Disaster-Driven Evacuation and Medication Loss : a Systematic Literature Review

AIM: The aim of this systematic literature review was to identify the extent and implications of medication loss and the burden of prescription refill on medical relief teams following extreme weather events and other natural hazards. METHOD: The search strategy followed the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA). Key health journal databases (Medline, Embase, PsycINFO, Maternity and Infant Care, and Health Management Information Consortium (HMIC)) were searched via the OvidSP search engine. Search terms were identified by consulting MeSH terms. The inclusion criteria comprised articles published from January 2003 to August 2013, written in English and containing an abstract. The exclusion criteria included abstracts for conferences or dissertations, book chapters and articles written in a language other than English. A total of 70 articles which fulfilled the inclusion criteria were included in this systematic review. RESULTS: All relevant information was collated regarding medication loss, prescription loss and refills, and medical aids loss which indicated a significant burden on the medical relief teams. Data also showed the difficulty in filling prescriptions due to lack of information from the evacuees. People with chronic conditions are most at risk when their medication is not available. This systematic review also showed that medical aids such as eye glasses, hearing aids as well as dental treatment are a high necessity among evacuees. DISCUSSION: This systematic review revealed that a considerable number of patients lose their medication during evacuation, many lose essential medical aids such as insulin pens and many do not bring prescriptions with them when evacuated.. Since medication loss is partly a responsibility of evacuees, understanding the impact of medication loss may lead to raising awareness and better preparations among the patients and health care professionals. People who are not prepared could have worse outcomes and many risk dying when their medication is not available.

with chronic conditions are most at risk when their medication is not available. This systematic review also showed that medical aids such as eye glasses, hearing aids as well as dental treatment are a high necessity among evacuees.
DISCUSSION: This systematic review revealed that a considerable number of patients lose their medication during evacuation, many lose essential medical aids such as insulin pens and many do not bring prescriptions with them when evacuated.. Since medication loss is partly a responsibility of evacuees, understanding the impact of medication loss may lead to raising awareness and better preparations among the patients and health care professionals. People who are not prepared could have worse outcomes and many risk dying when their medication is not available.
This project was not funded. The authors have declared that no competing interests exist.
After an extreme weather event or other natural hazard, the continuity of routine care is one of many challenging aspects of post disaster healthcare. Although a disaster can impact on all available services, healthcare facilities can be overwhelmed reducing their ability to maintain normal function. Structural and nonstructural damage to their buildings, creating an insecure environment for hospital staff, and disruption of supply chains, all lead to closure of wards for new admissions, or even evacuation of patients and staff, at a time when they are critical for those who are injured by the disaster. Recognising the impact of extreme events on healthcare facilities, normal functioning is still required for people needing routine health management such as oncological treatment, dialysis and maternity care, as well as those people with chronic conditions who require daily medication to maintain their wellbeing.
Longterm noncommunicable diseases are increasing with a growing and ageing world population and in 30 years from 1990, NCD's are estimated to increased 1.8 times . Ensuring continuation of routine care for chronic conditions will be an increasing burden during and post disaster periods both in developing and developed countries . Interruption of routine medication lead can lead to an exacerbation of chronic conditions such as insulindependent diabetes and infectious diseases for example tuberculosis . This can also potentially cause secondary lifethreatening outcomes as a result of the deterioration of chronic conditions such as ischemic heart diseases among patients with hypertension , and low compliance to medication regimens in the future .
During any disaster, medication maintenance is problematic due to people not having adequate dosages for a sufficient period of time, not having prescriptions with them, not remembering the medication they are on and more likely not having any medication with them at all. These people have been described as 'drug refugees'. In the Great East Japan Earthquake in 2011, for example, a large number of 'drug refugees' were reported , and at least 283 people were reported to have died from the exacerbation of preexisting conditions due to lack of access to healthcare . The health impacts on drug refugees has had, little research conducted among the affected population.
