The aim of this study was to shed light on the care situations of patients with initial SCS therapy and to look at the effects on pain diagnoses and comorbidities as well as cost structures in a period from one year before to three years after the start of therapy. Due to the variety of international literature on the cost effectiveness of different types of devices, the further question was whether a difference could also be recognized in the German context.

methodology
Study design
The analysis was carried out retrospectively in a longitudinal section using an anonymous, age- and gender-adjusted GKV routine data set, which with ~five million insured persons represents an approximately six percent, representative sample of the German population. The InGef research database is roughly comparable to the German population in terms of morbidity, mortality and drug consumption (Anderson/Walker 2015). The aim of this study was to follow the patients with an initial implanted neurostimulator from the InGef research database. For this purpose, insured persons were taken from the database with the OPS codes 5-039.e0, 5-0339.e1 and 5-039.e2 for the pooled index years 2013/2014 and, for each patient, three years from the day of implantation (FU1, FU2 and FU3 , 365 days each). Patients were not allowed to have implantations, changes, explantations or revisions of a stimulator in the previous year of the index (T-1) in order to ensure that it was an initial start of therapy. Patients had to be continuously insured throughout the observation period to avoid loss-to-follow-up. The study population was divided into a rechargeable device group and a non-rechargeable device group.

Statistical analysis
In order to describe the study population, age- and gender-specific information was collected for the patients. Furthermore, patient characteristics as underlying diagnoses at the time of implantation and comorbidities in the follow-up period were determined using the ICD-10-GM classification. The health costs of the study population were recorded in the following areas: total costs, inpatient, outpatient, medication, medical supplies and aids and sick pay. For the inpatient area, pain-related costs were queried using inpatient main diagnoses. Percentage increases and decreases were calculated comparing the follow-up period per year and between the year before implantation (T-1) and the third year after implantation (FU3). Significances were calculated using McNemar's test for time comparison between T-1 and FU3 for diagnosis development. For the cost development, significance was determined using the t-test for the time comparison between T-1 and FU3 and for the subgroup comparison for the follow-up year FU3. Patient numbers below five were not given for data protection reasons.
Results
study population
There were a total of 4.9 million insured people in the InGef research database between 2012 and 2017; Of these, 3.1 million were consistently insured. 153 patients received an implantation of a neurostimulator in the index years 2013/2014, three patients had already undergone a corresponding intervention in the previous year and were excluded. The final study population was 150 patients with an initial neurostimulator, of which 77 patients (51%) received a non-rechargeable neurostimulator and 73 patients (49%) received a rechargeable device (Fig. 2). Ten patients died during the follow-up period (6.7%), five of them in each of the subgroups (Fig. 1).
The mean age of the patients was 62.5 years (SD ± 12). Patients with a non-rechargeable device were on average one year older than the other group. The gender distribution was almost balanced with 51% women and 49% men. The age of the female patients was, on average, 62.5 years, one year younger than that of the males. More women received a rechargeable neurostimulator (55%) and more men (58%) a non-rechargeable device. The diagnoses at the time of implantation showed a heterogeneous picture when comparing the care sectors (Table 1). In the inpatient sector, the diagnosis of postlaminectomy syndrome/FBSS (ICD-10-GM M96.1) was mainly made on the index day (implantation) in 49 patients (33%). In the outpatient sector, most diagnoses in the quarter were for unclassified pain (R52) (104 patients). Other frequently made diagnoses in the outpatient sector related to back pain (64%) and other intervertebral disc damage (73%) (Table 1).

Pain and comorbidities
The overall population showed a decline in all diagnoses after implantation (Fig. 2/Table 2). For those in outpatient and inpatient
For the pain diagnoses M54 (back pain) and R52 (unspecific pain) that were most frequently made in the nary sector, the percentage decrease in the period before implantation (T-1) up to three years after implantation (FU3) was -18% and -10%. The diagnosis of G47 (lower extremity mononeuropathies) was made pre-index in 31 patients and recorded the highest percentage decrease with -58%. In the case of back pain, the number of patients with the diagnosis M54.4 (lumbar sciatica) fell from 85 patients to 48 patients (-44%; p<0.001). The percentage decrease for the diagnoses M54.5 (low back pain) from -43% (p<0.001) and M54.16 (lumbar radiculopathy) from -35% (p<0.05) was significant. The majority of patients with "unclassified pain" (R52) received the diagnosis R52.2 (other chronic pain). The percentage decrease was significantly lower with -3% (p=0.86) than with the diagnosis R52.1 (chronic uncontrollable pain) with -40% (p<0.05). The greatest reduction in back disorders and comorbidities was seen in M96.1 (FBSS) with -42%. In the case of back diseases, "other disc damage" (M51) was the most common diagnosis (103 patients, 69%). The percentage drop to FU3 was -24%. The number of patients diagnosed with spondylosis (M47) decreased from 89 to 75 (-27%; p=0.002). "Lower extremity pain" (M79.6) was seen in 14% (T-1) of the study population, this patient number hardly changed over the years. A majority of patients with a rechargeable neurostimulator showed a continuous decline in diagnosis, whereas patients with non-rechargeable devices frequently experienced a rebound in FU3.
