Trends in Noninvasive and Invasive Mechanical Ventilation Among Medicare Beneficiaries at the End of Life | Cardiology | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Figure 1.  Trends in Ventilatory Support at the End of Life, 2000-2017
Trends in Ventilatory Support at the End of Life, 2000-2017

Overall, 401 419 (16.3%) Medicare beneficiaries received invasive mechanical ventilation (IMV), 90 700 (3.7%) received noninvasive ventilation (NIV), and 25 689 (1.0%) received both IMV and NIV. The use of NIV increased almost 9-fold from 0.8% in 2000 to 7.1% in 2017, whereas the use of IMV was stable and then increased slightly from 15.0% in 2000 to 18.5% in 2017 in the last 30 days of life.

Figure 2.  Trends in Use of Noninvasive Ventilation (NIV), Invasive Mechanical Ventilation (IMV), or Both at the End of Life by Admitting Diagnosis, 2000-2017
Trends in Use of Noninvasive Ventilation (NIV), Invasive Mechanical Ventilation (IMV), or Both at the End of Life by Admitting Diagnosis, 2000-2017

Among decedents with congestive heart failure (CHF) and chronic obstructive pulmonary disease (COPD), increases in NIV use are balanced by reciprocal decreases and then stabilization in IMV use. Among decedents with cancer and dementia, statistically significant increases in NIV use occurred with concomitant increases in IMV use.

Figure 3.  Rates of In-Hospital Death, Hospice Enrollment at Death, and Late Hospice Enrollment in the Last 3 Days of Life by Ventilatory Support and Admitting Diagnosis, 2013-2017
Rates of In-Hospital Death, Hospice Enrollment at Death, and Late Hospice Enrollment in the Last 3 Days of Life by Ventilatory Support and Admitting Diagnosis, 2013-2017

Among all diagnosis subgroups (congestive heart failure [CHF], chronic obstructive pulmonary disease [COPD], cancer, and dementia), noninvasive ventilation (NIV) use compared with invasive mechanical ventilation (IMV) was associated with lower rates of in-hospital death in acute inpatient sites (50.3% [95% CI, 49.3%-51.3%] vs 76.7% [95% CI, 75.9%-77.5%]) and hospice enrollment in the last 3 days of life (57.7% [95% CI, 56.2%-59.3%] vs 63.0% [95% CI, 60.9%-65.1%]) and with a higher rate of hospice enrollment at death (41.3% [95% CI, 40.4%-42.3%] vs 20.0% [95% CI, 19.2%-20.7%]). Decedents with CHF and COPD who received NIV had similar rates of in-hospital death, hospice enrollment at death, and hospice enrollment in the last 3 days of life compared with decedents with cancer and dementia. In-hospital death and hospice enrollment outcomes are expressed as % of hospitalizations, whereas hospice enrollment in the last 3 days of life outcome is expressed as % of hospice enrollees. Error bars represent 95% CIs.

