Fusello Massimo,Scalisi Andrea,Battaggia Alessandro,Cancian Maurizio,Franco Novelletto Bruno,Michieli Raffaella,
Specialist Advice Does Not Modify the Risk of Death of Diabetic 2 Patients
Journal of Integrative Cardiology Open Access
2019
2674-2489
http://dx.doi.org/10.31487/j.JICOA.2019.04.02
https://www.sciencerepository.org/specialist-advice-does-not-modify-the-risk-of-death-of-diabetic-2-patients_JICOA-2019-4-102 
Abstract: Context: A recent meta-analysis (Bonora and coll.) reports benefits on death-risk for Italian diabetic
patients mainly followed by the diabetic clinics of the National Health Service.
Aims: A) to do a critical appraisal of the meta-analysis by Bonora and coll. B) to verify its results conducting
a controlled cohort study based on clinical records of a primary care setting.
Methods: (A) We evaluated the meta-analysis by Bonora through AMSTAR II checklist and the trials
recruited in the review through ROBINS-I tool. (B) We analysed a cohort of diabetes 2 patients living in
Veneto (Italy) and followed from 1/1/2009 to 12/31/2017 to compare the risk of death of a control group
(i.e. never followed by specialists) with that of another two groups (i.e. respectively, followed by one
specialist visit or by at least two visits in the last three years). We used a time-to-event approach (Cox
model) for the main analysis; complementary designs were also tested (Restricted design and Matched
design). Statistical adjustments were made both through Multivariate Cox regression and Propensity score.
For the adjustments, the covariates considered were: age, sex, severity of diabetes, comorbidity, laboratory
values, duration of diabetes and drugs use.
Results: (A) The meta-analysis by Bonora shows to be affected by serious pitfalls (B) A cohort of 6530
diabetic patients (none visit: n=3441; one visit: n=947; two or more visits: n=2142) was followed for a mean
of 7.32y. Main multivariate analysis was not able to demonstrate any difference in mortality between groups
exposed or not exposed to specialist advice: one visit HR=1.01 (0.98-1.03); two or more visits HR=1.12
(0.88-1.43). These results were confirmed by all other analytical approaches.
Conclusion: Mortality in diabetes2 is not influenced by specialist consultant. Our results differ by those
reported by the meta-analysis because of our better adjustment for prognostic and confounding factors. Most
of diabetes 2 patients should be entrusted with confidence to primary care facilities.Keywords: Primary care, referral and consultation, Diabetes mellitus type 2, cox model, propensity score, specialist advices, mortality