Abstract
BACKGROUND:
Postoperative 30-day readmissions have been shown to negatively affect survival and other important outcomes in patients with glioblastoma (GBM).
OBJECTIVE:
To further investigate patient readmission risk factors of primary and recurrent patients with GBM.
METHODS:
The authors retrospectively reviewed records of 418 adult patients undergoing 575 craniotomies for histologically confirmed GBM at an academic medical center. Patient demographics, comorbidities, and clinical characteristics were collected and compared by patient readmission status using chi-square and Mann–Whitney U testing. Multivariable logistic regression was performed to identify risk factors that predicted 30-day readmissions.
RESULTS:
The cohort included 69 (12%) 30-day readmissions after 575 operations. Readmitted patients experienced significantly lower median overall survival (11.3 vs 16.4 months, P = .014), had a lower mean Karnofsky Performance Scale score (66.9 vs 74.2, P = .005), and had a longer initial length of stay (6.1 vs 5.3 days, P = .007) relative to their nonreadmitted counterparts. Readmitted patients experienced more postoperative deep vein thromboses or pulmonary embolisms (12% vs 4%, P = .006), new motor deficits (29% vs 14%, P = .002), and nonhome discharges (39% vs 22%, P = .005) relative to their nonreadmitted counterparts. Multivariable analysis demonstrated increased odds of 30-day readmission with each 10-point decrease in Karnofsky Performance Scale score (odds ratio [OR] 1.32, P = .002), each single-point increase in 5-factor modified frailty index (OR 1.51, P = .016), and initial presentation with cognitive deficits (OR 2.11, P = .013).
CONCLUSION:
Preoperatively available clinical characteristics strongly predicted 30-day readmissions in patients undergoing surgery for GBM. Opportunities may exist to optimize preoperative and postoperative management of at-risk patients with GBM, with downstream improvements in clinical outcomes.
KEY WORDS: Brain tumor, Care quality, Glioblastoma, Readmission, Value
ABBREVIATIONS:
- DVTs/PEs
deep vein thrombosis/pulmonary embolisms
- ED
emergency department
- EMR
electronic medical record
- HR
hazard ratio
- HRRP
Hospital Readmissions Reduction Program
- IDH1
isocitrate-dehydrogenase-1
- KPS
Karnofsky Performance Scale
- LOS
length of stay
- mFI-5
5-factor modified frailty index
- MGMT
O(6)-methylguanine-DNA-methyltransferase
- SAR
subacute rehabilitation
- OS
overall survival
- QI
quality improvement.
The Agency for Healthcare Research and Quality estimates that individual readmissions cost $14 400 on average and, in 2011, were associated with $41.2 billion in healthcare costs.1 In 2012, the Centers for Medicare and Medicaid Services initiated the Hospital Readmissions Reduction Program (HRRP), to promote higher care quality through increased transparency of hospital readmission statistics, along with reimbursement penalties for medical centers with high readmission rates.2 Subsequent increases in these penalties have incentivized healthcare systems to identify and appropriately manage patients at greatest risk of early readmission to reduce overall patient and hospital financial burden.2
The prediction and impact of unplanned hospital readmissions have been investigated in nonoperative cohorts3-5 as well as in the surgical oncology literature.6-8 However, there is a relative paucity of literature on readmissions in neuro-oncology, including in glioblastoma (GBM), the most common primary malignant brain tumor.9-12 To the best of our knowledge and of the available literature related to GBM, one study has considered recurrent (ie, progressive) GBM13 and one10 has been largely confined to Medicare-linked databases.9,11 Although requirements for Medicare vary, most Medicare patients are older than 65 years14; these databases may therefore not adequately capture a large proportion of patients with GBM, given that the median age of those afflicted is approximately 56 years at diagnosis.15
Identification and risk stratification of patients with GBM at risk of readmission have the potential to improve care quality, minimize financial burden, and potentially affect downstream survival. Upfront identification of high-risk patients opens the door for early involvement of extended care teams in discharge disposition planning and mobilization of necessary social support systems. To address this unmet need, our study aimed to describe predictors of early readmission after GBM resection, including patients with recurrent/progressive disease, and elucidate the impact of early readmission on overall survival (OS).
METHODS
Patient Selection
Patients undergoing resection of primary or recurrent GBM between July 01, 2009, and September 30, 2019, were identified through an institutional database created with institutional review board approval and without requiring patient consent. Senior neuropathologists determined tumor pathology and grade according to current World Health Organization classification systems.16 Adult (18 years or older) patients with a histopathologically confirmed diagnosis of World Health Organization Grade IV glioblastoma were included. Patients who underwent biopsy only were excluded.
