A total of 6719 type 2 diabetes patients were identified in the pre-intervention phase and 5298 in the post-intervention phase. Table 1 shows the pre-intervention characteristics of patients in the intervention and control groups. Type 2 diabetes patients in both groups were predominantly women with a mean age of 60 years. The intervention group had more patients of Malay and Chinese ethnicity and fewer patients with dyslipidaemia as comorbidity. There were slightly more patients with newly diagnosed type 2 diabetes in the control group (11.0%, vs 9.0% in the intervention group). The median HbA1c was also higher in the control group (8.0% or 64 mmol/mol) compared to the intervention group (7.7% or 61 mmol/mol). The median HbA1c was 7.8% (61 mmol/mol) in the control group and 7.5% (58 mmol/mol) in the intervention group in the post-intervention period (not shown in table). The mean BMI of patients in both groups was 28kg/m2, which falls into the obese category.
Table 1: Pre-intervention patient characteristics
Patient Characteristics
|
N (%)
|
N (%)
|
P-value
|
|
Intervention group
|
Control group
|
|
|
3283
|
3436
|
|
Age ( years)a
|
60.0
|
(11.0)
|
59.8
|
(10.8)
|
0.526
|
Male
|
1183
|
(36.0)
|
1302
|
(37.9)
|
0.115
|
Ethnicity
|
|
|
|
|
|
Chinese
|
561
|
(17.1)
|
637
|
(18.5)
|
<0.001
|
Indian
|
364
|
(11.1)
|
469
|
(13.6)
|
|
Malay
|
2343
|
(71.4)
|
2303
|
(67.0)
|
|
Others
|
15
|
(0.5)
|
27
|
(0.8)
|
|
Weight (kg)a
|
69.6
|
(15.3)
|
69.9
|
(15.6)
|
0.499
|
BMI (kg/m2)a
|
28.1
|
(5.6)
|
28.3
|
(5.9)
|
0.175
|
BMI
|
|
|
|
|
0.685
|
<18.5
|
46
|
(1.6)
|
46
|
(1.5)
|
|
18.5-22.9
|
385
|
(13.3)
|
386
|
(12.8)
|
|
23-27.4
|
1046
|
(36.1)
|
1058
|
(35.1)
|
|
>27.4
|
1421
|
(49.0)
|
1527
|
(50.6)
|
|
No target organ damage
|
2317
|
(70.6)
|
2430
|
(70.7)
|
0.895
|
Smoking Status
|
|
|
|
|
<0.001
|
Current
|
201
|
(6.1)
|
209
|
(6.1)
|
|
Ex-smoker
|
39
|
(1.2)
|
56
|
(1.6)
|
|
Non-smoker
|
1111
|
(33.8)
|
1615
|
(47.0)
|
|
Unknown
|
1932
|
(58.8)
|
1556
|
(45.3)
|
|
Comorbidities
|
|
|
|
|
|
Dyslipidaemia
|
1539
|
(46.9)
|
1733
|
(50.4)
|
0.004
|
Hypertension
|
2562
|
(78.0)
|
2693
|
(78.4)
|
0.738
|
Newly diagnosed diabetes
|
294
|
(9.0)
|
379
|
(11.0)
|
0.005
|
Diabetes duration (years )*
|
5
|
(2.6,9.8)
|
5
|
(2.4,9.9)
|
0.644
|
Hypertension duration (years ) *
|
7
|
(3.5,11.2)
|
7
|
(3.3,11.3)
|
0.306
|
Dyslipidaemia duration (years ) *
|
5
|
(2.9,8.0)
|
5
|
(2.7,8.0)
|
0.531
|
Systolic blood pressure ( mmHg) a
|
137
|
(18.9)
|
138
|
(18.8)
|
0.750
|
Diastolic blood pressure ( mmHg) a
|
77
|
(11.2)
|
78
|
(10.7)
|
0.144
|
LDL ( mmol/l) a
|
3.0
|
(1.0)
|
3.0
|
(1.1)
|
0.760
|
HbA1c (%)*
|
7.7
|
(6.6,9.6)
|
8.0
|
(6.7, 9.7)
|
0.021
|
Data are presented as N (%) unless otherwise indicated, a mean (standard deviation) , * median (interquartile range)
After the EnPHC interventions, there were significant relative changes in seven out of 14 process of care measures in the intervention group as shown in Table 2. Compared to controls, the intervention clinics showed significant improvement in performance of four of six laboratory investigations: HbA1c tests in the past three months (OR 3.31, 95% CI, 2.13, 5.13), lipid test in the past one year (OR 4.59, 95% CI, 2.64, 7.97), LDL test in the past one year (OR 4.33, 95% CI, 2.16, 8.70), and UMA test in the past one year (OR 1.99, 95% CI, 1.12, 3.55). Conversely, patients in intervention clinics were three times less likely to have a blood glucose test on the day of visit. There was no significant change for LFT (OR 1.08, 95% CI, 0.51, 2.27).
There were two improvement observed out of five clinical management measures. BMI measured in the past six months showed improvement with a relative odds of 15.80 (95% CI, 4.78, 52.24) and CVD risk assessment in the past one year showed a marked increase in the intervention group with a relative odds of 174.65 (95% CI, 16.84, 1810.80). BP measurement at every visit showed high baseline values ranging from 97%-98.9% in both groups with little room for improvement.. As for counselling and prescription, the intervention group exhibited a significant post-intervention increase in exercise counselling with an odds ratio of 1.18 (95% CI, 1.04, 1.33). Although a higher odds of dietary counselling was observed in the intervention group, it was not significant (OR 1.86, 95% CI, 1.02, 3.38). Use of lipid-lowering medication exhibited high baseline values of over 98% in both groups with no significant change after the intervention (OR 0.97, 95% CI, 0.22, 4.31).
