Background and aim: To assess the validity of hospital discharge diagnoses (HDDs) as a means for the surveillance of diabetes-related lower extremity amputations on a population basis, and to compare the demographic and clinical characteristics of diabetic and non-diabetic amputees.
Methods and results: All of the 1996 hospital discharge diagnoses reporting ICD-9-CM codes 84.10-84.19 were reviewed in the Campania Region, an area in Southern Italy with 5.7 million inhabitants. Diabetes was defined as any concomitant 250 ICD-9 code. The completeness of ascertainment and accuracy in reporting diabetes were validated in a sub-sample of cases by matching their HDDs with the register of buried limbs and the clinical records of the hospital. During the study period, 701 individuals underwent 738 amputations, 330 (47.1%) of whom had diabetes mentioned in their discharge diagnosis. All of the amputations reported in the buried limbs register were identified by the HDDs, whereas some minor amputations recorded in the HDDs were not reported in the registers. On the basis of the clinical charts, the diagnosis of diabetes was underreported by 10%. The proportion of females was significantly higher among the diabetic amputees (41% vs 35%), whereas the proportion of people aged less than 50 years was significantly lower (5% vs 13%) (p < 0.002). The duration of hospitalisation was similar for the diabetics and non-diabetics (23 +/- 20 vs 22 +/- 18 days). Minor amputations were significantly more frequent in the individuals with diabetes (38.8% vs 29.1%; p < 0.001), as was reamputation (7.2% vs 2.9%; p < 0.01).
Conclusion: These findings confirm the markedly increased risk of amputation among diabetics and demonstrate that HDDs represent a suitable procedure for the surveillance of lower extremity amputations.