Can Messages Make a Difference The Association Between E Mail Messages and Health Outcomes in Diabetes Patients

Human Communication Research ISSN 0360-3989

ORIGINAL ARTICLE

Can Messages Make a Difference? The
Association Between E-Mail Messages and
Health Outcomes in Diabetes Patients
Jeanine Warisse Turner1 , James D. Robinson2 , Yan Tian3 , Alan Neustadtl4 ,
Pam Angelus5 , Marie Russell6 , Seong K. Mun5 , & Betty Levine5
1 Communication, Culture and Technology Program, Georgetown University, Washington, DC 20057, USA
2 Department of Communication, University of Dayton, OH 45469, USA
3 Department of Communication, University of Missouri–St. Louis, MO 63121, USA
4 Department of Sociology, University of Maryland, College Park, MD 20742, USA
5 Imaging Science and Information Systems Center, Georgetown University, Washington, DC 20057, USA
6 Phoenix Indian Medical Center, Indian Health Service, Phoenix, AZ 85004, USA

This investigation examined the impact of social support messages on patient health
outcomes. Forty-one American Indian, Alaska Native, and Native Hawaiian patients
received a total of 618 e-mail messages from their healthcare provider (HCP). The e-mail
messages were divided into 3,565 message units and coded for instances of emotional social
support. Patient glycosulated hemoglobin scores (HbA1c) showed significantly improved

glycemic control and emotional social support messages were associated with significant
decreases in HbA1c values. Patient involvement with the system, measured by system login
frequency and the frequency of uploaded blood glucose scores to the HCP, did not predict
change in HbA1c.
doi:10.1111/j.1468-2958.2012.01437.x

Web-based monitoring and messaging systems allow healthcare providers (HCPs)
to monitor closely changes in patient health, can be implemented at relatively low
cost, can reduce healthcare costs, and offer both patients and HCPs convenience
(Field & Grigsby, 2002; Krishna, Boren, & Balaas, 2009). These telemedicine systems
employ ‘‘store and forward technologies’’ and have been successfully used to monitor
blood pressure and hypertension (Green et al., 2008), gestational diabetes (PerezGerre et al., 2010), and diabetes (Levine, Turner, Robinson, Angelus, & Hu, 2009;
McMahon et al., 2005; Smith et al., 2004).
Some online monitoring systems provide patients with e-mail alerts about
appointments, annual exams, and/or reminders to check blood glucose levels. These
simple reminder systems have been shown to work, but their effectiveness wanes over

Corresponding author: Jeanine Warisse Turner; e-mail: TURNERJW@georgetown.edu
Seong K. Mun is presently at the College of Science, Virginia Tech University.
252

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

time (Hanauer, Wentzell, Laffel, & Laffel, 2009; Mollon et al., 2008). Other online
monitoring systems allow patients and providers to interact via e-mail. Given 96%
of adult Internet users use e-mail (Purcell, 2011), it is no surprise that adults with
Internet access want the opportunity to communicate with their HCPs via e-mail
(Liederman, Zimmerman, Athanasoulis, & Young, 2004). Although e-mail messages
are becoming a more accepted way to interact with HCPs, the impact of those e-mail
messages on patient health needs closer examination.
E-mail communication between practitioners and patients

A growing body of research has examined the viability of e-mail as a communication
vehicle (Bergmo & Wangberg, 2007; Delbanco & Sands, 2004; Patt, Houston, Sands, &
Ford, 2003; Roter, Larson, Sands, Ford, & Houston, 2008; White, Moyer, Stern, & Katz,
2004) and found that e-mail can augment face-to-face interaction. Research suggests

that patients find e-mail less intimidating than face-to-face interaction because it
allows time to reflect and compose questions about their health or treatment regimen
in advance (Borowitz & Wyatt, 1998). A study by Rosen and Kwoh (2007) suggested
pediatricians spend less time answering medical questions using e-mail than they
spent answering questions over the telephone. In addition, the 300 patients surveyed
reported being satisfied with e-mail access to their physician and they were satisfied
with the content of the e-mail sent by their pediatrician (Rosen & Kwoh, 2007).
Whereas some scholars (Baur, 2000) have cautioned that the lack of nonverbal cues
available through e-mail could reduce the effectiveness of patient–HCP interaction,
Roter et al. (2008) suggested e-mail mimics communication within traditional
doctor and patient dialog in that it contains both task and relational content. Roter
et al. contended that e-mail provides opportunities for patients to communicate
their problems and worries while allowing HCPs the opportunity to provide timely
patient-specific responses.
Online telemedicine systems can be used for a variety of things including serving
as a portal that provides patient access to their HCP. Researchers have suggested
online telemedicine systems can be used to encourage interaction and ideally enhance
the patient–HCP relationship (Kassirer, 1995; Mun & Turner, 1999). Until recently,
however, few studies have explored the integration of secure e-mail messaging within
a system of care to manage a specific disease.

One such study by Levine et al. (2009) examined the relationship between
blood glucose monitoring and online messaging with their HCP. They found that
the number of times a patient transferred blood glucose readings to a web-based
disease management system was correlated with the (a) number of messages sent
to HCP, (b) total number of messages received from HCP, and (c) number of
patient-centered messages received from HCP. This study supported the notion that
interaction via web-based disease management systems should be encouraged, given
the relationship between interaction and monitoring of blood glucose in patients
suffering from diabetes mellitus.
Human Communication Research 39 (2013) 252–268  2013 International Communication Association

253

Can Messages Make a Difference?

