Chordoma and chondrosarcoma of the craniospinal axis are challenging neurosurgical conditions. The management paradigm includes maximum safe surgical resection and radiotherapy. The treatment of these conditions can be costly because of the multitude of services required. For example, the treatment of spinal chordoma generally requires a complex spine procedure such as en-bloc surgical excision and multilevel instrumented fusion.[5, 12] This generally is associated with increased length of stay, need for rehabilitation, complications, risk for emergency department visits, hospital readmission, need for pain prescription refills, and cost.[5, 11, 13] Both diseases have high risk for recurrence (~ 57% for cranial and 27% for spinal disease) which may necessitate further treatment.[12, 14–16] These factors make these two conditions valuable to explore the cost and patterns of expenditure over time and the feasibility of adopting the BPCI model for reimbursement.
Our analyses showed that spinal CC patients had increased length of stay and complications rate, which was associated with higher median payment for the index hospitalization in comparison to the cranial group. This could be related to surgical pain and the presence of surgical drains which is common for complex spine surgery, which should be considered when planning the bundled payments for cranial and spinal patients. About 90% of the index hospitalization payments were hospital payments and the remaining 10% were physicians’ payment which was not different among the 3 groups. The length of stay has been linked before to increased costs.
Ninety days post-discharge period, the readmission rates were higher for spinal CC (group 2, 38% and group 3, 45%) in comparison to cranial CC (21%), which was associated with higher median payments ($40,227 and $42,242 for group 2 and 3 vs $24,116 for group 1). The higher readmission rates for the mobile spine and sacral groups could have been due to pain or wound problems when compared to the cranial group. Fry et al. reported a 90 days readmission rate of 25% after elective craniotomy for a mass lesion, while Lau et al. reported a 90 days readmission rate of 13% and 20% after surgery for spinal chordoma and chondrosarcoma, respectively. The difference between our study and Lau et al, their reported rate could have been due to a smaller sample size (23 chordomas, 10 chondrosarcomas). Also, they reported a mean direct hospital cost of ($65,413) and ($59,113) for spinal chordoma and chondrosarcoma, respectively. This was higher than the cost in our study ( mean cost for hospital payment $50,524). The observed difference can be due to a difference in data capture, since our cost data is based of insurer reimbursement to hospital claims, while their cost data was retrieved directly from hospital records. One of reasons for the aforementioned comparison between the cranial and spinal CC is to highlight the fact that disease location should be considered in calculating the bundled payment.
Jain et al. reported that back pain, leg pain, wound problems, cardiorespiratory complaints, gastrointestinal complaints, deep venous thrombosis, and systemic infections were the most common causes for ED visit and hospital readmission within 30 days of lumbar spine surgery. Fry et al. reported that seizures, sepsis, wound complications, pneumonia, and postoperative infections were the most common causes for readmission after elective cranial surgery for mass lesion. Lau et al. reported that wound infection, tumor recurrence requiring decompression, postoperative pain control, and proximal junctional kyphosis requiring revision procedure were the most common causes for readmission after spinal chordoma and chondrosarcoma surgery. These reports demonstrated a significant overlap between the causes for readmission for cranial and spinal surgery. Therefore, the use of certain preventative measures by hospitals to minimize occurrence of these complications might help to reduce the total cost of care. [21, 22] For the 4th − 12th - month post-discharge, groups 2 and 3 had higher readmissions rate and median payments in comparison to group 1.
Over time, each individual group showed a downward trend in hospital readmission rate, outpatient service utilization, prescription refills, and median payments. Cranial CC, for the first 6 months post index hospitalization, the readmission rate was 30%, the outpatient services were 113, and the overall median payments were $48,508. For the second 6 months there was a drop-in readmission rate to 14%, outpatient services to 53, and overall median payments to $23,786. Mobile spine CC, during the first 6 months post index hospitalization, the readmission rate was 45%, the outpatient services were 105, and the overall median payments were $49,425. During the second 6 months there was a drop-in readmission rate to 10%, outpatient services to 60, and overall median payments to $22,869. Sacral CC, during the first 6 months post index hospitalization, the readmission rate was 55%, the outpatient services were 119, and the overall median payments were $60,853. During the second 6 months there was a drop-in readmission rate to 9%, outpatient services to 92, and overall median payments to $40,622. These trends are assumed to be due to healing, reduction in postoperative pain and improved functional status with rehabilitation, which lead to less utilization of healthcare services and cost reduction. Off note, hospital readmission was the main factor for the costs incurred during the first 12 months post-discharge and to a lesser extent outpatient services utilization.
There was a significant variability in payments based on insurance type, Medicaid was associated with increased odds for smaller payment for all groups, and Medicare was associated with increased odds for smaller payment only for cranial CC when compared to commercial insurers. Higher EI value (multiple comorbidities) was associated with increased odds for larger payments for all groups, which re-emphasize that managing patients with comorbidities increases the cost of care. To that point, using a tool like the CMS – Human Health Services (HHS) Hierarchical Condition Category (HCC) risk adjustment model can be helpful. This model uses patients’ demographic data and coded diagnoses to produce a risk score that will help with financial estimation. Turcot et al. recently published a report where they tested this model on patients that underwent different spinal surgical interventions. They found that there was a significant association between the HCC score and readmission rates, length of stay, need for reoperation, and cost. 
Notably, the current bundles, as indicated on the CMS website include fairly specific clinical situations such as acute Myocardial Infarction, Sepsis, CABG, etc., which are frequent, and well-defined clinical entities. This is important as the calculation of the bundle must be very accurate and based on solid data from a prior experience. Nowhere in the current list of bundles, are any skull base tumors or malignancies, probably because of the difficulty in accurately defining them as individual clinical entities. The choice of chordomas and chondrosarcomas in this study may represent a limitation, given both the rarity of those tumors and the different outcome of those diagnoses. But it can be considered a strength, since the BPCI will eventually be the standard method for payment for all diseases, and there are no reports in the literature that investigated skull base or spinal conditions like chordomas and chondrosarcomas.
The higher cost needed to manage craniospinal chordoma and chondrosarcoma is not limited to a single episode of care, but rather maintained over a period of time. Most of the expenses were during the first 6 months post index hospitalization. The success of BPCI requires a joint effort between insurers and hospitals/providers. Complex neurosurgical conditions like craniospinal CC which has an inherent increased risk for complications, readmissions, and the need for outpatient services. Based on the results of our analyses and to increase the viability and acceptability of BPCI model 2, we suggest the following; BPCI should consider bundling the payments for the index hospitalization and the anticipated services during first 6 months after initial discharge together. Also, it should consider the patients’ comorbidities, and for diseases with no curative therapy, the BP model should consider the variability in treatment regimens like the use of experimental and off label treatments. In addition, streamlining payments through minimizing variability in reimbursements between Medicare/ Medicaid and commercial insurers is important. It’s well documented that treatment at a center with high case volume is associated with better outcomes and lower complications rate which in turn lead to lower cost. [25–27] Therefore, it may be of importance for the BPCI to stipulate that treatment of certain rare and complex neurosurgical conditions ought to be done at centers of excellence. Hospitals and providers should consider measures to further improve outcomes and decrease cost. Adoption of programs like Enhanced Recovery After Surgery (ERAS) or Enhanced Perioperative Care (EPOC) which aimed at decreasing length of stay, complications rate and readmissions can be valuable to achieve that. [21, 22]