Since medication loss is partly a responsibility of evacuees, understanding the impact of medication loss may lead to raising awareness and better preparations among the patients and health care professionals. The aim of this systematic literature review was to identify the extent and implications of medication loss and the burden of prescription refill on medical relief teams following extreme weather events and other natural hazards.  The search strategy followed the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) statement where applicable, and this checklist was used in designing and reporting our review .

Identification
The key health journal databases (Medline, Embase, PsycINFO, Maternity and Infant Care, and Health Management Information Consortium (HMIC)) were searched via the OvidSP search engine. Search terms were identified by consulting MeSH terms. The validity of the search was confirmed by comparing the generated results to articles obtained from expert consultation and 'snowball' search). After this pilot search, it was revealed that using only MeSH terms was not sufficient to identify all relevant articles. Therefore, key words related to disaster and medication were added to MeSH term search as shown in medication for chronic conditions within relief activities after disasters, and/or (iii) disruption of medications due to evacuees not bringing their medications.

2) Exclusion criteria
Articles and papers were excluded if they were (i) abstracts for conferences or dissertations; (ii) chapters of books; and (iii) articles written in a language other than English.

Study selection
The search was conducted on 5 September, 2013 and generated 5,382 results of which 1,652 were duplicates and removed, leaving 3,730 records, Out of these records 2,961 were apparently irrelevant when screened by title exclusion criteria. Hence, the initial screening by title identified 811 records as relevant and for these remaining articles, abstracts were checked independently by SO and other coauthors (SH and LM). Abstract eligibility screening excluded a further 513, however, given the wide range of sources searched, an additional 66 abstracts were added as a result of citation searching, shown as 'secondary screening' in Figure 1. Overall, a total of 364 articles were identified. The fulltext articles eligibility screening identified those not fulfilling the inclusion criteria which totalled 294, thus leaving a total of 70 articles for this systematic review ( Figure 1). Risk of bias and data synthesis Most of the data collected were from observational studies using convenience population samples because obtaining robust data from appropriate population samples is almost impossible immediately after an extreme event. Due to the heterogeneity of the data, no formal assessment of bias in each study was made. In addition, due to the wide variety of the targeted populations, statistical data synthesis was considered inappropriate.
This is the first systematic review that has addressed the topic of medication needs in disaster driven evacuation. Most of these publications did not focus on the medications needs from disaster driven evacuation thus the information was limited and often not covered comprehensively. Although the results are provided in detail below the findings are used principally to inform a detailed commentary of the assimilated results from the many publications identified.
From the MeSH terms, 70 papers were identified of which 69 articles were related to extreme events. These events showed a wide range of disasters occurring from 1992 to 2011 and included 14 hurricanes, typhoons or cyclones; eight earthquakes; two flooding; one wildfire; and one power outage and one conflict ( Table 2). There were 29 (44%) articles that reported on the impacts from the 2005 Hurricane Katrina. Additionally 54 (78%) of the disasters identified occurred in the United States of America (USA), a total of nine countries were found to have reported these extreme events. Table 3 is a summary of articles related to medication loss and interruption of care listed by evacuees, condition type and population based studies and lists chronologically the type of disaster, study method, sample and sample size for this systematic review where relevant. There are three sections within Table 3 and within each section there is an account of the outcome of each report relating to medication loss, medication unavailability and not having an adequate supply.  The remaining publications in this section related to a Californian wildfire, the Bam earthquake and flash floods in Japan. Data from California showed that patients did not have their medications with them during evacuation for one family member in 28.6% of households . Information from a cluster sample of households following the Bam earthquake reported that 25% of people in in official camps, 40% in larger rural unofficial camps and 85% in the smallest unofficial camp had no available drugs to manage their chronic medical conditions . After the flash floods in Japan in 2006, 48% of the evacuees left their medication and 88% left their prescription records behind .

Table 3B Patients with specific condition types
In this section, the focus was on people with specific medical conditions and availability of continued medication treatment following an extreme event. Eight articles focussed on hurricanes in the USA, seven were related to Hurricane Katrina and one article covered the impact of three consecutive hurricanes in Florida. Other articles in this section covered a power outage in New York, earthquakes in Japan and a cyclone in Queensland, Australia.