Recorded Variables
We manually reviewed physician and other provider notes for all included patients through the electronic medical record (EMR) from both within our institution and sister institutions throughout our state by using the Chesapeake Regional Information System for our Patients.17 Preoperative data on patient comorbidities (Supplement, http://links.lww.com/NEU/D227) and functional status were used to generate a 5-factor modified frailty index (mFI-5) score.18 Discharge disposition was defined as home or nonhome.19 Stupp et al19 protocol completion, representing the standard of care after primary surgery consisting of postoperative temozolomide and radiation therapy, was also recorded. Readmission was defined as any unplanned readmission within 30 days of postoperative discharge and classified as preventable or nonpreventable, as defined by prior literature in the field.10,20
Statistical Analysis
Statistical analyses were conducted using SAS 9.4 (SAS Institute). P values <0.05 were considered statistically significant (Supplement, http://links.lww.com/NEU/D227). The Kaplan-Meier method was used to analyze survival, and a multivariable Cox proportional hazards model was used to determine covariate associations with risk of death. Multivariable logistic regression modeling for unplanned readmission was also used (Supplement, http://links.lww.com/NEU/D227).
Reporting Guidelines
Strengthening the Reporting of Observational Studies in Epidemiology guidelines were used for this study.
RESULTS
Patient Characteristics
Our cohort included 418 patients undergoing 575 craniotomies for GBM. Overall, the cohort was majority Caucasian (81%) and male (85%). The mean preoperative age, Karnofsky Performance Scale (KPS) score, and mFI-5 score were 59.4 ± 12.9 years, 73.3 ± 16.3, and 0.59 ± 0.80, respectively. Most patients had private health insurance (208, 50%), whereas 172 (41%) had Medicare, 15 (4%) had Medicaid, 12 (3%) were uninsured, and 11 (3%) were self-pay. The mean and median length of stay (LOS) was 5.5 and 3 days, respectively. Patients with nonhome discharge were older relative to their home discharge counterparts (65.2 vs 58.2 years, P < .001). A total of 69 (12.0%) cases were readmitted within 30 days, with a median time to readmission totaling 14.33 days. The median OS was 15.4 months. The overall rate of O(6)-methylguanine-DNA-methyltransferase (MGMT) methylation was 17% and isocitrate-dehydrogenase-1 (IDH1) mutation was 7%. MGMT and IDH1 mutation status were known in 233 (56%) patients.
Comparative Analyses
The mean preoperative age, KPS score, and mFI-5 score for readmitted vs nonreadmitted patients were 56.2 ± 12.8 vs 60 ± 12.9 years (P = .009), 66.9 ± 16.4 vs 74.2 ± 16.1 (P = .005), and 0.77 ± 0.92 vs 0.57 ± 0.78 (P = .093), respectively (Table 1). There were no significant differences in readmission status by tumor location, race, sex, or insurance status (P = .364, P = .359, P = .114, and P = .364, respectively.) Repeat resection was not associated with a higher rate of 30-day readmission relative to primary resection (P = .616). Readmitted patients experienced a similar rate of gross total resections (26% vs 35%, P = .259) relative to nonreadmitted patients.
TABLE 1.
Patients With and Without 30-Day Readmissions—All Surgeries
| Characteristic | 30-day readmission category | P value | |
|---|---|---|---|
| Yes, n = 63 | No, n = 355 | ||
| Demographics | |||
| Age in years at surgery, mean ± SD | 56.2 ± 12.8 | 60 ± 12.9 | .0090 |
| Male | 48 (76) | 306 (86) | .1141 |
| Race | .3589 | ||
| White/Caucasian | 50 (79) | 290 (82) | |
| African American | 5 (8) | 26 (7) | |
| Asian or other | 8 (13) | 39 (11) | |
| Total resections | .1332 | ||
| 1 | 39 (62) | 263 (74) | |
| 2 | 23 (37) | 69 (19) | |
| 3 | 1 (1) | 16 (5) | |
| ≥4 | 0 | 7 (2) | |
| (+) MGMT methylation | 8 (13) | 64 (18) | .3943 |
| IDH1 mutation | 2 (3) | 26 (7) | .3271 |
| Case characteristics | n = 69 | n = 506 | |
| Surgery order | .6163 | ||
| 1st | 48 (70) | 371 (73) | |
| 2nd | 18 (26) | 95 (19) | |
| 3rd | 3 (4) | 28 (6) | |
| ≥4th | 0 (0) | 12 (2) | |
| Admitted through ED | 6 (9) | 67 (13) | .3404 |
| KPS score, mean ± SD | 66.9 ± 16.4 | 74.2 ± 16.1 | .0050 |
| mFI-5, mean ± SD | 0.77 ± 0.92 | 0.57 ± 0.78 | .0926 |
| Cognitive deficit at presentation | 22 (32) | 104 (21) | .0036 |
| Motor deficit at presentation | 29 (42) | 155 (31) | .0058 |
| Seizure at presentation | 22 (32) | 122 (24) | .1650 |
| Headache at presentation | 19 (28) | 157 (31) | .5483 |
| Op characteristics | |||
| GTR at surgery | 18 (26) | 178 (35) | .2587 |
| Received carmustine wafers | 8 (12) | 52 (10) | .7192 |
| Postoperative characteristics | |||
| LOS, mean ± SD | 6.1 ± 5.8 | 5.3 ± 6.7 | .0074 |
| Intracranial hematoma | 2 (3) | 5 (0.9) | .1941 |
| DVT/PE | 8 (12) | 20 (4) | .0058 |
| Seizure | 6 (9) | 23 (5) | .1407 |
| New motor deficit | 20 (29) | 71 (14) | .0015 |
| Palliative care consultation | 4 (6) | 24 (5) | .6588 |
| Nonhome disposition | 27 (39) | 114 (22) | .0051 |
DVT/PE, deep vein thrombosis/pulmonary embolism; ED, emergency department; GTR, gross total resection; IDH1, isocitrate-dehydrogenase-1; KPS, Karnofsky Performance Scale; LOS, length of stay; mFI-5, 5-factor modified frailty index.
