RLMAR2008 Layout


SA JOURNAL OF PHYSIOTHERAPY 2008 VOL 64 NO 3          7

High Levels of Self-efficacy
in Patients with Type 2 Diabetes
Attending a Tertiary Level Clinic 

R e s e a r c h

A r t i c l e

INTRODUCTION
Type 2 diabetes is a chronic disease that
affects two percent of the South African
population and is one of the major causes
of premature death worldwide largely
due to the complications caused by
incorrect or ineffective management of
the disease (Diabetes SA 2008).  (Sarkar
et al 2006) believe that the keystone of
diabetes care is self-management and
propose that self-efficacy plays a major
role in a patient’s self-management.  

“The theory of self-efficacy proposes
that patients’ confidence in their ability
to perform health behaviours influences
which behaviours they will engage in.”
Sarkar et al 2006, p. 824).  Certain factors
can either act as barriers or facilitators 
to a patient’s self-efficacy.  These factors
include demographic variables such as
age, gender, educational level, per -
sonality types and other health related
problems, e.g. depression and cancer
(Ciechanowski and Katon 2006).  The

relationship between self-management
and self-efficacy is shown in Figure 1.

It has been proposed that a reciprocal
relationship exists between self-manage-
ment and self-efficacy (Sarkar et al
2006).  Self-efficacy in this article is
conceptualised as the patient’s personal
judgement of his/her confidence in per-
forming activities in different domains
as listed in Figure 1, e.g. making and
staying with a regular exercise pro-
gramme.  In turn, self-efficacy may be
influenced by certain demographic 
variables, such as a low income and
increasing age, which is also associated
with physical inactivity (Nelson et al
2002).  Improvement in a patient’s self-
efficacy may lead to improvement in
self-management of diabetes, where
self-management refers to the daily per-
formance of a diabetes self-care regime.  

Diabetes management at Pretoria
Academic Hospital is conducted within
a multidisciplinary team, consisting of

Correspondence to:
Karien Mostert
Room 3-71, Department of Physiotherapy
School of Health Care Sciences
Dr Savage Road,
University of Pretoria,
South Africa
Tel: +27 12 354-1353
Fax: +27 12 354-1226
Email: karien.mostert@up.ac.za

A BST R A CT:  Self-management is a vital element in the care of type 2
diabetes patients.  In turn, self-efficacy plays a major role in patients’ self-
management. Self-efficacy is the patient’s personal judgement of his/her
confidence in performing aspects of diabetes self-management.  This
study investigated the level of self-efficacy of patients attending the
Pretoria A cademic Hospital Diabetes Clinic, in the light of high levels of
re-admission due to complications, suggesting low self-efficacy levels. 

Eighty type 2 diabetes patients, mean age of 59 years, completed the
published IDEA LL baseline questionnaire, to establish a self-efficacy
score.  Relationships between self-efficacy and demographic factors were
investigated using the chi-square test. The mean self-efficacy level of the
sample population is excellent (mean = 85.44%).

A lthough self-reported self-efficacy levels are excellent, in comparison to the Sarkar study (2006) in which parti  -
ci pants only scored “fair”, it is speculated that self-efficacy is not transferred to self-management behaviour in this
population.  A frikaans and English speaking participants score significantly better than those from other language 
categories. There is a positive relationship between self-efficacy and level of education and employment status (tended
towards significance with p values of 0.06 and 0.07 respectively). A lthough self-efficacy scores of clients at this 
tertiary level outpatient clinic are excellent, further research is necessary to quantify self-management strategies and
to correlate these with self-efficacy levels. 

KEY W ORDS:  TY PE 2 DIA BETES, SELF-EFFICA CY, PHY SIOTHERA PY / PHY SICA L THERA PY.

Mostert Wentzel K1;
Nel C1;

van Rooijen AJ1;
Francis J1;
Gibbs G1;
Hacker K1;
Gebert S1.

