Original Investigation Dialysis Care and Death Following Hurricane Sandy Jeffrey Kelman, MD, MMSc,1 Kristen Finne, BA,2 Alina Bogdanov, MA,3 Chris Worrall, BS,1 Gregg Margolis, PhD, NREMT-P,2 Kristin Rising, MD, MS,4 Thomas E. MaCurdy, PhD,3 and Nicole Lurie, MD, MSPH2 Background: Hurricane Sandy affected access to critical health care infrastructure. Patients with end-stage renal disease (ESRD) historically have experienced problems accessing care and adverse outcomes during disasters. Study Design: Retrospective cohort study with 2 comparison groups. Setting & Participants: Using Centers for Medicare & Medicaid Services claims data, we assessed the frequency of early dialysis, emergency department (ED) visits, hospitalizations, and 30-day mortality for patients with ESRD in Sandy-affected areas (study group) and 2 comparison groups: (1) patients with ESRD living in states unaffected by Sandy during the same period and (2) patients with ESRD living in the Sandy-affected region a year prior to the hurricane (October 1, 2011, through October 30, 2011). Factor: Regional variation in dialysis care patterns and mortality for patients with ESRD in New York City and the State of New Jersey. Measurements: Frequency of early dialysis, ED visits, hospitalizations, and 30-day mortality. Results: Of 13,264 study patients, 59% received early dialysis in 70% of the New York City and New Jersey dialysis facilities. The ED visit rate was 4.1% for the study group compared with 2.6% and 1.7%, respectively, for comparison groups 1 and 2 (both P , 0.001). The hospitalization rate for the study group also was significantly higher than that in either comparison group (4.5% vs 3.2% and 3.8%, respectively; P , 0.001 and P , 0.003). 23% of study group patients who visited the ED received dialysis in the ED compared with 9.3% and 6.3% in comparison groups 1 and 2, respectively (both P , 0.001). The 30-day mortality rate for the study group was slightly higher than that for either comparison group (1.83% vs 1.47% and 1.60%, respectively; P , 0.001 and P 5 0.1). Limitations: Lack of facility level damage and disaster-induced power outage severity data. Conclusions: Nearly half the study group patients received early dialysis prior to Sandy’s landfall. Poststorm increases in ED visits, hospitalizations, and 30-day mortality were found in the study group, but not in the comparison groups. Am J Kidney Dis. 65(1):109-115. 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: Dialysis; end-stage renal disease (ESRD); emergency preparedness; disaster planning; Kidney Community Emergency Response (KCER) Program; Hurricane Sandy; natural disaster; vulnerable population.
W
hen Hurricane Sandy made landfall in the United States on Monday, October 29, 2012, it affected 17 million individuals across numerous states, led to at least 162 deaths, and caused widespread destruction to thousands of homes and critical health care, transit, energy, water, and communication infrastructure.1-3 Sandy was the second deadliest storm in US history and, similar to Hurricane Katrina, had the potential to disproportionately affect chronically ill populations by limiting access to critical lifemaintaining health care services.4-8 One group of particular concern was patients with end-stage renal disease (ESRD) because many of them are reliant on regular dialysis services and have historically experienced increased morbidity related to disaster-induced disruptions in their dialysis schedules.4,5,7,9 Disaster preparedness practices for patients with ESRD have improved significantly since Hurricane Katrina. For example, many dialysis facilities have backup generators and have implemented practices to expand access to short-term dialysis care in times of Am J Kidney Dis. 2015;65(1):109-115
crisis.10 The Kidney Community Emergency Response (KCER) Program now assists ESRD networks to be better prepared and supports response by facilitating the identification and transportation of patients needing dialysis in emergencies.6 Whether improved From the 1Centers for Medicare & Medicaid Services and Office of the Assistant Secretary for Preparedness and Response, US Department of Health and Human Services, Washington, DC; 3 Acumen, LLC, Burlingame, CA; and the 4Department of Emergency Medicine, Thomas Jefferson University, Philadelphia, PA. Received April 9, 2014. Accepted in revised form July 9, 2014. Originally published online August 23, 2014. 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/$0.00 http://dx.doi.org/10.1053/j.ajkd.2014.07.005 2
109
Kelman et al
preparedness practices have improved dialysis-related care and adverse outcomes has been difficult to assess. The decentralized nature of dialysis care, broad variability and timeliness of renal registry data, and small survey samples have limited population-based research on dialysis care and outcomes related to disasters.5,6,9,11,12 We used claims data from the Centers for Medicare & Medicaid Services (CMS) Datalink Project to characterize patterns of care and mortality of patients with ESRD who live and receive dialysis in the areas that were most affected by Sandy. Understanding how patients with ESRD receive care during disasters is critical to informing preparedness actions and mitigating adverse outcomes for this vulnerable population.
METHODS Data Source and Study Population We used Medicare Parts A and B fee-for-service claims from October 1, 2012, through November 30, 2012, to identify patients with ESRD who received facility-based dialysis in the State of New Jersey and in the 5 New York City boroughs (Brooklyn, Manhattan, Queens, Staten Island, and the Bronx)—the areas most affected by Sandy. We included patients in our study group if they were ESRD Medicare beneficiaries receiving dialysis and enrolled in Medicare Parts A and B in October 2012, as determined by the Enrollment Database; alive as of October 28; and had a claim for at least one maintenance dialysis treatment between October 1 and October 28, as identified through Medicare Part B Outpatient fee-for-service claims with Type of Bill “72x.” We excluded patients receiving at-home hemodialysis or peritoneal dialysis. We defined 2 comparison groups to help us better understand whether patterns of care for the study group were associated with Sandy versus other factors, such as regional and seasonal variability. Comparison group 1 comprised patients with ESRD receiving care during the same week but living in areas unaffected by Sandy. This group included patients with ESRD in all states except New York, New Jersey, Connecticut, Delaware, Massachusetts, Maryland, Maine, North Carolina, New Hampshire, Pennsylvania, Rhode Island, Virginia, Vermont, West Virginia, and the District of Columbia. Comparison group 2 included patients with ESRD receiving care in the same New Jersey and New York City areas during the comparable week (October 31, 2011, through November 6, 2011) a year prior to Sandy. ESRD treatment facilities were identified using Medicare’s Dialysis Facility Compare and Certification and Survey Provider Enhanced Reporting data sets. We considered facilities open on a specific date if they submitted at least one claim with a treatment date of service for that date.
