https://orcid.org/0000-0002-4705-0262
Figures
Abstract
Background
Treatment options for mitral regurgitation range from diuretic therapy, to surgical and interventional strategies including TMVR in high-risk surgical candidates. Frailty has been associated with inferior outcomes following hospitalizations for heart failure and in open cardiac surgery.
Objective
The purpose of the present study was to evaluate the impact of frailty on clinical outcomes and resource use following transcatheter mitral valve repair (TMVR).
Methods
Adults undergoing TMVR were identified using the 2016–2018 Nationwide Readmissions Database, and divided into Frail and Non-Frail groups. Frailty was defined using a derivative of the Johns Hopkins Adjusted Clinical Groups frailty indicator. Generalized linear models were used to assess the association of frailty with in-hospital mortality, complications, nonhome discharge, hospitalization costs, length of stay, and non-elective readmission at 90 days. Average marginal effects were used to quantify the impact of frailty on predicted mortality.
Results
Of 18,791 patients undergoing TMVR, 11.6% were considered frail. The observed mortality rate for the overall cohort was 2.2%. After adjustment, frailty was associated with increased odds of in-hospital mortality (AOR 1.8, 95% CI 1.2–2.6), corresponding to an absolute increase in risk of mortality of 1.1%. Frailty was associated with a 2.7-day (95% CI 2.1–3.2) increase in postoperative LOS, and $18,300 (95% CI 14,400–22,200) increment in hospitalization costs. Frail patients had greater odds (4.4, 95% CI 3.6–5.4) of nonhome discharge but similar odds of non-elective 90-day readmission.
Citation: Hadaya J, Tran Z, Sanaiha Y, Aguayo E, Dobaria V, Calfon Press M, et al. (2021) Impact of frailty on short term outcomes, resource use, and readmissions after transcatheter mitral valve repair: A national analysis. PLoS ONE 16(11): e0259863. https://doi.org/10.1371/journal.pone.0259863
Editor: George C.M. Siontis, Bern University Hospital, SWITZERLAND
Received: August 22, 2021; Accepted: October 27, 2021; Published: November 18, 2021
Copyright: © 2021 Hadaya et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data cannot be shared publicly because of the Agency for Healthcare Research and Quality's policies. Data are available from the Agency for Healthcare Research and Quality for researchers who meet the criteria for access to confidential data after completion of a data-use agreement. Data use agreements are available at https://www.hcup-us.ahrq.gov/team/NationwideDUA.jsp and can be submitted to the Healthcare Cost and Utilization Project Central Distributor through email at HCUP@AHRQ.gov.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Mitral regurgitation (MR) is the second most common valvular disease in modern countries and is strongly associated with atrial fibrillation, congestive heart failure, and poor quality of life [1, 2]. Over three decades ago, Carpentier formalized the classification of mitral valve pathologies and proposed durable and tailored repair methods [3]. However, many patients at high surgical risk, such as those with reduced left ventricular function and pulmonary hypertension, have historically not been offered surgical therapy. Such patients often suffer from multiple heart failure episodes and require repeat hospitalizations for diuretic therapy.
Transcatheter mitral valve repair (TMVR) has recently emerged as an alternative to surgery for the treatment of severe symptomatic MR, using catheter-based methods to appose the two leaflets in regions of malcoaptation [4, 5]. Although randomized trials have not demonstrated an impact on life expectancy, TMVR has been found efficacious in reducing repeat hospitalizations [6]. Compared to surgical candidates, patients undergoing TMVR are often older and carry a significant burden of comorbidities, leading to suboptimal outcomes following this intervention [7, 8].
