DEVELOPMENT OF TOOLS TO ASSESS THE COMPLICATION RISK IN
MALTESE TYPE 2 DIABETIC PATIENTS
Sarah Baldacchino1, Liberato Camilleri2, Mario J Cachia3, Stephen Fava3, Anthony Serracino-Inglott1,
Lilian M Azzopardi1
1Department of Pharmacy, Faculty of Medicine & Surgery 2Department of Statistics and Operations Research,
3Mater Dei General Hospital, Msida, Malta.
Email: sbal0009@um.edu.mt
INTRODUCTION
RESULTS
Future hospital admission costs due to long-term diabetic complications could be
A statistically significant positive correlation was obtained between the scores
cut down if patients are effectively treated to improve health outcomes.1 However
assigned by three resident specialists (Pearson correlation=0.888, 0.844, 0.812),
treatment differentiation between high risk and low risk patients is necessary to
indicating good reliability for the DCRI.
ensure the sustainability of an intensive diabetes management program.2
Multivariate computer models enable a population to be classified in order of
complication risk and aid clinical judgement in the assignment of intensive
treatment to high risk
Diabetes Complication Models
The significant predictors which featured in the parsimonious models (Table 1)
were age, genetic predisposition, alcohol abuse, BMI, waist circumference,
patients.3
systolic BP, HbA1c level, serum total cholesterol level, serum fasting triglyceride
level, serum urea level, urinary glucose level and albumin-creatinine ratio.
OBJECTIVES
 To identify significant predictors of complication risk and;
neuropathy, retinopathy, nephropathy and macrovascular complications of that
Table 1. Predictors present in the parsimonious models and their regression coefficients
ranked by their contribution in explaining total variance of the dependent variables
(n=92).
Model
Parameter
B
Std. Error
p
Intercept
-2.180
0.415
0.000
Body Mass Index (kg/m2)
0.045
0.011
0.0001
Haemoglobin A1c Level (%)
0.107
0.027
0.0002
Serum Fasting Triglycerides (mmol/L)
0.140
0.044
0.002
DNeurM
No Alcohol Abuse
-0.401
0.172
0.022
Alcohol Abuse
0*
Systolic Blood Pressure (mmHg)
0.006
0.003
0.041
Age (yrs)
0.009
0.005
0.070
Intercept
-2.367
0.425
0.000
Systolic Blood Pressure (mmHg)
0.012
0.003
0.0004
Serum Fasting Triglycerides (mmol/L)
0.193
0.055
0.001
DRM
Haemoglobin A1c Level (%)
0.089
0.034
0.010
Albumin-Creatinine Ratio (mg/g)
0.006
0.003
0.040
Waist Circumference (cm)
0.005
0.003
0.095
Intercept
-3.543
0.517
0.000
Systolic Blood Pressure (mmHg)
0.016
0.003
3x10-7
Trace Urinary Glucose
-0.296
0.377
0.435
No Urinary Glucose
-0.370
0.219
0.095
+1 Urinary Glucose
0.183
0.235
0.438
+2 Urinary Glucose
0.149
0.246
0.545
+3 Urinary Glucose
-0.171
0.248
0.492
DNephrM +4 Urinary Glucose
0*
Albumin-Creatinine Ratio (mg/g)
0.010
0.003
0.0009
Waist Circumference (cm)
0.009
0.003
0.0012
Age (yrs)
0.016
0.005
0.006
No Genetic Predisposition
-0.309
0.136
0.026
Genetic Predisposition
0*
Serum Urea (mmol/L)
0.078
0.039
0.050
Serum Fasting Triglycerides (mmol/L)
0.090
0.048
0.062
Intercept
-3.056
0.368
0.000
Waist Circumference (cm)
0.015
0.003
1x10-6
MVM
Systolic Blood Pressure (mmHg)
0.011
0.003
0.00026
Total Serum Cholesterol (mmol/L)
0.129
0.050
0.011
Haemoglobin A1c Level (%)
0.062
0.033
0.060
particular participant. Once inter-rater reliability was established, one resident
*This
 To develop local diabetic neuropathy (DNeurM), retinopathy (DRM),
nephropathy (DNephrM) and macrovascular (MVM) models which determine
treatment effectiveness in Maltese type 2 diabetes patients.