Utilization and Outcome Measures We identified visits to dialysis facilities, emergency department (ED) visits, hospitalizations, and patient deaths 30 days after Sandy. Hospitalizations included those that originated in the ED or were direct admissions. ED visits included visits for patients who were treated in the ED and subsequently discharged. To determine whether patients received early dialysis, we compared 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
110
were categorized based on their prior week treatment pattern as receiving dialysis on a Monday, Wednesday, and Friday (MWF) or Tuesday, Thursday, and Saturday (TThS) schedule. 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 excluded patients who did not have a detectable MWF or TThS dialysis schedule the week prior to the storm (1,948 of 15,212), resulting in a final sample size for the early dialysis analysis of 13,264 (Fig S1, available as online supplementary material). For comparison group 2, we only examined care received on Sunday because most facilities are not open on Sunday and we could not ascertain whether a Monday treatment was for routine care or early dialysis in 2011. We calculated weekly rates of ED visits and hospitalizations from Sunday, October 7, through Saturday, November 17, for patients in the study and comparison groups. A patient was included in the weekly numerator if they were admitted to the hospital or visited the ED and were counted only once regardless of length of stay. A patient was not counted in the denominator if they spent the entire week in the hospital (Fig S1). We also reviewed ED and hospital discharge diagnoses to determine whether the visits likely were dialysis related (eg, fluid and electrolyte disorders and volume overload) or secondary disasterrelated conditions (eg, trauma). Finally, we counted the number of deaths in the 30 days from Sandy’s landfall onward (October 29, 2012, through November 27, 2012) for the study group and comparison group 1 and during the comparable period in 2011 (October 31, 2011, through November 29, 2011) for comparison group 2 (Fig S1).
Analysis We used frequency counts to compare receipt of early dialysis, ED visits, and hospitalizations in the weeks before, during, and after Sandy. We also calculated 30-day mortality after Sandy, using the number of patients with ESRD in the relevant location and time period as the denominator. We used c2 tests to compare ED visits, hospitalizations, and 30-day mortality for patients in the study group with those in the 2 comparison groups. All analyses were conducted using SAS, version 9.2 (SAS Institute Inc).
RESULTS Patient and Facility Characteristics Table 1 lists demographic information for all Medicare fee-for-service ESRD patients receiving care in the study and comparison group areas. During Sandy, 15,212 study patients were seen in 221 dialysis facilities in the affected study area. Patients in comparison group 1 less often were older, nonwhite, and dual-eligible for Medicare and Medicaid; patients in comparison group 2 were demographically similar to the study group. Of 221 facilities caring for the study group, 70% provided early dialysis on the Sunday preceding Sandy compared with 0.04% of facilities caring for comparison group 1 and 0.03% caring for comparison group 2. In the study area, 36 facilities were closed on Monday (October 29, 2012) and 120 were closed on Tuesday (October 30, 2012), with all except 12 facilities resuming some level of service by Wednesday (October 31, 2012). In comparison, no facilities were Am J Kidney Dis. 2015;65(1):109-115
Dialysis Following Hurricane Sandy Table 1. Patient Demographics and Facilities Study Group (n 5 15,212)
Region Bronx Brooklyn Manhattan Queens Staten Island State of New Jersey Treated in multiple regions Unaffected region Age category ,65 y 65-84 y $85 y
1,315 2,266 1,474 1,865 333 7,947 12
Comparison Group 1 (n 5 188,778)
(8.6) (14.9) (9.7) (12.3) (2.2) (52.2) (0.1)
Comparison Group 2 (n 5 15,661)
— — — — — — —
1,363 2,204 1,520 1,809 322 8,423 20
(8.7) (14.1) (9.7) (11.6) (2.1) (53.8) (0.1)
—
188,778 (100.0)
—
7,064 (46.4) 6,966 (45.8) 1,182 (7.8)
102,614 (54.4) 76,784 (40.7) 9,380 (5.0)
7,295 (46.6) 7,127 (45.5) 1,239 (7.9)
Female sex
6,647 (43.7)
86,448 (45.8)
6,923 (44.2)
Nonwhite race
9,591 (63.0)
101,525 (53.8)
9,812 (62.7)
Dual eligiblea
8,243 (54.2)
98,504 (52.2)
8,570 (54.7)
221
4,227
218
No. of ESRD facilities
Note: Unless otherwise indicated, values are given as number (percentage). Study group 5 affected region 2012; comparison group 1 5 unaffected region 2012; comparison group 2 5 affected region 2011. P values not provided because these patients represent the entire population, not a sample. Abbreviation: ESRD, end-stage renal disease. a Dual eligible for Medicare and Medicaid.