Unless otherwise indicated, values are expressed as the number (%). Boldface type indicates statistical significance.
Preoperatively, readmitted patients were more likely to have demonstrated upfront cognitive (32% vs 21%, P = .004) or motor (42% vs 31%, P = .006) deficits relative to nonreadmitted counterparts. Postoperatively, readmitted patients experienced more deep vein thrombosis/pulmonary embolisms (DVTs/PEs) (12% vs 4%, P = .006) and more new motor deficits (29% vs 14%, P = .002) before discharge (after surgery) relative to nonreadmitted patients. Readmitted patients also experienced longer mean predischarge hospital LOS (after surgery) (6.1 vs 5.3 days, P = .007) and more frequent nonhome discharges (39% vs 22%, P = .005) relative to nonreadmitted patients. Of all discharge dispositions (Supplement, http://links.lww.com/NEU/D227), patients discharged to subacute rehabilitation (SAR) had a significantly higher readmission rate (21% vs 9.7%, P = .003) compared with all other dispositions (Table 2). Stupp protocol completion rates were similar between readmitted and nonreadmitted patients (36 [68%] vs 210 [59%], P = .824).
TABLE 2.
Cases of 30-Day Readmission by Discharge Disposition After Resection
| Postresection discharge disposition | KPS score, mean ± SDa | 30-day readmission category, no. (%)b | P value | |
|---|---|---|---|---|
| Yes | No | |||
| Home | 76.7 ± 15 | 42 (10) | 392 (90) | .1680 |
| Inpatient rehabilitation | 75.2 ± 14 | 1 (14) | 6 (86) | .8523 |
| Subacute rehabilitation facility | 63.4 ± 16 | 25 (21) | 96 (79) | .0034 |
| Skilled nursing facility | 54.0 ± 18 | 1 (20) | 4 (80) | .5820 |
| Hospice | 50.0 ± 15 | 0 (0) | 8 (100) | .2963 |
KPS, Karnofsky Performance Scale.
KPS score differences were statistically significant at P < .0001.
Sixty-nine cases were readmitted within 30 days, and 506 were not.
Boldface type indicates statistical significance.
Readmision Effect on Overall Survival
Patients who experienced a 30-day readmission had significantly lower median OS compared with nonreadmitted patients (11.3 vs 16.4 months, P = .014) (Figure) despite similar survival rates at 30 days (95.2% vs 97.4%, P = .413). In a Cox proportional hazards model for death adjusting for age, mFI-5 score, KPS score, extent of tumor resection, and total number of surgeries, 30-day readmission status was associated with an increased risk of death (hazard ratio [HR]: 1.961, P < .0001). Similarly, older age (HR: 1.020, P < .0001), lower KPS score (HR: 1.161, P < .0001), and subtotal resection relative to gross total resection (HR: 1.803, P < .0001) were also associated with an increased risk of death (Table 3). By contrast, the increasing number of total surgeries (HR: 0.873, P = .0211) was associated with a decreased risk of death.
TABLE 3.
Cox Proportional Hazard Model for Death—All Patients
| Variable | Multivariable HR (95% CI) | P value |
|---|---|---|
| STR; ref: GTR | 1.803 (1.453-2.237) | <.0001 |
| 30-day readmission | 1.961 (1.456-2.641) | <.0001 |
| KPS score; decreasing, 10 points | 1.161 (1.085-1.244) | <.0001 |
| Age; increase, 1 year | 1.020 (1.012-1.029) | <.0001 |
| Total surgery no.; continuous, increasing | 0.873 (0.752-0.981) | .0211 |
| (+) MGMT methylation; ref: unmethylated | 0.506 (0.376-0.680) | <.0001 |
GTR, gross total resection; HR, hazard ratio; KPS, Karnofsky Performance Scale; MGMT, O(6)-methylguanine-DNA-methyltransferase; STR, subtotal resection.
Boldface type indicates statistical significance.
FIGURE.
Overall survival of readmitted vs nonreadmitted patients with glioblastoma—Kaplan-Meier curve.