While several trials and registries have examined patient outcomes based on general preexisting conditions and laboratory data, few have studied the impact of frailty in this cohort. Despite the lack of a universal definition or assessment, frailty has been associated with poor surgical and procedural outcomes [9–11]. This may be attributable, in part, to factors such as age, chronic diseases, and an inability to withstand acute physiologic stress [12, 13]. Frailty has been shown to increase 1-year postoperative mortality in several operative categories, ranging from oncologic operations to open cardiac surgery [14–18]. However, frailty may have a less profound impact on outcomes following less invasive procedures such as TMVR. Kundi et al. used the Hospital Frailty Risk Score (HFRS) in Medicare patients undergoing TMVR and found an association between frailty and 1-year mortality rates [19]. However, this study was limited by a small sample size, as well as a frailty indicator that included many diagnoses that correspond to in-hospital complications, hindering the study’s interpretation and limiting its applicability in a priori determination of risk. Recent studies examining frailty in surgical patients have utilized the Johns Hopkins Adjusted Clinical Groups frailty indicator (ACG) and its derivatives, which are based on frailty-defining diagnoses with minimal overlap with complications and traditional procedural risk factors [14, 20, 21]. The present study aimed to assess the impact of frailty as determined by the ACG on in-hospital mortality, complications, and resource use in a national cohort of patients undergoing TMVR.
Methods
Data source and cohort definitions
We performed a cohort study of all adult patients who underwent TMVR from 2016 to 2018 using the Nationwide Readmissions Database (NRD) [22]. The NRD is an all-payer inpatient database maintained by the Agency for Healthcare Research and Quality (AHRQ) as part of the Healthcare Cost and Utilization Project. The NRD provides nationally representative estimates of >57% of all inpatient hospitalizations in the United States annually [22]. The NRD contains linkage numbers for all sample patients, allowing for readmissions within each calendar year to be tracked across hospitals.
International Classification of Disease, Tenth Revision, Procedure Coding System (ICD-10-PCS) codes were used to identify adult patients (ages ≥ 18 years) who underwent TMVR (02UG3JZ, 02UG4JZ, 02QG4ZZ, 02QG3ZZ) from 2016–2018. Years prior to 2016 were not studied due to low sample size and the transition from International Classification of Disease, Ninth Revision to ICD-10-PCS codes. Patients were divided into Frail and Non-Frail cohorts using International Classification of Disease, Tenth Revision, Clinical Modification (ICD-10-CM) codes corresponding to the Johns Hopkins Adjusted Clinical Groups (ACG) frailty-defining diagnoses. The indicator is derived from ten groupings of frailty-defining diagnoses including malnutrition, dementia, vision impairment, decubitus ulcer, urinary and fecal incontinence, weight loss, falls, difficulty walking, poverty, and barriers to healthcare access. The presence of one or more diagnosis groupings was used to categorize a patient as frail [20, 21].
Variable definitions and outcomes
Patient and admission level characteristics were defined in accordance with the NRD Data Dictionary including age, sex, urgency of admission (elective versus urgent or emergent) and primary payer [22]. Similarly, hospital level variables available in NRD included bed size and teaching status. Comorbidities were further defined using ICD-10-CM codes and the Elixhauser Comorbidity Index, a previously validated numeric burden of 30 common conditions [23]. Complications were categorized into cardiac (cardiac arrest, myocardial infarction, cardiac tamponade, ventricular fibrillation, and ventricular tachycardia), pulmonary (pneumonia, pneumothorax, acute respiratory distress syndrome, respiratory failure), infectious (sepsis, septicemia, wound infection) and renal (acute kidney injury) as previously described [24]. Hospitalization costs were determined in accordance with methodology reported by HCUP [25]. Briefly, total hospital charges were converted to costs using hospital-specific cost-to-charge ratios published by the AHRQ and then inflation-adjusted to 2018 using the Bureau of Labor Statistics Consumer Price Index [26].
The primary outcomes of the study were in-hospital mortality, perioperative complications, non-home discharge and nonelective readmissions, defined as within 90-days of index hospitalization. Secondary outcomes included postoperative length of stay (LOS), adjusted hospitalization costs and diagnoses for rehospitalization. Diagnosis Related Groups (DRG) in combination with ICD codes were used to identify principal readmission diagnoses as previously described [27].