Following Hurricane Katrina evacuees on peritoneal dialysis were asked to bring with them approximately 1 week supply of personal medical supplies ; however, some dialysis patients evacuated from Cairns to Brisbane during cyclone Yasi arrived without personal identification, medications or medical information . experience severe side effects from not having this medication. Rath et al. reported that 19.7% of paediatric outpatients ran out of medications, while 7% of patients with hypertension did not bring their medication with them and 28% ran out of their supply .
In Florida following the three Hurricanes of Charley, Frances and Jeanne there was an increase in the number patients presenting at hospitals because of loss of power leading to lack of oxygen supplies and home haemodialysis .
During the New York blackout, within 48 hours the Emergency Department of Montefiore Medical Center reported that 56 out of 65 visits were related to respiratory device failure and that medication related problems were due to people unable to find or reach medications or were unable to obtain tests needed to adjust medications such as warfarin .
In Japan during the following the Great East Earthquake 4.2% of people ran out of oxygen before a new supply of oxygen cylinders could be delivered . Kishimoto and Noda noted that many diabetic patients had lost their medication records, and the medical team could not determine what treatments they had received before the disaster. After the MidNiigata earthquake in Japan in 2004, hospitalbased survey reported that 65% of patients with type 1 diabetes with insulin therapy were unable to continue the use of insulin pens due to medication loss immediately after the earthquake .   Table 4 is a summary of articles reporting on prescription refills, by study type and lists chronologically the disaster, type of research, number of patients and the relevant outcome from each study.
Twentythree articles described patients coming to field clinics or emergency hospitals for prescription refills. In each article listed, where available, the frequency of refill request percentage by study is documented. Details of some of the studies and their most relevant findings are summarised briefly below: After Hurricanes Katrina and Rita, 88 out of 198 evacuees (44%) at one shelter required a prescription refill ; according to a rapid needs assessment among the Katrina evacuees, medication refill was ranked third in terms of 'immediate health needs' following 'doctor' and 'dentist' .
Gavagan et al. reported on the post Hurricane Katrina evacuation complex in the Houston Astrodome. They found that chronic disease problems or medication refills prompted most adult visits to the clinic and that these refill requests were the fourth most common health and health related issue identified. Of note they showed that obtaining prescriptions for narcotic pain relief medication or refills of implanted narcotic and baclofen pumps proved difficult.
Medications are not limited to orallytaken pills -after the earthquake in El Salvador in 2001, one of the most popular services offered by the US Army Reserve hospital unit was the dispensing of 1,600 pairs of eyeglasses and the loss and breakage of eyeglasses was also identified as an issue by Sareen and Shoaf following the Northridge earthquake Following the Haiti earthquake a health care team survey showed that many chronic disease exacerbations (eg, hypertension, diabetes, congestive heart failure, asthma, and chronic obstructive pulmonary disease) were brought about by lack of medication and regular care . The most challenging part to bringing medication is ensuring a patient carries the full range of medication they need, including medical records, emergency drugs and lifesupport devices. Examples of the range of medications required following a disaster and evacuation was drawn from the reviewed articles, and is summarised in Table 5.
Medical records are critical in some situations, such as CD4 cell count and HIV RNA levels for HIV patients medication logs for cancer patients and Tb test results . Even for other patients, medical records including allergy to medications are essential. For patients with specific medical devices such as pacemakers, style and serial numbers of the devices is important . It is recommended that individuals keep a list of these essential items , which should be reviewed updated periodically and the list preferably kept it in wallet or purse, which is likely to be brought along during an emergency . Prescriptions specific to emergency situations should also be considered for each patient. For example, dialysis patients need to bring a potassiumexchanging resin, which is essential to reduce the potassium level when the access to dialysis is limited . Devices for insulin injection (vials, needles or pens with replaceable cartridges of insulin) , glucometres , fluids and devices for peritoneal dialysis, nebuliser machines, CPAP machines, oxygen cylinders, batteries for aspirators and artificial ventilators , suctioning and tube feedings, and canned nutritional supplements for the tube feedings should also be considered as part of the emergency pack for patients to bring. For the families with children who are dependent on technology and electrical devices, it seems more difficult to prepare for all the lifesupporting equipment required, such as power generator or car/vans that can be used to generate electricity .
