- Email: [email protected]
Early Dialysis and Adverse Outcomes After Hurricane Sandy
Original Investigation Early Dialysis and Adverse Outcomes After Hurricane Sandy Nicole Lurie, MD, MSPH,1 Kristen Finne, BA,1 Chris Worrall, BS,2 Maria Jauregui, BA,3 Tanayott Thaweethai, BS,3 Gregg Margolis, PhD, NREMT-P,1 and Jeffrey Kelman, MD, MMSc2 Background: Hemodialysis patients have historically experienced diminished access to care and increased adverse outcomes after natural disasters. Although “early dialysis” in advance of a storm is promoted as a best practice, evidence for its effectiveness as a protective measure is lacking. Building on prior work, we examined the relationship between the receipt of dialysis ahead of schedule before the storm (also known as early dialysis) and adverse outcomes of patients with end-stage renal disease in the areas most affected by Hurricane Sandy. Study Design: Retrospective cohort analysis, using claims data from the Centers for Medicare & Medicaid Services Datalink Project. Setting & Participants: Patients receiving long-term hemodialysis in New York City and the state of New Jersey, the areas most affected by Hurricane Sandy. Factor: Receipt of early dialysis compared to their usual treatment pattern in the week prior to the storm. Outcomes: Emergency department (ED) visits, hospitalizations, and 30-day mortality following the storm. Results: Of 13,836 study patients, 8,256 (60%) received early dialysis. In unadjusted logistic regression models, patients who received early dialysis were found to have lower odds of ED visits (OR, 0.75; 95% CI, 0.63-0.89; P 5 0.001) and hospitalizations (OR, 0.77; 95% CI, 0.65-0.92; P 5 0.004) in the week of the storm and similar odds of 30-day mortality (OR, 0.80; 95% CI, 0.58-1.09; P 5 0.2). In adjusted multivariable logistic regression models, receipt of early dialysis was associated with lower odds of ED visits (OR, 0.80; 95% CI, 0.67-0.96; P 5 0.01) and hospitalizations (OR, 0.79; 95% CI, 0.66-0.94; P 5 0.01) in the week of the storm and 30-day mortality (OR, 0.72; 95% CI, 0.52-0.997; P 5 0.048). Limitations: Inability to determine which patients were offered early dialysis and declined and whether important unmeasured patient characteristics are associated with receipt of early dialysis. Conclusions: Patients who received early dialysis had significantly lower odds of having an ED visit and hospitalization in the week of the storm and of dying within 30 days. Am J Kidney Dis. -(-):---. Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. INDEX WORDS: Disaster preparedness; emergency response; natural disaster; vulnerable population; early dialysis; hemodialysis; end-stage renal disease (ESRD); emergency department (ED) visit; hospitalization; mortality; missed dialysis session; adverse outcome; Hurricane Sandy.
D
evastating natural disasters disproportionately affect vulnerable populations, including children, those who are poor, and those who are medically fragile.1-4 Patients with end-stage renal disease (ESRD) are particularly at risk during disasters.1-6 Studies following Hurricane Katrina identified a number of adverse outcomes for dialysis patients, including increased hospitalizations.7,8 Since that time, considerable progress has been made in emergency preparedness for dialysisdependent patients. Many states require dialysis facilities to have back-up power plans and/or generators so they can remain open in the face of power outages.9 Dialysis facilities commonly have practices in place to facilitate access to urgent dialysis care in times of crisis, including directing patients to facilities that remain open, use of shorter dialysis treatments to accommodate more patients, remain open for a third shift, and in some cases will operate 24 hours a day using staff from other facilities to ensure that additional patients affected by the emergency can receive timely dialysis.9-11 The Kidney Community Emergency Response (KCER) Coalition, formed in response to Am J Kidney Dis. 2015;-(-):---
Hurricane Katrina, routinely activates during potential disasters to assist ESRD Networks in identifying and connecting patients to facilities at which they can receive dialysis if their usual source of dialysis is not available.12,13 Additionally, some dialysis providers seek to provide dialysis ahead of a scheduled visit, also known as “early dialysis,” so patients will not have to
From the 1Office of the Assistant Secretary for Preparedness and Response and 2Centers for Medicare & Medicaid Services, US Department of Health and Human Services, Washington, DC; and 3Acumen LLC, Burlingame, CA. Received December 3, 2014. Accepted in revised form April 29, 2015. Address correspondence to Nicole Lurie, MD, MSPH, Assistant Secretary for Preparedness and Response, US Department of Health and Human Services, Office of the Secretary, 200 Independence Ave SW, Rm 638G, Washington, DC 20201. E-mail: [email protected] Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. 0272-6386 http://dx.doi.org/10.1053/j.ajkd.2015.04.050 1
Lurie et al
miss a dialysis session if a major storm or hurricane disrupts electricity or transportation in a given area.14 Because surge dialysis care in advance of a storm is not precluded by federal law or regulation, state health officials commonly encourage dialysis facilities to implement this protective practice in advance of an anticipated emergency or disaster. For example, in advance of Hurricane Sandy, nearly 70% of ESRD facilities in the affected areas provided Sunday dialysis for patients who typically received dialysis on a Monday-Wednesday-Friday (MWF) schedule.15 In advance of a rare snowstorm in New Orleans in 2013, nearly all dialysis patients were contacted and advised to receive early dialysis. Fortunately, ,5% of patients with ESRD had an emergency department (ED) visit or were hospitalized following Hurricane Sandy, but it is not clear whether those rates would have been higher without the widespread receipt of early dialysis.15 If early dialysis is to be promoted as a best practice and protective measure in renal care, evidence for its effectiveness in avoiding subsequent adverse outcomes, such as ED visits, hospitalizations, and 30-day mortality, is needed. Building on prior work, we used claims data from the Centers for Medicare & Medicaid Services (CMS) Datalink Project to examine the relationship between the receipt of early dialysis and adverse outcomes of patients with ESRD who received hemodialysis in New York City and the state of New Jersey, the areas most affected by Hurricane Sandy.