Table 1.  Characteristics of Decedents With 1 or More Hospitalizations in Their Last Month of Life, by Year
Characteristics of Decedents With 1 or More Hospitalizations in Their Last Month of Life, by Year
Table 2.  Characteristics of Decedents With 1 or More Hospitalizations in Their Last Month of Life, by Ventilatory Support
Characteristics of Decedents With 1 or More Hospitalizations in Their Last Month of Life, by Ventilatory Support
1.
Teno  JM, Gozalo  P, Trivedi  AN,  et al.  Site of death, place of care, and health care transitions among US Medicare beneficiaries, 2000-2015.   JAMA. 2018;320(3):264-271. doi:10.1001/jama.2018.8981 PubMedGoogle ScholarCrossref
2.
Barnato  AE, Herndon  MB, Anthony  DL,  et al.  Are regional variations in end-of-life care intensity explained by patient preferences?: a study of the US Medicare population.   Med Care. 2007;45(5):386-393. doi:10.1097/01.mlr.0000255248.79308.41 PubMedGoogle ScholarCrossref
3.
Fisher  ES, Wennberg  DE, Stukel  TA, Gottlieb  DJ, Lucas  FL, Pinder  EL.  The implications of regional variations in Medicare spending. Part 2: health outcomes and satisfaction with care.   Ann Intern Med. 2003;138(4):288-298. doi:10.7326/0003-4819-138-4-200302180-00007 PubMedGoogle ScholarCrossref
4.
Teno  JM, Mor  V, Ward  N,  et al.  Bereaved family member perceptions of quality of end-of-life care in U.S. regions with high and low usage of intensive care unit care.   J Am Geriatr Soc. 2005;53(11):1905-1911. doi:10.1111/j.1532-5415.2005.53563.x PubMedGoogle ScholarCrossref
5.
Wright  AA, Zhang  B, Ray  A,  et al.  Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment.   JAMA. 2008;300(14):1665-1673. doi:10.1001/jama.300.14.1665 PubMedGoogle ScholarCrossref
6.
Tucker-Seeley  RD, Abel  GA, Uno  H, Prigerson  H.  Financial hardship and the intensity of medical care received near death.   Psychooncology. 2015;24(5):572-578. doi:10.1002/pon.3624 PubMedGoogle ScholarCrossref
7.
Wright  AA, Keating  NL, Ayanian  JZ,  et al.  Family perspectives on aggressive cancer care near the end of life.   JAMA. 2016;315(3):284-292. doi:10.1001/jama.2015.18604 PubMedGoogle ScholarCrossref
8.
Wunsch  H, Wagner  J, Herlim  M, Chong  DH, Kramer  AA, Halpern  SD.  ICU occupancy and mechanical ventilator use in the United States.   Crit Care Med. 2013;41(12):2712-2719. doi:10.1097/CCM.0b013e318298a139 PubMedGoogle ScholarCrossref
9.
Wunsch  H, Linde-Zwirble  WT, Angus  DC, Hartman  ME, Milbrandt  EB, Kahn  JM.  The epidemiology of mechanical ventilation use in the United States.   Crit Care Med. 2010;38(10):1947-1953. doi:10.1097/CCM.0b013e3181ef4460 PubMedGoogle ScholarCrossref
10.
Rothberg  MB, Herzig  SJ, Pekow  PS, Avrunin  J, Lagu  T, Lindenauer  PK.  Association between sedating medications and delirium in older inpatients.   J Am Geriatr Soc. 2013;61(6):923-930. doi:10.1111/jgs.12253 PubMedGoogle ScholarCrossref
11.
Bo  M, Porrino  P, Di Santo  SG,  et al; Italian Study Group on Delirium (ISGoD).  The association of indwelling urinary catheter with delirium in hospitalized patients and nursing home residents: an explorative analysis from the “Delirium Day 2015.”   Aging Clin Exp Res. 2019;31(3):411-420. doi:10.1007/s40520-018-0974-1 PubMedGoogle ScholarCrossref
12.
Gacouin  A, Jouneau  S, Letheulle  J,  et al.  Trends in prevalence and prognosis in subjects with acute chronic respiratory failure treated with noninvasive and/or invasive ventilation.   Respir Care. 2015;60(2):210-218. doi:10.4187/respcare.03467 PubMedGoogle ScholarCrossref
13.
Stefan  MS, Shieh  MS, Pekow  PS, Hill  N, Rothberg  MB, Lindenauer  PK.  Trends in mechanical ventilation among patients hospitalized with acute exacerbations of COPD in the United States, 2001 to 2011.   Chest. 2015;147(4):959-968. doi:10.1378/chest.14-1216 PubMedGoogle ScholarCrossref
14.
Lindenauer  PK, Stefan  MS, Shieh  MS, Pekow  PS, Rothberg  MB, Hill  NS.  Outcomes associated with invasive and noninvasive ventilation among patients hospitalized with exacerbations of chronic obstructive pulmonary disease.   JAMA Intern Med. 2014;174(12):1982-1993. doi:10.1001/jamainternmed.2014.5430 PubMedGoogle ScholarCrossref
15.
Nava  S, Grassi  M, Fanfulla  F,  et al.  Non-invasive ventilation in elderly patients with acute hypercapnic respiratory failure: a randomised controlled trial.   Age Ageing. 2011;40(4):444-450. doi:10.1093/ageing/afr003 PubMedGoogle ScholarCrossref
16.
Tallman  TA, Peacock  WF, Emerman  CL,  et al; ADHERE Registry.  Noninvasive ventilation outcomes in 2,430 acute decompensated heart failure patients: an ADHERE Registry Analysis.   Acad Emerg Med. 2008;15(4):355-362. doi:10.1111/j.1553-2712.2008.00059.x PubMedGoogle ScholarCrossref
17.
Mehta  AB, Syeda  SN, Wiener  RS, Walkey  AJ.  Epidemiological trends in invasive mechanical ventilation in the United States: a population-based study.   J Crit Care. 2015;30(6):1217-1221. doi:10.1016/j.jcrc.2015.07.007 PubMedGoogle ScholarCrossref
18.
Azoulay  E, Demoule  A, Jaber  S,  et al.  Palliative noninvasive ventilation in patients with acute respiratory failure.   Intensive Care Med. 2011;37(8):1250-1257. doi:10.1007/s00134-011-2263-8 PubMedGoogle ScholarCrossref
19.
Curtis  JR, Cook  DJ, Sinuff  T,  et al; Society of Critical Care Medicine Palliative Noninvasive Positive Ventilation Task Force.  Noninvasive positive pressure ventilation in critical and palliative care settings: understanding the goals of therapy.   Crit Care Med. 2007;35(3):932-939. doi:10.1097/01.CCM.0000256725.73993.74 PubMedGoogle ScholarCrossref
20.
Nava  S, Ferrer  M, Esquinas  A,  et al.  Palliative use of non-invasive ventilation in end-of-life patients with solid tumours: a randomised feasibility trial.   Lancet Oncol. 2013;14(3):219-227. doi:10.1016/S1470-2045(13)70009-3 PubMedGoogle ScholarCrossref
21.
Hui  D, Morgado  M, Chisholm  G,  et al.  High-flow oxygen and bilevel positive airway pressure for persistent dyspnea in patients with advanced cancer: a phase II randomized trial.   J Pain Symptom Manage. 2013;46(4):463-473. doi:10.1016/j.jpainsymman.2012.10.284 PubMedGoogle ScholarCrossref
22.
Azoulay  E, Kouatchet  A, Jaber  S,  et al.  Noninvasive mechanical ventilation in patients having declined tracheal intubation.   Intensive Care Med. 2013;39(2):292-301. doi:10.1007/s00134-012-2746-2 PubMedGoogle ScholarCrossref
23.
Rochwerg  B, Brochard  L, Elliott  MW,  et al.  Official ERS/ATS clinical practice guidelines: noninvasive ventilation for acute respiratory failure.   Eur Respir J. 2017;50(2):1602426. doi:10.1183/13993003.02426-2016 PubMedGoogle Scholar
24.
Sukul  D, Hoffman  GJ, Nuliyalu  U,  et al.  Association between Medicare policy reforms and changes in hospitalized Medicare beneficiaries’ severity of illness.   JAMA Netw Open. 2019;2(5):e193290. doi:10.1001/jamanetworkopen.2019.3290 PubMedGoogle Scholar
25.
Tsugawa  Y, Figueroa  JF, Papanicolas  I, Orav  EJ, Jha  AK.  Assessment of strategies for managing expansion of diagnosis coding using risk-adjustment methods for Medicare data.   JAMA Intern Med. 2019;179(9):1287-1290. doi:10.1001/jamainternmed.2019.1005 PubMedGoogle Scholar
26.
Department of Health & Human Services; Centers for Medicare and Medicaid Services. 5010 Implementation–processing additional International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis and procedure codes in Pricer, Grouper, and the Medicare Code Editor (MCE)–JA7004. Published August 13, 2010. Accessed February 1, 2020. https://www.cms.gov/Medicare/Medicare-Contracting/ContractorLearningResources/Downloads/JA7004.pdf
27.
de Miguel-Díez  J, Jiménez-García  R, Hernández-Barrera  V,  et al.  National trends in hospital admissions for asthma exacerbations among pediatric and young adult population in Spain (2002-2010).   Respir Med. 2014;108(7):983-991. doi:10.1016/j.rmed.2014.04.008 PubMedGoogle ScholarCrossref
28.
Stefan  MS, Priya  A, Pekow  PS,  et al.  The comparative effectiveness of noninvasive and invasive ventilation in patients with pneumonia.   J Crit Care. 2018;43:190-196. doi:10.1016/j.jcrc.2017.05.023 PubMedGoogle ScholarCrossref
29.
Valley  TS, Walkey  AJ, Lindenauer  PK, Wiener  RS, Cooke  CR.  Association between noninvasive ventilation and mortality among older patients with pneumonia.   Crit Care Med. 2017;45(3):e246-e254. doi:10.1097/CCM.0000000000002076 PubMedGoogle ScholarCrossref
30.