Readmission Predictors
Our multivariable logistic regression model adjusted for age, KPS score, mFI-5 score, cognitive deficit at presentation, postoperative DVTs/PEs occurring before discharge, new postoperative motor deficits, and nonhome discharge disposition demonstrated that each 10-point decline in KPS score (OR: 1.322, P = .002), each 1-point increase in mFI-5 score (OR: 1.506, P = .016), and preoperative cognitive deficits (OR: 2.11, P = .013) were independently associated with an increased risk of 30-day readmission. New postoperative motor deficits and DVTs/PEs were significant predictors of readmission on bivariate analysis but not on multivariable analysis (P = .060 and P = .084, respectively). Interestingly, older age was associated with a decreased risk of readmission (OR: 0.960, P = .0004) (Table 4). This model was well-calibrated (Hosmer-Lemeshow test, P = .926), performed well on area under the curve analysis (c-statistics 0.717), and on cross-validation (mean 5-fold c-statistic = 0.710). There was no collinearity (variance inflation factors range 1.03-1.33).
TABLE 4.
ORs for 30-Day Readmissions—All Surgeries
| Variable | Logistic regression | |||
|---|---|---|---|---|
| Univariate OR (95% CI) | P value | Multivariable OR (95% CI) | P value | |
| Older age at surgery | 0.984 (0.965-1.002) | .0850 | 0.960 (0.939-0.982) | .0004 |
| Male; ref: female | 1.471 (0.855-2.533) | .1635 | ||
| Surgery no.; continuous, increasing | 0.825 (0.559-1.217) | .3319 | ||
| Preoperative characteristics | ||||
| Greater mFI-5 | 1.322 (0.996-1.757) | .0536 | 1.506 (1.081-2.098) | .0155 |
| Admitted through ED; ref: elective admission | 0.623 (0.259-1.495) | .2890 | ||
| KPS score; decreasing, 10 points | 1.291 (1.115-1.499) | .0007 | 1.322 (1.112-1.582) | .0016 |
| Cognitive deficit at presentation | 1.805 (1.041-3.129) | .0355 | 2.114 (1.168-3.824) | .0134 |
| Seizure at presentation | 1.470 (0.852-2.537) | .1665 | ||
| Headache at presentation | 0.842 (0.481-1.476) | .5487 | ||
| Operative characteristics | ||||
| STR; ref: GTR | 1.603 (0.902-2.847) | .1075 | ||
| Received carmustine wafers | 1.215 (0.574-2.575) | .6104 | ||
| Postoperative characteristics | ||||
| Greater LOS | 1.017 (0.986-1.049) | .3389 | ||
| DVT/PE | 3.181 (1.343-7.533) | .0085 | 2.370 (0.890-6.308) | .0841 |
| Intracranial hematoma | 2.986 (0.568-15.694) | .1964 | ||
| Seizure | 1.996 (0.783-5.089) | .1479 | ||
| New motor deficit | 2.495 (1.401-4.445) | .0019 | 1.925 (0.972-3.814) | .0603 |
| Palliative care consultation | 1.278 (0.429-3.803) | .6593 | ||
| Nonhome disposition | 1.991 (1.183-3.350) | .0095 | 0.944 (0.476-1.870) | .8686 |
| Tumor characteristics | ||||
| (+) MGMT methylation | 0.614 (0.271-1.388) | .2411 | ||
DVT/PE, deep vein thrombosis/pulmonary embolism; ED, emergency department; GTR, gross total resection; KPS, Karnofsky Performance Scale; LOS, length of stay; mFI-5, 5-factor modified frailty index; MGMT, O(6)-methylguanine-DNA-methyltransferase; OR, odds ratio; STR, subtotal resection.
Boldface type indicates statistical significance.
DISCUSSION
Study Rationale
Evolutions in the healthcare landscape, including the introduction of the HRRP and novel payment models such as bundled payments, have incentivized healthcare systems to improve care quality while reducing healthcare costs.2,21 A key driver of costs is unplanned 30-day readmissions, which have cost healthcare systems more than $2 billion in penalties since the HRRP's inception.2 Beyond financial implications, early readmissions have been associated with decreased health-related quality of life and increased mortality rates.9,10,22-24 Despite the financial and clinical importance of unplanned readmissions, they remain a relatively understudied area within neurosurgical oncology, including in primary and recurrent patients with GBM. We undertook this study to identify predictors of unplanned readmissions, elucidate the clinical impact of readmissions, and provide a foundational evidence base for quality improvement (QI) efforts to reduce future readmissions.
Interpretation and Prior Literature
Our data augment the extant literature on readmissions in patients with GBM and offer multiple new findings that may help improve patient management. Our readmission rate (12%) aligns with large neurosurgical services demonstrating readmission rates ranging between 7.5% and 15.8%,10,11,20 suggesting generalizability of our findings and a general benchmark to use for QI initiatives. Specifically, the IDH1 mutation rate among our patient cohort was similar to previous multicenter studies, whereas the MGMT methylation rate was lower.25,26 Drivers of preventable readmissions in our cohort were similar to those described in other studies,9,10,20,27-30 suggesting processes related to underlying disease and current treatment paradigms rather than institutional differences in management.