closed in the comparison group 1 area and only one facility was closed in the comparison group 2 area on Monday. Similarly, on Tuesday, no facilities were closed in the comparison group 1 area and only 2 facilities were closed in the comparison group 2 area. Early Dialysis Of 13,264 study group patients, 7,791 (58.7%) received early dialysis on either Saturday, Sunday, or Monday. Regional differences in the receipt of early dialysis were evident across all New York City boroughs, as well as between New York City and New Jersey (Fig 1A). As shown by the multiple superimposed lines in Fig 1B, the marked surge in early dialysis for the study group was not detected in either of the 2 comparison groups, and there was no regional variation. ED Visits and Hospitalization Patients in the study group had significantly more ED visits and hospitalizations during the week of the storm than either comparison group (Figs 1C and D). As shown in Table 2, a total of 4.1% of study group patients had an ED visit compared with 2.6% and 1.7% for comparison groups 1 and 2, respectively (both P , 0.001). Similarly, 4.5% of study group patients were hospitalized compared with 3.2% in comparison group 1 (P , 0.001) and 3.8% in comparison group 2 (P 5 0.003). Twenty-three percent of study group patients who had an ED visit also Am J Kidney Dis. 2015;65(1):109-115
received dialysis in the ED compared with 9.3% of patients in comparison group 1 (P , 0.001) and 6.3% of patients in comparison group 2 (P , 0.001). For those in the study group, the overwhelming majority of ED visits and hospitalizations had primary discharge diagnoses related to dialysis or ESRD (Table S1), as opposed to common disaster-induced conditions, such as trauma. Mortality The 30-day mortality rate for patients in the study group (1.83%) was higher than that for either comparison group (1.47% and 1.60%, respectively; P , 0.001 and P 5 0.1; Table 3).
DISCUSSION We used Medicare claims data to assess the utilization experience and mortality of patients with ESRD affected by Sandy and found evidence of widespread use of early dialysis prior to Sandy’s landfall, as well as increases in ED visits, hospitalizations, and mortality after the storm. Although claims data are being used with increasing frequency to conduct surveillance for potential adverse outcomes in at-risk and medically vulnerable populations,13-16 to our knowledge, this is the first time that claims for the entire Medicare ESRD population have been used to examine the experience of a population following a disaster. Prior postdisaster studies of patients with ESRD generally have been based on 111
Kelman et al
Figure 1. Services and outcomes of study and comparison group patients. (A) Study group: early dialysis on Sunday by borough, state; (B) comparison groups 1 and 2: early dialysis on Sunday by borough, state; (C) study group: emergency department visits by borough, state; (D) comparison groups 1 and 2: emergency department visits by borough, state; (E) study group: hospitalization by borough, state; and (F) comparison groups 1 and 2: hospitalization by borough, state. Abbreviation: ESRD, end-stage renal disese.
small patient- and facility-reported data sets or national renal registry data that vary in timeliness, completeness, and quality.5,6,9,11,12 An important strength of this study is that Medicare data cover nearly all the nation’s patients with ESRD. 112
Further, data used for this analysis became available within a month of the storm, demonstrating the feasibility of quickly characterizing Medicare beneficiaries’ care and adverse outcomes. Similar information from Medicare could be accessed prior to, during, and after a Am J Kidney Dis. 2015;65(1):109-115
Dialysis Following Hurricane Sandy
P , 0.001 versus the study group. P 5 0.1 versus the study group.
storm damage and severity of power outages, and we were unable to correct for this confounding. Our findings support the feasibility of using early dialysis as a potential standard of care and protective measure when a hurricane or other “notice event” is anticipated to interrupt dialysis treatment. Emergency preparedness and response depend on good planning and strong day-to-day systems. Many dialysis providers routinely organize early dialysis for patients in advance of major holidays, suggesting that systems are in place to provide such care in advance of a major storm. However, organizing such treatment on a large scale requires integrated emergency planning and clear communications among state and local health officials, facilities, and the patient population.8,9,11,17-20 Prior to Sandy’s landfall, state health officials encouraged dialysis facilities to dialyze patients ahead of schedule and rapidly activated the KCER Coalition to provide additional assistance for coordinating notification and transportation services for patients and to activate additional staff and resources to provide treatment at numerous facilities.6,21,22 Facility preparedness likely played a significant role in the ability to mitigate adverse events as well. Many dialysis facilities in the affected region previously had conducted risk assessments and established emergency plans and alternative energy sources, such as backup generators, for events that could result in prolonged power outages.6,21,22 Such planning similarly may mitigate potential treatment disruptions for “no-notice events” when early dialysis is not possible. For example, despite such planning, we are aware of numerous anecdotal reports that power problems affected the telephones and answering machines at dialysis facilities; although some facility operators reported leaving detailed instructions for patients if they called, the electricity-dependent telephones and answering machines failed. Abir et al20 recently documented strategies used by dialysis facilities to ensure that their patients could continue to receive dialysis after the 2012 mid-Atlantic derecho that caused widespread prolonged power outages. Dialysis patients represent a chronically ill population that may be unusually well connected to an organized care system. Further research should examine whether the practices within the ESRD community that ensure that patients get their needed care can be applied similarly to other vulnerable populations, such as those who are electricity dependent or lose access to refrigerated medications during disasters. In conclusion, Medicare data provided timely and accurate data about access to care and adverse outcomes of a particularly vulnerable population in a disaster context. Administrative data from health systems in countries with universal health care
Am J Kidney Dis. 2015;65(1):109-115
113
Table 2. Hospitalization and ED Visit Rates No. Hospitalized (%)
Study group (n 5 14,928)
No. With ED Visit (%)
672 (4.50)
612 (4.10)
Comparison group 1 (n 5 186,712)
6,029 (3.23)a
4,864 (2.61)a
Comparison group 2 (n 5 15,397)
589 (3.83)b
268 (1.74)a
Note: A patient was not counted in the denominator if they were admitted to the hospital prior to October 28 and not discharged until after November 3 (hospitalized for entire week of storm). Abbreviation: ED, emergency deparment. a P , 0.001 versus the study group. b P 5 0.003 versus the study group.
