Journal of Endocrinology and Metabolism, ISSN 1923-2861 print, 1923-287X online, Open Access
Article copyright, the authors; Journal compilation copyright, J Endocrinol Metab and Elmer Press Inc
Journal website http://www.jofem.org

Original Article

Volume 9, Number 3, June 2019, pages 43-50


Development and Validation of a Primary Care Tool to Identify Patients With Type 2 Diabetes Mellitus at High Risk of Hypoglycemia-Related Inpatient Admissions

Figure

Figure 1.
Figure 1. Receiver operating characteristic (ROC) curves for performance of risk score in development and testing set.

Tables

Table 1. Characteristic of the Development and Validation of Type 2 Diabetes Mellitus Sets
 
CharacteristicDevelopment set (n = 31,653)Testing set (n = 15,751)
aCandidate risk factors for predicting 6-month hypoglycemia admission. SD: standard deviation; CKD: chronic kidney disease.
Agea, mean (SD), years64.1 (11.0)64.1 (11.1)
Age category, n (%)
  < 60 years10,562 (33.3%)5,278 (33.5%)
  60 - 64 years5,836 (18.4%)2,901 (18.4%)
  65 - 69 years5,139 (16.2%)2,509 (15.9%)
  70 - 74 years4,495 (14.2%)2,243 (14.2%)
  ≥ 75 years5,621 (17.8%)2,820 (17.9%)
Race, n (%)
  Chinese23,947 (75.7%)11,856 (75.3%)
  Malay3,255 (10.3%)1,622 (10.3%)
  Indian3,404 (10.8%)1,681 (10.7%)
  Others1,047 (3.3%)592 (3.8%)
Sex, n (%)
  Male15,242 (48.2%)7,453 (47.3%)
  Female16,411 (51.8%)8,298 (52.7%)
Last HbA1ca, mean (SD), %7.4 (1.3)7.4 (1.3)
HbA1c category, n (%)
  ≥ 75 mmol/mol (≥ 9.0%)2,977 (9.4%)1,461 (9.3%)
  64 - 74 mmol/mol (8.0-8.9%)4,088 (12.9%)2,018 (12.8%)
  53 - 63 mmol/mol (7.0-7.9%)11,464 (36.2%)5,748 (36.5%)
  42 - 52 mmol/mol (6.0-6.9%)11,274 (35.6%)5,650 (35.9%)
  < 42 mmol/mol (< 6.0%)1,850 (5.8%)874 (5.5%)
Largest reduction in HbA1c within 1 yeara, mean (SD), %0.7 (1.0)0.7 (1.0)
Largest reduction in HbA1c within 1 year category, n (%)
  No reduction observed9,011 (28.5%)4,504 (28.6%)
  Reduction of 1 - 10 mmol/mol (0.1-0.9%)16,489 (52.1%)8,137 (51.7%)
  Reduction of 11 - 21 mmol/mol (1.0-1.9%)3,883 (12.3%)1,940 (12.3%)
  Reduction of 22 - 32 mmol/mol (2.0-2.9%)1,133 (3.6%)602 (3.8%)
  Reduction of 33 - 43 mmol/mol (3.0-3.9%)515 (1.6%)256 (1.6%)
  Reduction of 44 - 55 mmol/mol (4.0-4.9%)257 (0.8%)126 (0.8%)
  Reduction of ≥ 55 mmol/mol (≥ 5.0%)365 (1.2%)185 (1.2%)
Last body mass index (BMI)a, kg/m226.5 (4.7)26.5 (4.7)
Body mass index category, n (%)
  Severely overweight (≥ 27.5 kg/m2)11,258 (35.6%)5,565 (35.3%)
  Moderately overweight (23.0 - 27.4 kg/m2)13,487 (42.6%)6,749 (42.8%)
  Acceptable weight (18.5 - 22.9 kg/m2)6,402 (20.2%)3,158 (20.0%)
  Underweight (< 18.5 kg/m2)506 (1.6%)279 (1.8%)
Last estimated glomerular filtration rate (eGFR)a, mL/min/1.73 m285.1 (28.4)85.4 (28.6)
Estimated glomerular filtration rate category, n (%)
  CKD stage 1 (≥ 90 mL/min/1.73 m2)13,172 (41.6%)6,684 (42.4%)
  CKD stage 2 (60 - 89 mL/min/1.73 m2)12,727 (40.2%)6,247 (39.7%)
  CKD stage 3 (30 - 59 mL/min/1.73 m2)5,053 (16.0%)2,511 (15.9%)
  CKD stage 4 (15 - 29 mL/min/1.73 m2)621 (2.0%)258 (1.6%)
  CKD stage 5 (< 15 mL/min/1.73 m2)80 (0.3%)51 (0.3%)
Hypoglycemia admission during past 1 yeara, n (%)
  Yes251 (0.8%)120 (0.8%)
  No31,402 (99.2%)15,361 (99.2%)
Usage of sulphonylurea and/or insulina, n (%)
  None15,750 (49.8%)7,912 (50.2%)
  Sulphonylurea only12,582 (39.7%)6,252 (39.7%)
  Sulphonylurea and insulin1,192 (3.8%)600 (3.8%)
  Insulin only2,129 (6.7%)987 (6.3%)