J. W. Turner et al.

In an effort to understand further the relationship between e-mail messages and
patient use of online health monitoring systems, Robinson, Turner, Levine, and
Tian (2011) analyzed e-mail messages sent by HCPs to their patients. These 618

messages were coded into 3,565 discrete message units, and then the message units
were coded into nine message types. The categories and the percentage of messages
for each category type were phatic (32.2%), informational social support (21.1%),
requests for health information (13%), social integration (9.0%), emotional social
support (7.2%), esteem social support (6.7%), tangible social support (6.3%), and
self-disclosure (< 1.0%). The remaining 3.5% of the messages focused on technical
support for the online system were omitted from this analysis.
Robinson et al. (2011) extended the work of Levine et al. (2009) by examining
the relationship between patient monitoring of blood glucose, online system usage,
and the types of messages they received. Specifically, they found blood glucose
monitoring and online system usage was related to substantial social support
messages (which included phatic messages, informational social support messages,
tangible social support messages, and messages containing references to previous
contact or interaction between the patient and the provider), as well as relational
affect messages (which included esteem social support messages and self-disclosure).
From this research, different types of messages appear to encourage or produce
different levels of patient blood glucose monitoring and online health system usage.
Previous research suggests social support is related to improved patient resistance
to infection and disease, longevity, patient mortality, self-efficacy, social functioning,
and enhanced psychological adjustment (Berkman & Syme, 1979; Blazer, 1982;

Bloom, 1982; Dimond, 1979; Hanson & Sauer, 1985; House, Robbins, & Metzner,
1982; Jemmott & Magloire, 1988; Litwak & Messeri, 1989; Major et al., 1990; OrthGomer & Johnson, 1987; Schoenbach, Kaplan, Freedman, & Kleinbau, 1986; Seeman
et al., 1993; Welin et al., 1985). In a review of 81 social support studies, Uchino,
Cacioppo, and Kiecolt-Glaser (1996) concluded that there were links between social
support and beneficial effects on the cardiovascular, endocrine, and the immune
systems of patients. This conclusion was reached even though 70% of these studies (57
of 81) focused on the cardiovascular system, more than half were correlational designs,
and the majority used only systolic and diastolic heart rate as their physiological
health outcome measure. Further these studies relied on self-report data that asked
about their network of friends, coworkers, kin, and/or social activity to measure
social support. None of the experimental, quasi-experimental, or prospective studies
reviewed by Uchino et al. employed actual or enacted social support messages—a
problem identified by Goldsmith (2004) that plagues this body of research.
In addition, only a small number of these studies reviewed by Uchino et al. (1996)
are experimental, quasi-experimental, or prospective in design. These studies focused
on single message effects on immediate stressor responses. Like their cross-sectional
brethren, these studies said little about the impact of actual social support messages
on long-term patient health and even less about the impact of messages between a
patient and their HCP and poorly controlled diabetes mellitus.
254

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

Types of social support

There is no consensus among researchers about the number and types of social
support but there is agreement that disaggregating the different types is difficult
(Berkman, Glass, Brissette, & Seeman, 2000). For example, Cutrona and Russell
(1990) suggested there are six types of social support. Sherbourne and Stewart (1991)
contended that there are only four types, and Westaway, Seager, Rheeder, and Van
Zyl (2005) argued for two dimensions of social support. Part of this controversy
stems from the fact that the dimensions of social support are so highly correlated
that they are nearly indistinguishable (Norbeck, Lindsey, & Carrieri, 1981; Robinson
et al., 2011; Semmer et al., 2008; Westaway et al., 2005).
Differences in types of measures further confound the issue. Different researchers
use either structural measures or functional measures of social support. Structural measures ask about the social network of the individual, only provide an

indirect measure of the availability of social support functions, and are only modestly related to functional measures of social support (Cohen & Willis, 1985;
Sherbourne & Stewart, 1991). Functional measures of social support rely heavily on self-report and the memory of the individual. This is problematic because
individuals in a single close relationship believe they receive more social support
than individuals who report having multiple superficial relationships (Cohen &
Willis, 1985).
Westaway et al. (2005) pointed out that inadequate socioemotional and tangible
support has been shown to be related to poorer functioning, general health, and
wellbeing (Littlefield, Rodin, Murray & Craven, 1990; Pouwer et al., 2003); higher
hospital admission rates (Kelly, Mahmood, Kelly, Turner, & Elliott, 1993); poorer
diabetes control, increased complications, and increased mortality (Edelstein & Linn,
1985; Robinson, Lloyd, & Stevens, 1998; Schwartz, Springer, Flaherty, & Kiani,
1986; Toth & James, 1992; Wang & Fiske, 1996; Wilson et al., 1986). Because this
research focuses on e-mail messages, the ability to provide tangible social support
via e-mail is limited. In previous efforts to code these e-mail messages (Robinson
et al., 2011), we found most of the messages contained references to tangible social
support rather than actual social support. Consequently, this investigation focuses on
enacted emotional social support messages (Robinson et al., 2011) and the following
hypothesis is posited:
H1: Enacted emotional social support messages will be directly related to improved
patient health outcomes.