Wunsch  H, Kramer  A, Gershengorn  HB.  Validation of intensive care and mechanical ventilation codes in Medicare data.   Crit Care Med. 2017;45(7):e711-e714. doi:10.1097/CCM.0000000000002316 PubMedGoogle ScholarCrossref
31.
Charlson  ME, Pompei  P, Ales  KL, MacKenzie  CR.  A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.   J Chronic Dis. 1987;40(5):373-383. doi:10.1016/0021-9681(87)90171-8 PubMedGoogle ScholarCrossref
32.
Shrank  WH, Rogstad  TL, Parekh  N.  Waste in the US health care system: estimated costs and potential for savings.   JAMA. 2019;322(15):1501-1509. doi:10.1001/jama.2019.13978 PubMedGoogle ScholarCrossref
33.
Centers for Medicare and Medicaid Services. Hospice quality reporting. Published 2018. Accessed December 1, 2019. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Hospice-Quality-Reporting
34.
Halpern  NA, Pastores  SM.  Critical care medicine in the United States 2000-2005: an analysis of bed numbers, occupancy rates, payer mix, and costs.   Crit Care Med. 2010;38(1):65-71. doi:10.1097/CCM.0b013e3181b090d0 PubMedGoogle ScholarCrossref
35.
Dasta  JF, McLaughlin  TP, Mody  SH, Piech  CT.  Daily cost of an intensive care unit day: the contribution of mechanical ventilation.   Crit Care Med. 2005;33(6):1266-1271. doi:10.1097/01.CCM.0000164543.14619.00 PubMedGoogle ScholarCrossref
36.
Burns  HP, Dayal  VS, Scott  A, van Nostrand  AW, Bryce  DP.  Laryngotracheal trauma: observations on its pathogenesis and its prevention following prolonged orotracheal intubation in the adult.   Laryngoscope. 1979;89(8):1316-1325. doi:10.1002/lary.1979.89.8.1316 PubMedGoogle ScholarCrossref
37.
Mutlu  GM, Factor  P.  Complications of mechanical ventilation.   Respir Care Clin N Am. 2000;6(2):213-252, v. doi:10.1016/S1078-5337(05)70069-1 PubMedGoogle ScholarCrossref
38.
Sinclair  SE, Altemeier  WA, Matute-Bello  G, Chi  EY.  Augmented lung injury due to interaction between hyperoxia and mechanical ventilation.   Crit Care Med. 2004;32(12):2496-2501. doi:10.1097/01.CCM.0000148231.04642.8D PubMedGoogle ScholarCrossref
39.
Ram  FS, Picot  J, Lightowler  J, Wedzicha  JA.  Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease.   Cochrane Database Syst Rev. 2004;(3):CD004104. doi:10.1002/14651858.CD004104.pub3 PubMedGoogle Scholar
40.
Weng  CL, Zhao  YT, Liu  QH,  et al.  Meta-analysis: noninvasive ventilation in acute cardiogenic pulmonary edema.   Ann Intern Med. 2010;152(9):590-600. doi:10.7326/0003-4819-152-9-201005040-00009 PubMedGoogle ScholarCrossref
41.
Keenan  SP, Sinuff  T, Cook  DJ, Hill  NS.  Does noninvasive positive pressure ventilation improve outcome in acute hypoxemic respiratory failure? a systematic review.   Crit Care Med. 2004;32(12):2516-2523. doi:10.1097/01.CCM.0000148011.51681.E2 PubMedGoogle ScholarCrossref
42.
Ferreyro  BL, Angriman  F, Munshi  L,  et al.  Association of noninvasive oxygenation strategies with all-cause mortality in adults with acute hypoxemic respiratory failure: a systematic review and meta-analysis.   JAMA. 2020;324(1):1-12. doi:10.1001/jama.2020.9524 PubMedGoogle Scholar
43.
Hill  NS, Brennan  J, Garpestad  E, Nava  S.  Noninvasive ventilation in acute respiratory failure.   Crit Care Med. 2007;35(10):2402-2407. doi:10.1097/01.CCM.0000284587.36541.7FPubMedGoogle ScholarCrossref
44.
Esteban  A, Anzueto  A, Frutos  F,  et al; Mechanical Ventilation International Study Group.  Characteristics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study.   JAMA. 2002;287(3):345-355. doi:10.1001/jama.287.3.345 PubMedGoogle ScholarCrossref
45.
Walkey  AJ, Wiener  RS.  Use of noninvasive ventilation in patients with acute respiratory failure, 2000-2009: a population-based study.   Ann Am Thorac Soc. 2013;10(1):10-17. doi:10.1513/AnnalsATS.201206-034OC PubMedGoogle ScholarCrossref
46.
Teno  JM, Gozalo  P, Khandelwal  N,  et al.  Association of increasing use of mechanical ventilation among nursing home residents with advanced dementia and intensive care unit beds.   JAMA Intern Med. 2016;176(12):1809-1816. doi:10.1001/jamainternmed.2016.5964 PubMedGoogle ScholarCrossref
47.
Kiehl  MG, Beutel  G, Böll  B,  et al; Consensus of the German Society of Hematology and Medical Oncology (DGHO), Austrian Society of Hematology and Oncology (OeGHO), German Society for Medical Intensive Care Medicine and Emergency Medicine (DGIIN), and Austrian Society of Medical and General Intensive Care and Emergency Medicine (ÖGIAIN).  Consensus statement for cancer patients requiring intensive care support.   Ann Hematol. 2018;97(7):1271-1282. doi:10.1007/s00277-018-3312-y PubMedGoogle ScholarCrossref
48.
Huaringa  AJ, Francis  WH.  Outcome of invasive mechanical ventilation in cancer patients: intubate or not to intubate a patient with cancer.   J Crit Care. 2019;50:87-91. doi:10.1016/j.jcrc.2018.11.014 PubMedGoogle ScholarCrossref
49.
Schettino  G, Altobelli  N, Kacmarek  RM.  Noninvasive positive pressure ventilation reverses acute respiratory failure in select “do-not-intubate” patients.   Crit Care Med. 2005;33(9):1976-1982. doi:10.1097/01.CCM.0000178176.51024.82 PubMedGoogle ScholarCrossref
50.
Azevedo  LCP, Caruso  P, Silva  UVA,  et al; Brazilian Research in Intensive Care Network (BRICNet).  Outcomes for patients with cancer admitted to the ICU requiring ventilatory support: results from a prospective multicenter study.   Chest. 2014;146(2):257-266. doi:10.1378/chest.13-1870 PubMedGoogle ScholarCrossref
51.
Azoulay  E, Mokart  D, Pène  F,  et al.  Outcomes of critically ill patients with hematologic malignancies: prospective multicenter data from France and Belgium–a groupe de recherche respiratoire en réanimation onco-hématologique study.   J Clin Oncol. 2013;31(22):2810-2818. doi:10.1200/JCO.2012.47.2365 PubMedGoogle ScholarCrossref
52.
Lagu  T, Zilberberg  MD, Tjia  J, Pekow  PS, Lindenauer  PK.  Use of mechanical ventilation by patients with and without dementia, 2001 through 2011.   JAMA Intern Med. 2014;174(6):999-1001. doi:10.1001/jamainternmed.2014.1179 PubMedGoogle ScholarCrossref
53.
Borjaille  CZ, Hill  AD, Pinto  R, Fowler  RA, Scales  DC, Wunsch  H.  Rates of mechanical ventilation for patients with dementia in Ontario: a population-based cohort study.   Anesth Analg. 2019;129(4):e122-e125. doi:10.1213/ANE.0000000000004003 PubMedGoogle ScholarCrossref
54.
Bouza  C, Martínez-Alés  G, López-Cuadrado  T.  Effect of dementia on the incidence, short-term outcomes, and resource utilization of invasive mechanical ventilation in the elderly: a nationwide population-based study.   Crit Care. 2019;23(1):291. doi:10.1186/s13054-019-2580-9 PubMedGoogle ScholarCrossref
55.
Delclaux  C, L’Her  E, Alberti  C,  et al.  Treatment of acute hypoxemic nonhypercapnic respiratory insufficiency with continuous positive airway pressure delivered by a face mask: a randomized controlled trial.   JAMA. 2000;284(18):2352-2360. doi:10.1001/jama.284.18.2352 PubMedGoogle ScholarCrossref
56.
Mitchell  SL.  Clinical practice: advanced dementia.   N Engl J Med. 2015;372(26):2533-2540. doi:10.1056/NEJMcp1412652 PubMedGoogle ScholarCrossref
57.
Carlucci  A, Richard  JC, Wysocki  M, Lepage  E, Brochard  L; SRLF Collaborative Group on Mechanical Ventilation.  Noninvasive versus conventional mechanical ventilation: an epidemiologic survey.   Am J Respir Crit Care Med. 2001;163(4):874-880. doi:10.1164/ajrccm.163.4.2006027 PubMedGoogle ScholarCrossref
58.
Di Marco  F, Centanni  S, Bellone  A,  et al.  Optimization of ventilator setting by flow and pressure waveforms analysis during noninvasive ventilation for acute exacerbations of COPD: a multicentric randomized controlled trial.   Crit Care. 2011;15(6):R283. doi:10.1186/cc10567 PubMedGoogle ScholarCrossref
59.
Mitchell  SL, Teno  JM, Kiely  DK,  et al.  The clinical course of advanced dementia.   N Engl J Med. 2009;361(16):1529-1538. doi:10.1056/NEJMoa0902234 PubMedGoogle ScholarCrossref
60.
St Germaine-Smith  C, Metcalfe  A, Pringsheim  T,  et al.  Recommendations for optimal ICD codes to study neurologic conditions: a systematic review.   Neurology. 2012;79(10):1049-1055. doi:10.1212/WNL.0b013e3182684707 PubMedGoogle ScholarCrossref
61.
Frat  JP, Thille  AW, Mercat  A,  et al; FLORALI Study Group; REVA Network.  High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.   N Engl J Med. 2015;372(23):2185-2196. doi:10.1056/NEJMoa1503326 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    October 19, 2020

