Key words
type 1 paediatric diabetes - diabetic ketoacidosis - costs
Introduction
Few data is available regarding medical expenditures associated with acute complications
in patients with paediatric diabetes. In the U.S., expenditures were found to be mainly
increased in patients with diabetic ketoacidotic events (DKA), in particular due to
hospitalisation, with up to 2-fold higher costs [1]. In Germany, all patients with paediatric diabetes have extensive diabetes education,
mainly performed as inpatient care in contrast to other countries [2]
[3]. The risk of DKA has been reported to be lower than in the U.S. [4]. We aimed to estimate the 1-year cumulative incidence of at least 1 DKA and analyse
differences in medical expenditures between paediatric patients with diabetes with
and without DKA in Germany ([Table 1]).
Table 1 Mean total annual diabetes-related costs1 in patients with paediatric diabetes with and without diabetic ketoacidosis (DKA).
DPV, Germany 2007.
|
Mean annual costs per patient [Euro]
|
Cost ratio2
|
|
1Including costs for inpatient and outpatient care, insulin, injection equipment and
pumps, blood glucose self control material
|
|
2Adjusted for age, sex, and diabetes duration
|
|
no DKA
|
3 330 (95%-CI 3 292–3 368)
|
1
|
|
one DKA
|
6 935 (95%-CI 6 627–7 244)
|
2.2 (95%-CI 2.1–2.3)
|
|
more than one DKA
|
10 728 (95%-CI 9 813–11 644)
|
3.6 (95%-CI 3.1–4.1)
|
Methods
We used 2007 data of the continuous diabetes data acquisition system for prospective
surveillance (Diabetes Prospective Documentation, DPV), a computer-based documentation
system which is used in more than 250 German centers providing paediatric diabetes
care. The methods have been described in detail [5]. Briefly, DPV provides all sociodemographic data, laboratory measures, and clinical
data of the routine diabetes care of patients. 2 times annually, all centers send
anonymized data sets to a central data registry. After plausibility checks, data are
centrally analysed. For all 12 001 diabetic patients 0–19 years of age with continuous
documentation in 2007 (52.6% male, mean age (SD) 12.6 (3.9) years, diabetes duration
5.5 (3.4) years, mean HbA1c (SD) 61 (8) mmol/mol (7.7 (1.4)%), 16.4% of the patients
with migration background), all DKAs (defined as hospital admission due to acute hyperglycemia
or pH <7.3) were assessed, as well as costs for diabetes-related treatment. To estimate
costs, the following health care utilization variables were assessed in detail: diabetes-related
hospitalisations, diabetes-related outpatient consultations, mean daily insulin dose,
type of insulin, continuous subcutaneous insulin infusion (CSII), daily blood glucose
self-measurements, treatment with antihypertensive drugs, lipid-lowering agents, and
biguanides. The study took the perspective of the statutory health insurance. Monetary
valuation was performed based on 2007 prices, using mean retail prices of the official
German Index of Medicines, DRGs, and medical service fees of the statutory health
insurance’s price scale for outpatient care. If the subjects took part in a Disease
Management Program (DMP), corresponding mean DMP fees were included. Associations
between costs and DKA were estimated using log-linear models, adjusting for age, sex,
and diabetes duration.
Results
457 (3.8%) patients had at least 1 DKA during 2007 (389 one, 68 2 or more). Patients
with DKA were more likely to be in pubertal age (proportion of patients aged 10–14
years: no DKA: 41%, 1 DKA: 50%, 2 or more DKAs: 54%, p<0.001), and to have a migration
background (16%, 21%, and 18%; p=0.02).
Mean total annual costs per patient in 2007 were € 3 489 (95%-CI 3 448–3 530) (US$
(30.06.2007) 4 701 (4 646–4 757). Costs were significantly increased in patients with
DKA. Total annual costs in patients without, with 1, or ≥ 2 DKAs were € 3 330 (3 292–3 368),
€ 6 935 (6 627–7 244), and € 10 728 (9 813–11 644), respectively, with largest differences
for hospitalisation costs (€ 693, € 4 145, € 8 092). Cost ratios of total costs in
patients with 1 or ≥2 DKAs compared to patients without DKA, adjusted for age, sex,
and diabetes duration, were 2.2 (95%-CI 2.1–2.3) and 3.6 (95%-CI 3.1–4.1), respectively.
Additional adjustment for migration background did not alter results.
Discussion
Several limitations have to be considered. Our analysis was not population-based but
represents the majority of German paediatric diabetes care facilities. As a consequence,
the results of the study may not be representative for all patients with type 1 diabetes
less than 20 years of age in Germany. However, the DPV documentation system is used
in all kinds of medical facilities. Also, when comparing our study population with
a population-based cohort, course and outcomes of diabetes care have been shown to
be very similar [3]. There may be some uncertainty in our cost estimation, e. g. we have applied the
general items of the medical fee schedule for calculating outpatient care costs. However,
we do not believe that this should have biased the association between DKA and costs.
The strength of our study lies in using a comprehensive computer database containing
data from continual, standardised documentation of routine care. Thus, our health
care utilization data can be assumed to be rather valid and unbiased. Nevertheless,
under- or overreporting cannot be excluded. On the other hand, many of the clinics
use this documentation system for administrative purposes and regular data validity
tests are being performed on the DPV data. Hence, we consider that our estimates are
valid.
Thus, in Germany, paediatric patients with DKA had up to 3.6-fold higher diabetes-related
costs compared to those without DKA. This cost excess was much higher compared to
the U.S. study [1]. In our study, the proportion of patients with DKA was much lower compared to the
U.S. analysis (3.8% vs. 14.9%). The lower frequency of DKA in Germany may be due to
a better and easier access to and higher utilisation of diabetes education. Increased
hospitalisation and higher costs among patients sustaining DKA may be due to poor
diabetes control. This hypothesis is supported by HbA1c values observed in patients
without, with 1, or ≥2 DKAs (61±9, 77±3, 89±4 mmol/mol, corresponding to 7.7±1.3,
9.2±1.9, and 10.3±1.8%, respectively). Intervention strategies should be developed
to reduce DKA and resulting hospital admission in paediatric patients in order to
reduce costs, but also improve quality of life.