Given our interest in demonstrating that enacted social support messages transmitted via e-mail are an effective tool for improving clinical outcomes, a second
goal for his study was to determine the relative impact of the various types of social
support on patient health outcomes. Further, we wanted to examine the impact
of system usage variables on patient health outcomes. Specifically, we examined
telemedicine system usage and patient blood glucose monitoring (or the uploading
of blood glucose levels) and enacted social support messages on patient health.
Previous research showed a relationship between blood glucose monitoring and
Human Communication Research 39 (2013) 252–268  2013 International Communication Association

255

Can Messages Make a Difference?

J. W. Turner et al.

the reception of personal e-mail messages (Robinson et al., 2011), so the following
hypothesis is forwarded:
H2: Enacted emotional social support messages will be a better predictor of improved
patient health outcomes than patient use of an online monitoring system.

Methods
Research design

A nonrandomized prospective study of 41 patients with poorly controlled diabetes was undertaken within four disparate geographical communities of American
Indian/Alaska Native and Native Hawaiian patients. Patients were eligible for this
study if they were over 18 years of age, diagnosed with type 1 or type 2 diabetes
mellitus (T1DM or T2DM), used a standard glucose meter regularly, had an HbA1c
level above 7 within 3 months prior to enrollment in the study, and were amenable
to using a computer. The medical protocol was designed such that patients were
enrolled in the program for 6 months and then given an exit HbA1c blood glucose
test to determine the program’s effectiveness.
Patients enrolled in the program were given access to a web-based diabetes management and messaging system called the MyCareTeam system. The MyCareTeam
system contains culturally appropriate diabetes education information and a software
program that helps users track their diet, exercise, and blood glucose levels. Additionally, patients were provided with a cable to connect their glucose meter to a computer
or to a modem that allowed them to transmit their blood glucose readings directly to
the remote secured database without a personal computer. Patients were instructed
on how to transmit their blood glucose readings, how to access the blood glucose
readings and other information via MyCareTeam using an Internet browser, and how
to interpret the results. Patients continued checking their blood glucose as directed

by their physician and uploading the readings at least once every 2 weeks or more
frequently if they chose. A complete description of the design of the MyCareTeam
web-based monitoring and messaging system has been reported elsewhere (Levine
et al., 2009; McMahon et al., 2005; Robinson et al., 2011).
Healthcare provider training

HCPs from each site were trained on the system and, in turn, trained the patients
on system usage. Depending on the particular clinic, a HCP could be a physician,
a nurse, a physician’s assistant, study or clinic coordinator, or diabetes educator.
Nearly all the e-mail contact was provided by HCPs that were not physicians.
These HCPs checked for e-mail and blood glucose uploads from their patients
daily. In addition, the HCPs used MyCareTeam to review their patients’ medical
information, to reply to their patients’ messages, to provide encouragement, and to
dispense medical advice. Additionally, although HCPs were trained in the use of the
telemedicine system they were neither sensitized nor trained on the type of messages
to use.
256

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

Finally, the staff members were not encouraged to send or reply to patient e-mail
frequently. As a baseline study, the goal was to see what would happen without the
introduction of staff motivation as an additional factor and potential confounding
factor within this study.
Participant population

A total of 140 patients were invited to participate in this study. These patients
were recruited from health clinics at three American Indian communities and two
Native Hawaiian health clinics. Human subjects’ approval was received from the
Indian Health Service Institutional Review Board (IRB), Native Hawaiian Health
IRB, Georgetown University IRB, and the U.S. Army Office of Research Protections before patients were recruited or enrolled. Of the 140 patients invited to
participate, a total of 41 patients with diabetes agreed to participate and completed
the study.
The remaining patients agreed to participate in the program but did not have
both baseline and follow-up HbA1c scores or were missing information on their
messaging. In some cases, the data were not entered into the database by their
HCP and, in other cases, the patient simply did not come in for their scheduled HbA1c test. Unfortunately, gaining HCP compliance in a complex data
gathering activity is difficult and further exacerbated by patients who are under
no obligation to have their follow-up HbA1c tests on time. These missing data
are just that—data missing because of the exigencies associated with gaining
compliance.
There was no mass exodus of patients and most of the patients sent and received email up until the end of the program—even if they did not complete the final HbA1c
test. A total of 25 patients (61%) were female and 33 (80%) were American Indians,
Alaskan Natives, or Aleuts. Patients’, and the average, age was 53 at enrollment
(Min = 30, Max = 76, SD = 11). Seventeen patients (42%) were from Idaho, 14
(34%) from Arizona, and 10 (24%) from Alabama.
Message sample

HCPs sent a total of 618 messages to the 41 patients between October 2005 and May
of 2008. All these messages were sent to a specific patient by the HCP and divided
into individual sentences. Compound sentences were divided into two sentence
fragments. For example, ‘‘Glad to hear you ran 2 miles, but I still need you to upload
your blood glucose scores’’ would be divided into two sentences—‘‘Glad to hear
you ran 2 miles’’ and ‘‘I still need you to upload your blood glucose.’’ This process
yielded 3,565 message units that were initially coded into the nine message types or
categories described elsewhere (Robinson et al., 2011).
Results from the post hoc two sample t-test indicated that the 41patients that
completed both the pre- and posttest HbA1c did not differ from those 80 patients
that did not complete the two HbA1c tests in terms of the number of messages
they received containing social support (t = −.57, df = 138, ns). In fact, 74% of the
Human Communication Research 39 (2013) 252–268  2013 International Communication Association

257

Can Messages Make a Difference?