METHODS Data Source and Study Population We used Parts A and B Medicare fee-for-service claims, processed as of October 10, 2014, and covering the period from September 29, 2012, to November 30, 2012, to identify patients with ESRD who received facility-based hemodialysis (dialysis) in New York City and the state of New Jersey, the areas most affected by Hurricane Sandy. We included patients in our study if they were ESRD Medicare beneficiaries alive as of October 28 and enrolled in Medicare Parts A and B in September, October, and November 2012, as determined by the Enrollment Database, and had a claim for at least one maintenance dialysis treatment for October 1 to October 28, as identified through Medicare Part B Outpatient fee-for-service claims with type of bill 72x. We excluded patients receiving athome hemodialysis or peritoneal dialysis. We also excluded patients who spent the entire week of the storm in the hospital, defined as being admitted on or prior to October 28 and not discharged until after November 3 (Fig S1, available as online supplementary material). We classified patients according to whether they received early dialysis. To determine whether patients received early dialysis, we compared study patient treatment patterns from the week prior to the storm (October 21, 2012, to October 27, 2012) to the week of the storm (October 28, 2012, to November 3, 2012). Patients were categorized as receiving dialysis on a MWF or Tuesday-ThursdaySaturday (TThS) schedule based on their prior week’s treatment pattern. We excluded patients who did not have a detectable MWF 2
or TThS dialysis schedule the week prior to the storm (n 5 1,663/ 15,499), resulting in a final sample size of 13,836 patients with a MWF or TThS schedule (Fig S1). We considered patients to have received early dialysis if they were in the MWF group and received dialysis on Saturday, October 27, or Sunday, October 28, 2012, or were in the TThS group and received dialysis on Sunday, October 28, or Monday, October 29, 2012. We used Medicare’s Dialysis Facility Compare (DFC) and Certification and Survey Provider Enhanced Reporting (CASPER) data sets to identify ESRD treatment facilities in the study areas. We considered facilities open on a specific date if they submitted at least one claim with a treatment date of service for that date.
Patient Characteristics and Adverse-Outcome Measures Demographic and health status characteristics of patients who did and did not receive early dialysis included age (,65, 65-84, or $85 years), sex, race (white or nonwhite), dual eligibility status (ie, eligible for Medicare and Medicaid; yes/no), and onset of dialysis care (dialysis vintage #1 or .1 year). To identify potential differences in health status, we also identified ED visits and hospitalizations in the 30 days before the storm, as well as evidence of underlying cardiovascular comorbid conditions because another study found elevated rates in New Jersey following Hurricane Sandy.16 Cardiovascular comorbid conditions were defined as a having an International Classification of Diseases, Ninth Revision diagnosis code of myocardial infarction (410), congestive heart failure (428), stroke (430-434), and arrhythmia (426-427) on an inpatient or outpatient Medicare fee-for-service claim 30 days (September 29, 2012, to October 28, 2012) prior to the storm. Outcome measures included ED visits and hospitalizations during the week of the storm (October 29, 2012, to November 3, 2012) and mortality 30 days after the storm (October 29, 2012, to November 27, 2012) for patients who did and did not receive early dialysis. Hospitalizations were determined by admission dates during the week of the storm and were counted once regardless of length of stay. ED visits were determined by service dates during the week of the storm, and a patient was only counted once regardless of frequency of visits. We identified patients who died within 30 days of the storm.
Analysis We compared patient-level variables of patients who did and did not receive early dialysis to identify potential demographic and health status differences between the 2 groups using c2 test. We then used an unadjusted logistic regression model to evaluate whether early dialysis was associated with ED visits or hospitalizations in the week of the storm and in 30-day mortality. Finally, we used multivariable-adjusted logistic regression, including age, sex, race, dual-eligibility status, onset of dialysis care, and health status characteristics (ED visits, hospitalizations, and evidence of underlying cardiovascular comorbid conditions) in the 30 days before the storm to calculate the adjusted odds of adverse outcomes and identify patient characteristics that were independently associated with the outcomes of interest. We considered P , 0.05 to be statistically significant. Analyses were conducted using SAS, version 9.2 (SAS Institute Inc), and STATA, version 12 (StataCorp LP).
Participant Protections The analysis was exempt from institutional review board review because it was performed as a public health quality review and deidentification methods were implemented in accordance with CMS policy and Health Insurance Portability and Accountability Act (HIPAA) requirements. Am J Kidney Dis. 2015;-(-):---
Early Dialysis and Hurricane Sandy
RESULTS Patient Characteristics Table 1 compares demographic and health status characteristics of patients who did and did not receive early dialysis. Of the total 13,836 study patients, 8,256 (60%) received early dialysis on Saturday (October 27), Sunday (October 28), or Monday (October 29) in 209 of 229 (92%) dialysis facilities in the study area. Patients who received early dialysis were more likely to be older (P , 0.001) and white (P , 0.001) and less likely to be dual eligible for Medicare and Medicaid (P , 0.001). ED Visits, Hospitalizations, and Mortality Unadjusted analyses show that patients with early dialysis had lower odds of ED visits (odds ratio [OR], 0.75; 95% confidence interval [CI], 0.63-0.89; P 5 0.001) and hospitalization (OR, 0.77; 95% CI, 0.65-0.92; P 5 0.004) than patients without early dialysis (Table 2). Patients with early dialysis had similar odds of 30-day mortality compared with patients without early dialysis (OR, 0.80; 95% CI, 0.581.09; P 5 0.2; Table 2). In adjusted analyses controlling for patient characteristics, early dialysis continued to be associated with
Table 1. Patient Characteristics Early Dialysis (n 5 8,256)
Age category ,65 y 65-84 y $85 y Male sex Nonwhite race Dual-eligibility statusa Incident dialysis onset ,1y Incident CV comorbidity 30 d before stormb Incident ED visit or hospitalization 30 d before stormc
No Early Dialysis (n 5 5,580)
c2 Test (P)
3,644 (44.1%) 2,685 (48.1%) ,0.001 3,894 (47.2%) 2,501 (44.8%) 718 (8.7%) 394 (7.1%) 4,658 4,946 3,922 1,257
(56.4%) 3,156 (56.6%) 0.9 (59.9%) 3,833 (68.7%) ,0.001 (47.5%) 3,608 (64.7%) ,0.001 (15.2%) 819 (14.7%) 0.4
938 (11.4%)
627 (11.2%)
0.8
1,342 (16.3%)
907 (16.3%)
0.9
Note: Of 229 facilities, 209 (92%) provided early dialysis to at least one patient. Abbreviations: CV, cardiovascular; ED, emergency department. a For both Medicare and Medicaid. b CV comorbidity was defined as having an International Classification of Diseases, Ninth Revision diagnosis code of myocardial infarction (410), congestive heart failure (428), stroke (430-434), or arrhythmia (426-427) on an inpatient or outpatient Medicare fee-for-service claim. c The 30-day period before the storm is September 29, 2012, through October 28, 2012. Am J Kidney Dis. 2015;-(-):---
decreased odds of ED visits (OR, 0.80; 95% CI, 0.670.96; P 5 0.01) and hospitalizations (OR, 0.79; 95% CI, 0.66-0.94; P 5 0.01) in the week of the storm (Table 2). Early dialysis was also found to be significantly associated with reduced odds of 30-day mortality (OR, 0.72; 95% CI, 0.52-0.997; P 5 0.048; Table 2). Patient Characteristics Associated With Adverse Outcomes As shown in Table 2, dual-eligible patients had increased odds for an ED visit and hospitalization, while being nonwhite was associated with decreased odds of both adverse outcomes. Cardiovascular comorbid conditions in the month prior to the storm were associated with increased odds of hospitalization and 30-day mortality. Hospitalization or ED visit in the month prior to the storm was associated with increased odds of hospitalization, ED visits, and 30-day mortality. In addition, being nonwhite and younger was associated with decreased odds of 30day mortality.
