Skip to main content

Table 5 Model comparisons for health outcomes in all-age diabetes cohorts

From: Predictive performance of comorbidity measures in administrative databases for diabetes cohorts

 

Cohort 1a(n= 29,058)

Cohort 2 (n= 41,925)

Model (df)b

c-statisticc (95% CI)

∆c d (%)

Brier score

LRTe

c-statisticc (95% CI)

∆c(%)

Brier score

LRTe

AMI

Base (7)

0.66 (0.64, 0.68)

.

0.02

.

0.68 (0.66, 0.70)

.

0.02

.

# diagnoses (10)

0.67 (0.65, 0.69)

0.01 (1.37)

0.02

21.1*

0.68 (0.67, 0.71)*

<0.01 (1.47)

0.02

32.1*

Charlson (10)

0.67 (0.65, 0.69)

0.01 (1.52)

0.02

21.5*

0.69 (0.67, 0.71)*

0.01 (1.92)

0.02

37.3*

Elixhauser (35)

0.68 (0.66, 0.69)*

0.02 (2.43)

0.02

44.2*

0.70 (0.68, 0.72)*

0.02 (2.95)

0.02

62.5*

# drugs (10)

0.68 (0.66, 0.69)*

0.02 (2.43)

0.02

36.0*

0.68 (0.66, 0.70)

<0.01 (0.88)

0.02

24.1*

CDS (11)

0.67 (0.65, 0.69)*

0.01 (2.28)

0.02

32.8*

0.68 (0.66, 0.70)

<0.01 (0.29)

0.02

6.1

Stroke

Base (7)

0.70 (0.68, 0.72)

.

0.02

.

0.71 (0.69, 0.72)

.

0.01

.

# diagnoses (10)

0.71 (0.70, 0.73)

0.01 (1.57)

0.02

37.8*

0.71 (0.69, 0.73)

<0.01 (0.85)

0.01

18.9*

Charlson (10)

0.72 (0.70, 0.73)

0.02 (1.85)

0.02

44.3*

0.71 (0.69, 0.73)

<0.01 (0.85)

0.01

17.2*

Elixhauser (35)

0.72 (0.70, 0.74)*

0.02 (2.56)

0.02

45.4*

0.71 (0.69, 0.73)

<0.01 (0.85)

0.01

20.6

# drugs (10)

0.71 (0.69, 0.73)

0.01 (0.85)

0.02

19.5*

0.71 (0.70, 0.73)

<0.01 (0.99)

0.01

24.2*

CDS (11)

0.71 (0.69, 0.73)

0.01 (1.28)

0.02

33.7*

0.71 (0.70, 0.73)

<0.01 (0.99)

0.01

21.2*

LEA

Base (7)

0.65 (0.62, 0.68)

.

0.01

.

0.68 (0.65, 0.70)

.

0.01

.

# diagnoses (10)

0.71 (0.69, 0.74)*

0.06 (10.36)

0.01

77.4*

0.74 (0.72, 0.77)*

0.06 (9.75)

0.01

128.7*

Charlson (10)

0.74 (0.71, 0.77)*

0.09 (14.06)

0.01

128.4*

0.76 (0.74, 0.79)*

0.08 (12.70)

0.01

201.7*

Elixhauser (35)

0.78 (0.75, 0.81)*

0.13 (20.71)

0.01

305.6*

0.79 (0.76, 0.82)*

0.11 (16.69)

0.01

372.8*

# drugs (10)

0.70 (0.67, 0.72)*

0.05 (7.42)

0.01

53.0*

0.74 (0.72, 0.76)*

0.06 (9.31)

0.01

110.2*

CDS (11)

0.69 (0.66, 0.72)*

0.04 (6.34)

0.01

42.4*

0.72 (0.70, 0.75)*

0.04 (6.65)

0.01

74.5*

ESRD

Base (7)

0.67 (0.63, 0.72)

.

<0.01

.

0.65 (0.61, 0.69)

.

<0.01

.

# diagnoses (10)

0.74 (0.70, 0.79)*

0.07 (10.73)

<0.01

49.0*

0.77 (0.73, 0.80)*

0.12 (18.52)

<0.01

128.0*

Charlson (10)

0.80 (0.76, 0.84)*

0.13 (19.08)

<0.01

138.6*

0.82 (0.79, 0.86)*

0.17 (27.01)

<0.01

276.6*

Elixhauser (35)

0.84 (0.80, 0.89)*

0.17 (25.78)

<0.01

328.6*

0.87 (0.84, 0.90)*

0.22 (34.41)

<0.01

555.1*

# drugs (10)

0.73 (0.69, 0.76)*

0.06 (8.20)

<0.01

36.6*

0.71 (0.68, 0.75)*

0.06 (9.88)

<0.01

50.2*

CDS (11)

0.75 (0.71, 0.79)*

0.08 (12.07)

<0.01

61.3*

0.72 (0.68, 0.76)*

0.07 (11.11)

<0.01

55.3*

Hospitalization, any reason

Base (7)

0.63 (0.62, 0.64)

.