disaster. An additional strength is the use of 2 different comparison groups, allowing us to conclude that secular trends, regional variation in care, and demographic differences were unlikely to account for the utilization patterns and mortality seen in the study group. The dialysis literature is clear that disruptions in care lead to higher rates of adverse outcomes.4,5,7,9,10,12,13 Our analysis indicates that the preparedness practice of providing early dialysis, in this case ahead of Sandy’s landfall, was widely implemented by facilities. However, the finding of regional variability in the receipt of early dialysis suggests that there may be room for improvement in the preparedness practices of patients with ESRD and dialysis facilities. Placed in the context of prior literature, our findings raise the question of whether the increases in ED use, hospitalizations, and mortality would have been far greater without early dialysis. Our study also is subject to limitations. Although Medicare data contain a rich set of variables and cover almost the entire ESRD population, we did not have access to sufficiently granular data on sustained power outages and facility damage, which may have enabled us to better understand the utilization patterns seen. For example, we would have liked to examine whether characteristics of dialysis facilities, such as ownership, differed between those that did and did not provide early dialysis. Unfortunately, facility ownership patterns differ significantly across the affected area in ways that may have been associated with the degree of Table 3. The 30-Day Mortality Rates No. Hospitalized (%)
Study group (n 5 15,212) Comparison group 1 (n 5 188,778) Comparison group 2 (n 5 15,661) a b
278 (1.83) 2,784 (1.47)a 250 (1.60)b
Kelman et al
systems likely could be put to similar use. From a preparedness perspective, dialysis facilities should consider having plans in place to provide early dialysis to patients in advance of predictable disasters such as hurricanes and to ensure their ability to function in a disaster. Mapping dialysis facility and network locations also can assist federal, state, and local governments to identify those who may be affected by a disaster or prolonged power outages and provide useful information to decision makers regarding priority power restoration. From a broader policy perspective, state licensing requirements and or changes to Conditions of Participation in Medicare and Medicaid could mandate the availability of backup power, emergency preparedness and communication plans, and training for all dialysis facility staff to increase the likelihood that a facility remains open and patients can access dialysis care during a disaster. Facilities and networks also should actively seek opportunities to participate in KCER and their local health care coalition, as well as regularly communicate emergency plan and practices information with their state and local health departments to better enable integration and coordination prior to and during an emergency. We note that CMS recently proposed such changes.23 Analyses of Medicare claims for non-ESRD populations may shed light on outcomes for those populations and identify other ways in which health care claims data can be used to prepare for and respond to future disasters for a broader range of at-risk and medically vulnerable populations and whether these protective measures can mitigate anticipated adverse outcomes.
ACKNOWLEDGEMENTS We thank Maria Jauregui, BA, Sasha Kapralov, MA, and Mark Page, BDSc(Hons), PhD, GDClinDent, for manuscript preparation assistance; Kacey Wulff, MPH, for editorial assistance; and Alicia Livinski, MPH, MA, for literature search assistance. The views expressed are solely those of the authors and do not necessarily represent those of the US Department of Health and Human Services. Support: This study was supported through the CMS DataLink contract (HHSM-500-2011-00115G) 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. Contributions: Research idea and study design: JK, KF, AB, CW, GM, KR, TEM, NL; data acquisition: AB, TEM; data analysis/ interpretation: JK, KF, AB, CW, GM, KR, TEM, NL; statistical analysis: AB, TEM; supervision or mentorship: JK, NL, TEM. 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. JK, NL, and TEM take responsibility that this study has been reported honestly, accurately, and transparently; that no important
114
aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
SUPPLEMENTARY MATERIAL Table S1: Top-10 primary diagnosis codes in hospitalization and ED claims for study group. Figure S1: Study participant tree. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2014.07.005) is available at www.ajkd.org
REFERENCES 1. Blatchford L. Written Testimony. Field Hearing with Senate Homeland Security Appropriations Subcommittee, March 1, 2013. http://www.appropriations.senate.gov/sites/default/files/hearings/ Laurel%20Blatchford%20Testimony.pdf. Accessed August 12, 2014. 2. Donovan S. Written Testimony. Hurricane Sandy: Getting the Recovery Right and the Value of Mitigation. March 20, 2013. http://www.hsgac.senate.gov/hearings/hurricane-sandy-gettingthe-recovery-right-and-the-value-of-mitigation. Accessed August 12, 2014. 3. Wikipedia. Hurricane Sandy. http://en.wikipedia.org/wiki/ Hurricane_sandy. Accessed May 31, 2013. 4. Zoraster R, Vanholder R, Sever MS. Disaster management of chronic dialysis patients. Am J Disaster Med. 2007;2(2):96-106. 5. Abdel-Kader K, Unrah ML. Disaster and end-stage renal disease: targeting vulnerable patients for improved outcomes. Kidney Int. 2009;75(11):1131-1133. 6. Lempert K, Kopp J. Hurricane Sandy as a kidney failure disaster. Am J Kidney Dis. 2013;61(6):865-868. 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. Redlener I, Reilly M. Lessons from Sandy—preparing health systems for future disasters. N Engl J Med. 2012;367(24):2269-2271. 9. Kopp JB, Ball LK, Cohen A, et al. Kidney patient care in disasters: lessons from the hurricanes and earthquake in 2005. Clin J Am Soc Nephrol. 2007;2(4):814-824. 10. 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. 11. Kleinpeter MA. Disaster preparedness of dialysis patients for hurricanes Gustav and Ike 2008. Adv Perit Dial. 2009;25:62-67. 12. 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. 13. Kutner NG, Muntner P, Huang Y, et al. Effect of Hurricane Katrina on the mortality of dialysis patients. Kidney Int. 2009;76(7):760-766. 14. Burwen D, Sandhu S, MaCurdy T, et al. Surveillance for Guillain-Barre syndrome after influenza vaccination among the Medicare population, 2009-2010. Am J Public Health. 2012;102(10):1921-1927. 15. Davis RL, Kolczak M, Lewis E, et al. Active surveillance of vaccine safety: a system to detect early signs of adverse events. Epidemiology. 2005;16(3):336-341. 16. Lieu TA, Kulldorf M, Davis RL, et al. Vaccine Safety Datalink Rapid Cycle Analysis Team. Real-time vaccine safety surveillance for the early detection of adverse events. Med Care. 2007;45(10)(suppl 2):S89-S95. 17. Kleinpeter MA, Normal LD, Krane NK. Disaster planning for peritoneal dialysis programs. Adv Perit Dial. 2006;22:124-129.