DISCUSSION This study confirms that early dialysis ahead of Hurricane Sandy’s landfall decreased the likelihood of ED visits, hospitalizations, and 30-day mortality for dialysis patients in the areas most affected. Such evidence affirms the importance of preparedness practices on the part of dialysis facilities to provide early dialysis, as well as the need for dialysis patients to receive early dialysis when access to routine dialysis may be threatened. Our prior study noted that while an impressive number of patients with ESRD in New York City and New Jersey received early dialysis before Hurricane Sandy made landfall, nearly 40% did not.15 Furthermore, that study found significant variation in the practice of early dialysis across the 5 New York City boroughs.15 Taken together, these findings suggest that routine availability of early dialysis in advance of a storm is generally achievable and should become standard practice. While further work is needed to better understand factors associated with the lack of receipt of early dialysis, it is likely that both patient and facility factors play a role. Patient factors likely include both measureable factors such as income and unmeasured variables such as social support, transportation access, and the capacity to plan well enough ahead to make arrangements for dialysis. We also were not able to determine which patients were offered early dialysis and declined. The KCER Coalition has had an important role in connecting patients with dialysis facilities, including arranging transportation when necessary, and has become an important national asset 3
Lurie et al Table 2. Unadjusted and Adjusted Odds of Adverse Outcomes in Hemodialysis Patients After Hurricane Sandy Unadjusted Regression Factors Associated With Outcome
Early dialysis Nonwhite race Male sex Age category ,65 y 65-84 y $85 y
OR (95% CI)
ED Visit 0.75 (0.63-0.89)
Dual eligiblea Dialysis vintage # 1 y CV comorbid condition diagnosis code in 30 d before storm Hospitalization or ED visit during 30 d before storm Hospitalization Early dialysis 0.77 (0.65-0.92) Nonwhite race Male sex Age category ,65 y 65-84 y $85 y Dual eligiblea Dialysis vintage # 1 y CV comorbid condition diagnosis code in 30 d before storm Hospitalization or ED visit 30 d before storm 30-Day Mortality Early dialysis 0.80 (0.58-1.09) Nonwhite race Male sex Age category ,65 y 65-84 y $85 y Dual eligiblea Dialysis vintage # 1 y CV comorbid condition diagnosis code in 30 d before storm Hospitalization or ED visit 30 d before storm
P
0.001
Adjusted Regression OR (95% CI)
0.80 (0.67-0.96) 0.73 (0.61-0.89) 1.02 (0.85-1.22)
0.01 0.001 0.8
1.01 (0.73-1.41) 0.87 (0.63-1.21) 1.00 (reference)
0.9 0.4
1.68 0.89 0.99 1.79 0.004
(1.38-2.03) (0.70-1.15) (0.75-1.32) (1.42-2.25)
,0.001 0.4 0.9 ,0.001
0.79 (0.66-0.94) 0.79 (0.65-0.96) 0.90 (0.75-1.07)
0.01 0.02 0.2
0.80 (0.58-1.11) 0.95 (0.69-1.31) 1.00 (reference)
0.2 0.8
1.30 0.90 1.39 2.30 0.2
P
(1.07-1.57) (0.70-1.16) (1.07-1.8) (1.84-2.88)
0.01 0.4 0.01 ,0.001
0.72 (0.52-0.997) 0.51 (0.36-0.73) 0.94 (0.68-1.3)
0.048 ,0.001 0.7
0.29 (0.18-0.46) 0.44 (0.29-0.66) 1.00 (reference)
,0.001 ,0.001
1.06 1.00 1.59 3.33
0.8 0.9 0.03 ,0.001
(0.75-1.50) (0.66-1.50) (1.05-2.42) (2.25-4.92)
Note: In the unadjusted logistic regression model, only early dialysis was included as a categorical variable, with the reference group not receiving early dialysis. Abbreviations: CI, confidence interval; CV, cardiovascular; ED, emergency department; OR, odds ratio. a For both Medicare and Medicaid.