Statistical analysis
All statistical analysis was performed with Stata 16.0 (StataCorp, College Station, TX) using survey-specific methods to account for clustering and stratification. Patients with missing age, sex, mortality data or discharge disposition, and costs were excluded from analysis (56, or 0.3% of final cohort). Continuous variables are reported as mean with standard deviation or median with interquartile range (IQR) if non-normally distributed. Categorical variables are reported as frequency or proportion. Patient and hospital characteristics were compared between cohorts using chi-square and adjusted Wald tests. Freedom from readmission was evaluated utilizing Kaplan-Meier survival analysis with the log-rank test used to assess the significance of frailty on readmission. Multivariable logistic and linear regression models were developed to evaluate the association of frailty with outcomes of interest. Elastic Net with retention of clinically-relevant characteristics was used for variable selection [28]. Briefly, Elastic Net utilizes a regressive least squares methodology to select explanatory variables aimed at reducing collinearity while applying penalties to decrease overfitting. Final models were evaluated by the area under the receiver operating characteristics curve and Akaike information criteria. Adjusted impact of independent variables are reported as adjusted odds ratios (AOR), beta-coefficients (β), and average marginal effects. To determine marginal effects, the Stata margins command was used to calculate point estimates and confidence intervals. Statistical significance was defined as α<0.05. This study was deemed exempt from full review by the Institutional Review Board at the University of California, Los Angeles. Specific consent from individual patients was not required due to the deidentified nature of the data set.
Results
Characteristics of frail and non-frail groups
Of an estimated 18,791 patients undergoing TMVR during the study period, 2,179 (11.6%) comprised the Frail group. Baseline characteristics of the Frail and Non-Frail groups are reported in Table 1. The most common frailty defining diagnoses were dementia (33.0%) and malnutrition (37.0%). Compared to Non-Frail, patients in the Frail group were older (78.9±10.5 years vs 77.3±10.8 years, P<0.001) and had a greater aggregate burden of comorbidities as defined by the Elixhauser Comorbidity Score (7.0±2.3 vs 5.6±2.0, P<0.001). Specifically, the Frail group had higher rates of congestive heart failure, coagulopathy, electrolyte disorders, but lower a lower incidence of peripheral vascular disease compared to the Non-Frail group (Table 1). Frail patients on average experienced a longer preoperative length of stay (5.3±9.4 days vs 1.3±4.0 days, P<0.001) and a greater proportion of non-elective admissions (51.7% vs 78.6%, P<0.001) compared to Non-Frail. There were no significant differences in hospital characteristics by frailty status with similar proportion treated by teaching status and bed size.
Unadjusted outcomes in frail versus non-frail patients following TMVR
Unadjusted outcomes for the two groups are presented in Table 2. Mean observed mortality rate for the entire cohort was 2.2%. Compared to their counterparts, Frail patients had higher unadjusted in-hospital mortality (6.0% vs 1.7%, P<0.001). In addition, Frail patients had higher unadjusted rates of cardiovascular, respiratory, renal and infectious complications (Table 2). Median costs of index hospitalization were significantly greater in the Frail relative to Non-Frail group ($55,200, IQR 38,300–85,400 vs $41,400, IQR 31,000–55,500, P<0.001). Postoperative length of stay was significantly greater in the Frail compared to the Non-Frail (6.3±8.8 days vs 2.7±3.8 days, P<0.001). Among those surviving the index admission, Frail patients had a higher proportion of nonhome destinations (41.1% vs 9.65%, P<0.001). Finally, Frail had a greater proportion of readmissions within 90 days (23.3% vs 17.1%, P<0.001) (Fig 1). As demonstrated in Table 3, the most common reasons for readmission were cardiovascular, fluid/electrolyte, and infectious reasons, with a similar distribution in Frail and Non-Frail groups.
Impact of frailty on risk adjusted outcomes following TMVR
After adjustment for patient and hospital characteristics, frailty was associated with increased odds of in-hospital mortality (AOR 1.7, 95% CI 1.2–2.6), corresponding to an average marginal effect of 1.1% (95% CI 0.3–1.8, Fig 1). Other factors predictive of in-hospital mortality included liver disease with an average marginal effect of 3.5% (95% CI 2.3–4.6), congestive heart failure with an average marginal effect of 1.9% (95% CI 0.5–3.3), and Elixhauser Comorbidity Index ≥ 10 with an average marginal effect of 4.0% (95% CI 1.2–6.9). As demonstrated in Fig 2, frailty was associated with increased odds of all studied perioperative complications.