1
Lecturers and students from the Department of
Physiotherapy, University of Pretoria, South Africa.

doctors, physiotherapists, occupational
therapists, podiatrists, dieticians, ophthal -
mologists and nurses.  This approach
ensures that patients receive education
about the disease process and its
manage ment from different health care
professionals’ perspectives.  Despite this
education, it has been observed by the
team that patients fail to adhere to 
self-management regimes and are often
admitted to hospital due to complica-
tions of diabetes, e.g. renal failure and



8 SA JOURNAL OF PHYSIOTHERAPY 2008 VOL 64 NO 3

gangrene. It was speculated that poor
levels of self-efficacy might be a possi-
ble reason for this. The current impor-
tance of self-efficacy in the management
of chronic diseases and from observa-
tions at the clinic deemed it necessary to
conduct this study with the aim of  deter-
mining the level of self-efficacy and its
influencing factors.  

METHOD
A cross-sectional descriptive study was
conducted using a survey. The study
population consisted of patients attend-
ing the diabetes clinic at the Pretoria
Academic Hospital, a public tertiary
care institution.  

Due to a limited number of patients,
consecutive sampling was done.  All the
patients who attended the clinic when
the researchers were present, and who
met the inclusion criteria, were included
in the study sample.  The inclusion 
criteria were: (i) a confirmed diagnosis
by the diabetes clinic physician of 
type 2 diabetes according to the World
Health Organization (WHO) criteria
(WHO 2006), and (ii) patients between
the ages of 25 and 80 years. (Statistics
from the Diabetes Clinic database show
that the majority of patients fall within
this age group.) Patients previously
diagnosed with a psychiatric disorder
were excluded due to ethical guidelines

pertaining to mentally-ill patients.
Biographic data were collected for each
participant covering the variables in
Figure 1.  Thereafter the IDEALL base-
line questionnaire was administered on 
a one to one basis by one of the
researchers, in the language of prefer-
ence of the participant (English or
Afrikaans).   

The IDEALL baseline questionnaire
is a valid self-efficacy scale adapted
from a reliable instrument measuring
self-efficacy in post myocardial infarc-
tion patients (Coyne and Smith 1994;
Skaaf et al 2003). The scale measures
diabetes-specific domains such as confi-
dence in self-monitoring blood glucose,
as well as general health domains such
as confidence in the ability to get medical
attention and to take care of one’s
health.  The questionnaire consists of
eight questions with a 4-point Likert-
type response from “1=not at all” to
“4=very sure”.  For each question patients
rate their confidence in their ability to
perform a recommended self-care rou-
tine.  The responses from the IDEALL
baseline survey questionnaire were
summed to obtain an overall self-efficacy
score.  The total scores are categorised
as follows:  8-13 = poor, 14-19 = fair,
20-25 = good and 26-32 = excellent,
with eight being the lowest possible
score and 32 the highest.  The score indi-
cates the self-reported self-efficacy 
of the participant, with a higher score
indicating better self-efficacy.  

Permission to conduct the study was
obtained from the superintendent of the
Pretoria Academic Hospital, the chief
coordinator of the Diabetes Clinic at
Pretoria Academic Hospital and the
Student Ethics Committee of the 
Faculty of Health Sciences (S21/2007),
University of Pretoria.

The patients read an information
leaflet about the study, explaining the
process and ethical considerations such
as confidentiality. Voluntary completion
of the questionnaire implied informed
consent for participation in the study.
Patient anonymity was ensured through-
out the study.  

Due to the qualitative nature of vari-
ables, categorical analysis was used to
examine the relationships between the
overall self-efficacy score and the demo-
graphic data. The chi-square test was

Figure 1:  Self-efficacy and factors influencing it, in relation to self-
management

Self-management
Confidence translated into behaviour

Self-efficacy
Confidence to perform self-care activities

Domains
•  General health 
•  Accessing medical services
•  Diet
•  Exercise
•  Blood-glucose testing
•  Medication
•  Support
•  Diabetes information

Variables
•  Age
•  Gender
•  Duration of diabetes
•  Level of education
•  Home language
•  Marital status
•  Employment status



SA JOURNAL OF PHYSIOTHERAPY 2008 VOL 64 NO 3          9

used with 0.05 as the level of signifi-
cance. Further categorising of the data
into fewer groups revealed no changes
in the level of significance for the differ-
ent relationships.

When the data were investigated with
other statistical tests, such as Kruskal-
Wallis, no further significant relationships

were found.  The scale reliability coeffi-
cient of the IDEALL baseline question-
naire (8 item scale) was 0.70 while the
coefficient of the demographic ques-
tions with the self-efficacy questions 
(15 item scale) was 0.67 in this study.
Data were analysed using the STATA 8.0
statistical analysis programme. 