    Trends in Noninvasive and Invasive Mechanical Ventilation Among Medicare Beneficiaries at the End of Life

    Author Affiliations
    • 1Division of Pulmonary and Critical Care Medicine, Department of Medicine, Oregon Health and Science University, Portland
    • 2Center to Improve Veteran Involvement in Care, Veterans Affairs Portland Healthcare System, Portland, Oregon
    • 3Division of General Internal Medicine and Geriatrics, School of Medicine, Oregon Health and Science University, Portland
    • 4Department of Health Services, Policy and Practice, Brown University School of Public Health, Providence, Rhode Island
    JAMA Intern Med. 2021;181(1):93-102. doi:10.1001/jamainternmed.2020.5640
    Key Points

    Question  What are the trends in the use of noninvasive ventilation at the end of life?

    Findings  In this cohort study of 2 470 435 Medicare beneficiaries who were hospitalized in the last 30 days of life, the use of noninvasive ventilation rapidly expanded from 2000 to 2017, with a slight increase in the use of invasive mechanical ventilation at the end of life. Use of noninvasive ventilation increased among patients with cancer and dementia, with concomitant increases in invasive mechanical ventilation use.

    Meaning  The findings from this study suggest that further research is needed to examine the outcomes of noninvasive ventilation at the end of life, especially for patients with cancer and dementia.

    Abstract

    Importance  End-of-life care is costly, and decedents often experience overtreatment or low-quality care. Noninvasive ventilation (NIV) may be a palliative approach to avoid invasive mechanical ventilation (IMV) among select patients who are hospitalized at the end of life.

    Objective  To examine the trends in NIV and IMV use among decedents with a hospitalization in the last 30 days of life.

    Design, Setting, and Participants  This population-based cohort study used a 20% random sample of Medicare fee-for-service beneficiaries who had an acute care hospitalization in the last 30 days of life and died between January 1, 2000, and December 31, 2017. Sociodemographic, diagnosis, and comorbidity data were obtained from Medicare claims data. Data analysis was performed from September 2019 to July 2020.

    Exposures  Use of NIV or IMV.

    Main Outcomes and Measures  Validated International Classification of Diseases, Ninth Revision, Clinical Modification or International Statistical Classification of Diseases, Tenth Revision, Clinical Modification procedure codes were reviewed to identify use of NIV, IMV, both NIV and IMV, or none. Four subcohorts of Medicare beneficiaries were identified using primary admitting diagnosis codes (chronic obstructive pulmonary disease [COPD], congested heart failure [CHF], cancer, and dementia). Measures of end-of-life care included in-hospital death (acute care setting), hospice enrollment at death, and hospice enrollment in the last 3 days of life. Random-effects logistic regression examined NIV and IMV use adjusted for sociodemographic characteristics, admitting diagnosis, and comorbidities.