Outcomes presented as odds ratio with 95% confidence interval for Frail relative to Non-Frail. C-statistic: mortality (0.83), cardiac complication (0.72), pulmonary complication (0.78), infectious complication (0.84), renal complication (0.80), non-home discharge (0.81) and 90-day readmission (0.63). All multivariable models included adjustment for age, year, sex, chronic lung disease, diabetes, hypothyroidism, end stage renal disease, malignancy, payer status, income quartile, hospital bed size, elective admission and Elixhauser Comorbidity Index.
Relative to Non-Frail, Frail patients had a 2.7-day (95% CI 2.1–3.2) increase in adjusted postoperative length of stay. Likewise, frailty was associated with an increase of $18,300 (95% CI 14,400–22,200) in hospitalization costs. Among patients who survived to discharge, frailty was associated with increased odds (AOR 4.4, 95% CI 3.6–5.4) of nonhome discharge. Finally, adjusted odds of 90-day readmission was similar among cohorts.
Discussion
An existing body of literature has reported frailty to be associated with worse outcomes in heart failure, open cardiac operations and transcatheter aortic valve replacement [16, 29, 30]. Although TMVR may be a viable alternative to surgical repair in high risk patients, the impact of frailty in this group has not been previously examined in detail. We used the largest publicly available database to investigate the effect of frailty, as assessed using a simple, administrative coding-based tool, on outcomes following TMVR and make several important observations. After adjusting for available patient and hospital characteristics, frailty remained a major independent predictor of mortality among patients undergoing TMVR. Similarly, frailty was associated with prolonged hospitalization and resource use as well as the need for post-discharge medical care but not readmissions. Our findings add to the existing body of literature on frailty and have important implications for patient selection and counseling. Furthermore, the present study demonstrates the ability of a coding-based frailty tool to provide additional prognostic value to existing administrative risk models.
Despite its recognized impact on postprocedural outcomes, measures of frailty have not been routinely incorporated into clinical practice. While frailty assessment tools exceed 30 in number, most are time intensive and require additional resources to administer [13, 31, 32]. Furthermore, the accuracy of such instruments has been questioned with a study demonstrating disagreement rates of 35–74% among various methods [33]. Administrative data provides a readily available alternate means to assess risk profiles including frailty. The commonly used HFRS score, derived from the National Health Services in England, provides a numeric score for frailty that has been examined in several studies [19]. While it provides acceptable discriminatory power and is associated with a differential in midterm outcomes, the HFRS has several shortcomings that may limit its utility in the preoperative setting. Firstly, dichotomization of this continuous score is arbitrary with thresholds that are likely to vary between procedure and investigators. Furthermore, the HFRS diagnostic codes include many that may be in fact postoperative complications. Thus, the prospective ability of this score to forecast procedural and surgical outcomes is uncertain. In the present study, we utilized a dichotomous frailty variable derived from a cluster of diagnostic codes as initially developed by the Johns Hopkins investigators. While the exact methodology is proprietary, we utilized codes validated in other studies to perform this analysis. This algorithm provided independent discriminatory power in our group of TMVR patients and was associated with mortality and several other endpoints of the present study. The ICD codes used in the ACG algorithm are conditions in several domains that are chronic, do not overlap with traditional surgical or procedural risk factors, and are unlikely to be related to acute hospitalization [14, 20, 21]. Whether a coding-derived frailty indicator can improve prospective determination of procedural risk warrants further investigation.