J. W. Turner et al.

respondents (n = 80) missing their HbA1c posttest continued uploading their blood
glucose scores throughout the 6-month period.
Coding procedure

The 3,565 message units were initially coded into nine categories: Emotional, esteem,
integrational, informational, and tangible social support, as well as technical support,
phatic communication, health information requests, self-disclosure. Three paid
research assistants underwent a series of three training sessions for a total of 4 hours.
Upon completion of the training session, each coder was asked to content analyze 200
messages to be coded independently. For this investigation, only emotional support
messages were included in the analysis. Cohen’s κ indicated acceptable intercoder
agreement for the emotional social support variable (κ = .846) (Cohen, 1960).
Any message sent by a HCP to a patient that communicated caring or concern
for the patient was coded as emotional social support. This included messages such
as ‘‘It is really hard for people to stay on their diet,’’ ‘‘You are doing great,’’ ‘‘I am
concerned with your blood sugar readings,’’ ‘‘Like the weight loss!!!,’’ and ‘‘You are
one of our best patients.’’ In addition, if the HCP made a general statement such
as ‘‘Call if you need anything’’ that was also coded as emotional social support.
Statements that were specific about the type of help provided were coded as tangible
social support or informational social support. The key here is the HCP’s willingness
to help—not the provision of specific medical information or task assistance.
Measurement

The dependent variable health outcome was measured using quarterly blood tests
known as glycosulated hemoglobin or HbA1c. An HbA1c test measures the long-term
glycemic control of individuals with diabetes. As a point of reference, an HbA1c of
6% is considered normal; a value above 6.5% may indicate a diagnosis of diabetes;
and a value above 7% may mean that one’s diabetes control is not as good as it
should be (NIH, 2012). For each participant, HbA1c was measured at baseline and at
a 4- to 6-month follow-up interval. Changes in HbA1c were calculated as follow-up
value minus baseline value (Min = −5.1, Max = 1.7, M = −0.77, SD = 1.8). Change
in HbA1c has been used as an outcome measure to support improvement in a patient
with diabetes (Glasgow, Boles, Kay, Feil & Barrera, 2003).
Emotional support was measured as the proportion of messages containing
emotionally supportive messages sent from HCPs to a patient during a 6-month
period (M = 0.29, SD = 0.33).
Patient system usage was measured by the number of times each patient logged
into MyCareTeam system and by the number of blood glucose scores each patient
uploaded to their HCP through MyCareTeam. These measures were recorded as 0
(no use), 1 (low use), 2 (high use for one month of login and upload activities because
some patients had problems initially that led to multiple login efforts and because
patients became involved in the program at different times). These discrete orderable
measures better reflect actual patient system usage.
258

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

Table 1 HbA1c Results at Baseline, Follow-Up, and Change
95% CI for M
HbA1c

N

M

SD

SE

Lower

Upper

Baseline
Follow-up
Change

41
41
41

8.9
8.1
−0.8

1.6
1.4
1.8

0.3
0.2
0.3

8.4
7.7
−1.3

9.4
8.6
−0.2

Results

Our primary outcome variable was HbA1c measured at baseline (when the patient
enrolled in the study) and at follow-up (approximately 4–6 months later). Table 1
shows the average HbA1c values at baseline, follow-up, and the change from baseline
to follow-up. Patient HbA1c values at study entry ranged from 6.5 to 13.4; and
follow-up values ranged from 5.4 to 11.3. The decrease in HbA1c was statistically
significant (t = −2.7, p = .009). Figure 1 shows the margins plot of the relationship
between change in HbA1c and support messages controlling for system usage. This
figure shows the importance of support messages on glycemic control. Other studies
have shown that people who used the MyCareTeam application the most, had
better glycemic control than infrequent users (McMahon et al., 2005; Smith et al.,
2004). These results indicated the general improvement of glycemic control over the
study period.
Table 2 shows the results of the change in HbA1c regressed on the emotional
support measure with uploads and logins as control measures. The results indicated

Figure 1 The relationship between emotional support messages and change in HbA1c.

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

259

Can Messages Make a Difference?

J. W. Turner et al.

Table 2 Regression Results for System Usage and Emotional Support
Variables

Coefficient

Monthly logins
Low
High
Monthly uploads
Low
High
Emotional support
Intercept

0.03
0.56
−1.18
−2.36
−0.03a
1.50

Note: R2 = .34
a
p = .0013, statistically significant α < .01

the importance of message content relative to MyCareTeam usage. For example, an
increase in the percentage of emotionally supportive messages of one is associated
with an average change in HbA1c of approximately −.03 (p = .0013). None of the
system usage measures were statistically significant, although the upload measures
were in the predicted direction.
H1 stated ‘‘Enacted emotional social support messages will be directly related to
improved patient health outcomes’’ and was supported. A percentage point increase
in emotionally supportive e-mail messages was associated with a change in HbA1c
of −.03 (p = .0013). That is, a larger percentage of social support messages tended to
be associated with greater average decreases in HbA1c controlling for system usage,
a desirable outcome. Presenting our results in another way, the percentage of e-mail
messages containing emotionally supportive content was 29% with an associated
adjusted change in HbA1c of −.26. If this average were increased by 25 percentage
points, the adjusted change in HbA1c would be approximately −.92.
Discussion