Alongside reducing readmissions, LOS has been a target for optimizing efficiency and cost. In our study, longer LOS was not associated with increased readmission risk which aligns with previous GBM-specific readmission literature9,10 but contrasts with larger neurosurgical reports of longer LOS predicting higher readmission rates.28,31,32 In addition, Lakomkin and Hadjipanayis33 previously demonstrated that nonroutine discharges were associated with increased postoperative complications and returns to the operating room in patients with brain tumor. We found that readmissions were associated with nonhome discharges on bivariate but not multivariable analysis. We speculate that these differences stem from our practice of discharging patients with GBM with a high readmission risk to inpatient rehabilitation where complications (eg surgical site infections and DVTs/PEs) are managed accordingly, thereby preventing some readmissions. This may also account for the association seen between older age and mildly decreased readmission risk (OR: 0.960, P = .0004). This paradox is further supported by our finding that patients with nonhome disposition were older relative to home disposition counterparts (62.5 vs 58.2 years, P = .001). The potential relationships between these outcomes highlight the need for QI initiatives targeting underlying drivers, which may be highly fruitful in reducing healthcare costs.
Our finding that readmitted patients experienced a 5.1-month reduction in OS aligns with those of Nuño et al9 and Dickinson,10 who found that readmitted patients experienced OS reductions of 1.6 and 9 months, respectively. We note that this relationship held across the 3 studies despite differences in patient cohorts, institutions, and geography. Whereas our data do not demonstrate whether this relationship is correlative or causative, prior surgical oncology literature posits the role of readmissions in delaying adjuvant treatment as a mechanism for the observed reduction in OS.34 Although intuitive, the effects of readmission on treatment delay have not been studied in GBM. Delays in radiotherapy, however, have demonstrated mixed effects on survival. A meta-analysis, conducted by Loureiro et al,35 found no relationship between time to radiotherapy and OS. Meanwhile, a 2018 study by Pollom et al36 found decreased OS in newly diagnosed patients with GBM who had >42-day delay in receiving radiotherapy. Ultimately, the relationship between readmissions and survival suggests that reducing readmissions is not only financially beneficial but may be critical for improving patient outcomes.
An important, novel contribution of our study was that repeat resection was not associated with increased risk of unplanned readmissions. Although the efficacy of repeat resections has been previously reviewed, the potential morbidity associated with additional surgery in these patients is less known.15,37 Thus, in addition to supporting the role of repeat resection in improving survival, our results provide confidence that repeat resection will not unnecessarily increase the risk of some adverse outcomes in appropriately selected patients.
Strategies to Reduce Readmissions in Patients With Glioblastoma
Interestingly, all factors in our multivariable analysis that predicted 30-day readmissions were based on data available in the preoperative period. This represents a potentially significant opportunity for targeted initiatives to reduce the risk of postoperative 30-day readmissions. Based on our data, we note some potential areas for QI initiatives to reduce readmissions in GBM: preoperative rehabilitation and improved discharge planning.
We postulate that patients' frailty and functional status could be used as key metrics for preoperative risk stratification. Prior literature has suggested that postoperative outcomes of frail, dependent patients could potentially be improved through preoperative prehabilitation.38 In the degenerative spine surgery literature, prehabilitation led to improved postoperative mobility and increased quality of life39 while also remaining cost-effective.40 Delaying primary resection to undergo prehabilitation, in patients with brain tumor, has unknown potential effects on survival41; however, Rivera-Rivera et al42 underscored the potential functional benefits of prehabilitation in patients with tumors in eloquent areas. Similarly, in surgical oncology, prehabilitation has been associated with reduced postoperative complications43 and early readmissions.44 Furthermore, 2 randomized clinical trials from the lung cancer literature demonstrated significant functional benefits of 2- and 3-week prehabilitation programs in patients undergoing lung lobectomy. Importantly, these delays in resection were not associated with an increased risk of mortality.45,46 In our cohort, each 1-point increase in mFI-5 score presented a 50% elevated risk of readmission on multivariable analysis (P = .016). The KPS score was lower in readmitted patients (P = .005), and each 10-point decrease elevated the risk of readmission by 32% (P = .002). Taken together, the demonstrated propensity of frail and lower functional status patients to be readmitted reinforces the potential opportunity for prehabilitation-related efforts in preoperative interdisciplinary neurosurgical care.
Reasons for nonpreventable readmissions (Supplement, http://links.lww.com/NEU/D227) varied widely but were primarily due to neurological symptoms concerning for disease-related progression. Although rapid disease progression requiring inpatient management is an unfortunate characteristic for some patients with GBM, knowledge of the predictors identified in this work may help enhance disposition planning while encouraging patient/family engagement and close outpatient follow-up. Appropriate discharge disposition may also optimize medication adherence, a contributory factor to 2 readmissions in our cohort. Patients discharged to SAR facilities had the highest readmission rate among all dispositions (P = .0034), a finding that aligns with prior literature by Lakomkin and Hadjipanayis.33 Although this probably reflects differences in patient functional status and treatment goals after resection, it also identifies a potential role for predischarge planning among those discharged to SAR. In a quasiexperimental pilot study, early predischarge planning, along with patient education and scheduled postoperative telephone follow-ups, minimized unplanned readmissions in patients who underwent craniotomy.47 Wider adoption of early and holistic discharge planning may help reduce a significant proportion of unplanned readmissions after craniotomy for GBM.
