Costs of Weaning Failure: A Prospective, Multicentre, Controlled, Non-Randomised, Interventional Study on Economic Implications for the German Health Care System

Background: Intensive care patients with respiratory failure often need invasive mechanical ventilation (IMV). With increasing population age and multimorbidity, the number of patients who cannot be weaned from IMV rises as well. Up to 85% of these patients have no access to a certified weaning centre. Their medical care is associated with impaired quality of life and high costs for the German health care system. Objectives: This study examined the weaning outcome of patients in certified weaning centres after a primarily unsuccessful weaning attempt in order to calculate saving expenses compared to patients on long-term IMV in an outpatient setting. Methods: In this multicentre, controlled, non-randomised, interventional, prospective study, 61 patients (16 from out-of-hospital long-term IMV, 49 from other hospitals) were referred to a certified weaning centre for a second weaning phase. The incurred costs after 1 year of the latter were compared to insurance claim data of patients who were discharged from an acute hospital stay to receive IMV in an outpatient setting. Results: In the intervention group, 50 patients (82%) could be completely weaned or partially weaned using non-invasive ventilation, thus not needing IMV any longer. The costs per patient for weaning and out-of-hospital care in the intervention group were EUR 114,877.08, and the costs in the comparison cohort were EUR 234,442.62. Conclusions: Early transfer to a certified weaning centre can increase weaning success and reduce total costs by approximately EUR 120,000 per patient in the first year. Given the existing structural prerequisites in Germany, every patient should have access to a weaning centre before being transferred to long-term IMV, from a medical and health economical point of view.

hospitals) were referred to a certified weaning centre for a second weaning phase.The incurred costs after 1 year of the latter were compared to insurance claim data of patients who were discharged from an acute hospital stay to receive IMV in an outpatient setting.Results: In the intervention group, 50 patients (82%) could be completely weaned or partially weaned using non-invasive ventilation, thus not needing IMV any longer.The costs per patient for weaning and out-of-hospital care in the intervention group were EUR 114,877.08, and the costs in the comparison cohort were EUR 234,442.62.Conclusions: Early transfer to a certified weaning centre can increase weaning success and reduce total costs by approximately EUR 120,000 per patient in the first year.Given the existing structural prerequisites in Germany, every patient should have access to a weaning centre before being transferred to long-term IMV, from a medical and health economical point of view.

Introduction
Medical progress, increased life expectancy [1,2], and higher multimorbidity lead to a growing number of patients with invasive mechanical ventilation (IMV) [3,4].In Germany, the number of patients using home IMV has doubled between 2008 and 2019, with 17,958 initiations and 49,140 controls in 2019 [5].Surviving acute intensive care therapy can lead to prolonged and complicated or impossible weaning from IMV [6].In Germany, affected patients with failed weaning from IMV are often discharged from hospital acute intensive care units to their own homes.There, they receive 24-h intensive medical care from outpatient intensive care teams.
In addition, care can be provided in special respiratory care homes, or the patients are discharged into specialised residential facilities that formally correspond to their own domesticity, where they are also treated by outpatient intensive care services.In many cases, the outpatient intensive care services or aid suppliers take the lead in the discharge management from the hospitals since most intensive care units are not familiar with the modalities of the transition to invasive out-of-hospital ventilatory therapy.Health care costs increase due to the need for intensive care [7].
According to internal analyses by some insurance providers, average costs for care per year and persons affected are between EUR 180,000 and 500,000.Thus, with approximately 20,000 currently suspected patients, around EUR 3-4 billion outpatient treatment costs are attributable to the treatment of the disease "chronic ventilatory insufficiency with invasive non-clinical ventilation."And this number is still increasing [8].
In the recent international, multicentre, observational study WEAN SAFE with 5,869 critically ill adult patients, only 65% of those who received IMV for more than 2 days were successfully weaned from IMV on day 90 [9].The quality of medical service for out-of-hospital intensive careindividual care at home or specialised residential groups for ventilated patientsvaries immensely since there is neither a standard of qualification for nursing staff nor the required equipment, nor is there a regular quality inspection infrastructure.Patients in the residential groups are often treated by general practitioners without special training or experience in respiratory medicine so that insufficient specialised medical care can be assumed [10].Due to a new law in Germany (Intensivpflege-und Rehabilitationsstärkungsgesetz, GKV -IPReG), changes to this situation are being made.But still the potential of some long-term ventilated patients to be weaned from the ventilator is neither rec-ognised nor financially favourable to the operator of a residential group or an outpatient care service.There are a number of specialised hospitals that offers units to wean patients off IMV following acute medical care [11].
These so-called weaning centres can be certified by the German Respiratory Society (DGP), if they meet the required criteria [12].To ensure that the weaning potential of each individual patient can be assessed prior to discharge to outpatient respiratory care, these centres will be included in the consideration of these new developments.
This multicentre study examines the efficiency of a structured and graded treatment concept for patients with prolonged weaning and compares the costs of outpatient care to this new weaning attempt.Results of clinical outcome in prolonged weaning have been published before with only secondary consideration of the pecuniary aspects [13].The focus of this study on economic aspects is therefore novel.