METHODOLOGY
Setting & Criteria
A retrospective cross-sectional study was carried out at the Endocrine and
Diabetes Centre at Mater Dei General Hospital (MDH), Malta.
The sample
population comprised of 120 randomly selected patients:
 Aged 25-70 years;
 Diagnosed with type 2 diabetes ≤1 year and;
 Taking metformin 500mg bd, perindopril 5mg od and simvastatin 40mg.
Written informed consent was obtained from the participants to record 20
different predictors from their medical files and computerised medical records.
After data collection the sample was reduced to 92 participants since certain
required data was not available.
Risk Assessment
Using the Diabetes Complication Risk Index (DCRI; figure 1), three resident
specialists assigned a total of 4 complication risk scores to 40 study participants
from the sample.
These 4 scores individually represent the current risk for
parameter is set to zero because it is redundant.
specialist continued the risk assessment for the rest of the participants.
Figure 1. Diabetes Complication Risk Index (DCRI) devised for risk assessment measure.
0
2
1
3
4
Although p-values for age (p=0.070) in DNeurM, waist circumference (p=0.095)
in DRM, serum fasting triglycerides (p=0.062) in DNephrM and HbA1c level
(p=0.060) in MVM exceed the 0.05 level of significance, they were included in the
model fit because their contribution was found to be considerable on the
corresponding R2 value.
No development
risk
Low development
risk
Moderate
development
risk
High
development
risk
Very high
development
risk
From the analyses, the presence of trace and 3+ urinary glucose was inversely
associated with the risk for nephropathy development.
In addition, since the
increase in the risk score associated with 1+ urinary glucose is 0.183, the
Data Processing and Analyses
detection of 2+ urinary glucose was expected to contribute more to the risk for
Bivariate Pearson correlation (p<0.05) was utilised to correlate complication risk
diabetic nephropathy than the resultant 0.149 increase. These outcomes conflict
scores assigned by the resident specialists. The data was modelled using ANCOVA
with literature and may be attributed to the small samples which fell in the
regression model analyses on SPSS® 17.0 and backward elimination variable
categories.
procedure (p<0.05) was then conducted to eliminate weak predictors.
LIMITATIONS
 Models have inadequate predictive power and a larger sample is required to draw out more accurately the existing relationships between the predictors and risk score.
 Complications which may have been present at the time of diagnosis have not been excluded from this study and therefore limit current models.
 Validation of models has yet to be conducted. Long-term follow-ups should be performed.
CONCLUSION
REFERENCES
Diabetes-specific models which stratify the diabetic population according to the
1. Sidorov J, Shull R, Tomcavage J, Girolami S, Lawton N, Harris R. Does diabetes disease management save money and improve
risk for complications were derived. Even though this study provides preliminary
evidence that the models could aid healthcare professionals identify the need for
intensive treatment, further studies and long-term follow-ups are required to
outcomes? Diabetes Care 2002; 25(4): 684-689.
2. Selby JV, Karter AJ, Ackerson LM, Ferrara A, Liu J. Developing a prediction rule from automated clinical databases to identify
high-risk patients in a large population with diabetes. Diabetes Care 2001;24(8):1547- 1555.
3. Clark CM, Snyder JW, Meek RL, Stutz LM, Parkin CG. A systematic approach to risk stratification and intervention within a
managed care environment improves diabetes outcomes and patient satisfaction. Diabetes Care 2001;24(6):1079-1086.
validate the models such that adequate risk assessment tools for primary
complication prevention are obtained.
DISCLOSURE: The authors declare no conflict of interest.
ACKNOWLEDGEMENTS: The authors thank Dr Alexia Abela, Dr Sandro Vella and Dr Mark Gruppetta for collaborating in risk assessment.