In research focusing on HCP—patient messages Smith et al. (2004) found that
patient involvement with a web-based diabetes management system resulted in
increased levels of interaction. Although specific messages were not analyzed in that
investigation, Smith et al. found system usage predicted improved patient glycemic
control. Previous research has also connected patient uploads of blood glucose
readings with messages sent and received from the HCP (Smith et al., 2004) and
specific message types (Robinson et al., 2011).
The current investigation is consistent with and extends this understanding of
the effectiveness of online monitoring systems by demonstrating that reception
of emotional social support messages is predictive of improved HbA1c levels and
ultimately patient health. These findings are particularly important because other
predictors such as system usage and frequency of patient uploading of BG scores were
260

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

not successful predictors of HbA1c change. This finding suggests that the messages
themselves and not merely involvement with the system may play an important role
in improving patient health.
Research employing short message service (SMS) technology (Hanauer et al.,
2009) suggested text message and e-mail reminders are effective but only in the short
term. Interaction via e-mail can be much more than a reminder. In addition to being
able to monitor their patients’ health more closely, the use of e-mail affords the HCP
the opportunity to interact more frequently, provide social support, and develop
relationships. This investigation suggests that the provision of social support may be
associated with improvements in patient health and can increase the likelihood of
patient health improvement. Exactly why emotional social support predicted HbA1c
improvements is not clear. It may be as Cutrona and Russell (1990) suggested that the
messages represented an optimal match between patient needs and the type of social
support being provided. It may also be that the interaction via e-mail represents an
opportunity of relationship development. Our next line of inquiry will focus on the
impact of the frequency of messages sent by patients’ and how this impact influences
the process. In addition we are examining patient levels of satisfaction with their
interaction.
Limitations of the present study

Although this study supports the notion that socially supportive e-mail messages
contribute to improved patient health, the fact that the study employed a volunteer sample of American Indian and Native Hawaiians was a limitation to be
addressed in future research. Similarly, the coding of the messages by trained coders
does not necessarily mean that patients perceived the e-mail messages as being
socially supportive. Determining what the patients believe they received from the
e-mail messages would be an interesting project for future researchers. Similarly,
the study was limited by the fact that patient message types were not examined.
Some types of messages may be the result of patient inquiries or health behaviors such as uploading of blood glucose readings. Again future research needs to
consider the relationship between sending messages to the HCP and patient health
improvements.
Additionally, although our findings began to demonstrate the effectiveness of
sending emotional social support messages within an online monitoring system to
improve management of diabetes, more studies are needed to determine whether
similar supportive messages can be effective in other types of web-based systems as
well as within the context of doctor and patient e-mail exchange. These studies are
needed for a variety of different reasons—not the least of which is one significant
limitation of this investigation. The use of a volunteer sample was problematic and
further exacerbated by the fact that data collection was impacted by our access
to patient records. As suggested earlier, the HCPs in charge of entering patient
data—in particular patient health records including HbA1c data—were not always
Human Communication Research 39 (2013) 252–268  2013 International Communication Association

261

Can Messages Make a Difference?

J. W. Turner et al.

willing or able to enter the patient records in a timely manner. Additionally, patients
did not always return to get their HbA1c tested on time. Changes in staff further
complicated the situation as the clinicians agreeing to participate in the study
changed and, once the study was ‘‘over,’’ we encountered difficulty in encouraging
new staff members to go through patient records for a program that was no longer
online.
Future research needs tighter controls on data entry and respondent participation.
Additionally, we would have found it helpful to establish a control group of patients
from each site—who did not have access to the telemedicine system—that we could
use for comparative purposes. However, because of the complication of the multisite
study and the need to recruit patients into the study, we were not able to collect
these data. The clinics maintained a relatively small patient census, and we were
advised at the outset of the research that because the population was wary of research,
a randomized study would be perceived poorly by the patient population. Finally,
many variables contribute to improvement in HbA1c, including improvement in
habits related to nutrition, exercise, changes in drug therapies, or adjustments of
life stressors. We did not specifically examine these variables and therefore do not
know the extent to which they might have contributed to the improvement in
HbA1c levels.
Implications for future research and practice

This investigation has several implications for both research and practice. First, it
explored the connection between content of e-mail messages and health outcomes.
Further, it showed that the value of online monitoring systems can extend (and possibly should) beyond simple reminder systems, opportunities for patient education,
or patient accountability. Recently, researchers have focused on the use of cell phones
for educational or informational interventions to improve health outcomes. In a
recent review, Krishna et al. (2009) found that 12 of 13 studies reported significant
improvements in clinical outcomes as a result of voice or text messages sent to a cell
phone. Of the 12 studies, 9 involved effective diabetes control and management but
none examined actual messages in their efforts to better understand how they might
contribute to outcome improvement.
Unfortunately, this research into Interactive Behavior Change Technology (IBCT)
has not been particularly successful in the long-term because it focuses on the
technology and not the functions that the technology serves (Piette, 2007). In a call to
reconceptualize the way we understand healthcare access in the 21st century, Fortney,
Burgess, Bosworth, Booth, and Kaboli (2011) suggested that digital communications
between patients and their practitioners would have the ability to ‘‘drastically improve
access’’ (Fortney et al., 2011, p. 645).
Our findings appear to support this call and that the content of the messages
themselves can affect patient outcomes. We believe that explicit guidance regarding
the types of messages most effective with patients will be critical to developing an
262