in emergency planning and response, particularly for the dialysis community.13,17 Facility preparedness is also critical.10,18 Many of the dialysis facilities in the Sandy-affected region had conducted facility-specific risk assessments and had established emergency plans and alternative energy sources (eg, back-up generators).11,14 All these activities better enabled them to rapidly activate and operate their emergency plans, prioritize and contact patients, and coordinate the surge in early dialysis prior to Hurricane Sandy’s landfall. Both patients and facilities can take additional steps to become prepared. For example, dialysis facilities might consider using a holiday as a proxy for a notice event and test 4
that portion of their emergency and communication plans. In doing so, they can better assess their staff’s ability to coordinate early dialysis in advance of an emergency and assess their patients’ ability to understand, implement, and arrive for their treatment and identify potential gaps in facility and patient personal preparedness. Patients should know how to request early dialysis and become better informed about emergency renal diet options that can assist them in tolerating a treatment delay.17,19,20 For example, during the November 2014 snowstorm in Buffalo, NY, patients were instructed to begin an emergency renal diet. They also need to understand how to seek care in an emergency, Am J Kidney Dis. 2015;-(-):---
Early Dialysis and Hurricane Sandy
preferably at a preidentified alternate dialysis facility, and that they should bring a copy of their dialysis treatment plan and list of their current medications with them.19 Finally, interoperable electronic health records would also help facilitate access to essential dialysis care information. Following Hurricane Sandy, a study conducted by Lin et al21 found that 12 of the 13 hospitals surveyed reported that the majority of the 347 dialysis patients seeking urgent dialysis treatment did not have documentation of their treatment plan or hepatitis status and overwhelmed EDs during the disaster. Access to medical records was identified as a problem in other post–Hurricane Sandy studies as well, further highlighting the importance of interoperable information and the need for facilities to provide their patients with updated treatment plans on a regular basis and in anticipation of potential disruptions.22,23 From a broader policy perspective, advances in data and geospatial mapping have begun to assist state and local health officials, KCER, ESRD Networks, dialysis facilities, and the local health care systems to better anticipate potential dialysis patient and facility needs and hospital surge patterns during emergencies.15,24,25 The US Department of Health and Human Services’ Office of the Assistant Secretary for Preparedness and Response and CMS have demonstrated the utility of mapping deidentified ESRD patient- and facility-level data at the local level to inform and support emergency planning and preparedness activities prior to an emergency.15 Furthermore, during an emergency, the US Department of Health and Human Services now has the capability of securely sharing privacy-protected information with local health officials to assist with direct outreach to dialysis patients.24 Ensuring that hemodialysis patients can maintain access to appropriate renal care through the duration of an emergency is important for both individual and community resilience. This study provides additional evidence that the routine provision of early dialysis in advance of major storms or other emergencies is likely to prevent bad outcomes and save lives.
ACKNOWLEDGEMENTS The views expressed are solely those of the authors and do not necessarily represent those of the US Department of Health and Human Services. We thank Thomas E. MaCurdy, PhD, and Alina Bogdanov, MA, for manuscript preparation assistance and Alicia Livinski, MPH, MA, for literature search assistance. Support: This study was supported through the CMS DataLink contract with Acumen LLC. The funders of this study had a role in the study design, interpretation of the data, writing the report, and the decision to submit the report for publication. Financial Disclosure: The authors declare that they have no other relevant financial interests.
Am J Kidney Dis. 2015;-(-):---
Contributions: Research idea and study design: NL, JK, KF, MJ, TT; data acquisition: JK, CW, MJ, TT; data analysis/interpretation: NL, JK, GM, KF, CW, MJ, TT; statistical analysis: MJ, TT; supervision or mentorship: NL, JK. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. NL and JK assert that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
SUPPLEMENTARY MATERIAL Figure S1: Study participant tree. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2015.04.050) is available at www.ajkd.org
REFERENCES 1. Miller A, Arquilla B. Chronic diseases and natural hazards: impact of disasters on diabetic, renal and cardiac patients. Prehosp Disaster Med. 2008;23(2):185-194. 2. Abdel-Kader K, Unrah ML. Disaster and end-stage renal disease: targeting vulnerable patients for improved outcomes. Kidney Int. 2009;75(11):1131-1133. 3. Zoraster R, Vanholder R, Sever MS. Disaster management of chronic dialysis patients. Am J Disaster Med. 2007;2(2):96-106. 4. Kopp JB, Ball LK, Cohen A, et al. Kidney patient care in disasters: lessons from the hurricanes and earthquakes of 2005. Clin J Am Soc Nephrol. 2007;2:814-824. 5. Vanholder RC, Van Biesen WA, Sever MS. Hurricane Katrina and chronic dialysis patients: better tidings than originally feared? Kidney Int. 2009;76(7):687-689. 6. Kleinpeter M. Disaster preparedness for dialysis patients. Clin J Am Soc Nephrol. 2011;6:2337-2339. 7. Howard D, Zhang R, Huang Y, Kutner N. Hospitalization rates among dialysis patients during Hurricane Katrina. Prehosp Disaster Med. 2012;27(4):1-5. 8. Anderson A, Cohen A, Kutner N, et al. Missed dialysis sessions and hospitalizations in hemodialysis patients after Hurricane Katrina. Kidney Int. 2009;75(11):1202-1208. 9. Abir M, Jan S, Jubelt L, et al. The impact of a large-scale power outage on hemodialysis center operations. Prehosp Disaster Med. 2013;28(6):543-546. 10. Kleinpeter MA. Disaster preparedness of dialysis patients for Hurricanes Gustav and Ike 2008. Adv Perit Dial. 2009;25: 62-67. 11. Lindsey H. Hurricane Sandy: dialysis community manages damage, outages, displacement. Nephrol Times. 2012;5(12):8-11. 12. Cary S, Schroeder K. Caring for patients on kidney dialysis in a disaster. Am J Nurs. 2008;108(1):26-32. 13. Kidney Community Emergency Response Coalition (KCER). About KCER. http://kcercoalition.com/kcer-home/aboutkcer/. Accessed December 16, 2014. 14. Lempert K, Kopp J. Hurricane Sandy as a kidney failure disaster. Am J Kidney Dis. 2013;61(6):865-868. 15. Kelman J, Finne K, Bogdanov A, et al. Dialysis care and death following Hurricane Sandy. Am J Kidney Dis. 2015;65(1): 109-115. 16. Swerdel JN, Janevic TM, Cosgrove NM, Kostis JB; Myocardial Infarction Data Acquisition System (MIDAS 24) Study Group. The effect of Hurricane Sandy on cardiovascular events in New Jersey. J Am Heart Assoc. 2014;3(6):e001354.
5
Lurie et al 17. Kidney Community Emergency Response Coalition (KCER). Hurricane Sandy after action report. December 2012. http://kcercoalition.com/wp-content/uploads/2015/06/KCER_ Hurricane_Sandy_After_Action_Report_12-28-12_508.pdf. Accessed May 31, 2013. 18. Kenney RJ. Emergency preparedness concepts for dialysis facilities: reawakened after Hurricane Katrina. Clin J Am Soc Nephrol. 2007;2(4):809-813. 19. Centers for Medicare & Medicaid Services. Preparing for emergencies: a guide for people on dialysis. CMS Pub No. 10150. 2002. http://kcercoalition.com. Accessed December 15, 2014. 20. Kleinpeter MA, Normal LD, Krane NK. Dialysis services in the hurricane-affected areas in 2005: lessons learned. Am J Med Sci. 2006;332(5):259-263. 21. Lin C-J, Pierce LC, Roblin PM, et al. Impact of Hurricane Sandy on hospital emergency and dialysis services: a retrospective study. Prehosp Disaster Med. 2014;29(4):374-379.