 

Table 2. Rare-Event Bias Corrected Logistic Regression Analysis of Hypoglycemia Admission During 6-Month Follow-Up
 
Risk factorFull model coefficient (OR)95% CIP valueParsimonious model coefficient (OR)95% CIP value
Bias correction factor, τ = 0.004. OR: odds ratio; CI: confidence interval.
Age (in 5-year increments)1.17(1.06, 1.28)0.0021.16(1.05, 1.28)0.003
Largest reduction in HbA1c within 1 year1.29(1.15, 1.44)< 0.0011.28(1.14, 1.43)< 0.001
Last BMI0.95(0.91, 0.99)< 0.0010.95(0.91, 0.99)< 0.001
Last eGFR (in 5-mL/min/1.73 m2 increments)0.91(0.87, 0.94)< 0.0010.91(0.87, 0.94)< 0.001
Usage of sulphonylurea and/or insulin
  NoneReferenceReference
  Sulphonylurea only2.13(1.33, 3.41)< 0.0012.29(1.44, 3.62)< 0.001
  Sulphonylurea and insulin4.75(2.31, 9.78)< 0.0015.69(2.89, 11.21)< 0.001
  Insulin only5.29(2.93, 9.55)< 0.0016.18(3.57, 10.70)< 0.001
Last HbA1c1.10(0.97, 1.26)0.14
Hypoglycemia admission during past 1 year (Yes)1.22(0.38, 3.94)0.74

 

Table 3. Hypoglycemia Admission Risk Scoring System
 
Risk factorCategoriesScore
Age, years< 600
60 - 646
65 - 698
70 - 7410
≥ 7514
Last BMI, kg/m2Severely overweight (≥ 27.5)0
Moderate overweight (23.0 - 27.4)6
Acceptable weight (18.5 - 22.9)9
Underweight (< 18.5)11
Last eGFR, mL/min/1.73 m2CKD stage 1 (≥ 90)0
CKD stage 2 (60 - 89)12
CKD stage 3 (30 - 59)20
CKD stage 4 (15 - 29)25
CKD stage 5 (< 15)28
Largest reduction in HbA1c within 1 yearNo change or increase0
Reduction of 1 - 10 mmol/mol (0.1-0.9%)2
Reduction of 11 - 21 mmol/mol (1.0-1.9%)5
Reduction of 22 - 32 mmol/mol (2.0-2.9%)8
Reduction of 33 - 43 mmol/mol (3.0-3.9%)11
Reduction of 44 - 55 mmol/mol (4.0-4.9%)14
Reduction of ≥ 55 mmol/mol (≥ 5.0%)23
Usage of sulphonylurea and/or insulinNone0
Sulphonylurea only11
Sulphonylurea and insulin23
Insulin only24

 

Table 4. Sensitivity and Specificity of the Hypoglycemia Risk Score at selected thresholds
 
ThresholdSensitivityNo. of positive cases correctly classifiedSpecificityNo. of negative cases correctly classified
0100.0%650.0%0
1898.5%6424.5%3,843
2893.8%6147.4%7,435
3090.8%5954.3%8,517
3284.6%5559.7%9,365
3476.9%5063.8%10,008
3676.9%5068.7%10,776
3873.8%4873.1%11,466
4069.2%4577.0%12,078
4266.2%4380.8%12,674
4460.0%3983.8%13,145
4653.8%3586.2%13,521
4843.1%2889.3%14,008
5824.6%1696.6%15,153
1000.0%0100.0%15,686