A previous single-center study of 213 patients by Metze et al. found the number of frail patients receiving TMVR was 10 times higher than rates of frail patients receiving surgical mitral valve repair, supporting the likelihood that patients not considered appropriate candidates for conventional surgical repair will be referred for percutaneous intervention [34]. While the study found frail patients to have comparable initial device success rates to non-frail patients, midterm mortality and heart failure were significantly greater in frail patients during follow-up. Poorer midterm outcomes in frail patients may be attributable to the increased risk for overall and individual postoperative complications observed in the present study. Furthermore, the association between frailty and increased LOS, hospitalization costs, and nonhome discharge observed in our study reflect the intensity of care required by frail patients beyond discharge. Previous studies have shown frailty to be associated with physical derangements such as malnutrition, weight loss, and dementia, factors that may reduce the ability to withstand perioperative stressors [11]. These factors may predispose frail patients to increased risk for complications, thereby increasing midterm mortality [31]. Interestingly, frailty was not associated with readmissions after risk-adjustment, suggesting that readmission in this patient population is driven by other factors, such as congestive heart failure, kidney dysfunction, and a greater burden of comorbidities. The present study provides the first nationwide analysis of index hospitalization for TMVR that may further forecast inferior long-term outcomes in frail patients receiving this procedure. Given the paucity of available data, further investigation of specific frailty-associated comorbidities that enhance procedural risk in the TMVR patient population is warranted.
The aforementioned study by Metze et al. identified 45.5% of patients as frail. Another recent analysis of TMVR outcomes using the Society of Thoracic Surgery Transcatheter Valve Therapy Registry by Sorajja et al. found a similarly high incidence of frailty, at 50.3% [35]. In contrast to these findings, our study identified frailty in comparatively less patients at approximately 12%. Beyond differences in sample size, this variation may be attributable to the differing methodologies used to identify frail patients. Metze et al. utilized the Fried criteria, which comprises five clinical characteristics including weakness, unintentional weight loss, exhaustion, slow gait, and low physical activity [36]. The Johns Hopkins ACG frailty indicator likewise provides a comprehensive and standardized procedure for identifying frail patients. In addition to substantiation by previous surgical studies of patient outcomes, it has been externally validated using the Vulnerable Elderly Scale and has been shown to accurately capture patients with limitations in activities of daily living [37]. Furthermore, its utilization of ICD- coding offers a uniform method of assessing frailty in the absence of granular clinical characteristics relied upon by other frailty indices, such as the Fried criteria. As frailty encompasses a broad range of clinical diagnoses, it is important that future studies of percutaneous repair adopt a consistent approach towards procedural risk in TMVR candidates. This standardization may better inform shared decision making in patients who are at particularly high general risk of deteriorating quality of life.
Limitations
The present study had several important limitations. The administrative nature of the NRD precludes access to granular physiologic and echocardiographic parameters such as ejection fraction and intracardiac pressures as well as New York Heart Association class. Moreover, diagnoses and procedures were identified using ICD-10 codes, which are dependent on hospital coding practices and may be prone to bias. However, we selected ICD-10-CM codes that were either previously validated or clinically relevant to inform multivariable models. Additionally, we utilized robust statistical methods, such as Elastic Net for variable selection, to reduce bias within the confines of available data. Furthermore, the data in the present study was limited to in-hospital outcomes and, as such, are unable to comment on the impact of frailty on long-term outcomes. Nonetheless, the NRD is unique in providing accurate estimates for all US hospitalizations and information on resource utilization, a facet not provided by registry data.
Conclusion
We used a large nationally representative cohort of TMVR patients and found frailty to be independently associated increased mortality, complications and resource utilization. Frailty further was associated with the need for post-discharge medical care but not readmission. Frailty derived from an administrative coding-based tool appears to serve as a powerful predictor of perioperative benchmarks, and may better inform shared decision making. With increased adoption of transcatheter therapies, incorporation of frailty into existing risk models should be strongly considered.
