, a resident anesthesiologist at Max Community Hospital, has the followi
3737
## Preference data
3838
Preferences for precision feedback are elicited through a preference survey that providers can take. The preference survey produces a preference model for each provider that, with the provider's permission, is shared with MPOG to maintain. MPOG analyses preference data that is shared to identify population-level preference segments. These segments are generated as preference profiles that can serve as a default preference model for an organization, or which can be selected by providers who do not take the preference survey, but who identify preferences that are close enough to their own in the settings menu for the precision feedback system.
3939

40-
Alice's preference data:
40+
Alice's preference data (Midwest Medicine group profile) :
4141

4242
|Motivating information |Utility value|Description |
4343
|-----------------------|-------------|--------------------------------------------------------------------------------|
44-
|Social gain |XX.XXXXXX |Performance was previously low, but it has improved to reach the peer benchmark.|
45-
|Social stayed better |6.164377 |Performance is consistently high (no recent change). |
46-
|Worsening |-12.23023 |Performance is worsening. |
47-
|Improving |3.325883 |Performance is improving. |
48-
|Goal approach |9.956127 |Performance was previously high, but it has dropped below the peer average. |
49-
|Social stayed worse |-7.710484 |Performance has remained below average (no recent change). |
50-
|Social better |-1.61124 |Performance is high this month. |
51-
|Social worse |-14.86794 |Performance is low this month. |
52-
|Social approach |13.99621 |Performance is improving, getting closer to the peer benchmark. |
53-
54-
55-
Chikondi's preference data:
44+
|Social better |-2 |Performance is high this month. |
45+
|Social gain |13 |Performance was previously low, but it has improved to reach the peer benchmark.|
46+
|Social stayed better |6 |Performance is consistently high (no recent change). |
47+
|Worsening |-22 |Performance is worsening. |
48+
|Improving |3 |Performance is improving. |
49+
|Social loss |10 |Performance was previously high, but it has dropped below the peer average. |
50+
|Social worse |-15 |Performance is low this month. |
51+
|Social stayed worse |-8 |Performance has remained below average (no recent change). |
52+
|Social approach |15 |Performance is improving, getting closer to the peer benchmark. |
53+
|Goal gain |13 |Performance was previously low, but it has improved to reach the goal. |
54+
|Goal loss |10 |Performance was previously high, but it has dropped below the goal. |
55+
|Goal approach |14 |Performance is improving, getting closer to the goal. |
56+
57+
|Display format |Utility value|
58+
|-----------------------|-------------|
59+
|Bar chart |1 |
60+
|Line chart |0 |
61+
|Text only |0 |
62+
63+
Chikondi's preference data (Midwest Medicine group profile) :
5664

5765
|Motivating information |Utility value|Description |
5866
|-----------------------|-------------|--------------------------------------------------------------------------------|
59-
|Social gain |-2.92114 |Performance was previously low, but it has improved to reach the peer benchmark.|
60-
|Social stayed better |-12.76936 |Performance is consistently high (no recent change). |
61-
|Worsening |1.06977 |Performance is worsening. |
62-
|Improving |-0.26266 |Performance is improving. |
63-
|Goal approach |2.72075 |Performance was previously high, but it has dropped below the peer average. |
64-
|Social stayed worse |9.97743 |Performance has remained below average (no recent change). |
65-
|Social better |-5.97766 |Performance is high this month. |
66-
|Social worse |-0.05277 |Performance is low this month. |
67-
|Social approach |-0.24384 |Performance is improving, getting closer to the peer benchmark. |
68-
67+
|Social better |-2 |Performance is high this month. |
68+
|Social gain |13 |Performance was previously low, but it has improved to reach the peer benchmark.|
69+
|Social stayed better |6 |Performance is consistently high (no recent change). |
70+
|Worsening |-22 |Performance is worsening. |
71+
|Improving |3 |Performance is improving. |
72+
|Social loss |10 |Performance was previously high, but it has dropped below the peer average. |
73+
|Social worse |-15 |Performance is low this month. |
74+
|Social stayed worse |-8 |Performance has remained below average (no recent change). |
75+
|Social approach |15 |Performance is improving, getting closer to the peer benchmark. |
76+
|Goal gain |13 |Performance was previously low, but it has improved to reach the goal. |
77+
|Goal loss |10 |Performance was previously high, but it has dropped below the goal. |
78+
|Goal approach |14 |Performance is improving, getting closer to the goal. |
79+
80+
|Display format |Utility value|
81+
|-----------------------|-------------|
82+
|Bar chart |1 |
83+
|Line chart |0 |
84+
|Text only |0 |
6985

7086
# Precision feedback message generation
7187
To generate precision feedback, MPOG sends de-identified performance and preference data to a precision feedback system that processes each provider's data with their peer comparator performance data. The precision feedback system produces and evaluates candidate messages using metadata from message templates, then selects an optimal precision feedback message to return to MPOG. The precision feedback system is a knowledge-based system that is comprised of the following parts: a knowledge base, a software pipeline, and a web service.