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

effective successful online environment as the walls of the doctor’s offices expand into
virtual spaces. As more online chronic conditions monitoring systems are developed,
HCPs need to be guided not only in the use of the system for assessing care, but also
in the role that messages can play in motivating patients to take a more active role in
managing their health.
Additionally, research revealed an association between perceived social support
and positive health outcomes (Barrera, Glasgow, McKay, Boles & Feil, 2002; Glasgow
et al., 2003; for meta-analysis see Rains & Young, 2009; Uchino et al., 1996).
However, these studies used a survey methodology to connect perceptions of social
support as revealed by a self-report measure to a specific measured health outcome.
The present research examined the relationship between enacted emotional social
support messages and health outcomes. It is the first study to connect HbA1c and
emotional support messages sent by HCPs to their patients via e-mail.
In a 2011 review of online vendor markets of diabetes applications, researchers
found that the most prevalent features in the 137 applications surveyed included:
insulin and medication recording, data export and communication, diet recording,
and weight management. In the same study, a literature search of studies of mobile
health applications found similar features (Chomutare, Fernandez-Luque, Arasand,
& Hartvigsen, 2011). These applications suggested that the connection between
messages and outcomes will continue to be possible as these types of technologies are
adopted.
Fortney et al. (2011) argued that the paradigm for healthcare delivery is evolving
with greater reliance on nonencounter-based digital communication between patients
and their healthcare teams, creating new opportunities to understand and measure
access to care. The changing healthcare environment provides new opportunities
for access to actual messages and connecting these messages to specific healthcare
outcomes. It also provides researchers the opportunities to address frameworks and
theories based on a face-to-face encounter with a HCP and find out how they can be
applied or adjusted to help understand digital encounters.
Similarly, social media technologies are providing new avenues for patients to
engage with HCPs and for those healthcare systems to use these tools to monitor
conversations and respond to patients (Thielst, 2011). Social media potentially
provides an additional tool for connecting actual messages to health outcomes.
The present research took an important step by identifying a literature of social
support that has indicated a connection to positive health outcomes and connecting
specific messages to those outcomes. In doing so, it built on past research as well
as provided an example of a new opportunity provided by digital environments
that capture messages and healthcare outcome information to communication
scholars.
Although this research suggested that supportive communication can enhance a
doctor and patient encounter in an online setting, considering the alternative may be
important. Specifically, in what settings might patients become stressed or bothered
by additional input from their HCP? Research has found that advice can foster
Human Communication Research 39 (2013) 252–268  2013 International Communication Association

263

Can Messages Make a Difference?

J. W. Turner et al.

feelings of gratitude and perceptions of caring but can also serve as a reminder of
a power imbalance or be perceived as intrusive (Goldsmith & Fitch, 1997). Future
research could explore the context within which patients receive e-mail messages
from a HCP to better understand the reception of these messages.
For example, receiving messages in an online system where a patient must log in
to communicate specifically with the HCP (similar to the system in this study) may
provide more agency to the patient than receiving a message within an e-mail inbox
that could interrupt the flow of other aspects of a patient’s daily life. Additionally,
social support can be complicated when the patient is not ready to receive specific
advice or is interested in preserving uncertainty (Brashers, Neidig, & Goldsmith,
2009). In this way, giving patients more agencies in the way they receive this support
may be important.
Therefore, as HCPs and their support staff begin communicating with patients in a
virtual environment, they should incorporate social support to develop relationships
with their patients. Some patients will undoubtedly respond to emotional and esteem
social support, whereas others may respond to other types of social support (Cutrona
& Russell, 1990). HCPs need to incorporate all types of social support into their
e-mail contact with patients. To do so may enhance patient adherence as well as
improve patient health.
Acknowledgments

Grant support from this research came from the U.S. Army Medical Research
and Materiel Command, award number W81XWH0420002. We thank the initial
reviewers of this article as well as the past and current editor of the journal for their
help and suggestions.
References
Barrera, M., Glasgow, R., McKay, H. G., Boles, S., & Feil, E. (2002). Do internet-based
support interventions change perceptions of social support?: An experimental trial of
approaches for supporting diabetes self-management. American Journal of Community
Psychology, 30(5), 637–654.
Baur, C. (2000). Limiting factors on the transformative powers of email in patient-physician
relationships: A critical analysis. Health Communication, 12, 239–259.
Bergmo, T. S., & Wangberg, S. C. (2007). Patients’ willingness to pay for electronic
communication with their general practitioner. The European Journal of Health
Economics, 8(2), 105–110.
Berkman, L. F., & Syme, S. L. (1979). Social networks, host resistance, and mortality: A nine
year follow up of Alameda County residents. American Journal of Epidemiology, 109,
186–204.
Berkman, L. F., Glass, T., Brissette, I., & Seeman, T. E. (2000). From social integration to
health: Durkheim in the new millennium. Social Science & Medicine, 51, 843–857.
Blazer, D. (1982). Social support and mortality in an elderly community population.
American Journal of Epidemiology, 115, 684–694.
264