    Results  A total of 2 470 435 Medicare beneficiaries (1 353 798 women [54.8%]; mean [SD] age, 82.2 [8.2] years) were hospitalized within 30 days of death. Compared with 2000, the adjusted odds ratio (AOR) for the increase in NIV use was 2.63 (95% CI, 2.46-2.82; % receipt: 0.8% vs 2.0%) for 2005 and 11.84 (95% CI, 11.11-12.61; % receipt: 0.8% vs 7.1%) for 2017. Compared with 2000, the AOR for the increase in IMV use was 1.04 (95% CI, 1.02-1.06; % receipt: 15.0% vs 15.2%) for 2005 and 1.63 (95% CI, 1.59-1.66; % receipt: 15.0% vs 18.2%) for 2017. In subanalyses comparing 2017 with 2000, similar trends found increased NIV among patients with CHF (% receipt: 1.4% vs 14.2%; AOR, 14.14 [95% CI, 11.77-16.98]) and COPD (% receipt: 2.7% vs 14.5%; AOR, 8.22 [95% CI, 6.42-10.52]), with reciprocal stabilization in IMV use among patients with CHF (% receipt: 11.1% vs 7.8%; AOR, 1.07 [95% CI, 0.95-1.19]) and COPD (% receipt: 17.4% vs 13.2%; AOR, 1.03 [95% CI, 0.88-1.21]). The AOR for increased NIV use was 10.82 (95% CI, 8.16-14.34; % receipt: 0.4% vs 3.5%) among decedents with cancer and 9.62 (95% CI, 7.61-12.15; % receipt: 0.6% vs 5.2%) among decedents with dementia. The AOR for increased IMV use was 1.40 (95% CI, 1.26-1.55; % receipt: 6.2% vs 7.6%) among decedents with cancer and 1.28 (95% CI, 1.17-1.41; % receipt: 5.7% vs 6.2%) among decedents with dementia. Among decedents with NIV vs IMV use, lower rates of in-hospital death (50.3% [95% CI, 49.3%-51.3%] vs 76.7% [95% CI, 75.9%-77.5%]) and hospice enrollment in the last 3 days of life (57.7% [95% CI, 56.2%-59.3%] vs 63.0% [95% CI, 60.9%-65.1%]) were observed along with higher rates of hospice enrollment (41.3% [95% CI, 40.4%-42.3%] vs 20.0% [95% CI, 19.2%-20.7%]).

    Conclusions and Relevance  This study found that the use of NIV rapidly increased from 2000 through 2017 among Medicare beneficiaries at the end of life, especially among persons with cancer and dementia. The findings suggest that trials to evaluate the outcomes of NIV are warranted to inform discussions about the goals of this therapy between clinicians and patients and their health care proxies.

    Introduction

    The quality of medical care for patients who are seriously ill at the end of life is concerning because they often receive potentially burdensome care, including excessive health care transitions and intensive care unit (ICU) admissions.1 This situation persists, even though most Medicare beneficiaries prefer treatment that is focused on palliation rather than life extension at the end of life.2 High-intensity care is not associated with better outcomes given that geographical regions with greater overall spending on end-of-life care do not have improved survival, care quality, patient satisfaction, or perceptions of quality among bereaved families.3,4 High-intensity care at the end of life may also be associated with adverse consequences, including reduced quality of life (QOL), increased financial hardship, and discordant care that is not aligned with the wishes of patients or their families.2,5-7

    A common diagnosis associated with ICU admission is acute respiratory failure, and many patients with this condition require ventilatory support. At any time, approximately 40% of patients admitted to an ICU receive invasive mechanical ventilation (IMV),8 which is costly, labor intensive, and accounts for a disproportionate amount of health care resources.9 In addition, IMV often necessitates sedating medications (and urinary catheters) that are associated with delirium, particularly among older patients.10,11 Use of noninvasive ventilation (NIV), such as continuous or bilevel positive airway pressure, has increased in select populations of patients with respiratory failure (eg, those with acute exacerbations of chronic obstructive pulmonary disease [COPD] or congestive heart failure [CHF])12,13 because of improved outcomes (ie, increased survival, shorter length of stay, and lower costs) compared with IMV.14-16 As a result of the expanded use of NIV, use of IMV among these populations has substantially decreased over time.13,17

    Use of NIV to improve survival has been established; however, its use has also been suggested to achieve palliation in persons with terminal illness.18,19 Although evidence is limited, under palliative circumstances, NIV may be introduced on a trial basis to reduce dyspnea and respiratory distress while allowing patients and families more time to address goals and finalize affairs. Compared with high-flow or supplemental oxygen therapy, NIV has been associated with reduced dyspnea and morphine needs among patients with cancer.20,21 In a mixed group of patients in the ICU with acute respiratory failure and a do-not-intubate order who received NIV, no statistically significant difference in QOL was observed between 90 days and baseline measurements among those who survived.22 As a result, the joint European Respiratory Society and American Thoracic Society guidelines on NIV for acute respiratory failure suggest offering NIV for palliation in the setting of terminal conditions.23 Therapy is considered successful if it improves breathlessness and respiratory distress without introducing adverse consequences, such as mask discomfort or prolonged agitation.

    This population-based cohort study aimed to address several important questions regarding ventilatory support for patients at the end of life using Medicare beneficiary data from 2000 to 2017. These questions were as follows: (1) what were the trends in NIV and IMV use among patients hospitalized at the end of life? (2) were there differences in NIV and IMV use among patients with select admitting diagnoses? and (3) were there differences in in-hospital death, hospice enrollment, and late hospice enrollment among these select patient populations?

    Methods

    This cohort study was approved by the institutional review boards of Oregon Health and Science University and Brown University. These institutional review boards waived patient consent requirements as the study involved deceased individuals.

    Data and Study Population

    We selected a 20% random sample of Medicare fee-for-service beneficiaries aged 66 years or older who had a hospital admission in the last 30 days of life and died between January 1, 2000, and December 31, 2017. In addition, within this cohort, we identified 4 subcohorts of Medicare beneficiaries. A diagnosis of dementia was ascertained using the primary or the first 9 secondary admission diagnosis codes for hospitalization from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) or International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM). The other subcohort diagnoses (CHF, COPD, and cancer) were identified using the primary admitting diagnosis codes from ICD-9-CM or ICD-10-CM (eMethods and eTable 1 in the Supplement). We truncated the number of secondary diagnosis coding slots after 2011 to reduce bias24,25 attributed to the Centers for Medicare & Medicaid Services expansion of the number of secondary diagnoses, which occurred in January 2011.26 Patients with a primary admitting diagnosis of cardiac arrest or with preexisting tracheostomy were excluded because of ventilatory requirements. The cohort included decedents in hospitals that provided 5 or more beneficiary hospitalizations per year in the last 30 days of life and 1 hospitalization per unique beneficiary. Because of the transition from ICD-9-CM to ICD-10-CM codes, deaths from October 1, 2015, to December 31, 2015, were excluded.

    Measures

    We reviewed validated procedure codes to identify the use of NIV (ICD-9-CM or ICD-10-CM codes 93.90, 93.91, 93.99/5A09357, 5A09457, 5A09557, 5A09358, 5A09458, 5A09558, 5A0935Z, 5A0945Z, and 5A0955Z), IMV (ICD-9-CM or ICD-10-CM codes 96.7x/5A1935Z, 5A1945Z, and 5A1955Z), both NIV and IMV, or none.27-30 The sociodemographic characteristics of the Medicare beneficiaries were based on the information contained in the Master Beneficiary Summary File, including age, sex, and race/ethnicity. Race/ethnicity data were based on information collected by the Social Security Administration. Comorbidities were based on ICD-9-CM or ICD-10-CM codes that were submitted as part of the Medicare claim for that hospitalization within 30 days of death. Measures of end-of-life care by ventilatory support and admitting diagnosis group included in-hospital death (acute care setting), hospice enrollment at death, and hospice enrollment in the last 3 days of life.

    Statistical Analysis

    Descriptive statistics were used to characterize patient variables by year of hospitalization and ventilatory support, and 95% CIs were calculated with the Clopper-Pearson method. Rates of NIV and IMV use were tabulated by year. We examined descriptions to ascertain whether the IMV trend for representative years varied by admission type (medical or surgical). Multivariable random-effects logistic regressions were performed to examine NIV and IMV use among decedents clustered by hospital, adjusting for age, sex, race/ethnicity, admitting diagnosis, and Charlson Comorbidity Index score (range, 0-40, with the highest score indicating the highest predicted risk of death).31 Dichotomous variables for the years were included to capture time trends, with the year 2000 used as a reference. For the NIV model, NIV and no ventilatory support groups were included; for the IMV model, IMV and no ventilatory support groups were included. Subgroup analyses (CHF, COPD, cancer, and dementia diagnoses) were conducted to investigate the secular trends among diagnostic groups. End-of-life care measures were tabulated by diagnostic group for descriptive comparisons using the most recent quartile (2013-2017), excluding Medicaid-eligible beneficiaries.

    Data analysis was performed with Stata, version 15.0 (StataCorp LLC) from September 2019 to July 2020. Comparisons between years were interpreted as different if their 95% CIs did not overlap.

