In total, 18.555 million patients were treated in German hospitals; of these, 4.591 million patients underwent elective surgery. The average cost of all patients was €3516 per patient and €5399 for the target population. The average hospital stay was 6.29 days for all patients and 6.40 days for the target population. A total of 3.151 million blood bags were given as RBC transfusion in 2015 (all patients). Of these, 1.026 million RBC blood bags were given to the target population. See Table 2 for full results for the ‘all patients’ and ‘elective surgery’ populations.
A total of 4,591,060 patients who had undergone elective surgery during 2015 were identified in DRG-Statistic 2015 for inclusion in the study. Of these, 308,190 patients were identified as having a diagnosis of anaemia (ICD D50.0, D50.8, D50.9, D62, and E61.1). In addition, 227,390 patients (both IDA and non-IDA) aged between 0 and 99+ years who underwent elective surgery and had received a RBC transfusion as part of the procedure were identified from the database.
The target population of interest comprised two groups: (1) 29,170 patients (patients with IDA: ICD D50.8 and D50.9) who had undergone elective surgery and were diagnosed as preoperatively iron deficient, who had either received (24%) or not received (76%) a RBC transfusion during their elective surgery procedure (a + b in the original allocation); (2) the expanded target population within the health economic model, which consists of the identified, undiagnosed 29,714 patients with IDA for whom a PAM was hypothetically implemented (Fig. 2).
Estimation of Key Epidemiological and Economic Parameters Using Database Information
Association Between IDA and Mortality: If a patient with IDA underwent elective surgery, their risk of death was calculated to have been 3.6-times higher than that of a patient without IDA (RR 3.630; 95% CI 3.401, 3.874). Table 3 details the patient numbers and full RR and odds ratio (OR) results.
Association Between RBC Transfusion and Mortality: If a patient received a RBC transfusion during a hospital stay after elective surgery, the risk of death was 24.6-times higher than in a non-transfused patient (RR 24.593; 95% CI 24.121, 25.075) (Table 3).
Association Between IDA and RBC Transfusion: If a patient with IDA underwent elective surgery, the risk of receiving a RBC transfusion was 5.0-times higher than that of a patient without IDA (RR 5.031; 95% CI 4.928, 5.136) (Table 3).
Average Cost per Patient: The average cost of treatment was €7883 in patients with IDA (without RBC transfusion) compared with €4560 for a patient without IDA; this equates to a 173% higher cost for patients with IDA (Table 4). The average cost of treating a patient with the two risk factors (IDA and RBC transfusion) was €21,744, compared with €4560 for a patient without either risk factor; this equates to a 477% higher cost for patients with both risk factors (Table 4).
Average Number of Days in Hospital per Patient: The average number of days in hospital for a patient with IDA (without RBC transfusion) was 11.45 days, compared with 5.46 days for a patient without risk factors (Table 4). The average number of hospital days as an inpatient for an individual with IDA and RBC transfusion was 27.83 days, compared with 5.46 days for a patient without risk factors (507%) (Table 4).
Obtaining New Allocation of Patients After Implementation of PAMs
The impact of implementing PAMs is shown in Fig. 2 and the final figures are shown in Table 5.
Step 1 Using the estimation that 13.5% of patients are undiagnosed with IDA, in Table 5 13.5% of the 220,100 patients who had received a RBC transfusion but were not treated for IDA would transfer to cell 'a' as diagnosed with IDA; 29,714 patients were reallocated to the IDA population row and were added to the existing 7080 patients (total of 36,794 patients).
Step 2 From the findings of Froessler et al. , a value of 87.5% of the 36,794 patients was considered to be the population that would no longer require a RBC transfusion after the implementation of PAMs. Consequently, from cell ‘a’ 32,194 patients were reallocated to cell ‘d’ (Table 5).
These estimations suggest that the implementation of PAMs would reduce the number of patients with IDA receiving a RBC transfusion to 4599, and the total number of patients receiving a RBC transfusion would reduce from 227,180 to 194,986 after the implementation of PAMs (Table 5). The undiagnosed and allocated patients were included in these model results.
Economic Impact of Implementing PAMs in Germany
As previously indicated, the estimation of potential hospital direct cost savings was derived from two sources (average cost per patient and average number of hospital days) and combined to give a total hospital cost saving of implementing PAMs.
Estimation of Hospital Cost Savings Using Mean Cost per Patient
The average costs per patient for elective surgery treatment both with and without a RBC transfusion in patients with and without IDA are shown in Table 6; this also details the total hospital direct costs of treatment of patients with and without IDA before and after the implementation of PAMs. The total hospital costs for 2015 were €65,247 million . It was estimated that the implementation of PAMs in the German population studied would have resulted in an annual hospital direct cost saving of €536 million.