6
22. Adalja A, Watson M, Bouri N, et al. Absorbing citywide patient surge during Hurricane Sandy: a case study in accommodating multiple hospital evacuations. Ann Emerg Med. 2014;64(1): 66-73. 23. Teperman S. Hurricane Sandy and the greater New York health care system. J Trauma Acute Care Surg. 2013;74(6):14011410. 24. DeSalvo K, Lurie N, Finne K, et al. Using Medicare data to identify individuals who are electricity-dependent to improve disaster preparedness and response. Am J Public Health. 2014;104(7): 1160-1164. 25. Lee DC, Smith SW, Carr BG, Goldfrank LR, Polsky D. Redistribution of emergency department patients after disasterrelated closures of a public versus private hospital in New York City [published online first March 17, 2015]. Disaster Med Public Health Prep. http://dx.doi.org/10.1017/dmp. 2015.11.
Am J Kidney Dis. 2015;-(-):---
D
evastating natural disasters disproportionately affect vulnerable populations, including children, those who are poor, and those who are medically fragile.1-4 Patients with end-stage renal disease (ESRD) are particularly at risk during disasters.1-6 Studies following Hurricane Katrina identified a number of adverse outcomes for dialysis patients, including increased hospitalizations.7,8 Since that time, considerable progress has been made in emergency preparedness for dialysisdependent patients. Many states require dialysis facilities to have back-up power plans and/or generators so they can remain open in the face of power outages.9 Dialysis facilities commonly have practices in place to facilitate access to urgent dialysis care in times of crisis, including directing patients to facilities that remain open, use of shorter dialysis treatments to accommodate more patients, remain open for a third shift, and in some cases will operate 24 hours a day using staff from other facilities to ensure that additional patients affected by the emergency can receive timely dialysis.9-11 The Kidney Community Emergency Response (KCER) Coalition, formed in response to Am J Kidney Dis. 2015;-(-):---
Hurricane Katrina, routinely activates during potential disasters to assist ESRD Networks in identifying and connecting patients to facilities at which they can receive dialysis if their usual source of dialysis is not available.12,13 Additionally, some dialysis providers seek to provide dialysis ahead of a scheduled visit, also known as “early dialysis,” so patients will not have to
From the 1Office of the Assistant Secretary for Preparedness and Response and 2Centers for Medicare & Medicaid Services, US Department of Health and Human Services, Washington, DC; and 3Acumen LLC, Burlingame, CA. Received December 3, 2014. Accepted in revised form April 29, 2015. Address correspondence to Nicole Lurie, MD, MSPH, Assistant Secretary for Preparedness and Response, US Department of Health and Human Services, Office of the Secretary, 200 Independence Ave SW, Rm 638G, Washington, DC 20201. E-mail: [email protected] Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. 0272-6386 http://dx.doi.org/10.1053/j.ajkd.2015.04.050 1
Lurie et al
miss a dialysis session if a major storm or hurricane disrupts electricity or transportation in a given area.14 Because surge dialysis care in advance of a storm is not precluded by federal law or regulation, state health officials commonly encourage dialysis facilities to implement this protective practice in advance of an anticipated emergency or disaster. For example, in advance of Hurricane Sandy, nearly 70% of ESRD facilities in the affected areas provided Sunday dialysis for patients who typically received dialysis on a Monday-Wednesday-Friday (MWF) schedule.15 In advance of a rare snowstorm in New Orleans in 2013, nearly all dialysis patients were contacted and advised to receive early dialysis. Fortunately, ,5% of patients with ESRD had an emergency department (ED) visit or were hospitalized following Hurricane Sandy, but it is not clear whether those rates would have been higher without the widespread receipt of early dialysis.15 If early dialysis is to be promoted as a best practice and protective measure in renal care, evidence for its effectiveness in avoiding subsequent adverse outcomes, such as ED visits, hospitalizations, and 30-day mortality, is needed. Building on prior work, we used claims data from the Centers for Medicare & Medicaid Services (CMS) Datalink Project to examine the relationship between the receipt of early dialysis and adverse outcomes of patients with ESRD who received hemodialysis in New York City and the state of New Jersey, the areas most affected by Hurricane Sandy.
METHODS Data Source and Study Population We used Parts A and B Medicare fee-for-service claims, processed as of October 10, 2014, and covering the period from September 29, 2012, to November 30, 2012, to identify patients with ESRD who received facility-based hemodialysis (dialysis) in New York City and the state of New Jersey, the areas most affected by Hurricane Sandy. We included patients in our study if they were ESRD Medicare beneficiaries alive as of October 28 and enrolled in Medicare Parts A and B in September, October, and November 2012, as determined by the Enrollment Database, and had a claim for at least one maintenance dialysis treatment for October 1 to October 28, as identified through Medicare Part B Outpatient fee-for-service claims with type of bill 72x. We excluded patients receiving athome hemodialysis or peritoneal dialysis. We also excluded patients who spent the entire week of the storm in the hospital, defined as being admitted on or prior to October 28 and not discharged until after November 3 (Fig S1, available as online supplementary material). We classified patients according to whether they received early dialysis. To determine whether patients received early dialysis, we compared study patient treatment patterns from the week prior to the storm (October 21, 2012, to October 27, 2012) to the week of the storm (October 28, 2012, to November 3, 2012). Patients were categorized as receiving dialysis on a MWF or Tuesday-ThursdaySaturday (TThS) schedule based on their prior week’s treatment pattern. We excluded patients who did not have a detectable MWF 2
or TThS dialysis schedule the week prior to the storm (n 5 1,663/ 15,499), resulting in a final sample size of 13,836 patients with a MWF or TThS schedule (Fig S1). We considered patients to have received early dialysis if they were in the MWF group and received dialysis on Saturday, October 27, or Sunday, October 28, 2012, or were in the TThS group and received dialysis on Sunday, October 28, or Monday, October 29, 2012. We used Medicare’s Dialysis Facility Compare (DFC) and Certification and Survey Provider Enhanced Reporting (CASPER) data sets to identify ESRD treatment facilities in the study areas. We considered facilities open on a specific date if they submitted at least one claim with a treatment date of service for that date.