RESULTS
Table 1 gives a profile of the sample
(n=80).  The mean and mode age was 59
years (SD±10.2), with a range of 29 to
76 years. Table 2 shows that the overall
self-efficacy was excellent among the
patients attending the clinic with a mean
self-efficacy score of 27.34 (SD±3.6)
out of 32, i.e. 85.44%.  The overall scores
ranged from 13 to 32.  The probability of
someone having a poor or fair score
(between 8 and 19) was 1:26.

The scores for the self-efficacy ques-
tions on health, diet and exercise were
relatively poor in comparison with the
other questions. (Refer to Tables 3 and
4.) These are areas in which health care
workers interact with patients and so
would require further emphasis.  

Generally without categorising the
patients into language groups (Group 1
consisting of Afrikaans and English 
and Group 2 consisting of all the other
language groups) there is no dependence
(although nearly significant) when 
considering the relationship between the
language of the patient and the overall
self-efficacy score. However, when the
grouping was done, it was found to be a
significant relationship (0.001). (See
Table 4.)  

Apart from the home language of the
patient, there was no other evidence to
show that the overall self-efficacy score
was dependent on the demographics of
the patient. (See Table 4.)  However,
participants with higher levels of 
education tended to have higher levels
of self-efficacy and employment status
demonstrated similar results (tended
towards significance with p values of
0.06 and 0.07 respectively).  

DISCUSSION
No previous research on self-efficacy in
patients with diabetes that has been con-
ducted in South Africa could be found
and therefore the results from this study
have been compared with results from
studies conducted elsewhere.  

The self-efficacy scores are unex-
pectedly high in the sample, with the
mean score in the “excellent” category.
The sample in the study done by Sarkar
et al (2006) have overall mean self-
efficacy scores in the “fair” category.
When a percentage for the mean self-
efficacy was calculated for the sample, it

Table 1: Summary of demographic data (n=80)

Variable Frequency  (%)

Gender
Male 39 (49.00)
Female 41 (51.00)

Duration of diabetes (in years)
1-4 13 (16.25)
5-9 15 (18.75)
10-20 32 (40.00)
>21 20 (25.00)

Level of education
None 6 (7.50)
Grade 1- 3 2 (2.50)
Grade 4-7 7 (8.75)
Grade 8-9 11 (13.75)
Grade 10-12 41 (51.25)
Post matric 3 (3.75)
College or University 10 (12.50)

Home language
Zulu 4 (5.00)
Sotho 1 (1.25)
Xhosa 0 (0.00)
Tsonga 3 (3.75)
Pedi 4 (5.00)
Tswana 4 (5.00)
Venda / Ndebele/ Swazi 4 (5.00%)
Afrikaans 46 (57.50%)
English 13 (16.25%)
Other (Portuguese) 1 (1.25%)

Marital status
Single 10 (12.50%)
Married or living together 49 (61.25%)
Widowed 10 (12.50%)
Separated or divorced 8 (13.75%)

Employment Status
Unemployed 16 (20.00%)
Retired 35 (43.75%)
Office Worker 3 (3.75%)
Homemaker 6 (7.50%)
Dressmaking 0 (0.00%)
Non-office worker 1 (1.25%)
Other 19 (23.75%)

Table 2:  Distribution of self-efficacy scores (n = 80)

Self-efficacy score category Frequency Percentage 

Poor 1 1.25    
Fair 2 2.50
Good 19 23.75
Excellent 58 72.50
Total 80 100.00



10 SA JOURNAL OF PHYSIOTHERAPY 2008 VOL 64 NO 3

was found to be 85.44% as compared to
74.00% in the study by Sarkar et al
(2006).  With a high self-efficacy score,
patients should have confidence in 
per forming  tasks, but the high number
of patients being re-admitted to the
Pretoria Academic Hospital with dia-
betes related complications, is suggestive
that self-efficacy does not translate 
into good health behaviour and self-
management of the disease irrespective
of the duration of diabetes. 

A barrier to achieving glycaemic 
control, as investigated by Apparico et al
(2006), was increasing age. Contrary 
to their finding, this study found that
self-efficacy was high regardless of age,
indicating that clients at the Pretoria
Academic Hospital clinic have high 
levels of confidence in managing the
disease. However, with increasing age,
even if one feels confident, i.e. have a
high self-efficacy score, and even when
adhering to the self-management pro-
gramme, complication may develop
(Foundation for Health in Aging 2007),
such as blindness or heart attacks.  In other
words, high self-efficacy may be seen in
older patients even though they present
with diabetes related complications.