    Results

    From 2000 to 2017, a total of 2 470 435 Medicare fee-for-service beneficiaries (1 353 798 women [54.8%] and 1 116 637 men [45.2%] with a mean [SD] age of 82.2 [8.2] years) were hospitalized within the last 30 days of life. Among these decedents, the top 3 admitting diagnoses were pneumonia or sepsis (n = 525 523 [21.3%]), cancer (n = 237 335 [9.6%]), and CHF (n = 155 273 [6.3%]). The sociodemographic characteristics for the cohort are shown in Table 1. Overall, IMV was used in 401 419 of 2 470 435 decedents (16.3%), NIV in 90 700 decedents (3.7%), and both IMV and NIV in 25 689 decedents (1.0%) (Table 2). Among decedents who received NIV, 51 038 (56.3%) had an ICU stay.

    From 2000 to 2017, an almost 9-fold absolute increase in NIV use from 0.8% to 7.1% occurred, whereas IMV use increased slightly from 15.0% to 18.5% and was twice as common as NIV use (Figure 1). Compared with 2000 (the reference group), the adjusted odds ratio (AOR) for the increase in NIV use was 2.63 (95% CI, 2.46-2.82; % receipt: 0.8% vs 2.0%) for 2005, 7.87 (95% CI, 7.38-8.39; % receipt: 0.8% vs 5.2%) for 2011, and 11.84 (95% CI, 11.11-12.61; % receipt: 0.8% vs 7.1%) for 2017 (eTable 2 in the Supplement). Compared with 2000, the AOR for the increase in IMV use was 1.04 (95% CI, 1.02-1.06; % receipt: 15.0% vs 15.2%) for 2005, 1.39 (95% CI, 1.36-1.42; % receipt: 15.0% vs 17.1%) for 2011, and 1.63 (95% CI, 1.59-1.66; % receipt: 15.0% vs 18.2%) for 2017. The IMV trend stratified by admission type was similar.

    NIV and IMV Use Among Select Populations

    All comparisons used 2000 as a reference group. In 2017, the AOR for the increase in NIV use was 14.14 (95% CI, 11.77-16.98; % receipt: 1.4% vs 14.2%) among decedents with CHF and 8.22 (95% CI, 6.42-10.52; % receipt: 2.7% vs 14.5%) among decedents with COPD (eTable 2 in the Supplement). The reciprocal AOR stabilization in IMV use for 2017 was 1.07 (95% CI, 0.95-1.19; % receipt: 11.1% vs 7.8%) among decedents with CHF and 1.03 (95% CI, 0.88-1.21; % receipt: 17.4% vs 13.2%) among decedents with COPD. In 2017, among decedents with cancer, the AOR for the increases was 10.82 (95% CI, 8.16-14.34; % receipt: 0.4% vs 3.5%) for NIV use and 1.40 (95% CI, 1.26-1.55; % receipt: 6.2% vs 7.6%) for IMV use. Among decedents with dementia, the AOR for the increases was 9.62 (95% CI, 7.61-12.15; % receipt: 0.6% vs 5.2%) for NIV use and 1.28 (95% CI, 1.17-1.41; % receipt: 5.7% vs 6.2%) for IMV use.

    Ventilatory support trends from 2000 to 2017 were plotted for select populations at the end of life (Figure 2). Among decedents with CHF and COPD, increases in NIV use in 2017 (AOR: 14.14 vs 8.22) were balanced by reciprocal decreases in 2010 (AOR: 0.72 vs 0.84) and then stabilization in 2017 (AOR: 1.07 vs 1.03) in IMV use. Among decedents with cancer and dementia, increases in NIV use in 2017 (AOR: 10.82 vs 9.62) were observed with concomitant increases in IMV use in 2017 (AOR: 1.40 vs 1.28).

    End-of-Life Care From 2013 to 2017

    In all subgroups of CHF, COPD, cancer, and dementia diagnoses, decedents who received NIV vs IMV had lower rates (expressed as % of hospitalizations or hospice enrollees) of in-hospital death (50.3% [95% CI, 49.3%-51.3%] vs 76.7% [95% CI, 75.9%-77.5%]) and hospice enrollment in the last 3 days of life (57.7% [95% CI, 56.2%-59.3%] vs 63.0% [95% CI, 60.9%-65.1%]) as well as a higher rate of hospice enrollment (41.3% [95% CI, 40.4%-42.3%] vs 20.0% [95% CI, 19.2%-20.7%]) (Figure 3). Decedents from CHF and COPD subgroups with NIV use had similar end-of-life care (in-hospital death, hospice enrollment in the last 3 days of life, and hospice enrollment) (CHF vs COPD, in-hospital death: 49.0% [95% CI, 47.3%-50.6%] vs 43.6% [95% CI, 40.7%-46.5%]; hospice enrollment in the last 3 days: 40.2% [95% CI, 38.6%-41.8%] vs 42.5% [95% CI, 39.6%-45.4%]; and hospice enrollment: 24.2% [95% CI, 22.9%-25.6%] vs 23.4% [95% CI, 21.0%-26.0%]). Meanwhile, decedents from cancer and dementia subgroups with NIV use had the following end-of-life care (in-hospital death: 60.8% [95% CI, 58.0%-63.5%] vs 50.1% [95% CI, 48.5%-51.8%]; hospice enrollment in the last 3 days: 38.9% [95% CI, 36.2%-41.7%] vs 43.0% [95% CI, 41.4%-44.6%]; and hospice enrollment: 24.3% [95% CI, 21.9%-26.8%] vs 24.0% [95% CI, 22.6%-25.4%]) (Figure 3).

    Discussion

    Critical decisions are made regarding the use of life-sustaining treatments associated with QOL, especially among those hospitalized within 30 days of death. For patients with CHF and COPD, use of NIV may improve outcomes and avoid the use of IMV. In this study, the finding of the rapid growth of NIV use among persons with cancer and dementia without reciprocal decreases in IMV use raises more questions than answers. Use of NIV may be associated with agitation and distress for patients with cancer and dementia at the end of life, prolonging their death. Given the rapid growth in NIV use, further research is needed to examine the goals of this therapy and whether it achieves those goals.

    Use of IMV among Medicare beneficiaries at the end of life was relatively stable from 2008 to 2015, a finding that was consistent with the reported stabilization of other measures of aggressive health care at the end of life (eg, ICU stay in the last 30 days of life).1 End-of-life care has been identified as an area for quality improvement32,33; however, it is difficult to attribute changes in this care to any single guideline or policy. We observed a slight increase in the use of IMV after 2015, which was a meaningful finding given that critical care accounted for a substantial proportion of US health care costs34 and that IMV was associated with considerably higher daily ICU costs per patient.35 In addition, IMV was associated with high rates of mortality and morbidity, which were partially attributable to the complications (eg, pneumonia, ventilator-associated lung injury) and adverse effects (eg, prolonged sedation, tracheal injury) of treatment.36,37 The introduction and expanded use of NIV, especially among patients with evidence-based indications, such as CHF or COPD, can decrease IMV use and its complications, shorten ICU length of stay, and improve survival; however, marked heterogeneity in outcomes has suggested that the result of this therapy varies among select populations.38

    The strongest evidence of the benefit of NIV exists among patients with acute exacerbations of CHF or COPD.39,40 The potential for improved outcomes in these patients likely explains the increases in NIV use and the reciprocal decreases in IMV use that we observed. However, outside of these indications (CHF and COPD), favorable outcome of NIV use in patients with hypoxemia was not established in randomized clinical trials41 unless trials of patients with CHF or COPD were included42; therefore, the rapid growth in NIV use that we observed among persons with cancer and dementia is concerning. In addition, NIV necessitates a higher level of care than provided in a hospital ward because of the closer monitoring required; therefore, the recommendation is to implement NIV in an ICU or a high-dependency unit (eg, ICU step-down unit) in hospitals that have these units.43 Other obstacles for NIV include the contraindications in patients who are agitated, are uncooperative, have substantial airway secretions, or are unable to protect their airway. Approximately 22% of the present cohort may have had NIV failure, a rate that was comparable to the one-quarter to one-third of patients who reportedly had NIV failure and required endotracheal intubation and IMV,13,44 which has been associated with increased risk of death.12,44

    The rapid increase in NIV use that was observed in this study was consistent with findings in previous studies of hospitalized patients,45 with an important distinction that the present population was analyzed at the end of life. Among patients with cancer and dementia, the expansion in NIV use was lower than among patients with CHF but higher than among patients with COPD. The most troubling aspect of this finding was that this rapid expansion occurred without a reciprocal reduction in IMV use. Instead, increases in IMV use were seen among patients with cancer and dementia at the end of life. This situation occurred despite evidence that IMV use was a factor in mortality in patients with cancer and was not associated with substantial improvements in survival in patients with dementia.46-49 However, improvements in outcomes among patients with cancer who received IMV may partly explain the expanded IMV use in this group.50,51 The trend that we observed in increased IMV use among patients with dementia was consistent with trends in similar populations (any dementia diagnosis) as reported in previous studies from the US,52 Canada,53 and parts of Europe.54 The rapid growth in NIV use among patients with cancer and dementia may represent another example of overtreatment or low-quality care (ie, care in which the risks outweigh the benefits) at the end of life.32 Given the potential for substantial patient and family burden, costs, and health care resources associated with NIV without demonstrable benefit, and even the potential to introduce harm,55 NIV use among patients with cancer and dementia at the end of life warrants a thorough discussion about the goals of therapy between clinicians and patients and their health care proxies.