Patient Characteristics and Adverse-Outcome Measures Demographic and health status characteristics of patients who did and did not receive early dialysis included age (,65, 65-84, or $85 years), sex, race (white or nonwhite), dual eligibility status (ie, eligible for Medicare and Medicaid; yes/no), and onset of dialysis care (dialysis vintage #1 or .1 year). To identify potential differences in health status, we also identified ED visits and hospitalizations in the 30 days before the storm, as well as evidence of underlying cardiovascular comorbid conditions because another study found elevated rates in New Jersey following Hurricane Sandy.16 Cardiovascular comorbid conditions were defined as a having an International Classification of Diseases, Ninth Revision diagnosis code of myocardial infarction (410), congestive heart failure (428), stroke (430-434), and arrhythmia (426-427) on an inpatient or outpatient Medicare fee-for-service claim 30 days (September 29, 2012, to October 28, 2012) prior to the storm. Outcome measures included ED visits and hospitalizations during the week of the storm (October 29, 2012, to November 3, 2012) and mortality 30 days after the storm (October 29, 2012, to November 27, 2012) for patients who did and did not receive early dialysis. Hospitalizations were determined by admission dates during the week of the storm and were counted once regardless of length of stay. ED visits were determined by service dates during the week of the storm, and a patient was only counted once regardless of frequency of visits. We identified patients who died within 30 days of the storm.
Analysis We compared patient-level variables of patients who did and did not receive early dialysis to identify potential demographic and health status differences between the 2 groups using c2 test. We then used an unadjusted logistic regression model to evaluate whether early dialysis was associated with ED visits or hospitalizations in the week of the storm and in 30-day mortality. Finally, we used multivariable-adjusted logistic regression, including age, sex, race, dual-eligibility status, onset of dialysis care, and health status characteristics (ED visits, hospitalizations, and evidence of underlying cardiovascular comorbid conditions) in the 30 days before the storm to calculate the adjusted odds of adverse outcomes and identify patient characteristics that were independently associated with the outcomes of interest. We considered P , 0.05 to be statistically significant. Analyses were conducted using SAS, version 9.2 (SAS Institute Inc), and STATA, version 12 (StataCorp LP).
Participant Protections The analysis was exempt from institutional review board review because it was performed as a public health quality review and deidentification methods were implemented in accordance with CMS policy and Health Insurance Portability and Accountability Act (HIPAA) requirements. Am J Kidney Dis. 2015;-(-):---
Early Dialysis and Hurricane Sandy
RESULTS Patient Characteristics Table 1 compares demographic and health status characteristics of patients who did and did not receive early dialysis. Of the total 13,836 study patients, 8,256 (60%) received early dialysis on Saturday (October 27), Sunday (October 28), or Monday (October 29) in 209 of 229 (92%) dialysis facilities in the study area. Patients who received early dialysis were more likely to be older (P , 0.001) and white (P , 0.001) and less likely to be dual eligible for Medicare and Medicaid (P , 0.001). ED Visits, Hospitalizations, and Mortality Unadjusted analyses show that patients with early dialysis had lower odds of ED visits (odds ratio [OR], 0.75; 95% confidence interval [CI], 0.63-0.89; P 5 0.001) and hospitalization (OR, 0.77; 95% CI, 0.65-0.92; P 5 0.004) than patients without early dialysis (Table 2). Patients with early dialysis had similar odds of 30-day mortality compared with patients without early dialysis (OR, 0.80; 95% CI, 0.581.09; P 5 0.2; Table 2). In adjusted analyses controlling for patient characteristics, early dialysis continued to be associated with
Table 1. Patient Characteristics Early Dialysis (n 5 8,256)
Age category ,65 y 65-84 y $85 y Male sex Nonwhite race Dual-eligibility statusa Incident dialysis onset ,1y Incident CV comorbidity 30 d before stormb Incident ED visit or hospitalization 30 d before stormc
No Early Dialysis (n 5 5,580)
c2 Test (P)
3,644 (44.1%) 2,685 (48.1%) ,0.001 3,894 (47.2%) 2,501 (44.8%) 718 (8.7%) 394 (7.1%) 4,658 4,946 3,922 1,257
(56.4%) 3,156 (56.6%) 0.9 (59.9%) 3,833 (68.7%) ,0.001 (47.5%) 3,608 (64.7%) ,0.001 (15.2%) 819 (14.7%) 0.4
938 (11.4%)
627 (11.2%)
0.8
1,342 (16.3%)
907 (16.3%)
0.9
Note: Of 229 facilities, 209 (92%) provided early dialysis to at least one patient. Abbreviations: CV, cardiovascular; ED, emergency department. a For both Medicare and Medicaid. b CV comorbidity was defined as having an International Classification of Diseases, Ninth Revision diagnosis code of myocardial infarction (410), congestive heart failure (428), stroke (430-434), or arrhythmia (426-427) on an inpatient or outpatient Medicare fee-for-service claim. c The 30-day period before the storm is September 29, 2012, through October 28, 2012. Am J Kidney Dis. 2015;-(-):---
decreased odds of ED visits (OR, 0.80; 95% CI, 0.670.96; P 5 0.01) and hospitalizations (OR, 0.79; 95% CI, 0.66-0.94; P 5 0.01) in the week of the storm (Table 2). Early dialysis was also found to be significantly associated with reduced odds of 30-day mortality (OR, 0.72; 95% CI, 0.52-0.997; P 5 0.048; Table 2). Patient Characteristics Associated With Adverse Outcomes As shown in Table 2, dual-eligible patients had increased odds for an ED visit and hospitalization, while being nonwhite was associated with decreased odds of both adverse outcomes. Cardiovascular comorbid conditions in the month prior to the storm were associated with increased odds of hospitalization and 30-day mortality. Hospitalization or ED visit in the month prior to the storm was associated with increased odds of hospitalization, ED visits, and 30-day mortality. In addition, being nonwhite and younger was associated with decreased odds of 30day mortality.
