Randomized Clinical Trial of Bidirectional
Text Messaging to Improve Adherence to
Recommended Lipid Testing
Laura-Mae Baldwin, MD, MPH
Caitlin Morrison, MPH
Nick Anderson, PhD
Kelly Edwards, PhD, MA
Jeff Green
Jonathan Griffin, MD, MHA
William Reiter, MD, FACP
Cleary Waldren, MS
A collaboration between the University of Washington Department of Family
Medicine, the Institute of Translational Health Sciences, Northwest EHR
Collaborative (Anaconda, MT), and a multidisciplinary group practice In rural MT.
Acknowledgements
• Chris Miller, Lalitha Ramesh – IT Support
• Anthony Beaver, Brenda Benefiet, Mitchell
Marrujo - Patient Contact
Acknowledgements
• This research was funded by the Office of the
National Coordinator for Health Information
Technology (Award Number 10775336) and by the
Northwest EHR Collaborative.
• The content is solely the responsibility of the authors
and does not necessarily represent the official views
of the Office of the National Coordinator for Health
Information Technology.
Background
• Increasing alliance of technology and health
care
• EHR most prevalent technology that is
increasing the ability to record and monitor
care
Background
• Patients simultaneously are adopting technologies that
may have health applications
• Mobile phone saturation is one of the most dramatic –
85% of adults have a cell phone *
–
–
–
–
–
–
95% ages 18-34
92% ages 35-46
86% ages 47-56
84% ages 57-65
68% ages 66-74
48% ages 75+
*Source: Zickuhr, Kathryn. Generations and their gadgets. Pew Internet & American Life Project, February 3, 2011,
http://pewinternet.org/Reports/2011/Generations-and-gadgets.aspxaccessed on June 5, 2013.
Study Questions
• Can patient text messaging reminders improve
adherence to recommended care?
• Will patients participate in bidirectional patientprovider text messaging?
• How do patients feel about bidirectional text
messaging with their medical providers?
Recommended Care: Lipid Testing
• Lipid testing among patients who require ongoing
monitoring:
–
–
–
–
Diabetes Mellitus
Ischemic Vascular Disease
Coronary Artery Disease
Hyperlipidemia
• PQRS and NCQA measure important to the
medical group
Organizational Structure
Clinical Care
Research
BAA
Medical
Group
NWEHR
IRB
UW
Research
UW
Informatics
Methods
Pilot randomized controlled trial
•
•
•
•
Identified study population
Randomized to intervention and control groups
Applied intervention for specified time
Compared study groups
Study Population
• 18 to 75 years old and assigned to one of two
physicians’ practices
• At least one face to face visit with a medical provider
(MD, DO, NP, PA) in the Medical Group in the prior 12
months.
• Diagnosis of Diabetes Mellitus (DM), Ischemic Vascular
Disease (IVD), Coronary Artery Disease (CAD)
• Diagnosis of Hyperlipidemia (HLD) and no DM, IVD,
CAD
Eligible for Reminder Text Message
IF:
• No LDL in the past 12 months
OR
• No LDL in the past 3 months
AND:
– The latest LDL between 3 and 12 months prior is >100 for
patients with a diagnosis of DM, IVD, or CAD.
– The latest LDL between 3 and 12 months prior is >130 for
patients with a diagnosis of HLD.
Methods
• Called eligible patients to determine whether they had a
cell phone, and were willing to accept text messaging from
the practice. At the call, recorded current cell phone
number.
• Randomized half of group to receive text message
reminders about lipid testing
• Sent up to 3 text message reminders 30 days apart
• Interviewed intervention group on experience with text
messaging after 2nd text message reminder sent
Text Message
“You are due for a cholesterol test. Reply 1
for YES, I will go to <MEDICAL GROUP
NAME>’s lab this week, open 7:30-6 M-F.
Reply 2 for NO. Fasting not needed.”
Project Timeline
• Text #1 – April 1, 2013
• Text #2 – April 29, 2013
• Research interviews conducted
• Text #3 – May 28, 2013
• End of observation period –June 25, 2013
Outcomes of Interest
1. Adherence to recommended testing
2. Patient response to text messages
3. Patient experience of text messaging as a
method for screening reminders
Enrollment Chart
248
Overdue for LDL based
on EHR data
65
112
Eligible – had cell
phone, willing to
participate
71
Not reached after 3
recruitment calls
Ineligible or declined
45
61
Eligible – and
randomized
4
Received LDL prior to
intervention
Ineligible (no cell phone,
no text plan, different
primary physician)
26
Refused
Study Groups
61
Eligible and
randomized
31
30
Randomized to text
message group
Randomized to control
group
1
2
Refused after first
text message
Duplicate phone
number –only one
text message
28
Received up to 3
text messages
3
Duplicate phone
number – only one
text message
Patient Characteristics
n = 61
% age 55 and older
49.2%
% female
32.8%
% with a Wellness Visit in the last 12
months
24.6%
Mean (SD) # ambulatory visits in
previous year
4.0 (2.8)
Chronic Disease Status
80%
72.1%
70%
60%
50%
40%
27.9%
30%
20%
10%
0%
HLD only
CAD, DM or IVD
Adherence to Recommended Testing
Persons who
came in for
LDL testing
Month
#1
%
Month
#2
%
Month
#3
%
Total
Overall
%
Interventio
n
(N=31)
7
(n=31)
22.6%
2
(n=24)
8.3%
2
(n=22)
9.1%
11
35.4%
Control
(N=30)
2
(n=30)
6.7%
4
(n=28)
14.3%
3
(n=24)
12.5%
9
30.0%
Adjusted
Relative
Risk (CIs)
4.96
(1.0811.35)
1.42,
(0.612.39)
Response Rate to Text Messages
Responded to messages
Participants (n=28)
Percent
Never responded
9
32.1%
Ever responded
19
67.9%
First only
8
28.6%
Second only
1
3.6%
Third only
2
7.1%
First and second only
4
14.3%
First and third only
2
7.1%
Second and third only
1
3.6%
All messages
1
3.6%
How Patients Responded to Text Messages
Response Type by Message
Message 1
Persons who responded
Percent of responses
N = 15
Yes
6
40.0%
No
6
40.0%
Other
3
20.0%
Message 2
N=7
Yes
3
42.9%
No
2
28.6%
Other
2
28.6%
Message 3
N=6
Yes
1
16.7%
No
5
83.3%
Other
0
0.0%
“Other” responses
• Who is this for?