Limitations
As a retrospective study from a single healthcare system, our findings would be strengthened by external and/or prospective validation. We also recognize the inability to capture readmissions omitted from the patient's EMR, which we attempted to minimize by judiciously reviewing provider notes both from within our institution and from sister institutions throughout our state using Chesapeake Regional Information System for our Patients, a shared EMR system. The increasing use of readmissions as an indicator of healthcare quality necessitates deeper investigation and external validation, and we are hopeful that our findings add this to the growing body of clinical research.
CONCLUSION
Early readmissions remain a widely used surrogate for hospital quality, and excess readmissions are a target for Centers for Medicare and Medicaid Services reimbursement penalties. Readmissions have been strongly associated with poor outcomes; however, predictors of readmission in patients with GBM and the impact of readmission on OS remain relatively understudied. We examined readmissions in a large GBM cohort, confirmed previous findings of decreased OS in readmitted patients with GBM, and demonstrated that patients with higher mFI-5 scores, lower KPS scores, and those presenting with cognitive deficits are at greatest risk for readmission after GBM resection. The preoperative nature of these risk factors underscores a potential opportunity for intervention with prehabilitation and early, thorough discharge planning. Understanding and enhancing risk stratification for readmission holds the potential to benefit the quality and value of care delivered to future patients with GBM.
Acknowledgments
We thank the Epic Faculty Scholars Program and the Johns Hopkins Institute for Clinical and Translational Research for their support.
Footnotes
Presented in part at the American Association of Neurological Surgeons 2021 Annual Scientific Meeting (August 21-25, 2021, Orlando, Florida) as a Poster Presentation.
Supplemental digital content is available for this article at neurosurgery-online.com.
Neurosurgery Speaks! Audio abstracts are available for this article at neurosurgery-online.com. Audio abstracts can be located under “Associated Multimedia.”
Funding
This work was supported by the Epic Faculty Scholars Program through the Johns Hopkins Institute for Clinical and Translational Research (UL1TR001079). Dr Brem receives research Funding from NIH, Johns Hopkins University, and Acuity Bio Corp* and philanthropy.
Disclosures
Dr Brem is a paid consultant to Insightec and chairman of the company's Medical Advisory Board. Insightec is developing focused ultrasound treatments for brain tumors. This arrangement has been reviewed and approved by the Johns Hopkins University in accordance with its conflict-of-interest policies. Dr Brem is a consultant for Consultant for AsclepiX Therapeutics, StemGen, InSightec*, Accelerating Combination Therapies*, Catalio Nexus Fund II, LLC*, LikeMinds, Inc*, Galen Robotics, Inc.* and Nurami Medical* (*includes equity or options). Dr Lim has grants/payments from Stryker. The other authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.
Supplemental Digital Content
Supplement. eMethods 1. Pertinent Data Points. eMethods 2. Statistical Analysis—Multivariable Logistic Regression. eMethods 3. Statistical Analysis – Multivariable Cox Proportional Hazards Model. eResults 1. Presenting Patient and Case Characteristics. eResults 2. Readmitted Patients and Preventability. eResults 3. Palliative Care Team Inclusion and Readmission.
REFERENCES
- 1.Bailey MK, Weiss AJ, Barrett ML, et al. Characteristics of 30-Day All-Cause Hospital Readmissions, 2010-2016: Statistical Brief #248; 2006. [PubMed] [Google Scholar]
- 2.Centers for Medicare & Medicaid Services. Hospital Readmissions Reduction Program. 2021. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Readmissions-Reduction-Program.