Study Design
In a multicentre, controlled, non-randomised, interventional study, the outcome of weaning in certified weaning centres after a previous primarily unsuccessful weaning attempt was prospectively investigated from December 8, 2015, until December 7, 2017 [13].The primary outcome was the percentage of successfully weaned patients.Secondary outcomes were 1-year survival, predictive analyses, and a comparison of health costs.The current analysis focuses on one of the secondary outcomes and compares the costs incurred after 1 year with health insurance claims of patients who have been unsuccessfully weaned.The study was conducted at the weaning centres of universities or universityassociated hospitals in Germany under pulmonological guidance of the University Clinic Greifswald, the Lung Center Cologne Merheim, and the Thorax Clinic Heidelberg.These are among the leading scientific weaning centres in Germany.

Patient Population
This prospective study included 2 kinds of patients in the intervention group.The first were to be discharged as nonweanable from medical facilities excluding DGP-certified weaning centres to out-of-hospital long-term IMV with already initiated discharge management into outpatient long-term IMV.These patients were referred to the weaning centre by the intensive care unit that classified the patients as non-weanable with the question whether weaning could be successful at the centre.
The second were secondarily transferred from out-of-hospital IMV to a DGP-certified weaning centre with the specific question of weaning potential.These patients were either seen by physicians of the centre as consultation or were admitted because of an acute complication.Patients' basic characteristics are found in Table 1, causes leading to ventilation in Table 2, and comorbidities in Table 3.In order to be able to make a cost calculation, data records from the Allgemeine Ortskrankenkasse (AOK), a German statutory health insurance, were used, consisting of patients who were not weaned after the acute weaning phase and were discharged with IMV in the out-of-hospital ventilation area (care at home or in residential groups).In the comparison group, 21 patients were intensively cared for at home, 12 patients in a ventilation resident group, and 28 patients in a respiratory care home.These were matched with the intervention group in terms of gender and age.The comparison group was randomly chosen.The allocation to the intervention group was made prospectively according to the defined inclusion criteria in the individual centres.

Inclusion and Exclusion Criteria
The prerequisite was completion of weaning classified as unsuccessful.Further inclusion criteria were the existence of a signed participation information and the consent of the patient, legal representative, or authorised representative, as well as sufficient knowledge of the German language.Exclusion criteria were illness with an estimated prognosis <6 months, severe irreversible neurological diseases (e.g., amyotrophic lateral sclerosis), coma, or loss of consciousness.There were no exclusion criteria for age, underlying diseases, or comorbidities.Patients who could not enter a prolonged weaning process due to need for emergency treatment or an unstable health status were excluded.

Intervention
Intervention took place in the certified weaning centres as a renewed weaning phase called "weaning phase 2," which was carried out under inpatient conditions, regular medical monitoring, intensive respiratory specialist care, physiotherapy treatment, and mobilisation.Weaning was carried out according to the national recommendations of the DGP [7].

Measurements and Outcomes
If patients did not require further IMV for at least 72 h, they were assigned to the weaning success groups.If ventilation was stopped completely, the patient was assessed as completely weaned (Group 3a).If non-invasive ventilation was used, this was considered partially weaned (Group 3b), according to the recommendations of the DGP.Patients who still required invasive ventilation via tracheal cannula were classified as weaning failure (Group 3c).
The patient status was recorded after 3, 6, and 12 months after conclusion of the weaning phase.The primary endpoint of the study was the proportion of patients who had been successfully weaned after 6 months and who did not need any further IMV (expected value 60%).Secondary endpoints were 1-year survival, predictive analysis, and health care cost comparison.