DISCUSSION This study confirms that early dialysis ahead of Hurricane Sandy’s landfall decreased the likelihood of ED visits, hospitalizations, and 30-day mortality for dialysis patients in the areas most affected. Such evidence affirms the importance of preparedness practices on the part of dialysis facilities to provide early dialysis, as well as the need for dialysis patients to receive early dialysis when access to routine dialysis may be threatened. Our prior study noted that while an impressive number of patients with ESRD in New York City and New Jersey received early dialysis before Hurricane Sandy made landfall, nearly 40% did not.15 Furthermore, that study found significant variation in the practice of early dialysis across the 5 New York City boroughs.15 Taken together, these findings suggest that routine availability of early dialysis in advance of a storm is generally achievable and should become standard practice. While further work is needed to better understand factors associated with the lack of receipt of early dialysis, it is likely that both patient and facility factors play a role. Patient factors likely include both measureable factors such as income and unmeasured variables such as social support, transportation access, and the capacity to plan well enough ahead to make arrangements for dialysis. We also were not able to determine which patients were offered early dialysis and declined. The KCER Coalition has had an important role in connecting patients with dialysis facilities, including arranging transportation when necessary, and has become an important national asset 3
Lurie et al Table 2. Unadjusted and Adjusted Odds of Adverse Outcomes in Hemodialysis Patients After Hurricane Sandy Unadjusted Regression Factors Associated With Outcome
Early dialysis Nonwhite race Male sex Age category ,65 y 65-84 y $85 y
OR (95% CI)
ED Visit 0.75 (0.63-0.89)
Dual eligiblea Dialysis vintage # 1 y CV comorbid condition diagnosis code in 30 d before storm Hospitalization or ED visit during 30 d before storm Hospitalization Early dialysis 0.77 (0.65-0.92) Nonwhite race Male sex Age category ,65 y 65-84 y $85 y Dual eligiblea Dialysis vintage # 1 y CV comorbid condition diagnosis code in 30 d before storm Hospitalization or ED visit 30 d before storm 30-Day Mortality Early dialysis 0.80 (0.58-1.09) Nonwhite race Male sex Age category ,65 y 65-84 y $85 y Dual eligiblea Dialysis vintage # 1 y CV comorbid condition diagnosis code in 30 d before storm Hospitalization or ED visit 30 d before storm
P
0.001
Adjusted Regression OR (95% CI)
0.80 (0.67-0.96) 0.73 (0.61-0.89) 1.02 (0.85-1.22)
0.01 0.001 0.8
1.01 (0.73-1.41) 0.87 (0.63-1.21) 1.00 (reference)
0.9 0.4
1.68 0.89 0.99 1.79 0.004
(1.38-2.03) (0.70-1.15) (0.75-1.32) (1.42-2.25)
,0.001 0.4 0.9 ,0.001
0.79 (0.66-0.94) 0.79 (0.65-0.96) 0.90 (0.75-1.07)
0.01 0.02 0.2
0.80 (0.58-1.11) 0.95 (0.69-1.31) 1.00 (reference)
0.2 0.8
1.30 0.90 1.39 2.30 0.2
P
(1.07-1.57) (0.70-1.16) (1.07-1.8) (1.84-2.88)
0.01 0.4 0.01 ,0.001
0.72 (0.52-0.997) 0.51 (0.36-0.73) 0.94 (0.68-1.3)
0.048 ,0.001 0.7
0.29 (0.18-0.46) 0.44 (0.29-0.66) 1.00 (reference)
,0.001 ,0.001
1.06 1.00 1.59 3.33
0.8 0.9 0.03 ,0.001
(0.75-1.50) (0.66-1.50) (1.05-2.42) (2.25-4.92)
Note: In the unadjusted logistic regression model, only early dialysis was included as a categorical variable, with the reference group not receiving early dialysis. Abbreviations: CI, confidence interval; CV, cardiovascular; ED, emergency department; OR, odds ratio. a For both Medicare and Medicaid.
in emergency planning and response, particularly for the dialysis community.13,17 Facility preparedness is also critical.10,18 Many of the dialysis facilities in the Sandy-affected region had conducted facility-specific risk assessments and had established emergency plans and alternative energy sources (eg, back-up generators).11,14 All these activities better enabled them to rapidly activate and operate their emergency plans, prioritize and contact patients, and coordinate the surge in early dialysis prior to Hurricane Sandy’s landfall. Both patients and facilities can take additional steps to become prepared. For example, dialysis facilities might consider using a holiday as a proxy for a notice event and test 4
that portion of their emergency and communication plans. In doing so, they can better assess their staff’s ability to coordinate early dialysis in advance of an emergency and assess their patients’ ability to understand, implement, and arrive for their treatment and identify potential gaps in facility and patient personal preparedness. Patients should know how to request early dialysis and become better informed about emergency renal diet options that can assist them in tolerating a treatment delay.17,19,20 For example, during the November 2014 snowstorm in Buffalo, NY, patients were instructed to begin an emergency renal diet. They also need to understand how to seek care in an emergency, Am J Kidney Dis. 2015;-(-):---
Early Dialysis and Hurricane Sandy
preferably at a preidentified alternate dialysis facility, and that they should bring a copy of their dialysis treatment plan and list of their current medications with them.19 Finally, interoperable electronic health records would also help facilitate access to essential dialysis care information. Following Hurricane Sandy, a study conducted by Lin et al21 found that 12 of the 13 hospitals surveyed reported that the majority of the 347 dialysis patients seeking urgent dialysis treatment did not have documentation of their treatment plan or hepatitis status and overwhelmed EDs during the disaster. Access to medical records was identified as a problem in other post–Hurricane Sandy studies as well, further highlighting the importance of interoperable information and the need for facilities to provide their patients with updated treatment plans on a regular basis and in anticipation of potential disruptions.22,23 From a broader policy perspective, advances in data and geospatial mapping have begun to assist state and local health officials, KCER, ESRD Networks, dialysis facilities, and the local health care systems to better anticipate potential dialysis patient and facility needs and hospital surge patterns during emergencies.15,24,25 The US Department of Health and Human Services’ Office of the Assistant Secretary for Preparedness and Response and CMS have demonstrated the utility of mapping deidentified ESRD patient- and facility-level data at the local level to inform and support emergency planning and preparedness activities prior to an emergency.15 Furthermore, during an emergency, the US Department of Health and Human Services now has the capability of securely sharing privacy-protected information with local health officials to assist with direct outreach to dialysis patients.24 Ensuring that hemodialysis patients can maintain access to appropriate renal care through the duration of an emergency is important for both individual and community resilience. This study provides additional evidence that the routine provision of early dialysis in advance of major storms or other emergencies is likely to prevent bad outcomes and save lives.