References
- 1. Glenn WW, Turk LN. The surgical treatment of mitral insufficiency: the fate of a vascularized transchamber intracardiac graft. Ann Surg. 1955;141: 510–518. pmid:14362383
- 2. El Sabbagh A, Reddy YNV, Nishimura RA. Mitral Valve Regurgitation in the Contemporary Era: Insights Into Diagnosis, Management, and Future Directions. JACC Cardiovasc Imaging. 2018;11: 628–643. pmid:29622181
- 3. Carpentier A. Cardiac valve surgery—the "French correction". J Thorac Cardiovasc Surg. 1983;86: 323–337. pmid:6887954
- 4. Stone GW, Lindenfeld J, Abraham WT, Kar S, Lim DS, Mishell JM, et al. Transcatheter Mitral-Valve Repair in Patients with Heart Failure. N Engl J Med. 2018;379: 2307–2318. pmid:30280640
- 5. Obadia J-F, Messika-Zeitoun D, Leurent G, Iung B, Bonnet G, Piriou N, et al. Percutaneous Repair or Medical Treatment for Secondary Mitral Regurgitation. N Engl J Med. 2018;379: 2297–2306. pmid:30145927
- 6. Kebler M, Seeger J, Muche R, Wöhrle J, Rottbauer W, Markovic S. Predictors of rehospitalization after percutaneous edge-to-edge mitral valve repair by MitraClip implantation. Eur J Heart Fail. 2019;21: 182–192. pmid:30178493
- 7. Velazquez EJ, Samad Z, Al-Khalidi HR, Sangli C, Grayburn PA, Massaro JM, et al. The MitraClip and survival in patients with mitral regurgitation at high risk for surgery: A propensity-matched comparison. Am Heart J. 2015;170: 1050–1059.e3. pmid:26542516
- 8. Mirabel M, Iung B, Baron G, Messika-Zeitoun D, Détaint D, Vanoverschelde JL, et al. What are the characteristics of patients with severe, symptomatic, mitral regurgitation who are denied surgery? Eur Heart J. 2007;28: 1358–1365. pmid:17350971
- 9. Adams P, Ghanem T, Stachler R, Hall F, Velanovich V, Rubinfeld I. Frailty as a predictor of morbidity and mortality in inpatient head and neck surgery. JAMA Otolaryngol—Head Neck Surg. 2013;139: 783–789. pmid:23949353
- 10. Makary MA, Segev DL, Pronovost PJ, Syin D, Bandeen-Roche K, Patel P, et al. Frailty as a Predictor of Surgical Outcomes in Older Patients. J Am Coll Surg. 2010;210: 901–908. pmid:20510798
- 11. Mrdutt MM, Papaconstantinou HT, Robinson BD, Bird ET, Isbell CL. Preoperative Frailty and Surgical Outcomes Across Diverse Surgical Subspecialties in a Large Health Care System. J Am Coll Surg. 2019;228: 482–490. pmid:30885474
- 12. Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the Concepts of Disability, Frailty, and Comorbidity: Implications for Improved Targeting and Care. Journals Gerontol Ser A Biol Sci Med Sci. 2004;59: M255–M263.
- 13. Morley JE, Vellas B, Abellan van Kan G, Anker SD, Bauer JM, Bernabei R, et al. Frailty consensus: A call to action. J Am Med Dir Assoc. 2013;14: 392–397. pmid:23764209
- 14. McIsaac DI, Bryson GL, Van Walraven C. Association of frailty and 1-year postoperative mortality following major elective noncardiac surgery: A population-based cohort study. JAMA Surg. 2016;151: 538–545. pmid:26791334
- 15. Koh LY, Hwang NC. Frailty in Cardiac Surgery. J Cardiothorac Vasc Anesth. 2019;33: 521–531. pmid:29580797
- 16. Lee DH, Buth KJ, Martin BJ, Yip AM, Hirsch GM. Frail patients are at increased risk for mortality and prolonged institutional care after cardiac surgery. Circulation. 2010;121: 973–978. pmid:20159833
- 17. Gilbert T, Neuburger J, Kraindler J, Keeble E, Smith P, Ariti C, et al. Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet. 2018;391: 1775–1782. pmid:29706364
- 18. McAlister F, Van Walraven C. External validation of the Hospital Frailty Risk Score and comparison with the Hospital-patient One-year Mortality Risk Score to predict outcomes in elderly hospitalised patients: a retrospective cohort study. BMJ Qual Saf. 2019;28: 284–288. pmid:30381331
- 19. Kundi H, Popma JJ, Reynolds MR, Strom JB, Pinto DS, Valsdottir LR, et al. Frailty and related outcomes in patients undergoing transcatheter valve therapies in a nationwide cohort. Eur Heart J. 2019;40: 2231–2239. pmid:30977798
- 20. Nieman CL, Pitman KT, Tufaro AP, Eisele DW, Frick KD, Gourin CG. The effect of frailty on short-term outcomes after head and neck cancer surgery. Laryngoscope. 2018;128: 102–110. pmid:28731497
- 21. Dobaria V, Hadaya J, Sanaiha Y, Aguayo E, Sareh S, Benharash P. The Pragmatic Impact of Frailty on Outcomes of Coronary Artery Bypass Grafting. Ann Thorac Surg. 2021;112: 108–115. pmid:33080240
- 22.