    As an alternative explanation to overtreatment, NIV use at the end of life has been suggested based on limited evidence for palliation on a trial basis to help alleviate respiratory distress and provide patients and families time to address goals.19 Small studies among patients with advanced cancer have suggested some symptom advantages20,21 without reductions in 90-day QOL22; however, findings from randomized clinical trials are nonexistent. In this study, evidence that supported the palliative intent of NIV, based on measures of end-of-life care, was lacking given that in-hospital death, hospice enrollment, and late hospice enrollment were similar for patients with cancer and dementia who received NIV compared with patients with CHF or COPD who received NIV. Furthermore, we observed worse rates of hospice enrollment among patients with cancer and dementia who received NIV than were previously reported among Medicare fee-for-service beneficiaries at the end of life,1 which was not consistent with palliative intent. Although palliative use of NIV at the end of life may offer some value to patients with advanced cancer, based on limited evidence, inherent differences among patients with dementia deserve consideration.56-59

    The finding that NIV use rapidly increased among patients with dementia in the last 30 days of life was striking because advanced dementia is characterized by profound memory deficits, reduced verbal abilities, and severely diminished cognition.56 Agitation and lack of cooperation are relative contraindications to NIV initiation, and up to 22% of all patients started on NIV were expected to discontinue use for these reasons.57 Patient-machine asynchrony or NIV asynchrony, which is characterized by air leaks, play a substantial role in NIV failure,58 which may be exacerbated by the absence of psychomotor skills and reflexes, and the generalized rigidity that can be seen in patients with dementia. Pneumonia is common in dementia; therefore, the use of NIV can present difficulties with clearing secretions. Agitation and copious secretions are both contraindications to NIV initiation and are expected to be associated with increased NIV failure and its associated mortality. Treatment decisions for patients with advanced dementia are often dependent on their health care proxies, who overwhelmingly prefer comfort-based care59; therefore, patient assent and participation with therapy could be another barrier to success. Use of NIV among patients with dementia at the end of life should be considered carefully given the potential harm and low likelihood of favorable outcome associated with this therapy.

    Limitations

    This study has limitations. We used Medicare claims data and did not have access to clinical data, such as disease severity or patient preferences for end-of-life care; IMV or NIV may represent goal-concordant care for some decedents. Medicare claims files were only included for fee-for service beneficiaries; thus, these results may not be generalizable to other populations or patients with Medicare managed plans (ie, Medicare Advantage beneficiaries) who are less likely to be hospitalized at the end of life.1 We relied on ICD-9-CM or ICD-10-CM diagnosis codes in Medicare claims to identify decedents with a diagnosis of CHF, COPD, cancer, or dementia; these methods are known to underdiagnose dementia in decedents.60 We applied methods to reduce bias25,27 from the Centers for Medicare & Medicaid Services expansion of secondary diagnoses in January 2011,24,26 but residual confounding may exist. Among nonsurgical patients, codes for mechanical ventilation using Medicare data have high specificity and positive and negative predictive value, but they may have low sensitivity.30 An increasing component of clinical practice since 2015, use of high-flow nasal cannula was unavailable; however, use of high-flow nasal cannula did not reduce intubation rates compared with NIV among older adults.61

    Conclusions

    From 2000 to 2017, the use of NIV in older adults who were hospitalized in the last 30 days of life rapidly increased. Although expanded use of NIV for CHF and COPD can be justified based on favorable outcomes, the increased use of NIV among patients with cancer and dementia, without a concomitant decrease in IMV use, suggests possible overtreatment or palliation. Given the rapid growth, the potential for patient harm and distress, and the substantial health care resources associated with NIV use, further research is warranted to evaluate its outcomes and to inform discussions about the goals of this therapy between clinicians and patients and their health care proxies so that they can make patient-centered choices.

    Back to top
    Article Information

    Accepted for Publication: July 29, 2020.

    Corresponding Author: Donald R. Sullivan, MD, MA, MCR, Division of Pulmonary and Critical Care Medicine, Department of Medicine, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd, Portland, OR 97239-3098 (sullivad@ohsu.edu).

    Published Online: October 19, 2020. doi:10.1001/jamainternmed.2020.5640

    Author Contributions: Drs Sullivan and Kim had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Sullivan, Gozalo, Bunker, Teno.

    Acquisition, analysis, or interpretation of data: Sullivan, Kim, Gozalo, Teno.

    Drafting of the manuscript: Sullivan, Teno.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Kim, Gozalo, Teno.

    Obtained funding: Sullivan, Teno.

    Administrative, technical, or material support: Sullivan, Bunker.

    Supervision: Sullivan, Teno.

    Conflict of Interest Disclosures: Dr Sullivan reported receiving grants from the National Cancer Institute and the National Institute on Aging (NIA) of the National Institutes of Health (NIH) during the conduct of the study. Drs Kim, Gozalo, and Teno and Ms Bunker reported receiving grants from the NIA of the NIH during the conduct of the study. No other disclosures were reported.

    Funding/Support: This study was funded by awards R56AG063748 and 2P01AG027296-11 from the NIA of the NIH.

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The content and views herein are those of the authors and do not reflect the official policy or position of the NIH. The Department of Veterans Affairs did not have a role in the conduct of the study; in the collection, management, analysis, and interpretation of data; or in the preparation of the manuscript. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States Government.

    Additional Information: The code-sharing document used in this study is part of the Brown University Digital Repository (https://doi.org/10.26300/pxh2-3790).