Supportive tools for daily life, such as wheel chairs, hearing aids , canes, walkers, dentures, glasses , extra batteries for wheelchairs and other assistive devices, and incontinence briefs for the elderly are often lost at the time of evacuation. For those who cannot speak, bringing a personal identifier is also critical . Overthecounter medicines, such as medication for fever or pain, antihistamine for allergy, denture adhesive, and sanitary products are also important when access to pharmacies are disrupted . For those with hand disabilities, openers for the medications are also an essential item .
This systematic review revealed that a considerable number of patients lose their medication during evacuation. As a result, medication refill is an immediate health need, making the prescription of medications for preexisting conditions an increasing burden of medical relief activities at a time when acute needs are also overwhelming. At the current time, preparedness with respect to medications for disasters is not fully appreciated nor given much attention by those requiring daily and constant medication; meaning that a large number of patients facing extreme events could have avoided prescription interruption had they not lost their medication and or medical devices. Until individuals, with the assistance of their healthcare providers, undertake preparative actions, those organising relief activities need to be prepared to cope with emerging treatment alongside the management of chronic illnesses, including medication refills and devices.
The discussion is presented by a) the impact of medication loss and interruption of care, b) the impact of prescription refill post disaster and c) the value of effective preparation actions.
a) The impact of medication loss and interruption of care Studies from developed countries show that the impact of medication loss and interruption of care can be a significant issue. From the results above, it is possible to demonstrate that surveys targeting paediatric patients in New Orleans after Hurricane Katrina revealed that 33.9% of evacuated children with preexisting chronic conditions ran out of medication; as a result, 58.4% experienced at least one disruption to care. In a questionnairebased survey of geriatric patients visiting a hospital in Florida one year after Hurricane Wilma, 3.4%6.7% reported that they had missed medication within two weeks of the event . Some studies found that although some people brought their medication with them upon evacuation, they had only brought enough supply for a limited period . After Hurricane Ivan in 2004, 10% of households in most affected counties had problem obtaining medication one week after the storm, thus bringing sufficient supply may have be difficult . After the Great East Japan Earthquake and the following tsunamis in 2011, many evacuees had no time to gather their belongings. Some of them were treated with unique medication, such as immuno suppressants, which were not obtainable at the disaster area. As a result, these patients had to stop their medication for weeks knowing that their medical conditions would deteriorate . 49 The problem regarding medication loss is not limited to developed countries. Among the internally displaced population from the 2005 Pakistan earthquake, 85% of the households in small unofficial camps had no available drugs to manage their chronic medical conditions. After the Haiti earthquake in 2010, a relief team reported treating many women for chronic disease exacerbations brought about by lack of medication and/or regular care . Some articles reported medication by specific chronic condition, which showed that the level of preparation may vary by conditions. People with mental illness are also at high risk of medication interruption .
Examples of loss of medication lists and medical devices were identified as medical and lifesupport devices are as important as technological tools like haemodialysis and oxygen. A questionnaire conducted on the evacuees two weeks after Hurricane Katrina revealed that medical services were only the fourth most commonly reported medical need. The most common was dental care (57% of the respondents), followed by eyeglasses (34%) and dentures (28%). Other devices needed were hearing aids, canes, wheelchairs, and walkers .
b) The impact of prescription refill postdisaster The high prevalence of the need of prescriptions becomes a burden on medical teams at the disaster area which was identified in Table 4. For example, an analysis of survey data after Hurricane Katrina revealed 7.2% of the patients visiting emergency treatment facilities within 2 months of the hurricane were attending for medication refill .