Am J Kidney Dis. 2015;65(1):109-115
Dialysis Following Hurricane Sandy 18. 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. 19. Kenney RJ. Emergency preparedness concepts for dialysis facilities: reawakened after Hurricane Katrina. Clin J Am Soc Nephrol. 2007;2(4):809-813. 20. Abir M, Jan S, Jubelt L, Merchant RM, Lurie N. The impact of a large-scale power outage on hemodialysis center operations. Prehosp Disaster Med. 2013;28(6):543-546.
Am J Kidney Dis. 2015;65(1):109-115
21. Kidney Community Emergency Response Coalition (KCER). Hurricane Sandy after action report. December 2012. http://www.kcercoalition.com. Accessed May 31, 2013. 22. Lindsey H. Hurricane Sandy: dialysis community manages damage, outages, displacement. Nephrol Times. 2012;5(12):8-11. 23. Centers for Medicare & Medicaid Services. Emergency preparedness requirements for Medicare and Medicaid participating providers and suppliers. Fed Regist. 2013;78(249):7908279200.
115
W
hen Hurricane Sandy made landfall in the United States on Monday, October 29, 2012, it affected 17 million individuals across numerous states, led to at least 162 deaths, and caused widespread destruction to thousands of homes and critical health care, transit, energy, water, and communication infrastructure.1-3 Sandy was the second deadliest storm in US history and, similar to Hurricane Katrina, had the potential to disproportionately affect chronically ill populations by limiting access to critical lifemaintaining health care services.4-8 One group of particular concern was patients with end-stage renal disease (ESRD) because many of them are reliant on regular dialysis services and have historically experienced increased morbidity related to disaster-induced disruptions in their dialysis schedules.4,5,7,9 Disaster preparedness practices for patients with ESRD have improved significantly since Hurricane Katrina. For example, many dialysis facilities have backup generators and have implemented practices to expand access to short-term dialysis care in times of Am J Kidney Dis. 2015;65(1):109-115
crisis.10 The Kidney Community Emergency Response (KCER) Program now assists ESRD networks to be better prepared and supports response by facilitating the identification and transportation of patients needing dialysis in emergencies.6 Whether improved From the 1Centers for Medicare & Medicaid Services and Office of the Assistant Secretary for Preparedness and Response, US Department of Health and Human Services, Washington, DC; 3 Acumen, LLC, Burlingame, CA; and the 4Department of Emergency Medicine, Thomas Jefferson University, Philadelphia, PA. Received April 9, 2014. Accepted in revised form July 9, 2014. Originally published online August 23, 2014. 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/$0.00 http://dx.doi.org/10.1053/j.ajkd.2014.07.005 2
109
Kelman et al
preparedness practices have improved dialysis-related care and adverse outcomes has been difficult to assess. The decentralized nature of dialysis care, broad variability and timeliness of renal registry data, and small survey samples have limited population-based research on dialysis care and outcomes related to disasters.5,6,9,11,12 We used claims data from the Centers for Medicare & Medicaid Services (CMS) Datalink Project to characterize patterns of care and mortality of patients with ESRD who live and receive dialysis in the areas that were most affected by Sandy. Understanding how patients with ESRD receive care during disasters is critical to informing preparedness actions and mitigating adverse outcomes for this vulnerable population.
METHODS Data Source and Study Population We used Medicare Parts A and B fee-for-service claims from October 1, 2012, through November 30, 2012, to identify patients with ESRD who received facility-based dialysis in the State of New Jersey and in the 5 New York City boroughs (Brooklyn, Manhattan, Queens, Staten Island, and the Bronx)—the areas most affected by Sandy. We included patients in our study group if they were ESRD Medicare beneficiaries receiving dialysis and enrolled in Medicare Parts A and B in October 2012, as determined by the Enrollment Database; alive as of October 28; and had a claim for at least one maintenance dialysis treatment between October 1 and October 28, as identified through Medicare Part B Outpatient fee-for-service claims with Type of Bill “72x.” We excluded patients receiving at-home hemodialysis or peritoneal dialysis. We defined 2 comparison groups to help us better understand whether patterns of care for the study group were associated with Sandy versus other factors, such as regional and seasonal variability. Comparison group 1 comprised patients with ESRD receiving care during the same week but living in areas unaffected by Sandy. This group included patients with ESRD in all states except New York, New Jersey, Connecticut, Delaware, Massachusetts, Maryland, Maine, North Carolina, New Hampshire, Pennsylvania, Rhode Island, Virginia, Vermont, West Virginia, and the District of Columbia. Comparison group 2 included patients with ESRD receiving care in the same New Jersey and New York City areas during the comparable week (October 31, 2011, through November 6, 2011) a year prior to Sandy. ESRD treatment facilities were identified using Medicare’s Dialysis Facility Compare and Certification and Survey Provider Enhanced Reporting data sets. We considered facilities open on a specific date if they submitted at least one claim with a treatment date of service for that date.