Statistical Analysis
For the primary endpoints, absolute numbers and percentages are given, the detailed statistical analysis of which has been published elsewhere [13].The results of the cost analysis are given in euros, as absolute totals.For that reason, the DRG (German Diagnosis-Related Groups) fee rates of the respective hospital were recorded, and further care costs mediated by the social service as part of the discharge management were determined.The initial plan was to document the cost of aids and remedies.However, depending on the health insurance, existing devices were sometimes new, used, or embedded in specialised discount contracts with aid and therapeutic agent providers so that comparability was not possible.It was also shown, that due to the complexity of required health care, it was very imprecise to record direct and indirect costs.A detailed cost comparison was therefore omitted here.Sole accommodation costs depending on further care were brought to consideration.This included accommodation costs of 61 patients, age-and gender-matched, who were discharged into the outpatient intensive care depending on ventilation without undergoing a structured weaning process in a certified weaning centre.An overview of the intervention and comparison groups is shown in Figure 1.

Primary Outcome
A total of 65 patients were included in the intervention group of the study.They either already had out-ofhospital care for their invasive ventilation (n = 16) or were considered unable to wean and, in the case of IMV dependency, were transferred from an intensive care unit of a non-specialised centre to the weaning centre for further care (n = 49).After the initial consent, 2 patients withdrew the previously given consent, 2 patients withdrew from the follow-up so that ultimately 61 patients participated in the study.After undergoing standardised weaning in the weaning centre, 29 (47.5%)patients were able to be completely weaned, 21 (31.4%) were partially weaned using non-invasive ventilation, and 11 (18%) patients could not be weaned and had to remain dependent on IMV.Summarising completely and partially weaned patients, the study succeeded in weaning in 50 patients (82%).

Secondary Outcomes
The inpatient stay in the weaning units was 24.7 days on average (R 5-124 d, SD ± 22 d).All patients were able to leave the weaning unit alive, but 1 patient died 28 days after discharge due to advanced carcinoma, so the 30-day mortality rate was 1.6%.After completing the inpatient phase, patients were transferred to out-ofhospital care.In 47 (77%) patients, this was done at home.Twelve (19.6%) patients were transferred to a nursing home and 2 (3.2%) to a residential group.Analysis regarding weaning success showed 100% (29 out of 29 patients) home discharges without additional ventilation aids after complete weaning.Following partial weaning (non-invasive mask ventilation), 18 out of 21 (85.7%)patients were discharged to their home, 2 (9.5%) patients into a residential group, and 1 (4.8%) patients to a nursing home.If weaning was unsuccessful, 2 (18%) patients were transferred to a domiciliary intensive care service, and 9 (82%) patients were transferred to a respiratory care home.An overview of the outcomes and accommodation of the intervention and comparison groups is shown in Figure 2. Eleven (18%) patients died during the 1-year follow-up.One-year survival was significantly higher in the successfully weaned groups.There was no difference between the completely or partially weaned groups [13].

Cost Analysis
Accommodation costs were set at EUR 0 for completely and partially weaned patients discharged home.No further ventilation care support was required.Contrarily, cost analysis of 1:1 domiciliary intensive care showed hourly rates between approx.EUR 30 and EUR 35 (annual costs of EUR 262,800 to EUR 306,600, monthly average EUR 28,470), depending on the health insurance company.Furthermore, patient contribution could include no additional payments or rates of EUR 500-1,000 per month, correlating with the level of needed care and insurance policiesmostly for basic care and housekeeping.These costs were equated for all patients because they could not be presented transparently for both the intervention group and the comparison group.
In a respiratory care home, hourly fees were rated between EUR 11.50 and EUR 22 (annual costs from EUR 100,740-EUR 192,720, monthly average EUR 14,670).If necessary, patients contributed additional rent and the ancillary and residual costs for basic care.A differentiated view was not possible here either.In a ventilation flat share (usually 1:3 care), the costs were between 13,900-17,900 EUR/month (monthly average 15,260 EUR).The patient had to pay rent, additional costs, and housekeeping.
To summarise, all costs referring to the encoded DRG fee rates, weaning of the 61 patients cost a total amount of EUR 4,126,782.15.This equals EUR 67,652.17per patient on average, with a range from EUR 7,015.62-EUR169,520.56depending on each patient and the additional procedures performed as necessary.This sum included all services, i.e., also the in-house follow-up care in the normal ward and  4).To enable comparison of the cohorts, the costs of patients who died in the year of observation were extrapolated over 12 months.The data of the comparison group were inevitable of 61 surviving patients as it was taken from living patients.However, since patients in the intervention group died in the course of the year and therefore did not need the full 12 months of out-ofhospital care costs, the actual costs incurred amounted to EUR 1,147,010.The costs for weaning and out-ofhospital care for the patients amounted to EUR 5,273,792.15 in actual terms, hypothetically for the comparison to the other cohort to EUR 7,007,502.15.In the comparison cohort, 21 patients were intensively cared for at home, 12 patients in a ventilation resident group, and 28 in a respiratory care home.This resulted in annual supply costs of EUR 14,301,000 (Table 5).Thus, the cost of patients treated in a weaning centre was significantly reduced: in the hypothetical approach by EUR 119,565 per patient (Table 6) and in actual terms by EUR 147.987 per patient.