0.24

.

0.63 (0.62, 0.63)

.

0.24

.

# diagnoses (10)

0.70 (0.69, 0.71)*

0.07 (11.13)

0.22

2115.1*

0.70 (0.69, 0.70)*

0.07 (11.31)

0.22

3101.3*

Charlson (10)

0.68 (0.67, 0.69)*

0.05 (8.11)

0.23

1424.1*

0.67 (0.67, 0.68)*

0.04 (7.32)

0.23

1794.8*

Elixhauser (35)

0.69 (0.68, 0.69)*

0.06 (9.06)

0.22

1663.7*

0.68 (0.68, 0.69)*

0.05 (8.60)

0.23

2165.5*

# drugs (10)

0.66 (0.66, 0.67)*

0.03 (5.09)

0.23

823.1*

0.65 (0.65, 0.66)*

0.02 (3.98)

0.23

996.2*

CDS (11)

0.65 (0.64, 0.66)*

0.02 (3.50)

0.23

539.2*

0.64 (0.64, 0.65)*

0.01 (2.23)

0.24

503.0*

Hospitalization, diabetes

Base (7)

0.60 (0.59, 0.61)

.

0.18

.

0.59 (0.59, 0.60)

.

0.15

.

# diagnoses (10)

0.65 (0.65, 0.66)*

0.05 (8.65)

0.18

889.9*

0.66 (0.65, 0.66)*

0.07 (10.83)

0.15

1251.0*

Charlson (10)

0.66 (0.65, 0.67)*

0.06 (9.65)

0.17

1194.1*

0.65 (0.64, 0.66)*

0.06 (9.98)

0.15

1141.8*

Elixhauser (35)

0.67 (0.66, 0.68)*

0.07 (11.65)

0.17

1554.3*

0.66 (0.66, 0.67)*

0.07 (12.35)

0.15

1715.9*

# drugs (10)

0.64 (0.63, 0.64)*

0.04 (5.66)

0.18

524.5*

0.63 (0.62, 0.63)*

0.04 (6.09)

0.15

633.9*

CDS (11)

0.63 (0.62, 0.63)*

0.03 (4.16)

0.18

365.1*

0.62 (0.61, 0.63)*

0.03 (4.57)

0.15

397.7*

Death

Base (7)

0.77 (0.76, 0.78)

.

0.08

.

0.79 (0.78, 0.79)

.

0.07

.

# diagnoses (10)

0.79 (0.78, 0.80)*

0.02 (3.39)

0.08

499.4*

0.80 (0.80, 0.81)*

0.01 (2.16)

0.07

475.1*

Charlson (10)

0.82 (0.81, 0.82)*

0.05 (6.53)

0.07

1014.0*

0.82 (0.82, 0.83)*

0.03 (4.57)

0.07

1018.0*

Elixhauser (35)

0.83 (0.82, 0.83)*

0.06 (7.96)

0.07

1379.9*

0.84 (0.83, 0.84)*

0.05 (6.10)

0.06

1618.7*

# drugs (10)

0.79 (0.78, 0.80)*

0.02 (2.74)

0.08

429.1*

0.80 (0.79, 0.81)*

0.01 (1.78)

0.07

401.4*

CDS (11)

0.78 (0.77, 0.79)*

0.01 (1.96)

0.08

306.1*

0.79 (0.79, 0.80)*

<0.01 (0.89)

0.07

216.9*

  1. aCohort 1 was defined using administrative data from fiscal years 1996/97 to 1997/98 and Cohort 2 was defined using data from 1996/97 to 2001/02; bBase model includes covariates of age, age2, sex, income quintile, geography, and recent diabetes diagnosis, and full models include these variables in addition to the specified comorbidity measure; cA c-statistic with * is significantly different (p < .01) from the c-statistic for the base model, d∆c = c-statistic for full model minus c-statistic for base model, eA likelihood ratio test (LRT) with * is statistically significant at α = .01; df=degrees of freedom, CDS Chronic Disease Score, AMI Acute Myocardial Infarction, LEA Lower Extremity Amputation, ESRD End-Stage Renal Disease.
Back to article page

BMC Health Services Research

ISSN: 1472-6963

Contact us