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

Bloom, J. R. (1982). Social support, accommodation to stress, and adjustment to breast
cancer. Social Science & Medicine, 16, 1329–1338.
Borowitz, S., & Wyatt, J. (1998). The origin, content, and workload of e-mail consultations.
Journal of the American Medical Association, 280(15), 1321–1324.
Brashers, D., Neidig, J., & Goldsmith, D. (2009). Social support and the management of
uncertainty for people living with HIV or AIDS. Health Communication, 16,
305–331.
Chomutare, T., Fernandez-Luque, L., Arasand, E., & Hartvigsen, G. (2011). Features of
mobile diabetes applications: Review of the literature and analysis of current applications
compared against evidence-based guidelines. Journal of Medical Internet Research, 13(3),
e65. doi: 10.2196/jmir.1874.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological
Measurement, 20, 37–46.
Cohen, S., & Willis, T. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98, 310–357.
Cutrona, C., & Russell, D. (1990). Type of social support and specific stress: Toward a theory
of optimal matching. In B. Sarason, I. Sarason, & G. Pierce (Eds.), Social support: An
interactional view. New York, NY: John Wiley & Sons.
Delbanco, T. L., & Sands, D. (2004). Electrons in flight−e-mail between doctors and patients.
The New England Journal of Medicine, 350, 1705–1707.
Dimond, M. (1979). Social support and adaptation to chronic illness: The case of
maintenance hemodialysis. Research in Nursing and Health, 2, 101–108.
Edelstein, J., & Linn, M. W. (1985). The influence of the family on control of diabetes. Social
Science & Medicine, 21, 541–544.
Field, M., & Grigsby, J. (2002). Telemedicine and remote patient monitoring. Journal of the
American Medical Association, 288(4), 423–425.
Fortney, J. C., Burgess, J. F., Bosworth, H. B., Booth, B. M., & Kaboli, P. J. (2011). A
reconceptualization of access for 21st century healthcare. Journal of General Internal
Medicine, 26, 639–647.
Green, B., Cook, A., Ralston, J., Fishman, P., Catz, S., Carlson, J., & Thompson, R. (2008).
Effectiveness of home blood pressure monitoring, web communication, and pharmacist
care on hypertension control: A randomized control trial. Journal of the American
Medical Association, 299(24), 2857–2868.
Glasgow, R., Boles, S., Kay, G., Feil, E., & Barrera, M. (2003). The D-Net diabetes
self-management program: Long-term implementation, outcomes, and generalization
results. Preventive Medicine, 36, 410–419.
Goldsmith, D. (2004). Communicating social support. Cambridge, England: Cambridge
University Press.
Goldsmith, D., & Fitch, K. (1997). The normative context of advice as social support. Human
Communication Research, 23(4), 454–476.
Hanauer, D. A., Wentzell, K., Laffel, N., & Laffel, L. M. (2009). Computerized automated
reminder diabetes system (CARDS): E-mail and SMS cell phone text messaging reminders
to support diabetes management. Diabetes Technology & Therapeutics, 11(2), 99–106.
Hanson, S. M., & Sauer, W. J. (1985). Children and their elderly parents. In W. J. Sauer, & R.
T. Coward (Eds.), Social support networks and the care of the elderly (pp. 41–66). New
York, NY: Springer.
Human Communication Research 39 (2013) 252–268  2013 International Communication Association

265

Can Messages Make a Difference?

J. W. Turner et al.

House, J. S., Robbins, C., & Metzner, H. L. (1982). The association of social relationships and
activities with mortality: Prospective evidence from the Tecumseh community health
study. American Journal of Epidemiology, 116, 123–140.
Jemmott, J. B., & Magloire, K. (1988). Academic stress, social support, and secretory
immunoglobin A. Journal of Personality and Social Psychology, 55, 803–810.
Kassirer, J. (1995). The next transformation in the delivery of care. The New England Journal
of Medicine, 332(1), 52–54.
Kelly, W. F., Mahmood, R., Kelly, M. J., Turner, S., & Elliott, K. (1993). Influence of social
deprivation on illness in diabetic patients. British Medical Journal, 307, 1115–1116.
Krishna, S., Boren, S. A., & Balaas, E. A. (2009). Healthcare via cell phones: A systematic
review. Telemedicine Journal of E-Health, 15(3), 231–240.
Levine, B. A., Turner, J. W., Robinson, J. D., Angelus, P., & Hu, M. J. (2009).
Communication plays a critical role in web-based monitoring. Journal of Diabetes Science
and Technology, 3(3), 461–467.
Liederman, E., Zimmerman, E., Athanasoulis, M., & Young, M. (2004). Systemwide rollout
of doctor-patient secure web messaging. In P. Whitten, & D. Cook (Eds.), Understanding
health communication technologies (pp. 244–250). San Francisco, CA: Jossey-Bass.
Littlefield, C., Rodin, G., Murray, M., & Craven, J. (1990). Influence of functional
impairment and social support on depressive symptoms in persons with diabetes. Health
Psychology, 9(6), 737–749.
Litwak, E., & Messeri, P. (1989). Organizational theory, social supports, and mortality rates:
A theoretical convergence. American Sociological Review, 54, 49–66.
Major, B., Cozzarelli, C., Sciaccitano, A. M., Cooper, M. L., Testa, M., & Mueller, O. M.
(1990). Perceived support, self-efficacy, and adjustment to abortion. Journal of
Personality and Social Psychology, 59, 452–463.
McMahon, G., Gomes, H., Hohne, S., Ming-Jye Hu, T., Levine, B., & Conlin, P. (2005).
Web-based care management in patients with poorly controlled diabetes. Diabetes Care,
28(7), 1624–1629.
Mollon, B., Holbrook, A. M., Keshavhjee, K., Troyan, S., Gaebel, K., Thabane, L., & Perera,
G. (2008). Automated telephone reminder messages can assist electronic diabetes care.
Journal of Telemedicine and Telecare, 14(1), 32–36.
Mun, S. K., & Turner, J. W. (1999). Telemedicine: The emergence of e-medicine. Annual
Review of Biomedical Engineering, 1, 589–610.
Norbeck, J. S., Lindsey, A. M., & Carrieri, V. L. (1981). The development of an instrument to
measure social support. Nursing Research, 30, 264–269.
NIH (2012). Medline Plus: HbA1c. In A.D.A.M. medical encyclopedia. Retrieved January 13,
2012 from http://www.nlm.nih.gov/medlineplus/ency/article/003640.htm
Orth-Gomer, K., & Johnson, J. V. (1987). Social network interaction and mortality. A six
year follow-up study of a random sample of the Swedish population. Journal of Chronic
Disease, 40, 949–957.
Patt, R. M., Houston, T., Sands, D., & Ford, D. (2003). Doctors who are using e-mail with
their patients: A qualitative exploration. Journal of Medical Internet Research, 5(2), e9.
doi: 10:2196/jmir.5.2.e9.
Perez-Gerre, N., Galinda, M., Fernandez, D., Velasco, V., de la Cruz, M., Martin, P., &
Calle-Pascual, A. (2010). A telemedicine system based on internet and short message
service as a new approach in the follow-up of patients with gestational diabetes. Diabetes
Research and Clinical Practice, 87, e15–e17.
266