Utilization and Outcome Measures We identified visits to dialysis facilities, emergency department (ED) visits, hospitalizations, and patient deaths 30 days after Sandy. Hospitalizations included those that originated in the ED or were direct admissions. ED visits included visits for patients who were treated in the ED and subsequently discharged. To determine whether patients received early dialysis, we compared 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
110
were categorized based on their prior week treatment pattern as receiving dialysis on a Monday, Wednesday, and Friday (MWF) or Tuesday, Thursday, and Saturday (TThS) schedule. 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 excluded patients who did not have a detectable MWF or TThS dialysis schedule the week prior to the storm (1,948 of 15,212), resulting in a final sample size for the early dialysis analysis of 13,264 (Fig S1, available as online supplementary material). For comparison group 2, we only examined care received on Sunday because most facilities are not open on Sunday and we could not ascertain whether a Monday treatment was for routine care or early dialysis in 2011. We calculated weekly rates of ED visits and hospitalizations from Sunday, October 7, through Saturday, November 17, for patients in the study and comparison groups. A patient was included in the weekly numerator if they were admitted to the hospital or visited the ED and were counted only once regardless of length of stay. A patient was not counted in the denominator if they spent the entire week in the hospital (Fig S1). We also reviewed ED and hospital discharge diagnoses to determine whether the visits likely were dialysis related (eg, fluid and electrolyte disorders and volume overload) or secondary disasterrelated conditions (eg, trauma). Finally, we counted the number of deaths in the 30 days from Sandy’s landfall onward (October 29, 2012, through November 27, 2012) for the study group and comparison group 1 and during the comparable period in 2011 (October 31, 2011, through November 29, 2011) for comparison group 2 (Fig S1).
Analysis We used frequency counts to compare receipt of early dialysis, ED visits, and hospitalizations in the weeks before, during, and after Sandy. We also calculated 30-day mortality after Sandy, using the number of patients with ESRD in the relevant location and time period as the denominator. We used c2 tests to compare ED visits, hospitalizations, and 30-day mortality for patients in the study group with those in the 2 comparison groups. All analyses were conducted using SAS, version 9.2 (SAS Institute Inc).
RESULTS Patient and Facility Characteristics Table 1 lists demographic information for all Medicare fee-for-service ESRD patients receiving care in the study and comparison group areas. During Sandy, 15,212 study patients were seen in 221 dialysis facilities in the affected study area. Patients in comparison group 1 less often were older, nonwhite, and dual-eligible for Medicare and Medicaid; patients in comparison group 2 were demographically similar to the study group. Of 221 facilities caring for the study group, 70% provided early dialysis on the Sunday preceding Sandy compared with 0.04% of facilities caring for comparison group 1 and 0.03% caring for comparison group 2. In the study area, 36 facilities were closed on Monday (October 29, 2012) and 120 were closed on Tuesday (October 30, 2012), with all except 12 facilities resuming some level of service by Wednesday (October 31, 2012). In comparison, no facilities were Am J Kidney Dis. 2015;65(1):109-115
Dialysis Following Hurricane Sandy Table 1. Patient Demographics and Facilities Study Group (n 5 15,212)
Region Bronx Brooklyn Manhattan Queens Staten Island State of New Jersey Treated in multiple regions Unaffected region Age category ,65 y 65-84 y $85 y
1,315 2,266 1,474 1,865 333 7,947 12
Comparison Group 1 (n 5 188,778)
(8.6) (14.9) (9.7) (12.3) (2.2) (52.2) (0.1)
Comparison Group 2 (n 5 15,661)
— — — — — — —
1,363 2,204 1,520 1,809 322 8,423 20
(8.7) (14.1) (9.7) (11.6) (2.1) (53.8) (0.1)
—
188,778 (100.0)
—
7,064 (46.4) 6,966 (45.8) 1,182 (7.8)
102,614 (54.4) 76,784 (40.7) 9,380 (5.0)
7,295 (46.6) 7,127 (45.5) 1,239 (7.9)
Female sex
6,647 (43.7)
86,448 (45.8)
6,923 (44.2)
Nonwhite race
9,591 (63.0)
101,525 (53.8)
9,812 (62.7)
Dual eligiblea
8,243 (54.2)
98,504 (52.2)
8,570 (54.7)
221
4,227
218
No. of ESRD facilities
Note: Unless otherwise indicated, values are given as number (percentage). Study group 5 affected region 2012; comparison group 1 5 unaffected region 2012; comparison group 2 5 affected region 2011. P values not provided because these patients represent the entire population, not a sample. Abbreviation: ESRD, end-stage renal disease. a Dual eligible for Medicare and Medicaid.
closed in the comparison group 1 area and only one facility was closed in the comparison group 2 area on Monday. Similarly, on Tuesday, no facilities were closed in the comparison group 1 area and only 2 facilities were closed in the comparison group 2 area. Early Dialysis Of 13,264 study group patients, 7,791 (58.7%) received early dialysis on either Saturday, Sunday, or Monday. Regional differences in the receipt of early dialysis were evident across all New York City boroughs, as well as between New York City and New Jersey (Fig 1A). As shown by the multiple superimposed lines in Fig 1B, the marked surge in early dialysis for the study group was not detected in either of the 2 comparison groups, and there was no regional variation. ED Visits and Hospitalization Patients in the study group had significantly more ED visits and hospitalizations during the week of the storm than either comparison group (Figs 1C and D). As shown in Table 2, a total of 4.1% of study group patients had an ED visit compared with 2.6% and 1.7% for comparison groups 1 and 2, respectively (both P , 0.001). Similarly, 4.5% of study group patients were hospitalized compared with 3.2% in comparison group 1 (P , 0.001) and 3.8% in comparison group 2 (P 5 0.003). Twenty-three percent of study group patients who had an ED visit also Am J Kidney Dis. 2015;65(1):109-115
received dialysis in the ED compared with 9.3% of patients in comparison group 1 (P , 0.001) and 6.3% of patients in comparison group 2 (P , 0.001). For those in the study group, the overwhelming majority of ED visits and hospitalizations had primary discharge diagnoses related to dialysis or ESRD (Table S1), as opposed to common disaster-induced conditions, such as trauma. Mortality The 30-day mortality rate for patients in the study group (1.83%) was higher than that for either comparison group (1.47% and 1.60%, respectively; P , 0.001 and P 5 0.1; Table 3).