However, this  study did not investigate
the relationship between age and com-
plications, neither is the relation between
age and re-admittance of this study 
population known.

Results from this study suggest that
self-efficacy is similar in males and
females.  However, Hawthorne and
Tomlinson (1999) found that only 24%
of diabetes participants knew how 
persistent hyperglycaemia should be
managed, with women being worse at
this than men (19% vs 31%).  Women
were less likely to understand why it is
necessary for glucose levels to be moni-
tored regularly, and therefore they 
ultimately had poorer control over their
glycaemic levels.

Insufficient attention has been given
in the literature to what the effect of the
duration of diabetes is on the level of
self-efficacy of a patient.  Research from
a cross sectional observational study
found that the duration of clinic atten-
dance was inversely related to glycaemic
control (Apparico et al 2007), suggesting
that self-management did not improve
as time went by. Self-management was
not investigated in this study and although
confidence in performing a task, accord-

ing to social cognitive theory, increases
proportionally with the duration of 
performing the task; this relationship
was not found.  It may be that health
education by the team at the Pretoria
Academic Hospital clinic is effective,
and that patients feel confident in the
domains tested in this study. 

Participants with higher educational
levels tended to have better self-efficacy
scores than those with lower levels of
education. Findings in another study 
on adult diabetic patients suggest that
“The low educational level of the parti -
cipants, illiteracy, and language profi-
ciency may have had an influence on 
the understanding of the educational
material about good dietary practices”
(van Rooijen et al 2004 p. 353).  This
finding emphasises that greater care
should be given when educating parti -
cipants with lower levels of education. 

Social support is important in the
management of any chronic disease.
Diabetes self-care is for example more
effective when using a family-centered
or church-based approach (Samuel-
Hodge et al 2000).  Widowed patients
have less emotional support and there-
fore experience more difficulty with the
management of diabetes (Westaway et al
2005).  When comparing self-efficacy
between patients who were married or
living together with all other marital 
status groups, no significant differences
were found.  This finding could be
explained by the increase in diabetes
support systems available to patients
suffering from diabetes.  These systems
include electronic support groups,
Diabetes SA and the support provided at
the health clinics and hospitals. 

Table 4:  Correlation between the scores for the self-efficacy categories
and demographic variables using the chi-square test

Variable p = value

Gender 0.78
Duration of diabetes 0.29
Levels of education 0.06*
Home language 0.0001**
Marital status 0.87
Employment status 0.07*

** Significant at p < 0.05 
* Tend towards significance

Table 3:  Distribution of self-efficacy score per IDEALL question by Likert category (n= 80) 

Dimension Not at all sure Just a little sure Fairly sure Very sure

Frequency (percentage)

Take care of health 6 (7.50%) 10 (12.50%) 26 (32.50%) 38 (47.50%)

Get medical attention when needed 2 (2.50%) 6 (7.50%) 20 (25.00%) 52 (65.00%)

Make and stay with changes in diet 5 (6.25%) 15 (18.75%) 29 (36.25%) 31 (38.75%)

Make and stay with a regular exercise plan 14 (17.50%) 16 (20.00%) 22 (27.50%) 28 (35.00%)

Test blood sugar regularly 4 (5.00%) 0 (0.00%) 10 (12.50%) 66 (82.50%)

Take all diabetes medicines correctly 2 (2.50%) 1 (1.25%) 9 (11.25%) 68 (85.00%)

Get people around to help with
diabetes when needed 2 (2.50%) 6 (7.50%) 16 (20.00%) 56 (70.00%)



SA JOURNAL OF PHYSIOTHERAPY 2008 VOL 64 NO 3          11

In terms of the relationship between
employment status and diabetes self-
efficacy, Barrett et al (2007) and
Westaway et al (2005) found that patients
from higher income groups tend to 
have better health and engage in more
regular leisure time physical activity.  In
this study a similar correlation was found
between employment status and high
levels of self-efficacy, which tended to
approach significance.