‎vignettes/goal_gain_vignette/readme.md‎

Lines changed: 39 additions & 21 deletions
Original file line numberDiff line numberDiff line change
@@ -32,33 +32,51 @@ Gaile, also a resident at Midwest Medicine, has the following performance data o
3232
## Preference data
3333
Preferences for precision feedback are elicited through a preference survey that providers can take. The preference survey produces a preference model for each provider that, with the provider's permission, is shared with MPOG to maintain. MPOG analyses preference data that is shared to identify population-level preference segments. These segments are generated as preference profiles that can serve as a default preference model for an organization, or which can be selected by providers who do not take the preference survey, but who identify preferences that are close enough to their own in the settings menu for the precision feedback system.
3434

35-
Eugene's preference data:
35+
Eugene's preference data (Individual profile from conjoint analysis survey) :
3636

3737
|Motivating information |Utility value|Description |
3838
|-----------------------|-------------|--------------------------------------------------------------------------------|
39-
|Social Gain |10.4659 |Performance was previously low, but it has improved to reach the peer benchmark.|
40-
|Social Stayed Better |-11.3618 |Performance is consistently high (no recent change). |
41-
|Worsening |11.6598 |Performance is worsening. |
42-
|Improving |7.58489 |Performance is improving. |
43-
|Social Loss |8.46823 |Performance was previously high, but it has dropped below the peer average. |
44-
|Social Stayed Worse |-3.64987 |Performance has remained below average (no recent change). |
45-
|Social Better |-8.64669 |Performance is high this month. |
46-
|Social Worse |12.1645 |Performance is low this month. |
47-
|Social Approach |-10.6889 |Performance is improving, getting closer to the peer benchmark. |
48-
49-
Gaile's preference data:
39+
|Social better |4 |Performance is high this month. |
40+
|Social gain |5 |Performance was previously low, but it has improved to reach the peer benchmark.|
41+
|Social stayed better |-16 |Performance is consistently high (no recent change). |
42+
|Worsening |5 |Performance is worsening. |
43+
|Improving |6 |Performance is improving. |
44+
|Social loss |2 |Performance was previously high, but it has dropped below the peer average. |
45+
|Social worse |5 |Performance is low this month. |
46+
|Social stayed worse |-13 |Performance has remained below average (no recent change). |
47+
|Social approach |2 |Performance is improving, getting closer to the peer benchmark. |
48+
|Goal gain |5 |Performance was previously low, but it has improved to reach the goal. |
49+
|Goal loss |2 |Performance was previously high, but it has dropped below the goal. |
50+
|Goal approach |2 |Performance is improving, getting closer to the goal. |
51+
52+
|Display format |Utility value|
53+
|-----------------------|-------------|
54+
|Bar chart |0 |
55+
|Line chart |1 |
56+
|Text only |0 |
57+
58+
Gaile's preference data (Midwest Medicine group profile) :
5059

5160
|Motivating information |Utility value|Description |
5261
|-----------------------|-------------|--------------------------------------------------------------------------------|
53-
|Social gain |xx.xxxxx |Performance was previously low, but it has improved to reach the peer benchmark.|
54-
|Social stayed better |-12.76936 |Performance is consistently high (no recent change). |
55-
|Worsening |1.06977 |Performance is worsening. |
56-
|Improving |-0.26266 |Performance is improving. |
57-
|Social loss |2.72075 |Performance was previously high, but it has dropped below the peer average. |
58-
|Social stayed worse |9.97743 |Performance has remained below average (no recent change). |
59-
|Social better |-5.97766 |Performance is high this month. |
60-
|Social worse |-0.05277 |Performance is low this month. |
61-
|Social approach |-0.24384 |Performance is improving, getting closer to the peer benchmark. |
62+
|Social better |-2 |Performance is high this month. |
63+
|Social gain |13 |Performance was previously low, but it has improved to reach the peer benchmark.|
64+
|Social stayed better |6 |Performance is consistently high (no recent change). |
65+
|Worsening |-22 |Performance is worsening. |
66+
|Improving |3 |Performance is improving. |
67+
|Social loss |10 |Performance was previously high, but it has dropped below the peer average. |
68+
|Social worse |-15 |Performance is low this month. |
69+
|Social stayed worse |-8 |Performance has remained below average (no recent change). |
70+
|Social approach |15 |Performance is improving, getting closer to the peer benchmark. |
71+
|Goal gain |13 |Performance was previously low, but it has improved to reach the goal. |
72+
|Goal loss |10 |Performance was previously high, but it has dropped below the goal. |
73+
|Goal approach |14 |Performance is improving, getting closer to the goal. |
74+
75+
|Display format |Utility value|
76+
|-----------------------|-------------|
77+
|Bar chart |1 |
78+
|Line chart |0 |
79+
|Text only |0 |
6280

6381
# Precision feedback message generation
6482
To generate precision feedback, MPOG sends de-identified performance and preference data to a precision feedback system that processes each provider's data with their peer comparator performance data. The precision feedback system produces and evaluates candidate messages using metadata from message templates, then selects an optimal precision feedback message to return to MPOG. The precision feedback system is a knowledge-based system that is comprised of the following parts: a knowledge base, a software pipeline, and a web service.

0 commit comments

Comments
 (0)