- 3.Zuckerman RB, Sheingold SH, Orav EJ, Ruhter J, Epstein AM. Readmissions, observation, and the hospital readmissions reduction program. N Engl J Med. 2016;374(16):1543-1551. [DOI] [PubMed] [Google Scholar]
- 4.Halfon P, Eggli Y, Prêtre-Rohrbach I, et al. Validation of the potentially avoidable hospital readmission rate as a routine indicator of the quality of hospital care. Med Care. 2006;44(11):972-981. [DOI] [PubMed] [Google Scholar]
- 5.Kansagara D, Englander H, Salanitro A, et al. Risk prediction models for hospital readmission: a systematic review. JAMA. 2011;306(15):1688-1698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Fernandez FG, Khullar O, Force SD, et al. Hospital readmission is associated with poor survival after esophagectomy for esophageal cancer. Ann Thorac Surg. 2015;99(1):292-297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Reddy DM, Townsend CM, Kuo YF, et al. Readmission after pancreatectomy for pancreatic cancer in medicare patients. J Gastrointest Surg. 2009;13(11):1963-1974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Greenblatt DY, Weber SM, O’Connor ES, et al. Readmission after colectomy for cancer predicts one-year mortality. Ann Surg. 2010;251(4):659-669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nuño M, Ly D, Ortega A, et al. Does 30-day readmission affect long-term outcome among glioblastoma patients? Neurosurgery. 2013;74(2):196-205. [DOI] [PubMed] [Google Scholar]
- 10.Holly D, Christine C, Miriam N, et al. Unplanned readmissions and survival following brain tumor surgery. J Neurosurg. 2015;122(1):61-68. [DOI] [PubMed] [Google Scholar]
- 11.Nuño M, Ly D, Mukherjee D, et al. Quality of surgical care and readmission in elderly glioblastoma patients. Neuro-Oncology Pract. 2014;1(2):33-39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Thakkar JP, Dolecek TA, Horbinski C, et al. Epidemiologic and molecular prognostic review of glioblastoma. Cancer Epidemiol Biomarkers Prev. 2014;23(10):1985-1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mallela AN, Agarwal P, Goel NJ, et al. An additive score optimized by a genetic learning algorithm predicts readmission risk after glioblastoma resection. J Clin Neurosci. 2020;80:1-5. [DOI] [PubMed] [Google Scholar]
- 14.US Department of Health and Human Services. Who Is Eligible for Medicare?. 2014. https://www.hhs.gov/answers/medicare-and-medicaid/who-is-eligible-for-medicare/index.html.
- 15.Botros D, Dux H, Price C, et al. Assessing the efficacy of repeat resections in recurrent glioblastoma: a systematic review. Neurosurg Rev. 2020;44(3):1259-1271. [DOI] [PubMed] [Google Scholar]
- 16.Louis DN, Perry A, Reifenberger G, et al. The 2016 World Health Organization Classification of Tumors of the central nervous system: a summary. Acta Neuropathol. 2016;131(6):803-820. [DOI] [PubMed] [Google Scholar]
- 17.Horrocks D. The Chesapeake Regional Information System (CRISP). Maryl Med. 2010;11(3):20. [PubMed] [Google Scholar]
- 18.Subramaniam S, Aalberg JJ, Soriano RP, et al. New 5-factor modified frailty index using American college of surgeons NSQIP data. J Am Coll Surg. 2018;226(2):173-181.e8. [DOI] [PubMed] [Google Scholar]
- 19.Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;352(10):987-996. [DOI] [PubMed] [Google Scholar]
- 20.Shah MN, Stoev IT, Sanford DE, et al. Are readmission rates on a neurosurgical service indicators of quality of care? J Neurosurg. 2013;119(4):1043-1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ferro EG, Secemsky EA, Wadhera RK, et al. Patient readmission rates for all insurance types after implementation of the hospital readmissions reduction program. Health Aff (Millwood). 2019;38(4):585-593. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Volz A, Schmid J-P, Zwahlen M, et al. Predictors of readmission and health related quality of life in patients with chronic heart failure: a comparison of different psychosocial aspects. J Behav Med. 2011;34(1):13-22. [DOI] [PubMed] [Google Scholar]
- 23.Ingraham AM, Cohen ME, Bilimoria KY, et al. Association of Surgical care improvement project infection-related process measure compliance with risk-adjusted outcomes: implications for quality measurement. J Am Coll Surg. 2010;211(6):705-714. [DOI] [PubMed] [Google Scholar]
- 24.Khera R, Dharmarajan K, Wang Y, et al. Association of the hospital readmissions reduction program with mortality during and after hospitalization for acute myocardial infarction, heart failure, and pneumonia. JAMA Netw Open. 2018;1(5):e182777–e182777. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Azoulay M, Santos F, Shenouda G, et al. Benefit of re-operation and salvage therapies for recurrent glioblastoma multiforme: results from a single institution. J Neurooncol. 2017;132(3):419-426. [DOI] [PubMed] [Google Scholar]
- 26.Molinaro AM, Hervey-Jumper S, Morshed RA, et al. Association of maximal extent of resection of contrast-enhanced and non-contrast-enhanced tumor with survival within molecular subgroups of patients with newly diagnosed glioblastoma. JAMA Oncol. 2020;6(4):495-503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kerezoudis P, Glasgow AE, Alvi MA, et al. Returns to operating room after neurosurgical procedures in a tertiary care academic medical center: implications for health care policy and quality improvement. Neurosurgery. 2018;84(6):E392–E401. [DOI] [PubMed] [Google Scholar]