NRD Overview. [cited 1 Aug 2021]. Available: https://www.hcup-us.ahrq.gov/nrdoverview.jsp
- 23. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity Measures for Use with Administrative Data. Med Care. 1998;36: 8–27. pmid:9431328
- 24. Sanaiha Y, Rudasill S, Sareh S, Mardock A, Khoury H, Ziaeian B, et al. Impact of hospital safety-net status on failure to rescue after major cardiac surgery. Surgery. 2019;166: 778–784. pmid:31307773
- 25.
HCUP Supplemental Files. [cited 1 Aug 2021]. Available: https://www.hcup-us.ahrq.gov/toolssoftware/supplemental.jsp#ccr
- 26.
Using Appropriate Price Indices for Expenditure Comparisons. [cited 1 Aug 2021]. Available: https://meps.ahrq.gov/about_meps/Price_Index.shtml
- 27. Sanaiha Y, Kavianpour B, Mardock A, Khoury H, Downey P, Rudasill S, et al. Rehospitalization and resource use after inpatient admission for extracorporeal life support in the United States. Surgery. 2019;166: 829–834. pmid:31277884
- 28. Zou H, Hastie T. Regularization and variable selection via the elastic net. J R Stat Soc B. 2005;67: 301–320.
- 29. Vidán MT, Blaya-Novakova V, Sánchez E, Ortiz J, Serra-Rexach JA, Bueno H. Prevalence and prognostic impact of frailty and its components in non-dependent elderly patients with heart failure. Eur J Heart Fail. 2016;18: 869–875. pmid:27072307
- 30. Green P, Woglom AE, Genereux P, Daneault B, Paradis JM, Schnell S, et al. The impact of frailty status on survival after transcatheter aortic valve replacement in older adults with severe aortic stenosis: A single-center experience. JACC Cardiovasc Interv. 2012;5: 974–981. pmid:22995885
- 31. Abellan Van Kan G, Rolland Y, Andrieu S, Bauer J, Beauchet O, Bonnefoy M, et al. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) task force. J Nutr Heal Aging. 2009;13: 881–889. pmid:19924348
- 32. Engelhardt KE, Reuter Q, Liu J, Bean JF, Barnum J, Shapiro MB, et al. Frailty screening and a frailty pathway decrease length of stay, loss of independence, and 30-day readmission rates in frail geriatric trauma and emergency general surgery patients. J Trauma Acute Care Surg. 2018;85: 167–173. pmid:29659475
- 33. Afilalo J, Lauck S, Kim DH, Lefèvre T, Piazza N, Lachapelle K, et al. Frailty in Older Adults Undergoing Aortic Valve Replacement: The FRAILTY-AVR Study. J Am Coll Cardiol. 2017;70: 689–700. pmid:28693934
- 34. Metze C, Matzik AS, Scherner M, Körber MI, Michels G, Baldus S, et al. Impact of Frailty on Outcomes in Patients Undergoing Percutaneous Mitral Valve Repair. JACC Cardiovasc Interv. 2017;10: 1920–1929. pmid:28917516
- 35. Sorajja P, Vemulapalli S, Feldman T, Mack M, Holmes DR, Stebbins A, et al. Outcomes With Transcatheter Mitral Valve Repair in the United States: An STS/ACC TVT Registry Report. J Am Coll Cardiol. 2017;70: 2315–2327. pmid:29096801
- 36. Bieniek J, Wilczyński K, Szewieczek J. Fried frailty phenotype assessment components as applied to geriatric inpatients. Clin Interv Aging. 2016;11: 453–459. pmid:27217729
- 37. Sternberg SA, Bentur N, Abrams C, Spalter T, Karpati T, Lemberger J, et al. Identifying frail older people using predictive modeling. Am J Manag Care. 2012;18: 392–397. pmid:23145847