The number of patients with specific diagnoses and rechargeable devices was higher on average. Diagnosis of FBSS, patients with rechargeable devices showed a percentage decrease of -47% (p<0.001), the other group recorded a smaller decrease of -35% (p<0.05). For the diagnosis of M51 - other intervertebral disc damage, the percentage decrease in both patient groups was -20% (RC: p<0.002; NRC: p<0.05). Patients with non-rechargeable neurostimulators and a diagnosis of M47 (spondylosis) experienced a slightly higher decrease in -29% (p<0.05). The number of patients diagnosed with M48 (other spondylopathies) again decreased more with the rechargeable devices (-38%; p<0.05). Both patient groups show a decrease in diagnoses for back pain (M54) (rechargeable -13%; nonrechargeable -23%), unclassified pain (R52) (f -12%; nw -7%), and lower extremity mononeuropathies (f -56%; nw-58%). There is a difference in the diagnoses R52.1 (chronic uncontrollable pain) and R52.2 (other chronic pain). The number of patients with rechargeable devices shows a steady decrease for both diagnoses, the number of the other patient group increases again for the diagnosis R52.1 in FU3. The diagnosis R52.2 shows a percentage increase of +7% (p=0.21) year-on-year (Fig. 2).
Cost development overall and by device type
The total costs per patient rose sharply in the year of implantation due to the device costs (Fig.3/Tab.3). In the following years, costs fell below the cost level before implantation (-6%; p=0.98). Pain-related costs accounted for a large proportion of expenditures in the inpatient sector. The year-on-year percentage decrease was -19% (p=0.3) in the inpatient sector and -24% (p<0.05) in the outpatient sector. Medication and device costs showed a percentage increase of +25% (p=0.27) and +138% (p<0.05) for the study population. Patients with rechargeable neurostimulators showed a greater decrease in overall costs and a weaker increase in medication and assistive device costs. While the percentage decrease in total costs in the rechargeable device group was -24% (p=0.10), the costs for the non-rechargeable devices increased again in the third year after implantation (+9%; p=0.33). This difference was significant at p<0.05.
A recurring pattern is the renewed increase in FU3 for total, inpatient and pain-related inpatient costs, related to non-rechargeable devices. The total costs of patients with rechargeable devices decreased to an average of 9,347 euros (± 10,158) in FU3, below the baseline level of 12,267 euros (± 7,710) in T-1. In contrast, the costs of patients with non-rechargeable neurostimulators increased to 14,878 euros (± 19,082) in FU3 above the original level (13,648 euros ± 11,514). With a percentage decrease of -4% (p=0.86) for inpatient costs for non-rechargeable neurostimulators and a corresponding decrease of -36% (p<0.05) for re-
A noticeable difference was observed between the rechargeable devices. The cost difference in FU3 of the stationary costs of 4,998 euros (± 7,701) for the rechargeable devices to 8,299 euros (± 11,906) for the non-rechargeable devices was not significant (p = 0.06). In T-1, 21 people (14%) of the study population were eligible to receive sick pay (Table 4). This number fell to 13 people for the entire follow-up period (FU1 to FU 3).
Sickness benefits fell from an average of 2,004 in year T-1 to an average of 561 euros in the entire follow-up period (-72%). A strong difference within the comparison groups was not observed (Table 3: Fig. 3, Table 4).
discussion
Compared to the international SCS literature, the study population determined from the data set appears to be slightly older, with an average age of 62.5 years. Regarding gender distribution, publications report a small female majority, which corresponds to the results shown (van Buyten et al. 2017; Hornberger et al. 2008; Zucco et al. 2015). A systematic review with 11 studies found an average age of 56 years for SCS patients and a female majority of patients with 53% (Odonkor et al. 2019). Due to the patient characteristics collected and the widespread use of the GKV routine data set, the results can be viewed as representative (Anderson/Walker 2015). The study population has various pain diagnoses in the inpatient and outpatient sectors at the time of implantation. The larger proportion of diagnoses with an unspecific or non-classifiable character in the outpatient sector is significant. This could be due to the diverse manifestations of chronic pain, but also to the primary family doctor care and the resulting lack of specialist attention to the underlying disease. With the exception of the FBSS, the number of diagnoses requested for SCS therapy is higher in the outpatient than in the inpatient area. The diagnosis of a frustrating back operation is made after a long period of suffering, with multiple back operations and many visits to the doctor. It is therefore understandable that patients arrived at the specialist inpatient sector shortly before the implantation of the neurostimulator and received the corresponding diagnosis as an indication for therapy. It can be observed that SCS therapy has positive effects on patients with back diseases and pain diagnoses and ensures a significant reduction in diagnoses.