Human Communication Research 39 (2013) 252–268  2013 International Communication Association

J. W. Turner et al.

Can Messages Make a Difference?

Piette, J. (2007). Interactive behavior change technology to support diabetes
self-management: Where do we stand? Diabetes Care, 30(10), 2425–2432.
Pouwer, F., Beekman, A., Nijpels, G., Dekker, J., Snoek, R., Kostense, P., & Deeg, D. (2003).
Rates and risks for co-morbid depression in patients with type 2 diabetes mellitus: Results
from a community-based study. Diabetologia, 46, 892–898.
Purcell, K. (2011). Search and email still top the list of most popular online activities: Two
activities nearly universal among adult internet users. Pew Research Center’s Internet &
American Life Project. Retrieved January 13, 2012 from http://pewinternet.org/Reports/
2011/Search-and-email.aspx.
Rains, S., & Young, V. (2009). A meta-analysis of research on formal computer-mediated
support groups: Examining group characteristics and health outcomes. Human
Communication Research, 35, 309–336.
Robinson, N., Lloyd, C. E., & Stevens, L. K. (1998). Social deprivation and mortality in adults
with diabetes mellitus. Diabetic Medicine, 15, 205–212.
Robinson, J. D., Turner, J. W., Levine, B. A., & Tian, Y. (2011). Expanding the walls of the
healthcare encounter: Support and outcomes for patients online. Health Communication,
27, 1–10.
Rosen, P., & Kwoh, C. K. (2007). Patient-physician e-mail: An opportunity to transform
pediatric health care delivery. Pediatrics, 120(4), 701–706.
Roter, D., Larson, S., Sands, D., Ford, D., & Houston, T. (2008). Can e-mail messages between
patients and physicians be patient-centered? Health Communication, 23(1), 80–86.
Schoenbach, V. J., Kaplan, B. G., Freedman, L., & Kleinbau, D. G. (1986). Social ties and
mortality in Evans County, Georgia. American Journal of Epidemiology, 123, 577–591.
Schwartz, L. S., Springer, J., Flaherty, J. A., & Kiani, R. (1986). The role of recent life events
and social support in the control of diabetes mellitus. General Hospital Psychiatry, 8,
212–216.
Seeman, T. E., Berkman, L. F., Kohout, F., Lacroix, A., Glynn, R., & Blazer, D. (1993).
Intercommunity variations in the association between social ties and mortality in the
elderly: A comparative analysis of three communities. Annals of Epidemiology, 3, 325–335.
Semmer, N., Elfering, A., Jacobshagen, N., Perrot, T., Beehr, T., & Boos, N. (2008). The
emotional meaning of instrumental support. International Journal of Stress Management,
15(3), 235–251.
Sherbourne, C. D., & Stewart, A. L. (1991). The MOS social support survey. Social Science &
Medicine, 32, 705–714.
Smith, K., Levine, B., Clement, S., Hu, M. J., Alaoui, A., & Mun, S. K. (2004). Impact of
MyCareTeam for poorly controlled diabetes mellitus. Diabetes Technology &
Therapeutics, 6(6), 828–835.
Thielst, C. B. (2011). Social media: Ubiquitous community and patient engagement. Frontiers
of Health and Service Management, 28(2), 3–14.
Toth, E. L., & James, I. (1992). Description of a diabetes support group: Lessons for diabetes
caregivers. Diabetic Medicine, 9, 773–778.
Uchino, B., Cacioppo, J., & Kiecolt-Glaser, J. (1996). The relationship between social support
and physiological processes: A review with emphasis on underlying mechanisms and
implications for health. Psychological Bulletin, 119(3), 488–531.
Wang, C. Y., & Fiske, M. M. (