Discussion
The aim of our intervention study was to compare the medium-to long-term costs incurred in the treatment of IMV patients with a group of patients treated in certified weaning centres after a primarily unsuccessful weaning attempt.In doing so, we were able to show that early transfer to a certified weaning centre can increase weaning success and reduce total costs by approximately EUR 120,000 per patient in the first year.
Several studies worldwide have shown that 60-80% of patients discharged from non-specialised ICUs as weaning failure could still be weaned from IMV after admission to a specialised weaning centre [14,15].However, this rate decreases when multiple comorbidities are present [16] and a prolonged ventilation period prior to transfer to a specialised centre [16,17].Up to 85% of patients discharged from hospital with IMV in Germany did not have access to a certified weaning centre.These affected patients are transferred directly to invasive out-of-hospital ventilation without special weaning attempts.This is associated with a considerable economic burden on the health system (costs approx.EUR 15,000-30,000 per patient/month for respiratory care) [18], which is congruent with the costs from the data records of the German statutory health insurance AOK with a monthly average of EUR 28,470.Patients with longterm IMV account for about 10% of intensive care cases in Germany but tie up about 50% of available resources [3].It is difficult to estimate the number of patients receiving care outside of hospitals after unsuccessful prolonged weaning without registries.Currently, it must be assumed that about 15,000-20,000 people are affected nationwide, with an annual increase of about 10% [19].
Due to the various funding sources in Germany, the actual post-inpatient costs can only be assessed and not determined transparently.The costs of medication, medicines, and aids could not be recorded but can be assumed not to affect the results to the disadvantage of the intervention group since it cannot be assumed that a tracheotomised long-term ventilated patient requires less medication than a person without invasive ventilation.The costs of medication should not differ significantly due to the matching of the   the weaning rate has even improved significantly [3].This was also comparable to the effectiveness of structured weaning in other studies [16,17,20,21].As the costs for completely weaned and partially weaned (non-invasive ventilation) patients did not differ significantly and the latter could be cared for at home without intensive care, this should be considered as an alternative to IMV.
Besides the limitations of omitting a detailed comparison of the incurred costs, the data were collected about 4-5 years ago and then the clinical outcome published 3 years ago.Furthermore, there could be a possible selection bias in the group transferred from other intensive care units as these patients were initially classified as non-weanable, but the physicians still asked for a second opinion from the weaning centre.Forty-nine patients belong to that group.This is a different collective than, for example, in WeanNet, the weaning register of the DGP.Only 16 were admitted from out-of-hospital long-term ventilation.The comparison group was matched from patients already in long-term ventilation, therefore being the better match for the 16 intervention patients.
Furthermore, 100% of the completely weaned patients could be cared for at home without intensive care.This is a better outcome than recently shown in the WeanNet data for COPD patients, in which only 53% have a closed tracheostoma [22].As well, the mortality rate in our study was rather low compared to the data from WeanNet [12].In conclusion, there could be a selection bias.
However, in addition to the resulting economic consequences of potentially unnecessary long-term IMV, the perceived quality of life of patients with out-of-hospital continuation of IMV is reduced, especially for elderly patients with chronic lung disease, regardless of whether they are cared for at home or in nursing facilities [23,24].Thus, future efforts should allow access to a structured and certified weaning process for patients at risk of prolonged weaning or weaning failure as early as possible.First steps are being made with the new law in Germany, "Intensivpflegeund Rehabilitationsstärkungsgesetz, GKV -IPReG."But there is still a long way to go.

Statement of Ethics
The study was approved by the Ethics Committee of the University of Witten/Herdecke with application number 48/2015.Written consent was obtained from all participants or their authorised representatives.

Table 2 .
Causes leading to ventilation

Table 4 .
Comparison of the costs incurred, actual and hypothetical

Table 5 .
The costs incurred in comparison

Table 6 .
Costs/patient/year This coincides with the previously published, mostly retrospective analyses.The comparison shows that the success rate of this weaning study is in the upper range.Compared to the last weaning survey in Germany in 2008,