The R52 pain diagnosis is the leading diagnosis and, according to the S3 guideline for spinal cord stimulation, the first indication for SCS therapy (Baron et al. 2013). The diagnostic acceptance of the FBSS must be emphasized because this is a patient client who has often had to go through a path of suffering and SCS therapy often represents a last resort. The proportion of FBSS diagnosis in the entire diagnostic spectrum at the time of implantation is consistent with international studies (Kumar et al. 2013, 1998).
The group comparison results indicate the special characteristics of the non-rechargeable devices. Patients using rechargeable stimulators show a continued decline in pain-related diagnoses and back disorders. Patients with non-rechargeable neurostimulators show a resurgence in the last year of observation, particularly in FBSS and chronic uncontrollable pain. Non-rechargeable neurostimulators have a lifespan of two to five years (Hornberger et al. 2008). The resurgence in diagnostic metrics and costs last year could be related to battery degradation and an associated loss of effectiveness. The costs incurred by SCS patients decrease continuously after an increase in the year of implantation and even fall below the level before therapy at the end of the observation period (-16%). Outpatient and inpatient costs are steadily decreasing, particularly inpatient pain-related costs falling significantly (-43%).
The decrease in total costs is more pronounced in patients with rechargeable devices and again, unlike in other patients, it is continuous. Overall costs and inpatient pain-related costs increase again for patients with non-rechargeable devices in FU3. This result strengthens the assumption of a loss of effectiveness due to the battery fading in year 3, as patients have to seek more inpatient treatment again. An increase in the cost of medication and expenses for medical devices and aids is more evident in patients with non-rechargeable devices. The total costs for patients with non-rechargeable neurostimulators in FU3 exceed the baseline level from the year before starting therapy. These results are consistent with Hornberger et al. to be brought. There, a cost-effect analysis found that rechargeable devices save costs in patients with FBSS, although the initial costs were higher than non-rechargeable devices. This was also due to the fact that rechargeable devices with a lifespan of 10 to 25 years were changed between 2.6 and 4.2 times less and patients with non-rechargeable devices experienced a change after just four years (Hornberger et al. 2008). Further cost-effectiveness analyzes showed that SCS therapy with rechargeable devices is more cost-effective in FBSS and complex regional pain syndrome (Kemler et al. 2010; Taylor et al. 2010). The present results also show that patients with SCS therapy are brought back to work because the sick pay for patients who are entitled to sick pay decreases.
limitations
The analysis does not include a closer look at the patient's history and the basis for decisions made by doctors. Decisions for or against a device type cannot therefore be traced. Aspects such as the number of back operations before starting therapy or patient compliance can influence doctors and patients in their choice of device type. In addition, there are certain codes that ensure allocations by the health insurance companies via the morbidity-oriented risk structure compensation and are therefore used differently. The analysis is based on secondary data, so a decrease in diagnoses represents an approximation for quality of life. Primary data, such as those generated in patient surveys, were not collected.
conclusion
The results of the analysis show that spinal cord stimulation therapy increases the quality of life of patients, as a significant reduction in relevant pain diagnoses and diseases can be seen. SCS therapy has a positive effect on the costs in the statutory health insurance system. The present results also show that patients with SCS therapy are brought back to work because the sick pay for patients who are entitled to sick pay decreases. Even if the costs for neurostimulation are initially high, the benefits of the devices for patients and health insurance companies can be clearly seen through the reduction in diagnosis and costs. An advantage can also be stated for rechargeable devices, as they show a steady and more significant decrease in diagnoses and costs. The increase in the key figures in the third year, which is characteristic of the non-rechargeable devices, strongly indicates the shorter life cycles and the associated declining freedom from pain of the patients. <<

Epidural spinal cord stimulation (SCS – spinal cord stimulation) is a therapeutic method for patients with various forms of neuropathic, sympathetically mediated or ischemic pain and PDN (peripheral diabetic neuropathy) (Baron et al. 2013; Deer et al. 2014; Gilner/Cegla 2019). ). It takes place via implantable systems consisting of electrodes and neurostimulators (IPG – implantable pulse generator). Electrical impulses, which are delivered from the generator via electrodes to the spinal cord in different frequency forms, lead to a reduction in pain in patients. Low-frequency stimulation occurs between 40 and 100 Hz and high-frequency up to 10 kHz (Fishman et al. 2019). The systems can be implanted in a one-step or two-step process. In the two-stage procedure, the electrode is first implanted and tested using an external stimulator between three days and three weeks. If the test is successful, the internal pulse generator is implanted (Baron et al. 2013). A distinction is made between rechargeable and non-rechargeable devices. The latter must be explanted again after some time and replaced with a new device. A neurostimulator is recharged externally through the skin. The lifetime of the devices varies depending on the manufacturer. The lifetime of non-rechargeable devices is between two and five years (Verin 2012). Rechargeable neurostimulators remain in the body for between 10 and 25 years (Elridge et al. 2011). Studies assigned a superiority to rechargeable neurostimulators. Although the initial costs were higher than the non-rechargeable devices, there was better cost-effectiveness and a reduced risk of complications due to fewer device changes required (Fishman et al. 2019; Hornberger et al. 2008).