CONCLUSION
The purposes of the study, to determine
the degree of self-efficacy of patients at
the Pretoria Academic Diabetes Clinic
and to correlate these with certain 
demographic variables, were achieved.
The mean self-efficacy level of the sam-
ple population is excellent, with  English
or Afrikaans as a home language being a
predictor of high self-efficiency. Higher
levels of education and employment sta-
tus are weakly related to self-efficiency.

RECOMMENDATIONS 
Although the IDEALL baseline ques-
tionnaire was reported to have been 
validated in diabetes patients, evidence
of published findings from reliability
testing could not be found on further
investigation. Although there is no reason
to doubt its test-retest reliability, it is
suggested that this be determined in the
South-African context before use in
future studies.  It appears that self-effi-
cacy does not necessarily translate into
health behaviour and therefore further
studies should investigate self-efficacy,
behaviour and health outcomes simulta-
neously. These investigations should
give a clearer understanding of which
diabetes domains patients struggle most
with in the South African context.  

Clinical implications from the results
suggest that clinicians should be aware
that self-efficacy as reported by patients,
does not necessarily lead to good self-
management behaviours and adherence
to diabetes programmes.  Adherence to
these programmes must be monitored in
relation to patients’ behaviours.    

More care should be taken when 
educating patients with lower levels of
education and employment status, as
well as those who do not have Afrikaans
or English as a home language.

ACKNOWLEDGEMENTS
We are thankful to Dr S Olorunju of 
the Medical Research Council, who
guided the statistical analysis and Sister
Burger, staff and patients at the Pretoria
Academic Diabetes Clinic.

REFERENCES

Apparico N, Clerk N, Henry G, Seale J, Sealy R,
Ward S, Mungrue K 2006 How well controlled
are our type 2 diabetic patients in 2002?  An
observational study in North and Central
Trinidad. Diabetes Research and Clinical
Practice 75:301-305

Barrett J E, Plotnikoff R C, Courneya K S, 
Raine K D 2007 Physical activity and type 2 
diabetes: exploring the role of gender and
income. The Diabetes Educator 33:128-143

Ciechanowski P, Katon W J 2006 The interper-
sonal experience of health care through the
eyes of patients with diabetes. Social Science
and Medicine 63:3067-3079

Coyne J C, Smith D A F 1994 Couples coping
with a myocardial infarction: Contextual per-
spective on patient self-efficacy. Journal of
Family Psychology 8:43-54

Diabetes SA 2008 About diabetes, viewed 27
March 2008 <http://www.diabetessa.org>

Hawthorne K, Tomlinson S 1999 Pakistani
Moslems with type 2 diabetes mellitus: Effect of
sex, literacy skills, known diabetic compli -
cations and place of care on diabetic knowl-
edge, reported self-monitoring management
and glycemic control. Diabetes Medicine
16:591-597

Nelson K M, Reiber G, Boyko E J 2002 Diet and
exercise among adults with type 2 diabetes.
Diabetes Care 25:1722-1728

Samuel-Hodge C D, Headen S W, Skelly A H,
Ingram A F, Keyserling T C, Jackson E J,
Ammerman A S, Elasy T A 2000 Influences on
day-to-day self-management of type 2 diabetes
among African-American women:  Spirituality,
the multi-caregiver role, and other social con-
text factors. Diabetes Care 23:928-933

Sarkar U, Fisher L, Schillinger D 2006 Is self-
efficacy associated with diabetes self-manage-
ment across race/ethnicity and health literacy?
Diabetes Care 29:823-829

Skaaf M M, Mullan J T, Fisher L, Chesla C A
2003 A contextual model of control beliefs,
behaviour, and health: Latino and European
Americans with type 2 diabetes. Psychology
and Health 18:295-312

Statics of South Africa Viewed on 2 February
2007 <http/www.diabetessa.org>

van Rooijen A J, Rheeder P, Eales C J, Becker P J
2004 Effect of exercise versus relaxation on
haemoglobin A1c in black females with type 2
diabetes mellitus.  QJM: monthly journal of the
Association of Physicians 97:343-351  

Westaway M S, Seager J R, Rheeder P, 
van Zyl D G 2005 The effects of social support
on health, well-being and management of 
diabetes mellitus: a black South African 
perspective. Ethnicity and Health 10:73-89

World Health Organization 2006 Diabetes Fact
Sheet No 312 Viewed on 25 March 2008
<http://www.who.int/mediacentre/factsheets/
fs312/en/index.html>