Table 3.

Study results of health outcomes (HbA1c, diabetes–related hospitalisation, primary endpoints)

Initiative, study design (first author and date published)	Result reported	Baseline (or basic service level) [n] a	Final visit or comparison group [n] a	Difference	95% Confidence interval or P value
CCT, pre – post cohort	HbA1c mean mmol/L% (CI)	9.0 (8.6, 9.4) [n = 137]	8.8 (8.3, 9.2) [n = 146]	−0.2	0.23 for trend
(Bailie et al. 2004) [20]
Improved care coordination, cohort follow-up	HbA1c mean mmol/L% (CI)	9.3 (8.8, 9.8) [n = 295]	8.9 (8.6, 9.3) [n = 252]	−0.4	−0.7, –0.1
(Bailie et al. 2007) [21]
IDERP – IHS, cross sectional	HbA1c mmol/L% (% patients <7.0)	not reported	not reported	OR = 1.1	0.8, 1.7
(Roubideaux 2008)b [23]
IHS, repeated cross section	HbA1c mean mmol/L% (SE)	8.9 (0.04) [n = 7110]	7.9 (0.03) [n = 15537]	−1.0	0.0001 (1995 vs 2001)
(Wilson et al. 2005) [22]
Care Plus, open prospective cohort	HbA1c mean mmol/L% (CI)c	8.1 (8.0, 8.2) [n = 354]	7.2 (6.7, 7.5) [n = 3]	−0.9	<0.05 (based on CIs)
(Kenealy et al. 2010) [26]
Get Checked, pre – post cohort	HbA1c mean mmol/L% (SD)	8.0 (1.6) [n = 298]	8.0 (1.6) [n = 298]	0	Not reportedd
(Smith et al. 2011) [27]
Integrated diabetes service, pre – post cohort	HbA1c mean mmol/L% (SD)	10.4 (2.2) [n = 30]	7.9 (1.9) [n = 30]	−2.5	<0.001
(Simmons et al. 2003) [28]
Foot program – IHS, pre – post incidence study (Schraer et al. 2003) [25]	Amputation incidence (per 1000 person years)	16.4 (1342 person years)	6.8 (1628 person years)	−59%	0.021
Special diabetes program, repeated cross section	HbA1c mean mmol/L%	8.4 [n = 1394]	7.4 [n = 1839]	−1	<0.001
(Ramesh et al. 2008) [24]
SLICK, pre – post cohort	HbA1c Mean mmol/L%	8.12 [n = 285]	8.01 [n = 285]	−0.11	0.176
(Virani et al. 2006) [32]
Evidence based management of diabetes, cluster randomised trial	HbA1c mmol/L% (% <7)	not reported [n = 555]	IG = 22 CG = 20 [n = 678]	RR: 1.26e	0.219
(McDermott et al. 2001) [29]	Persons hospitalised for diabetes reasons (%)	IG = 20 CG = 22 [n = 555]	IG = 12 CG = 20 [n = 678]	IG = –8% CG = –2%	IG = 0.012 CG = 0.514f
	Diabetes hospital episodes (%)	IG =23 CG =30 [n = 555]	IG =19 CG =29 [n = 678]	IG = –4% CG = –1%	IG = 0.015 CG = 0.746f
Hypertension management, randomised unblinded control trial	HbA1c mean mmol/L% (SD)	IG = 7.9 (1.9)g CG = 7.7 (1.8) [n = 99]	IG 7.8 (2.1) CG 7.7 (1.9) [n = 95]	IG = –0.1 (1.7) CG = –0.0 (1.3)	>0.05h
(Tobe et al. 2006) [30]
DOVE, pre – post cohort	HbA1c mean mmol/L% (CI)	7.4 (7.0–7.8) [n = 94]	7.7 (7.4–7.9) [n = not reported]	+0.3	>0.05 (based on CIs)
(Ralph-Campbell et al. 2006) [31]

Notes:

^aNumber of Indigenous participants.

^bRoubideaux et al. [23] compared patient health outcomes of primary health care services awarded 1) service recognition of educational or integrated quality and 2) those that were in developmental stages of achieving an integrated service. Those in stage 2, more advanced stages of development, were 10% more likely to have a higher proportion of patients with a HbA1c < 7. This finding was not a statistically significant finding.

^cHbA1c predicted from multivariate model.

^dSmith et al. [27] report a net effect (measured from baseline to, at 5 years) of 0.03% increase in HbA1c for Maori and a 0.18% increase for European, both adjusted for age and gender.

^eInterpreted as: Individuals at intervention sites are 26% more likely to have a HbA1c of <7 mmol% compared to those at control sites.

^fP value for difference between intervention and control group not reported.

^gTypographical error in Table 2 of original study manuscript reporting HbA1c level of intervention at baseline.

^hSame result for comparison between groups over time and within group over time.

CCT – Coordinated Care Trial; CI – confidence interval; IDERP – Integrated Diabetes Education Recognition Program; IG – intervention group; IHS – Indian Health Service; mmol/L – millimoles per litre; SLICK – Screening for Limb, I-Eye, Cardiovascular and Kidney; DOVE – Diabetes Outreach Van Enhancement; CG – control group; RR – risk ratio.