DISCUSSION We used Medicare claims data to assess the utilization experience and mortality of patients with ESRD affected by Sandy and found evidence of widespread use of early dialysis prior to Sandy’s landfall, as well as increases in ED visits, hospitalizations, and mortality after the storm. Although claims data are being used with increasing frequency to conduct surveillance for potential adverse outcomes in at-risk and medically vulnerable populations,13-16 to our knowledge, this is the first time that claims for the entire Medicare ESRD population have been used to examine the experience of a population following a disaster. Prior postdisaster studies of patients with ESRD generally have been based on 111
Kelman et al
Figure 1. Services and outcomes of study and comparison group patients. (A) Study group: early dialysis on Sunday by borough, state; (B) comparison groups 1 and 2: early dialysis on Sunday by borough, state; (C) study group: emergency department visits by borough, state; (D) comparison groups 1 and 2: emergency department visits by borough, state; (E) study group: hospitalization by borough, state; and (F) comparison groups 1 and 2: hospitalization by borough, state. Abbreviation: ESRD, end-stage renal disese.
small patient- and facility-reported data sets or national renal registry data that vary in timeliness, completeness, and quality.5,6,9,11,12 An important strength of this study is that Medicare data cover nearly all the nation’s patients with ESRD. 112
Further, data used for this analysis became available within a month of the storm, demonstrating the feasibility of quickly characterizing Medicare beneficiaries’ care and adverse outcomes. Similar information from Medicare could be accessed prior to, during, and after a Am J Kidney Dis. 2015;65(1):109-115
Dialysis Following Hurricane Sandy
P , 0.001 versus the study group. P 5 0.1 versus the study group.
storm damage and severity of power outages, and we were unable to correct for this confounding. Our findings support the feasibility of using early dialysis as a potential standard of care and protective measure when a hurricane or other “notice event” is anticipated to interrupt dialysis treatment. Emergency preparedness and response depend on good planning and strong day-to-day systems. Many dialysis providers routinely organize early dialysis for patients in advance of major holidays, suggesting that systems are in place to provide such care in advance of a major storm. However, organizing such treatment on a large scale requires integrated emergency planning and clear communications among state and local health officials, facilities, and the patient population.8,9,11,17-20 Prior to Sandy’s landfall, state health officials encouraged dialysis facilities to dialyze patients ahead of schedule and rapidly activated the KCER Coalition to provide additional assistance for coordinating notification and transportation services for patients and to activate additional staff and resources to provide treatment at numerous facilities.6,21,22 Facility preparedness likely played a significant role in the ability to mitigate adverse events as well. Many dialysis facilities in the affected region previously had conducted risk assessments and established emergency plans and alternative energy sources, such as backup generators, for events that could result in prolonged power outages.6,21,22 Such planning similarly may mitigate potential treatment disruptions for “no-notice events” when early dialysis is not possible. For example, despite such planning, we are aware of numerous anecdotal reports that power problems affected the telephones and answering machines at dialysis facilities; although some facility operators reported leaving detailed instructions for patients if they called, the electricity-dependent telephones and answering machines failed. Abir et al20 recently documented strategies used by dialysis facilities to ensure that their patients could continue to receive dialysis after the 2012 mid-Atlantic derecho that caused widespread prolonged power outages. Dialysis patients represent a chronically ill population that may be unusually well connected to an organized care system. Further research should examine whether the practices within the ESRD community that ensure that patients get their needed care can be applied similarly to other vulnerable populations, such as those who are electricity dependent or lose access to refrigerated medications during disasters. In conclusion, Medicare data provided timely and accurate data about access to care and adverse outcomes of a particularly vulnerable population in a disaster context. Administrative data from health systems in countries with universal health care
Am J Kidney Dis. 2015;65(1):109-115
113
Table 2. Hospitalization and ED Visit Rates No. Hospitalized (%)
Study group (n 5 14,928)
No. With ED Visit (%)
672 (4.50)
612 (4.10)
Comparison group 1 (n 5 186,712)
6,029 (3.23)a
4,864 (2.61)a
Comparison group 2 (n 5 15,397)
589 (3.83)b
268 (1.74)a
Note: A patient was not counted in the denominator if they were admitted to the hospital prior to October 28 and not discharged until after November 3 (hospitalized for entire week of storm). Abbreviation: ED, emergency deparment. a P , 0.001 versus the study group. b P 5 0.003 versus the study group.
disaster. An additional strength is the use of 2 different comparison groups, allowing us to conclude that secular trends, regional variation in care, and demographic differences were unlikely to account for the utilization patterns and mortality seen in the study group. The dialysis literature is clear that disruptions in care lead to higher rates of adverse outcomes.4,5,7,9,10,12,13 Our analysis indicates that the preparedness practice of providing early dialysis, in this case ahead of Sandy’s landfall, was widely implemented by facilities. However, the finding of regional variability in the receipt of early dialysis suggests that there may be room for improvement in the preparedness practices of patients with ESRD and dialysis facilities. Placed in the context of prior literature, our findings raise the question of whether the increases in ED use, hospitalizations, and mortality would have been far greater without early dialysis. Our study also is subject to limitations. Although Medicare data contain a rich set of variables and cover almost the entire ESRD population, we did not have access to sufficiently granular data on sustained power outages and facility damage, which may have enabled us to better understand the utilization patterns seen. For example, we would have liked to examine whether characteristics of dialysis facilities, such as ownership, differed between those that did and did not provide early dialysis. Unfortunately, facility ownership patterns differ significantly across the affected area in ways that may have been associated with the degree of Table 3. The 30-Day Mortality Rates No. Hospitalized (%)
Study group (n 5 15,212) Comparison group 1 (n 5 188,778) Comparison group 2 (n 5 15,661) a b
278 (1.83) 2,784 (1.47)a 250 (1.60)b
Kelman et al
systems likely could be put to similar use. From a preparedness perspective, dialysis facilities should consider having plans in place to provide early dialysis to patients in advance of predictable disasters such as hurricanes and to ensure their ability to function in a disaster. Mapping dialysis facility and network locations also can assist federal, state, and local governments to identify those who may be affected by a disaster or prolonged power outages and provide useful information to decision makers regarding priority power restoration. From a broader policy perspective, state licensing requirements and or changes to Conditions of Participation in Medicare and Medicaid could mandate the availability of backup power, emergency preparedness and communication plans, and training for all dialysis facility staff to increase the likelihood that a facility remains open and patients can access dialysis care during a disaster. Facilities and networks also should actively seek opportunities to participate in KCER and their local health care coalition, as well as regularly communicate emergency plan and practices information with their state and local health departments to better enable integration and coordination prior to and during an emergency. We note that CMS recently proposed such changes.23 Analyses of Medicare claims for non-ESRD populations may shed light on outcomes for those populations and identify other ways in which health care claims data can be used to prepare for and respond to future disasters for a broader range of at-risk and medically vulnerable populations and whether these protective measures can mitigate anticipated adverse outcomes.