ACKNOWLEDGEMENTS The views expressed are solely those of the authors and do not necessarily represent those of the US Department of Health and Human Services. We thank Thomas E. MaCurdy, PhD, and Alina Bogdanov, MA, for manuscript preparation assistance and Alicia Livinski, MPH, MA, for literature search assistance. Support: This study was supported through the CMS DataLink contract with Acumen LLC. The funders of this study had a role in the study design, interpretation of the data, writing the report, and the decision to submit the report for publication. Financial Disclosure: The authors declare that they have no other relevant financial interests.
Am J Kidney Dis. 2015;-(-):---
Contributions: Research idea and study design: NL, JK, KF, MJ, TT; data acquisition: JK, CW, MJ, TT; data analysis/interpretation: NL, JK, GM, KF, CW, MJ, TT; statistical analysis: MJ, TT; supervision or mentorship: NL, JK. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. NL and JK assert that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
SUPPLEMENTARY MATERIAL Figure S1: Study participant tree. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2015.04.050) is available at www.ajkd.org
REFERENCES 1. Miller A, Arquilla B. Chronic diseases and natural hazards: impact of disasters on diabetic, renal and cardiac patients. Prehosp Disaster Med. 2008;23(2):185-194. 2. Abdel-Kader K, Unrah ML. Disaster and end-stage renal disease: targeting vulnerable patients for improved outcomes. Kidney Int. 2009;75(11):1131-1133. 3. Zoraster R, Vanholder R, Sever MS. Disaster management of chronic dialysis patients. Am J Disaster Med. 2007;2(2):96-106. 4. Kopp JB, Ball LK, Cohen A, et al. Kidney patient care in disasters: lessons from the hurricanes and earthquakes of 2005. Clin J Am Soc Nephrol. 2007;2:814-824. 5. Vanholder RC, Van Biesen WA, Sever MS. Hurricane Katrina and chronic dialysis patients: better tidings than originally feared? Kidney Int. 2009;76(7):687-689. 6. Kleinpeter M. Disaster preparedness for dialysis patients. Clin J Am Soc Nephrol. 2011;6:2337-2339. 7. Howard D, Zhang R, Huang Y, Kutner N. Hospitalization rates among dialysis patients during Hurricane Katrina. Prehosp Disaster Med. 2012;27(4):1-5. 8. Anderson A, Cohen A, Kutner N, et al. Missed dialysis sessions and hospitalizations in hemodialysis patients after Hurricane Katrina. Kidney Int. 2009;75(11):1202-1208. 9. Abir M, Jan S, Jubelt L, et al. The impact of a large-scale power outage on hemodialysis center operations. Prehosp Disaster Med. 2013;28(6):543-546. 10. Kleinpeter MA. Disaster preparedness of dialysis patients for Hurricanes Gustav and Ike 2008. Adv Perit Dial. 2009;25: 62-67. 11. Lindsey H. Hurricane Sandy: dialysis community manages damage, outages, displacement. Nephrol Times. 2012;5(12):8-11. 12. Cary S, Schroeder K. Caring for patients on kidney dialysis in a disaster. Am J Nurs. 2008;108(1):26-32. 13. Kidney Community Emergency Response Coalition (KCER). About KCER. http://kcercoalition.com/kcer-home/aboutkcer/. Accessed December 16, 2014. 14. Lempert K, Kopp J. Hurricane Sandy as a kidney failure disaster. Am J Kidney Dis. 2013;61(6):865-868. 15. Kelman J, Finne K, Bogdanov A, et al. Dialysis care and death following Hurricane Sandy. Am J Kidney Dis. 2015;65(1): 109-115. 16. Swerdel JN, Janevic TM, Cosgrove NM, Kostis JB; Myocardial Infarction Data Acquisition System (MIDAS 24) Study Group. The effect of Hurricane Sandy on cardiovascular events in New Jersey. J Am Heart Assoc. 2014;3(6):e001354.
5
Lurie et al 17. Kidney Community Emergency Response Coalition (KCER). Hurricane Sandy after action report. December 2012. http://kcercoalition.com/wp-content/uploads/2015/06/KCER_ Hurricane_Sandy_After_Action_Report_12-28-12_508.pdf. Accessed May 31, 2013. 18. Kenney RJ. Emergency preparedness concepts for dialysis facilities: reawakened after Hurricane Katrina. Clin J Am Soc Nephrol. 2007;2(4):809-813. 19. Centers for Medicare & Medicaid Services. Preparing for emergencies: a guide for people on dialysis. CMS Pub No. 10150. 2002. http://kcercoalition.com. Accessed December 15, 2014. 20. Kleinpeter MA, Normal LD, Krane NK. Dialysis services in the hurricane-affected areas in 2005: lessons learned. Am J Med Sci. 2006;332(5):259-263. 21. Lin C-J, Pierce LC, Roblin PM, et al. Impact of Hurricane Sandy on hospital emergency and dialysis services: a retrospective study. Prehosp Disaster Med. 2014;29(4):374-379.
6
22. Adalja A, Watson M, Bouri N, et al. Absorbing citywide patient surge during Hurricane Sandy: a case study in accommodating multiple hospital evacuations. Ann Emerg Med. 2014;64(1): 66-73. 23. Teperman S. Hurricane Sandy and the greater New York health care system. J Trauma Acute Care Surg. 2013;74(6):14011410. 24. DeSalvo K, Lurie N, Finne K, et al. Using Medicare data to identify individuals who are electricity-dependent to improve disaster preparedness and response. Am J Public Health. 2014;104(7): 1160-1164. 25. Lee DC, Smith SW, Carr BG, Goldfrank LR, Polsky D. Redistribution of emergency department patients after disasterrelated closures of a public versus private hospital in New York City [published online first March 17, 2015]. Disaster Med Public Health Prep. http://dx.doi.org/10.1017/dmp. 2015.11.
Am J Kidney Dis. 2015;-(-):---