• I already had the above test done prior to
seeing Dr. XXXXX in early March of this year.
• I also need to get a PSA test for {another
provider}. I will waite and coordinate into one
lab visit.
• Already had one!
• .
“Yes” / “No” with more information
• 2. I will use the state clinic for lab tests.
• 1/…
• Not sure who is requesting or if you are
meaninh / 2 i will call you
Patient Experience with Text Messaging
• Invited the 28 individuals who received text
message reminders to participate in a brief
interview about their experience
• $10 incentive for participating
• 11/28 agreed to participate
• 8/11 were reached for the interview
Patient Experience with Text Messaging
<MEDICAL GROUP NAME> is trying to find ways that help people get
the health care they need. A little over 5 weeks ago, <MEDICAL
GROUP NAME> sent you the first text message with a reminder to
get a cholesterol test. Do you remember getting this text message?
What did you think about getting a reminder like this via text
message?
•
•
7 of 8 thought text messaging was a positive experience
1 of 8 remembered receiving the message, but did not
regularly text, so did not successfully retrieve the message
Patient Comments about Text Messaging
“Text is an easy way to communicate and can also be
automated.”
“I think it’s a good idea, people make appointments
and it helps to recall when to do….It would be helpful if
there are extenuating circumstances, like mine, to
include a phone number.”
Patient Experience with Text Messaging
• Of the 7 interviewees who received the text
message:
– 3 came in for testing
• 2 noted that the second message prompted them to
come in
– The 4 who did not come had the test done with
another doctor (n=1), were waiting to test at their
annual appointment (n=1), or did not think they
were due (n=2)
Patient Experience with Text Messaging
How concerned are you, if at all, about the
privacy of your health information in text
messages?
– 5 did not want actual health information
communicated in accessible formats
– 1 had no concern
– 1 person was very concerned
– 1 person did not know
There are many ways to get health reminders
like this – phone, text message, postal mail,
email. What do you think about getting a
health reminder via text message compared to
these other ways?
Method Preferred
Participants (n=8)
Text
2
Email
2
Text or email
1
Text, email or phone
1
Mail
1
Phone
1
Main Findings
• A first text message reminder significantly increased
LDL testing among rural patients overdue for this test.
• After three months, patients with no text message
reminders nearly “caught up” with LDL testing
compared with patient who had received reminders.
• This pilot study had a small sample size, so a larger
study is needed.
Main Findings
• A majority of patients engaged in text
messaging with their provider (67.9%)
• The minority of those who text messaged back
said they would be coming in for their LDL
tests
– 40% after message 1
– 42.9% after message 2
– 16.7% after message 3
Main Findings
• Most patients had a positive experience with
text message reminders, but did not want
health information transmitted by text
message.
• Half of those interviewed rated text messaging
as a preferred method for getting health
reminders, others preferred other methods (email, phone, postal mail)
Limitations
• Excluded adults over 75 years of age
• Small sample size
• Special procedures gathered cell phone
numbers – not always available
Conclusions
• Text message reminders may result in
clinically meaningful improvements in
adherence to recommended laboratory
monitoring
Conclusions
• In this study, the greatest testing effect was
after the first text message.
• Some of those who came in for LDL testing
after the first text message may have been
tested within a few months anyway.
• Though not statistically significant, more
individuals who received text messages did
get LDL testing after 3 months.
Conclusions
• Given this study’s findings, among 1000 patients
needing an LDL test, the number tested within 3
months would be:
– 354 if reminded by text message
– 300 if not reminded
• Over 50 more patients would come in for
testing.
Conclusions
• Patients have a positive experience with text
messaging
• Patients will use text messaging to relay
additional information
Next Steps
• Testing text message reminders in a larger
study population
• Better understanding of patient concerns and
limits regarding health care applications that
use mobile technology
For more information,
contact:
Laura-Mae Baldwin
lmb@uw.edu
Caitlin Morrison
cm46@uw.edu