ACKNOWLEDGEMENTS We thank Maria Jauregui, BA, Sasha Kapralov, MA, and Mark Page, BDSc(Hons), PhD, GDClinDent, for manuscript preparation assistance; Kacey Wulff, MPH, for editorial assistance; and Alicia Livinski, MPH, MA, for literature search assistance. The views expressed are solely those of the authors and do not necessarily represent those of the US Department of Health and Human Services. Support: This study was supported through the CMS DataLink contract (HHSM-500-2011-00115G) 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. Contributions: Research idea and study design: JK, KF, AB, CW, GM, KR, TEM, NL; data acquisition: AB, TEM; data analysis/ interpretation: JK, KF, AB, CW, GM, KR, TEM, NL; statistical analysis: AB, TEM; supervision or mentorship: JK, NL, TEM. 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. JK, NL, and TEM take responsibility that this study has been reported honestly, accurately, and transparently; that no important
114
aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
SUPPLEMENTARY MATERIAL Table S1: Top-10 primary diagnosis codes in hospitalization and ED claims for study group. Figure S1: Study participant tree. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2014.07.005) is available at www.ajkd.org
REFERENCES 1. Blatchford L. Written Testimony. Field Hearing with Senate Homeland Security Appropriations Subcommittee, March 1, 2013. http://www.appropriations.senate.gov/sites/default/files/hearings/ Laurel%20Blatchford%20Testimony.pdf. Accessed August 12, 2014. 2. Donovan S. Written Testimony. Hurricane Sandy: Getting the Recovery Right and the Value of Mitigation. March 20, 2013. http://www.hsgac.senate.gov/hearings/hurricane-sandy-gettingthe-recovery-right-and-the-value-of-mitigation. Accessed August 12, 2014. 3. Wikipedia. Hurricane Sandy. http://en.wikipedia.org/wiki/ Hurricane_sandy. Accessed May 31, 2013. 4. Zoraster R, Vanholder R, Sever MS. Disaster management of chronic dialysis patients. Am J Disaster Med. 2007;2(2):96-106. 5. Abdel-Kader K, Unrah ML. Disaster and end-stage renal disease: targeting vulnerable patients for improved outcomes. Kidney Int. 2009;75(11):1131-1133. 6. Lempert K, Kopp J. Hurricane Sandy as a kidney failure disaster. Am J Kidney Dis. 2013;61(6):865-868. 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. Redlener I, Reilly M. Lessons from Sandy—preparing health systems for future disasters. N Engl J Med. 2012;367(24):2269-2271. 9. Kopp JB, Ball LK, Cohen A, et al. Kidney patient care in disasters: lessons from the hurricanes and earthquake in 2005. Clin J Am Soc Nephrol. 2007;2(4):814-824. 10. 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. 11. Kleinpeter MA. Disaster preparedness of dialysis patients for hurricanes Gustav and Ike 2008. Adv Perit Dial. 2009;25:62-67. 12. 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. 13. Kutner NG, Muntner P, Huang Y, et al. Effect of Hurricane Katrina on the mortality of dialysis patients. Kidney Int. 2009;76(7):760-766. 14. Burwen D, Sandhu S, MaCurdy T, et al. Surveillance for Guillain-Barre syndrome after influenza vaccination among the Medicare population, 2009-2010. Am J Public Health. 2012;102(10):1921-1927. 15. Davis RL, Kolczak M, Lewis E, et al. Active surveillance of vaccine safety: a system to detect early signs of adverse events. Epidemiology. 2005;16(3):336-341. 16. Lieu TA, Kulldorf M, Davis RL, et al. Vaccine Safety Datalink Rapid Cycle Analysis Team. Real-time vaccine safety surveillance for the early detection of adverse events. Med Care. 2007;45(10)(suppl 2):S89-S95. 17. Kleinpeter MA, Normal LD, Krane NK. Disaster planning for peritoneal dialysis programs. Adv Perit Dial. 2006;22:124-129.
Am J Kidney Dis. 2015;65(1):109-115
Dialysis Following Hurricane Sandy 18. 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. 19. Kenney RJ. Emergency preparedness concepts for dialysis facilities: reawakened after Hurricane Katrina. Clin J Am Soc Nephrol. 2007;2(4):809-813. 20. Abir M, Jan S, Jubelt L, Merchant RM, Lurie N. The impact of a large-scale power outage on hemodialysis center operations. Prehosp Disaster Med. 2013;28(6):543-546.
Am J Kidney Dis. 2015;65(1):109-115
21. Kidney Community Emergency Response Coalition (KCER). Hurricane Sandy after action report. December 2012. http://www.kcercoalition.com. Accessed May 31, 2013. 22. Lindsey H. Hurricane Sandy: dialysis community manages damage, outages, displacement. Nephrol Times. 2012;5(12):8-11. 23. Centers for Medicare & Medicaid Services. Emergency preparedness requirements for Medicare and Medicaid participating providers and suppliers. Fed Regist. 2013;78(249):7908279200.
115