- 28.Senders JT, Muskens IS, Cote DJ, et al. Thirty-day outcomes after craniotomy for primary malignant brain tumors: a national surgical quality improvement program analysis. Neurosurgery. 2018;83(6):1249-1259. [DOI] [PubMed] [Google Scholar]
- 29.Liu J, Gormley N, Dasenbrock HH, et al. Cost-benefit analysis of transitional care in neurosurgery. Neurosurgery. 2019;85(5):672-679. [DOI] [PubMed] [Google Scholar]
- 30.Dasenbrock HH, Yan SC, Smith TR, et al. Readmission after craniotomy for tumor: a national surgical quality improvement program analysis. Neurosurgery. 2017;80(4):551-562. [DOI] [PubMed] [Google Scholar]
- 31.Ansari SF, Yan H, Zou J, et al. Hospital length of stay and readmission rate for neurosurgical patients. Neurosurgery. 2017;82(2):173-181. [DOI] [PubMed] [Google Scholar]
- 32.Dasenbrock HH, Liu KX, Devine CA, et al. Length of hospital stay after craniotomy for tumor: a National Surgical Quality Improvement Program analysis. Neurosurg Focus. 2015;39(6):E12. [DOI] [PubMed] [Google Scholar]
- 33.Lakomkin N, Hadjipanayis CG. Non-routine discharge disposition is associated with post-discharge complications and 30-day readmissions following craniotomy for brain tumor resection. J Neurooncol. 2018;136(3):595-604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Chen SY, Stem M, Gearhart SL, et al. Readmission adversely affects survival in surgical rectal cancer patients. World J Surg. 2019;43(10):2506-2517. [DOI] [PubMed] [Google Scholar]
- 35.Loureiro LVM, Victor EDS, Callegaro-Filho D, et al. Minimizing the uncertainties regarding the effects of delaying radiotherapy for glioblastoma: a systematic review and meta-analysis. Radiother Oncol. 2016;118(1):1-8. [DOI] [PubMed] [Google Scholar]
- 36.Pollom EL, Fujimoto DK, Han SS, et al. Newly diagnosed glioblastoma: adverse socioeconomic factors correlate with delay in radiotherapy initiation and worse overall survival. J Radiat Res. 2018;59(suppl_1):i11-i18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Lu VM, Jue TR, McDonald KL, et al. The survival effect of repeat surgery at glioblastoma recurrence and its trend: a systematic review and meta-analysis. World Neurosurg. 2018;115:453-459.e3. [DOI] [PubMed] [Google Scholar]
- 38.Silver JK, Baima J, Mayer RS. Impairment-driven cancer rehabilitation: an essential component of quality care and survivorship. CA Cancer J Clin. 2013;63(5):295-317. [DOI] [PubMed] [Google Scholar]
- 39.Lindbäck Y, Tropp H, Enthoven P, et al. PREPARE: presurgery physiotherapy for patients with degenerative lumbar spine disorder: a randomized controlled trial. Spine J. 2018;18(8):1347-1355. [DOI] [PubMed] [Google Scholar]
- 40.Nielsen PR, Andreasen J, Asmussen M, et al. Costs and quality of life for prehabilitation and early rehabilitation after surgery of the lumbar spine. BMC Health Serv Res. 2008;8:209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Worrell SL, Kirschner ML, Shatz RS, et al. Interdisciplinary approaches to survivorship with a focus on the low-grade and benign brain tumor populations. Curr Oncol Rep. 2021;23(2):19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Rivera-Rivera PA, Rios-Lago M, Sanchez-Casarrubios S, et al. Cortical plasticity catalyzed by prehabilitation enables extensive resection of brain tumors in eloquent areas. J Neurosurg. 2017;126(4):1323-1333. [DOI] [PubMed] [Google Scholar]
- 43.Silver JK, Baima J. Cancer prehabilitation: an opportunity to decrease treatment-related morbidity, increase cancer treatment options, and improve physical and psychological health outcomes. Am J Phys Med Rehabil. 2013;92(8):715-727. [DOI] [PubMed] [Google Scholar]
- 44.Dewberry LC, Wingrove LJ, Marsh MD, et al. Pilot prehabilitation program for patients with esophageal cancer during neoadjuvant therapy and surgery. J Surg Res. 2019;235:66-72. [DOI] [PubMed] [Google Scholar]
- 45.Liu Z, Qiu T, Pei L, et al. Two-Week multimodal prehabilitation program improves perioperative functional capability in patients undergoing thoracoscopic lobectomy for lung cancer: a randomized controlled trial. Anesth Analg. 2020;131(3):840-849. [DOI] [PubMed] [Google Scholar]
- 46.Laurent H, Aubreton S, Galvaing G, et al. Preoperative respiratory muscle endurance training improves ventilatory capacity and prevents pulmonary postoperative complications after lung surgery. Eur J Phys Rehabil Med. 2020;56(1):73-81. [DOI] [PubMed] [Google Scholar]
- 47.Dai YT, Chang Y, Hsieh CY, et al. Effectiveness of a pilot project of discharge planning in Taiwan. Res Nurs Heal. 2003;26(1):53-63. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplement. eMethods 1. Pertinent Data Points. eMethods 2. Statistical Analysis—Multivariable Logistic Regression. eMethods 3. Statistical Analysis – Multivariable Cox Proportional Hazards Model. eResults 1. Presenting Patient and Case Characteristics. eResults 2. Readmitted Patients and Preventability. eResults 3. Palliative Care Team Inclusion and Readmission.