improving patient access and adherence to an endocrine program€¦ · pranjal singh amy cohn, phd...

Improving Patient Access and

Adherence to an Endocrine Program

Pranjal Singh

Amy Cohn, PhD

Amy Rothberg, MD

Collaborators of the Project

2

William Herman, MD3 Amy Rothberg, MD3

Student

Cara Cheshire1

1 Center for Healthcare Engineering and Patient Safety, University of Michigan, Ann Arbor

2 Department of Industrial and Operations Engineering, University of Michigan, Ann Arbor 3 Department of Internal Medicine, University of Michigan Health System

Staff

Yu-Hei Moses Chan, MSE1

Amy Cohn, PhD1,2

Background on the WMP Program

3

[1]. Rothberg et al. BMC Obesity (2015) 2:11 DOI 10.1186/s40608-015-0041-9

• The University of Michigan Weight Management Program (WMP) is a two-year program

o Intense energy restriction o Behavioral change

• 26 total visits required over two years • Patients must attend ≥80% of scheduled appointments

• First month of program requires more frequent visits

0 6 12 18 24 months

Problem Statement

4

Patients waiting too long to get into program Access

Patients are not able to adhere to the program’s structured visit timeline

Adherence

Approaching the Problem

5

• We designed a temporal database to analyze:

o Capacity Dynamics o Patient Behavior

• The stages of the database are: 1. Slot 2. Appointment Opportunity 3. Provider Template 4. Appointment Schedule 5. Appointment Schedule Snapshots

Slot

• Basic building block of the database

• Represents single 15 minute time period

• Corresponds to a single record in the database

• Defined by:

o Provider Name

o Slot Date

o Slot Time

6

Slot Provider “A” 7/17/2014 9:15 AM

Appointment Opportunity

• Appointment Opportunity represents single placeholder for a patient appointment

• Appointment Opportunities are created expecting:

o Return Visit (Single Slot)

o New Patient (Multiple Slots)

• Defined by:

o Appointment Length

o Appointment Type

o Slot Number in Appointment

o Total Number of Slots in Appointment

7

8:00

8:00

8:15

8:30

Slot Appointment Opportunity

1/1

1/3

2/3

3/3

Appointment Opportunity

OR

Slot

Provider Template

• Represents all possible Appointment Opportunities over a given timeframe

• This template is an aggregated schedule of each provider’s general availability to see patients

• This excludes information about intermittent unavailability, e.g.

o Out of office for conference

o Vacation leave

o Administrative duty

8

9/14/2015 9/15/2015 9/16/2015

Appointment Schedule

• Represents Provider Template overlaid with:

o Each provider’s intermittent unavailability

o Patient appointment data

• Some Appointment Opportunities are occupied by patient appointments

9

9/14/2015 9/15/2015 9/16/2015

Appointment Schedule Snapshots

• On each day (M-F), we view the Appointment Schedule

• Each day’s view represents an Appointment Schedule Snapshot

• We compare two consecutive Appointment Schedule Snapshots and capture changes from one to another, e.g.

10

What can we learn from the database?

11

From 7/15/2014 – 4/29/2016: New Patients • 25% of scheduled appointments were cancelled • Of cancelled appointments, 90% were rescheduled Return Visits • 34% of scheduled appointments were cancelled • Of cancelled appointments, 76% were rescheduled

Major Questions for Analysis

12

Capacity Dynamics

Patient

Behavior

1. How long is an appointment held before it gets cancelled?

1. How easy is it to schedule an appointment “X” weeks into the future?

2. How far into the future are “X” slots available for scheduling?

How easy is it to schedule an appointment “X” weeks into

the future?

13

How far into the future are “X” slots available for

scheduling?

14

0

5

10

15

20

25

30

35

40

45

0 5 10 15 20 25

Day

s

Number of Empty Slots

How far into the future are “X” slots available for scheduling? (New Patients)

Mean

Median

Analysis (cont’d)

15

How long are people holding their appointments before they cancel? Appointment Create

Date Appointment Cancellation

Date Appointment Date

Appointment Holding Time

Appointment Lifetime

Ratio = Appointment Holding Time

Appointment Lifetime

16

N = 56

Analysis (cont’d)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0

2

4

6

8

10

12

14

16

Freq

uen

cy

Appointment Holding Time Ratio

Appointment Lifetime ≤ 1 Week

Analysis (cont’d)

17

N = 765

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0

50

100

150

200

250

Fre

qu

en

cy

Appointment Holding Ratio

1 Week < Appointment Lifetime ≤ 1 Month

Analysis (cont’d)

18

N = 2522

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0

100

200

300

400

500

600

700

800

900

Fre

qu

en

cy

Appointment Holding Ratio

1 Month < Appointment Lifetime ≤ 3 Months

Analysis (cont’d)

19

N = 1306

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0

100

200

300

400

500

600

Fre

qu

en

cy

Appointment Holding Ratio

3 Months < Appointment Lifetime ≤ 6 Months

Analysis (cont’d)

20

N = 541

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0

50

100

150

200

250

300

Fre

qu

en

cy

Appointment Holding Ratio

Appointment Lifetime > 6 Months

Analysis Summary

21

• Difficult to see new patients within the first month of the clinic’s schedule o Variability reduced in long term

• Utilization impacted by appointment holding times

Ongoing Work

22

• Look at the impact of seasonality on utilization metrics

• Track utilization of a particular slot over a specified duration of time o How is that slot changing?

• Look at priority of waitlists

Acknowledgements

23

• The Seth Bonder Foundation • The Metabolism, Endocrinology & Diabetes Clinic • Center for Healthcare Engineering and Patient Safety • University of Michigan College of Engineering • University of Michigan Medical School

24

Thank you!

Pranjal Singh pranjsin@umich.edu

Amy Cohn amycohn@umich.edu

Amy Rothberg arothber@umich.edu