Three of four intermediate clinical outcomes showed improvement (i.e., HbA1c, LDL and HDL within target range), but none of the changes were statistically significant. Detailed results are shown in Table 3.
Table 2: Difference-in-difference (DID) analysis of EnPHC interventions on process of care in Type 2 Diabetes patients
Outcome (%)
|
Na
|
Intervention Group
|
Control Group
|
Difference
(C=A-B)
|
ORb
|
95% CI
(Lower CI)
|
95% CI
(Upper CI)
|
|
Pre
|
Change (A)
(Post-Pre)
|
Pre
|
Change (B)
(Post-Pre)
|
Adjusted P-valuee
|
Lab investigations
|
|
|
|
|
|
|
|
|
|
|
HbA1c test in the past 3 months d
|
10803
|
38.4
|
23.2
|
36.2
|
-5.6
|
28.8
|
3.306
|
2.131
|
5.130
|
<0.001
|
Blood glucose test at every visit d
|
10821
|
82.4
|
-30.2
|
84.9
|
-4.5
|
-25.7
|
0.323
|
0.142
|
0.737
|
0.018
|
Lipid test in the past one year d
|
10821
|
83.3
|
13.6
|
78.8
|
3.5
|
10.1
|
4.587
|
2.640
|
7.970
|
<0.001
|
LDL test in the past one year
|
10820
|
70.2
|
22.4
|
69
|
2.6
|
19.8
|
4.331
|
2.157
|
8.700
|
<0.001
|
UMA test in the past one year
|
10802
|
65.9
|
15.9
|
67.8
|
2.1
|
13.8
|
1.994
|
1.119
|
3.553
|
0.038
|
Liver function test in the past one year d
|
10821
|
58.2
|
3.0
|
51.3
|
1.8
|
1.2
|
1.084
|
0.518
|
2.269
|
0.894
|
Clinical assessments
|
|
|
|
|
|
|
|
|
|
|
Blood pressure at every visit
|
10821
|
97.4
|
1.1
|
98.9
|
0.4
|
1.5
|
3.207
|
0.794
|
12.962
|
0.159
|
BMI measured in the past six monthc,d
|
11945
|
29.0
|
49.5
|
50.8
|
-12.5
|
62.4
|
15.80
|
4.78
|
52.24
|
<0.001
|
Fundus examination in the past one year d
|
10479
|
38.5
|
-1.3
|
37.7
|
-8.9
|
7.6
|
1.325
|
0.778
|
2.256
|
0.420
|
Foot examination in the past one year d
|
10695
|
42.5
|
-8.1
|
52.6
|
-3.8
|
-4.3
|
0.742
|
0.333
|
1.655
|
0.593
|
CVD risk assessment the past one year d
|
10821
|
0.6
|
86.3
|
0.1
|
0.8
|
85.5
|
174.654
|
16.840
|
1810.800
|
<0.001
|
Counselling and prescription
|
|
|
|
|
|
|
|
|
|
|
Exercise counselling d
|
10821
|
44.9
|
21.8
|
44.9
|
4.7
|
17.1
|
1.18
|
1.044
|
1.330
|
0.018
|
Diet counselling d
|
10821
|
67.6
|
14
|
63.0
|
3.5
|
10.5
|
1.862
|
1.026
|
3.377
|
0.071
|
Lipid lowering drug prescription d
|
8835
|
98.0
|
0.3
|
98.4
|
-0.3
|
0.6
|
0.969
|
0.218
|
4.308
|
0.968
|
OR, Odds ratio, UMA, Urine microalbumin, CVD, Cardiovascular disease
a Complete case analyses were performed
b Adjusted for covariates (age, sex, ethnicity, duration of Type 2 Diabetes, body mass index, presence of Hypertension, presence of Hyperlipidemia, presence of target organ damage, state, urban/rural, clinic type)
c Adjusted for covariates (age, sex, ethnicity, duration of Type 2 Diabetes, presence of Hypertension, presence of Hyperlipidemia, presence of target organ damage, state, urban/rural, clinic type)
d Parallel assumption met
e P-value adjustment using Benjamin & Hochberg (1995) method
Table 3: Difference-in-difference (DID) analysis of EnPHC interventions on intermediate clinical outcomes in Type 2 Diabetes patients
Outcome (%)
|
Na
|
Intervention Group
|
Control Group
|
Difference
(C=A-B)
|
ORb
|
95% C
(Lower CI)
|
95% CI
(Upper CI)
|
|
|
Pre
|
Change (A)
(Post-Pre)
|
Pre
|
Change (B)
(Post-Pre)
|
P-valued
|
|
HbA1c ≤ 7% c
|
9195
|
35.1
|
3.2
|
33.0
|
0.8
|
2.4
|
1.056
|
0.8394
|
1.328
|
0.672
|
BP ≤ 135/75mmHg c
|
10821
|
27.0
|
-2.9
|
27.0
|
-0.8
|
-2.1
|
0.900
|
0.705
|
1.253
|
0.672
|
LDL ≤ 2.6mmol/L c
|
8050
|
39.2
|
5.5
|
40.1
|
1.9
|
3.6
|
1.168
|
0.8698
|
1.57
|
0.603
|
HDL within control
|
10821
|
47.9
|
7.6
|
44.1
|
-2.6
|
10.2
|
1.445
|
1.003
|
2.083
|
0.192
|
OR, Odds ratio, BP, Blood pressure
a Complete case analyses were performed
b Adjusted for covariates (age, sex, race, duration of Type 2 Diabetes, body mass index, presence of Hypertension, presence of Hyperlipidemia, presence of Target organ damage, state, urban/rural, clinic type)
c Parallel assumption met
d P-value adjustment using Benjamin & Hochberg (1995) method