    References
    1.
    Teno  JM, Gozalo  P, Trivedi  AN,  et al.  Site of death, place of care, and health care transitions among US Medicare beneficiaries, 2000-2015.   JAMA. 2018;320(3):264-271. doi:10.1001/jama.2018.8981 PubMedGoogle ScholarCrossref
    2.
    Barnato  AE, Herndon  MB, Anthony  DL,  et al.  Are regional variations in end-of-life care intensity explained by patient preferences?: a study of the US Medicare population.   Med Care. 2007;45(5):386-393. doi:10.1097/01.mlr.0000255248.79308.41 PubMedGoogle ScholarCrossref
    3.
    Fisher  ES, Wennberg  DE, Stukel  TA, Gottlieb  DJ, Lucas  FL, Pinder  EL.  The implications of regional variations in Medicare spending. Part 2: health outcomes and satisfaction with care.   Ann Intern Med. 2003;138(4):288-298. doi:10.7326/0003-4819-138-4-200302180-00007 PubMedGoogle ScholarCrossref
    4.
    Teno  JM, Mor  V, Ward  N,  et al.  Bereaved family member perceptions of quality of end-of-life care in U.S. regions with high and low usage of intensive care unit care.   J Am Geriatr Soc. 2005;53(11):1905-1911. doi:10.1111/j.1532-5415.2005.53563.x PubMedGoogle ScholarCrossref
    5.
    Wright  AA, Zhang  B, Ray  A,  et al.  Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment.   JAMA. 2008;300(14):1665-1673. doi:10.1001/jama.300.14.1665 PubMedGoogle ScholarCrossref
    6.
    Tucker-Seeley  RD, Abel  GA, Uno  H, Prigerson  H.  Financial hardship and the intensity of medical care received near death.   Psychooncology. 2015;24(5):572-578. doi:10.1002/pon.3624 PubMedGoogle ScholarCrossref
    7.
    Wright  AA, Keating  NL, Ayanian  JZ,  et al.  Family perspectives on aggressive cancer care near the end of life.   JAMA. 2016;315(3):284-292. doi:10.1001/jama.2015.18604 PubMedGoogle ScholarCrossref
    8.
    Wunsch  H, Wagner  J, Herlim  M, Chong  DH, Kramer  AA, Halpern  SD.  ICU occupancy and mechanical ventilator use in the United States.   Crit Care Med. 2013;41(12):2712-2719. doi:10.1097/CCM.0b013e318298a139 PubMedGoogle ScholarCrossref
    9.
    Wunsch  H, Linde-Zwirble  WT, Angus  DC, Hartman  ME, Milbrandt  EB, Kahn  JM.  The epidemiology of mechanical ventilation use in the United States.   Crit Care Med. 2010;38(10):1947-1953. doi:10.1097/CCM.0b013e3181ef4460 PubMedGoogle ScholarCrossref
    10.
    Rothberg  MB, Herzig  SJ, Pekow  PS, Avrunin  J, Lagu  T, Lindenauer  PK.  Association between sedating medications and delirium in older inpatients.   J Am Geriatr Soc. 2013;61(6):923-930. doi:10.1111/jgs.12253 PubMedGoogle ScholarCrossref
    11.
    Bo  M, Porrino  P, Di Santo  SG,  et al; Italian Study Group on Delirium (ISGoD).  The association of indwelling urinary catheter with delirium in hospitalized patients and nursing home residents: an explorative analysis from the “Delirium Day 2015.”   Aging Clin Exp Res. 2019;31(3):411-420. doi:10.1007/s40520-018-0974-1 PubMedGoogle ScholarCrossref
    12.
    Gacouin  A, Jouneau  S, Letheulle  J,  et al.  Trends in prevalence and prognosis in subjects with acute chronic respiratory failure treated with noninvasive and/or invasive ventilation.   Respir Care. 2015;60(2):210-218. doi:10.4187/respcare.03467 PubMedGoogle ScholarCrossref
    13.
    Stefan  MS, Shieh  MS, Pekow  PS, Hill  N, Rothberg  MB, Lindenauer  PK.  Trends in mechanical ventilation among patients hospitalized with acute exacerbations of COPD in the United States, 2001 to 2011.   Chest. 2015;147(4):959-968. doi:10.1378/chest.14-1216 PubMedGoogle ScholarCrossref
    14.
    Lindenauer  PK, Stefan  MS, Shieh  MS, Pekow  PS, Rothberg  MB, Hill  NS.  Outcomes associated with invasive and noninvasive ventilation among patients hospitalized with exacerbations of chronic obstructive pulmonary disease.   JAMA Intern Med. 2014;174(12):1982-1993. doi:10.1001/jamainternmed.2014.5430 PubMedGoogle ScholarCrossref
    15.
    Nava  S, Grassi  M, Fanfulla  F,  et al.  Non-invasive ventilation in elderly patients with acute hypercapnic respiratory failure: a randomised controlled trial.   Age Ageing. 2011;40(4):444-450. doi:10.1093/ageing/afr003 PubMedGoogle ScholarCrossref
    16.
    Tallman  TA, Peacock  WF, Emerman  CL,  et al; ADHERE Registry.  Noninvasive ventilation outcomes in 2,430 acute decompensated heart failure patients: an ADHERE Registry Analysis.   Acad Emerg Med. 2008;15(4):355-362. doi:10.1111/j.1553-2712.2008.00059.x PubMedGoogle ScholarCrossref
    17.
    Mehta  AB, Syeda  SN, Wiener  RS, Walkey  AJ.  Epidemiological trends in invasive mechanical ventilation in the United States: a population-based study.   J Crit Care. 2015;30(6):1217-1221. doi:10.1016/j.jcrc.2015.07.007 PubMedGoogle ScholarCrossref
    18.
    Azoulay  E, Demoule  A, Jaber  S,  et al.  Palliative noninvasive ventilation in patients with acute respiratory failure.   Intensive Care Med. 2011;37(8):1250-1257. doi:10.1007/s00134-011-2263-8 PubMedGoogle ScholarCrossref
    19.
    Curtis  JR, Cook  DJ, Sinuff  T,  et al; Society of Critical Care Medicine Palliative Noninvasive Positive Ventilation Task Force.  Noninvasive positive pressure ventilation in critical and palliative care settings: understanding the goals of therapy.   Crit Care Med. 2007;35(3):932-939. doi:10.1097/01.CCM.0000256725.73993.74 PubMedGoogle ScholarCrossref
    20.
    Nava  S, Ferrer  M, Esquinas  A,  et al.  Palliative use of non-invasive ventilation in end-of-life patients with solid tumours: a randomised feasibility trial.   Lancet Oncol. 2013;14(3):219-227. doi:10.1016/S1470-2045(13)70009-3 PubMedGoogle ScholarCrossref
    21.
    Hui  D, Morgado  M, Chisholm  G,  et al.  High-flow oxygen and bilevel positive airway pressure for persistent dyspnea in patients with advanced cancer: a phase II randomized trial.   J Pain Symptom Manage. 2013;46(4):463-473. doi:10.1016/j.jpainsymman.2012.10.284 PubMedGoogle ScholarCrossref
    22.
    Azoulay  E, Kouatchet  A, Jaber  S,  et al.  Noninvasive mechanical ventilation in patients having declined tracheal intubation.   Intensive Care Med. 2013;39(2):292-301. doi:10.1007/s00134-012-2746-2 PubMedGoogle ScholarCrossref
    23.
    Rochwerg  B, Brochard  L, Elliott  MW,  et al.  Official ERS/ATS clinical practice guidelines: noninvasive ventilation for acute respiratory failure.   Eur Respir J. 2017;50(2):1602426. doi:10.1183/13993003.02426-2016 PubMedGoogle Scholar
    24.
    Sukul  D, Hoffman  GJ, Nuliyalu  U,  et al.  Association between Medicare policy reforms and changes in hospitalized Medicare beneficiaries’ severity of illness.   JAMA Netw Open. 2019;2(5):e193290. doi:10.1001/jamanetworkopen.2019.3290 PubMedGoogle Scholar
    25.
    Tsugawa  Y, Figueroa  JF, Papanicolas  I, Orav  EJ, Jha  AK.  Assessment of strategies for managing expansion of diagnosis coding using risk-adjustment methods for Medicare data.   JAMA Intern Med. 2019;179(9):1287-1290. doi:10.1001/jamainternmed.2019.1005 PubMedGoogle Scholar
    26.
    Department of Health & Human Services; Centers for Medicare and Medicaid Services. 5010 Implementation–processing additional International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis and procedure codes in Pricer, Grouper, and the Medicare Code Editor (MCE)–JA7004. Published August 13, 2010. Accessed February 1, 2020. https://www.cms.gov/Medicare/Medicare-Contracting/ContractorLearningResources/Downloads/JA7004.pdf
    27.
    de Miguel-Díez  J, Jiménez-García  R, Hernández-Barrera  V,  et al.  National trends in hospital admissions for asthma exacerbations among pediatric and young adult population in Spain (2002-2010).   Respir Med. 2014;108(7):983-991. doi:10.1016/j.rmed.2014.04.008 PubMedGoogle ScholarCrossref
    28.
    Stefan  MS, Priya  A, Pekow  PS,  et al.  The comparative effectiveness of noninvasive and invasive ventilation in patients with pneumonia.   J Crit Care. 2018;43:190-196. doi:10.1016/j.jcrc.2017.05.023 PubMedGoogle ScholarCrossref
    29.
    Valley  TS, Walkey  AJ, Lindenauer  PK, Wiener  RS, Cooke  CR.  Association between noninvasive ventilation and mortality among older patients with pneumonia.   Crit Care Med. 2017;45(3):e246-e254. doi:10.1097/CCM.0000000000002076 PubMedGoogle ScholarCrossref
    30.
    Wunsch  H, Kramer  A, Gershengorn  HB.  Validation of intensive care and mechanical ventilation codes in Medicare data.   Crit Care Med. 2017;45(7):e711-e714. doi:10.1097/CCM.0000000000002316 PubMedGoogle ScholarCrossref
    31.
    Charlson  ME, Pompei  P, Ales  KL, MacKenzie  CR.  A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.   J Chronic Dis. 1987;40(5):373-383. doi:10.1016/0021-9681(87)90171-8 PubMedGoogle ScholarCrossref
    32.
    Shrank  WH, Rogstad  TL, Parekh  N.  Waste in the US health care system: estimated costs and potential for savings.   JAMA. 2019;322(15):1501-1509. doi:10.1001/jama.2019.13978 PubMedGoogle ScholarCrossref
    33.
    Centers for Medicare and Medicaid Services. Hospice quality reporting. Published 2018. Accessed December 1, 2019. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/Hospice-Quality-Reporting
    34.
    Halpern  NA, Pastores  SM.  Critical care medicine in the United States 2000-2005: an analysis of bed numbers, occupancy rates, payer mix, and costs.   Crit Care Med. 2010;38(1):65-71. doi:10.1097/CCM.0b013e3181b090d0 PubMedGoogle ScholarCrossref
    35.
    Dasta  JF, McLaughlin  TP, Mody  SH, Piech  CT.  Daily cost of an intensive care unit day: the contribution of mechanical ventilation.   Crit Care Med. 2005;33(6):1266-1271. doi:10.1097/01.CCM.0000164543.14619.00 PubMedGoogle ScholarCrossref
    36.
    Burns  HP, Dayal  VS, Scott  A, van Nostrand  AW, Bryce  DP.  Laryngotracheal trauma: observations on its pathogenesis and its prevention following prolonged orotracheal intubation in the adult.   Laryngoscope. 1979;89(8):1316-1325. doi:10.1002/lary.1979.89.8.1316 PubMedGoogle ScholarCrossref
    37.
    Mutlu  GM, Factor  P.  Complications of mechanical ventilation.   Respir Care Clin N Am. 2000;6(2):213-252, v. doi:10.1016/S1078-5337(05)70069-1 PubMedGoogle ScholarCrossref
    38.
    Sinclair  SE, Altemeier  WA, Matute-Bello  G, Chi  EY.  Augmented lung injury due to interaction between hyperoxia and mechanical ventilation.   Crit Care Med. 2004;32(12):2496-2501. doi:10.1097/01.CCM.0000148231.04642.8D PubMedGoogle ScholarCrossref
    39.
    Ram  FS, Picot  J, Lightowler  J, Wedzicha  JA.  Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease.   Cochrane Database Syst Rev. 2004;(3):CD004104. doi:10.1002/14651858.CD004104.pub3 PubMedGoogle Scholar
    40.
    Weng  CL, Zhao  YT, Liu  QH,  et al.  Meta-analysis: noninvasive ventilation in acute cardiogenic pulmonary edema.   Ann Intern Med. 2010;152(9):590-600. doi:10.7326/0003-4819-152-9-201005040-00009 PubMedGoogle ScholarCrossref
    41.
    Keenan  SP, Sinuff  T, Cook  DJ, Hill  NS.  Does noninvasive positive pressure ventilation improve outcome in acute hypoxemic respiratory failure? a systematic review.   Crit Care Med. 2004;32(12):2516-2523. doi:10.1097/01.CCM.0000148011.51681.E2 PubMedGoogle ScholarCrossref
    42.
    Ferreyro  BL, Angriman  F, Munshi  L,  et al.  Association of noninvasive oxygenation strategies with all-cause mortality in adults with acute hypoxemic respiratory failure: a systematic review and meta-analysis.   JAMA. 2020;324(1):1-12. doi:10.1001/jama.2020.9524 PubMedGoogle Scholar
    43.
    Hill  NS, Brennan  J, Garpestad  E, Nava  S.  Noninvasive ventilation in acute respiratory failure.   Crit Care Med. 2007;35(10):2402-2407. doi:10.1097/01.CCM.0000284587.36541.7FPubMedGoogle ScholarCrossref
    44.
    Esteban  A, Anzueto  A, Frutos  F,  et al; Mechanical Ventilation International Study Group.  Characteristics and outcomes in adult patients receiving mechanical ventilation: a 28-day international study.   JAMA. 2002;287(3):345-355. doi:10.1001/jama.287.3.345 PubMedGoogle ScholarCrossref
    45.
    Walkey  AJ, Wiener  RS.  Use of noninvasive ventilation in patients with acute respiratory failure, 2000-2009: a population-based study.   Ann Am Thorac Soc. 2013;10(1):10-17. doi:10.1513/AnnalsATS.201206-034OC PubMedGoogle ScholarCrossref
    46.
    Teno  JM, Gozalo  P, Khandelwal  N,  et al.  Association of increasing use of mechanical ventilation among nursing home residents with advanced dementia and intensive care unit beds.   JAMA Intern Med. 2016;176(12):1809-1816. doi:10.1001/jamainternmed.2016.5964 PubMedGoogle ScholarCrossref
    47.
    Kiehl  MG, Beutel  G, Böll  B,  et al; Consensus of the German Society of Hematology and Medical Oncology (DGHO), Austrian Society of Hematology and Oncology (OeGHO), German Society for Medical Intensive Care Medicine and Emergency Medicine (DGIIN), and Austrian Society of Medical and General Intensive Care and Emergency Medicine (ÖGIAIN).  Consensus statement for cancer patients requiring intensive care support.   Ann Hematol. 2018;97(7):1271-1282. doi:10.1007/s00277-018-3312-y PubMedGoogle ScholarCrossref
    48.
    Huaringa  AJ, Francis  WH.  Outcome of invasive mechanical ventilation in cancer patients: intubate or not to intubate a patient with cancer.   J Crit Care. 2019;50:87-91. doi:10.1016/j.jcrc.2018.11.014 PubMedGoogle ScholarCrossref
    49.
    Schettino  G, Altobelli  N, Kacmarek  RM.  Noninvasive positive pressure ventilation reverses acute respiratory failure in select “do-not-intubate” patients.   Crit Care Med. 2005;33(9):1976-1982. doi:10.1097/01.CCM.0000178176.51024.82 PubMedGoogle ScholarCrossref
    50.
    Azevedo  LCP, Caruso  P, Silva  UVA,  et al; Brazilian Research in Intensive Care Network (BRICNet).  Outcomes for patients with cancer admitted to the ICU requiring ventilatory support: results from a prospective multicenter study.   Chest. 2014;146(2):257-266. doi:10.1378/chest.13-1870 PubMedGoogle ScholarCrossref
    51.
    Azoulay  E, Mokart  D, Pène  F,  et al.  Outcomes of critically ill patients with hematologic malignancies: prospective multicenter data from France and Belgium–a groupe de recherche respiratoire en réanimation onco-hématologique study.   J Clin Oncol. 2013;31(22):2810-2818. doi:10.1200/JCO.2012.47.2365 PubMedGoogle ScholarCrossref
    52.
    Lagu  T, Zilberberg  MD, Tjia  J, Pekow  PS, Lindenauer  PK.  Use of mechanical ventilation by patients with and without dementia, 2001 through 2011.   JAMA Intern Med. 2014;174(6):999-1001. doi:10.1001/jamainternmed.2014.1179 PubMedGoogle ScholarCrossref
    53.
    Borjaille  CZ, Hill  AD, Pinto  R, Fowler  RA, Scales  DC, Wunsch  H.  Rates of mechanical ventilation for patients with dementia in Ontario: a population-based cohort study.   Anesth Analg. 2019;129(4):e122-e125. doi:10.1213/ANE.0000000000004003 PubMedGoogle ScholarCrossref
    54.
    Bouza  C, Martínez-Alés  G, López-Cuadrado  T.  Effect of dementia on the incidence, short-term outcomes, and resource utilization of invasive mechanical ventilation in the elderly: a nationwide population-based study.   Crit Care. 2019;23(1):291. doi:10.1186/s13054-019-2580-9 PubMedGoogle ScholarCrossref
    55.
    Delclaux  C, L’Her  E, Alberti  C,  et al.  Treatment of acute hypoxemic nonhypercapnic respiratory insufficiency with continuous positive airway pressure delivered by a face mask: a randomized controlled trial.   JAMA. 2000;284(18):2352-2360. doi:10.1001/jama.284.18.2352 PubMedGoogle ScholarCrossref
    56.
    Mitchell  SL.  Clinical practice: advanced dementia.   N Engl J Med. 2015;372(26):2533-2540. doi:10.1056/NEJMcp1412652 PubMedGoogle ScholarCrossref
    57.
    Carlucci  A, Richard  JC, Wysocki  M, Lepage  E, Brochard  L; SRLF Collaborative Group on Mechanical Ventilation.  Noninvasive versus conventional mechanical ventilation: an epidemiologic survey.   Am J Respir Crit Care Med. 2001;163(4):874-880. doi:10.1164/ajrccm.163.4.2006027 PubMedGoogle ScholarCrossref
    58.
    Di Marco  F, Centanni  S, Bellone  A,  et al.  Optimization of ventilator setting by flow and pressure waveforms analysis during noninvasive ventilation for acute exacerbations of COPD: a multicentric randomized controlled trial.   Crit Care. 2011;15(6):R283. doi:10.1186/cc10567 PubMedGoogle ScholarCrossref
    59.
    Mitchell  SL, Teno  JM, Kiely  DK,  et al.  The clinical course of advanced dementia.   N Engl J Med. 2009;361(16):1529-1538. doi:10.1056/NEJMoa0902234 PubMedGoogle ScholarCrossref
    60.
    St Germaine-Smith  C, Metcalfe  A, Pringsheim  T,  et al.  Recommendations for optimal ICD codes to study neurologic conditions: a systematic review.   Neurology. 2012;79(10):1049-1055. doi:10.1212/WNL.0b013e3182684707 PubMedGoogle ScholarCrossref
    61.
    Frat  JP, Thille  AW, Mercat  A,  et al; FLORALI Study Group; REVA Network.  High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure.   N Engl J Med. 2015;372(23):2185-2196. doi:10.1056/NEJMoa1503326 PubMedGoogle ScholarCrossref
    ×