Looking at the timeline of health needs following a disaster, the proportion of medication refills does not appear to change between the early and late stage of the relief activities. For example, according to a retrospective review on the patients seen by a Disaster Medical Assistance Team during four extreme events in New Mexico, US, the proportion of the patients visiting for medication refill was 6.0% within 7 days from the events and 7.6% after 7 days . Even for specialty care, medication refill often shares a significant part of relief activities. After Katrina, of 421 patients who were seen by mental health professionals, 119 (28.3%) were attending for medication refill .
Despite medication refills being a common need among the disasters studied, the proportions of patients coming to health facilities for medication refill vary within and between events. For example 20.9% in a Louisiana clinic site two weeks after the event, , within 2 months 20.8% in a temporary clinic in New Orleans , 12.6% in the mobile medical units within 3 weeks, and 48% in medical units in Chicago between 14 weeks after the event.
Additionally, a comparison between disasters suggested the burden differs depending on the scale and types of the disasters For example, Cookson et al. reported that nonsignificant increases were seen with medication refill request after the Katrina. In other cases, the proportion requiring medication refill was reported to be 3.53.6% after the Hurricane Iniki in 1992 , 6.7%10.0% after the Hurricane Andrew in 1992 , 1.33.

c) The value of effective preparation actions
Although the main scope of the review was to identify patients' reactions relating to bringing their medications during a disaster, several implications for effective preparedness were identified.
Twelve articles described possible effective preparation actions for patients. Having a personal 44  stockpile is recommended in many articles, though the recommended personal stockpile ranges from 34 days to 1 month . In a survey study conducted in California, US, the proportion of those who had a 2week supply of medication ranged from 60.1% among nonveteran women to 81.9% among veteran men . It was customary in India for pregnant women (81.1% reported) to have individual stockpile of their routine medications .
However, just keeping extra doses of medication stockpiled is likely to have limited impact. In a hospitalbased survey study targeting evacuated outpatients from the Japanese flash flood in 2005, keeping a personal stockpile did not increase the likelihood of bringing medications to the evacuated sites . On the other hand, those who had prepared an emergency pack were 5.7 times more likely to bring medications to the evacuated sites . Therefore, the researchers recommended that the stockpile is packed in a bag for easy access. Even so, the compliance for making an emergency pack seems to be low. For example, 63% of the haemodialysis patients in California had a 2week supply of medicines but only 31% stored the items in an emergency pack. In a study targeting patients with rheumatoid arthritis in Japan, 46% reported they had a personal stockpile of their medications but only 25% had packed an emergency bag . Other researchers reported that although 82.8% of households with children had stocked common medication for fever or pain for 3days, only 60.6% kept a first aid box and only 14.6% thought they would take their medications during evacuation .
Carrying medications at all time or keeping extra medication in multiple places, such as schools and offices , are described as the most robust and effective emergency plan for patients. However, low compliance with this action has been reported. Among the rheumatoid arthritis patients in Japan, only 53% of those who had a personal stockpile carried their medication all the time . Among HIV patients, 33% of patients did not have individual health cards at the time of interview, potentially making the provision of therapy to these patients difficult .
Although having a personal stockpile is recommended by the Centres for Disease Control and Prevention (CDC) and the American Red Cross , our review implies that just having a personal stockpile might not be effective . Many onsite workers recommend patients should have an emergency bag. For example, a station manager at The Hampshire Fire & Rescue Service in the United Kingdom (UK) remembered: 'after fires, many people had to go back to their home to get medications and medical devices they need.' He said that if communities can be educated in advance they will be able to prepare an emergency 'grab bag' containing vital personal items such as prescription medicines and medical devices and glasses/contact lenses, and suggests that this increases the likelihood of bring medicines/devices during a disaster .
To achieve patients' preparedness, healthcare professionals play an essential role in establishing effective emergency planning for patients.
First of all, they should provide patients with medications and other resources sufficient for disaster preparedness as well as uptodate medical records. A successful case study can be taken from the time of the election violence in Kenya , during which the dispensing of greater quantities of prescriptions were protective against treatment interruption among HIV patients.