Estimation of Hospital Cost Savings Using Average Number of Hospital Days
The average cost of a hospital stay for a patient having elective surgery in Germany in 2015 was €844 per day (RDC of the Federal Statistical Offices of the Länder, DRG-Statistic 2015; own calculations ). The estimated number of annual hospital days for patients having elective surgery was projected to be reduced by 596,070 days after the implementation of PAMs (Table 7). This equates to an estimated annual additional cost saving of €503 million for 2015 if PAMs had been implemented. These additional cost savings result from a lower utilization/occupancy rate of hospital beds. The cost savings from a lower occupancy rate can then be used to improve efficiency. There is no double counting of costs, as the avoidable direct hospital costs refer to reduced treatment costs (e.g. fewer complications) and the avoidable costs relate to reduced inpatient infrastructure use.
Estimation of Total Hospital Cost Savings of Implementing PAMs
The aggregated impact on hospital cost savings accounts for €1039 million (€536 million + €503 million). It was estimated there were 58,884 undiagnosed and existing patients with IDA in the German population during 2015. The average cost per patient of IDA therapies in Germany during 2016 was €176.68 (with no differences in costs reported between 2015 and 2016, see Table 1) [57, 58]. The annual cost of implementing PAMs in Germany was estimated to be €10.4 million. Therefore, the net hospital direct cost saving of implementing PAMs in Germany for 2015 would have been €1029 million. This accounts for approximately 1.58% of total hospital direct costs of the German healthcare system.
The annual cost of implementing PAMs in Germany is estimated to be €10.4 million, whereby 58,884 patients with IDA would receive IDA therapy preoperatively. If, instead of a 500 mg IV iron dose, 1000 mg was used throughout (see footnote to Table 1), the cost would be €17.3 million.
Epidemiological Impact of Implementing PAMs in Germany
On the basis of the new patient allocation after implementation of PAMs, it was projected that a total of 3036 hospital deaths would have been avoided if PAMs were implemented in Germany in 2015 (Table 8).
Univariate Sensitivity Analysis
Estimation of Patients Undiagnosed/Untreated for Preoperative IDA
A univariate sensitivity analysis was conducted to assess the robustness of the results regarding the evidence-based data used to inform the estimate of patients having elective surgery in 2015 who were undiagnosed/untreated for preoperative IDA. In order to be able to estimate the sensitivity, ranges were defined for the individual assumptions. The entry into and exit from the bandwidth are always 10% below and 10% above the known value, respectively. The following ranges were formulated for the three evidence-based assumptions used in Fig. 1:
Range for assumption 1 [0.10; 0.50], step size [0.05]
Range for assumption 2 [0.40; 0.70], step size [0.05]
Range for assumption 3 [0.80; 1.00], step size [0.05]
If the lowest and highest values of the ranges of assumptions 1–3 are multiplied by each other (lowest range: 0.10 × 0.40 × 0.80; highest range: 0.50 × 0.70 × 1.00), a common range is obtained: [0.032; 0.350]. The following values are assumed for the univariate sensitivity analysis for calculation: [0.032; 0.135; 0.350].
For these parameter values, results were then derived from the new distribution of patients after the implementation of PAMs. This means that the patient numbers were recalculated with the values 0.032–0.350. According to the patient redistribution model, the patients with undiagnosed IDA were redistributed from population ‘c’ to population ‘a’. The results are presented in Table 9.
Estimation of Reallocation of Patients with IDA Who No Longer Need RBC Transfusion After Receiving PAMs
A sensitivity analysis was also conducted on the data supporting the assumption used to estimate the reallocation of patients with IDA who no longer need a RBC transfusion after receiving PAMs (Fig. 2). The following range was formulated for this:
The following values are assumed for the univariate sensitivity analysis: [0.80; 0.85; 0.875; 0.90; 0.95]. The minimum scenario (0.80), the base case scenario (0.875), and the maximum scenario (0.95) were determined. The results for these three scenarios are presented in Table 9.
The results of the sensitivity analysis (minimum, base case, and maximum scenario) for IDA treatment costs, avoidable direct hospital costs, avoidable hospital days (additional cost savings), total cost savings, and total cost savings as a percentage of the total national hospital budget for Germany are presented in Table 10.
Avoidable Deaths Following Implementation of PAMs
A sensitivity analysis was conducted to establish the minimum, base case, and maximum scenario for the number of deaths which could have potentially been avoided if PAMs had been implemented in Germany in 2015. As described earlier, the following ranges were formulated for this:
The minimum scenario (0.032 and 0.80), the base case scenario (0.135 and 0.875), and the maximum scenario (0.350 and 0.95) were determined. The percentage mortality for each of the groups was then used to project the number of avoided deaths in the new allocation of patients based on the implementation of PAMs. In the minimum scenario, 1083 deaths were avoidable, in the base case scenario the number of avoidable deaths was 3036, and in the maximum scenario it is estimated that 7518 lives could have been saved with the implementation of PAMs.