Secondly, health professionals have a responsibility to educate patients about the potential health impacts of medication interruption, and can emphasise that bringing medication and medical records may be the only way to enable them to continue normal care in an emergency.
Thirdly, they must help patients to design an individualised and practical emergency plan that takes in to account patientspecific barriers such as forgetfulness, side effects , and allergy . In addition, they can help train patients in practices of safe storing medicines and packing necessary medical devices : the medication and medical devices in the emergency pack should be effectively protected from contamination by toxins from flood waters or mechanical damage 20 Fourthly, in disasters for which there is some advance warning, such as hurricanes, health professionals may make contact with patients when a disaster alert has been made, reminding them to bring their medication and medical records when they evacuate .
Finally, and most importantly, frontline public health workers and the members of rescue teams should have adequate medication for their own medical conditions to sustain them for the duration of their rescue efforts during a disaster .
Other stakeholders, such as policy makers and researchers, should also coordinate around patients' emergency planning (Figure 2). Recently, an increasing effort has been made by national and local governments in several countries to encourage patients to prepare an emergency bag. For example, in the UK, the National Health Service provides discharged patients with a carry bag ('green bag') for their medications to encourage patients to bring their medication in emergency situations , which is applicable to the time of evacuation. In the US, the CDC and American Red Cross raised 'gather emergency supplies' as the first step for emergency preparedness . 43  Even so, evidence is still weak with regard to preparedness actions by patients and the efficacy of intervention to encourage preparation. Researchers should be actively involved in disaster plans to leverage the preparedness among patients. For example, health impacts caused by loss of medication should be assessed using feasible and standardised methods to enable targeted aid following a disaster. Baseline data should also be measured and made available to appropriate agencies, including the burden of chronic diseases in each community , health care disparities , vulnerable populations for whom preparedness is a challenge , preparedness of general or specific groups of people as well as factors that affect emergency preparedness . Intensive research following disasters is also critical, and should include rapid health needs assessment among the evacuees, health impact assessment including medication adherence , and evaluation of the efficacy of preparedness actions to inform future planning and preparation.
The key findings and recommendations are summarised in Box 1. 87

Limitations
The most significant limitation of this study is lack of comparative data. There is no standardised way of measuring the impact of bringing medication at evacuation. Most frequently, survey is conducted on convenience samples, which makes it hard to generalise beyond that particular population. After a disaster, obtaining quality data is challenging due to the flow of evacuees and temporary nature of their status, lack of personnel, and ethical concerns accompanying the conduct of research on suffering people. Simple, unobtrusive and feasible approaches of monitoring preparedness and health outcomes should be carefully designed and established before disasters occur, especially in those regions subject to frequent disasters.
Another limitation is publication bias; most of the relevant articles were from the US, and a large proportion specifically focused on hurricane Katrina. Whether the issue of medication loss is less a problem in developing countries or simply less frequently studied and/or published is not clear. There is a clear need for evidence from all over the world, and from the most marginalised, thus rarely reported, populations.
This research does not focus on longer term crises, such as drought or political and economic failures, in which restoration of healthcare provision may take many months . In such disasters, other issues may predominate, and preparing and bringing a stockpile of medication/medical devices may not be the best solution.

Conclusion
To achieve patients' preparedness, healthcare professionals play an essential role in establishing effective emergency planning for patients should provide patients with medications and other resources sufficient for disaster preparedness as well as uptodate medical records. Health professionals have a responsibility to educate patients about the potential health impacts of medication interruption, emphasising that bringing medication and medical records may be the only way to enable them to continue normal care in an emergency. They must help patients to design an individualised and practical emergency plan that takes in to account patientspecific barriers such as forgetfulness. In disasters for which there is some advance warning, such as hurricanes, health professionals may make contact with patients reminding them to bring their medication and medical records when they evacuate. Frontline public health workers and the members of rescue teams should have adequate medication for their own medical conditions to sustain them for the duration of their rescue efforts during a disaster.
People may survive the initial disaster but if they are not educated or appropriately prepared in particular when medication is involved they may not survive the aftermath.