Living and health conditions of Palestinian refugees in an unofficial camp in the Lebanon: a cross-sectional survey

Transcription

1 Living and health onditions of Palestinian refugees in an unoffiial amp in the Lebanon: a ross-setional survey J E Zabaneh, 1 G C M Watt, 2 C A O Donnell 2 Evidene-based publi health poliy and pratie 1 Arab Resoure Colletive, Beirut, Lebanon; 2 General Pratie and Primary Care, Division of Community-based Sienes, University of Glasgow, Glasgow, UK Correspondene to: C A O Donnell, kate.o donnell@ linmed.gla.a.uk Aepted 9 Marh 2007 ABSTRACT Objetive: To determine the living onditions and selfreported health of Palestinian refugees living in an unoffiial amp in Lebanon. Design: Cross-setional survey. Setting: Gaza displaement entre, Beirut, Lebanon. Partiipants: 97 Households and 437 residents. Main outome measures: Household harateristis, inluding the number of rooms per household; aess to outside air; the presene of mould and dampness. Resident harateristis, inluding age; eduational attainment; and hroni onditions. Results: Half of the households surveyed had only one room; 44% had three or more people per room; 11% had no external ventilation; 49% had no heating; 54% had mould and dampness. The use of wood or haroal for heating was assoiated with an inrease in mould and dampness (p = 0.015). 135 Members of the population (31%) were aged under 15 years; 130 (30%) had a hroni ondition. Logisti regression results showed that overrowding (odds ratio (OR) 3.26) and a member of the household living in Gaza buildings for more than 15 years (OR 0.48) were signifiantly assoiated with hildren under 15 years. Age over 45 years (OR 5.32), a member of the household in full-time employment (OR 0.58) and a member of the household living in Gaza buildings for more than 15 years (OR 1.71) were signifiantly assoiated with hroni disease. Conlusion: This study demonstrates the poor onditions under whih Palestinian refugees in unoffiial amps live, resembling the slum housing of the United Kingdom in the last entury. In the absene of routine data olletion, researh may be the only way to obtain suh data for future publi and environmental health planning. Fored migration, partiularly as a result of war or internal onflit, has been a major publi health issue for muh of the 20th entury. 1 2 The olletion of routine data in suh situations to inform publi health responses and planning is diffiult, 3 however, and an ontinue after the initial risis is over. One suh area is the Middle East, with Palestinian refugees and their desendents displaed for almost 60 years in the West Bank, the Gaza Strip and the neighbouring ountries of Lebanon, Syria, Jordan and Egypt. To deal with this, the United Nations established the United Nations Relief and Works Ageny for Palestine Refugees in the Near East (UNRWA) in 1949, whih ontinues to arry out eduation, health, relief and soial programmes for Palestinian refugees registered with UNRWA (Box 1). 4 Those original refugees and desendents now number , approximately 10% of the Lebanese population. 5 Approximately live in one of the 12 offiial UNRWA amps. The remainder live either in Lebanese towns and villages or in what are alled unoffiial or unregistered amps or displaement entres. In addition to this population, there is an undefined number of Palestinian refugees who do not qualify for UNRWA registration beause they, or their anestors, were not displaed during the original 1948 onflit (Box 1). These individuals also live in the unoffiial amps or displaement entres, but are deprived of UNRWA servies. Whether registered with UNRWA or not, however, Palestinian refugees are marginalised in Lebanese soiety by the imposition of restritions on their ivil rights. They are urtailed in the extent to whih they an travel through restritions on the availability of visas; they are denied the right to beome Lebanese itizens; they are exluded from over 72 professions, foring them into low-paid, often unregulated work; they are denied aess to higher eduation, health servies and soial seurity; they suffer restritions on building and reonstrution work within the amps; and they are prevented from owning land in Lebanon. 6 8 A marginalised population thus ontains an even more marginalised group with almost no aess to basi servies other than those provided by non-governmental organisations (NGO), resulting in few opportunities to monitor their health and living onditions. It is one suh site that is the setting for the researh reported here. Families oupied the Gaza displaement entre during the war of the amps in , when Palestinian refugees were attaked by a Lebanese militia group. Loated in the Sabra area, surrounding the Shatila amp in Beirut, the site was formerly a hospital but was semi-destroyed in The entre is a luster of four buildings (known as Gaza 1 to 4). As an unoffiial amp, it is exluded from UNRWA ativities. Instead a loal NGO, Popular Aid for Relief and Development (PARD), provides basi medial are, health eduation and sanitation servies. Links between poor environmental onditions and ill health have been postulated, with overrowding, old, dampness, mould and poor sanitation suggested to ontribute to poor physial and mental health There are aknowledged limitations with researh in this area, inluding diffiulties with the generalisability of small-sale studies, a lak of larity around the diret (eg the physial harateristis of the housing) and indiret (eg the J Epidemiol Community Health 2008;62: doi: /jeh

2 Box 1 Classifiation of Palestinian refugees in the Lebanon 1. Resident and registered Palestinian in 1949 Inludes Palestinians and their desendents who have resided in Lebanon sine 1948 and who were registered in the Census onduted by UNRWA and the International Red Cross. These people have legal identity ards and travel douments and are eligible for all UNRWA servies. They also have reords in the Lebanese Surete Generale and the Diretorate of Arab Palestinian Refugees. 2. Resident but unregistered Palestinian in 1949 Inludes Palestinians and their desendents who have resided in Lebanon sine 1948, but were not registered in the Census. It also inludes refugees who ame to Lebanon after the 1967 war. Upon arrival in Lebanon, the Ministry of Internal Affairs issued them with identity ards and travel douments, but not being registered with UNRWA, they annot benefit from the organisation s servies. 3. Resident after 1967 Inludes Palestinians and their desendents who ame to Lebanon after the 1976 Arab Israeli war and have not been registered with the Lebanese Ministry of Internal Affairs or with UNRWA. They arry no identifiation ards or travel douments and do not benefit from UNRWA s servies. demographi and soial harateristis of the oupants) effets on health, and the lak of agreement on definitions suh as overrowding Even with suh limitations, however, there is evidene to suggest a link between poorer housing and health with long-term impats on hildren, although most of this has been onduted in the United Kingdom, Europe and Ameria. To date, there has been little work on the impat of housing on the health of refugee populations. One study, the LIPRIL survey (Living Conditions among Palestinian Refugees Living in Camps and Gatherings in Lebanon), has looked at the health and living onditions of refugees in the offiial amps, but muh less is known about the onditions faing individuals living in the unoffiial amps. Here, we report on the living onditions and self-reported health of a sample of families living within the Gaza displaement amp, with partiular referene to the impat of housing onditions on the health of hildren and on hroni disease. METHODS Sample seletion A ross-setional survey was onduted in May 2003, using a speially developed questionnaire designed to investigate the physial environment of the buildings and the health of the residents. All households in the displaement entre s four buildings (Gaza 1 to 4) were eligible for inlusion in the study. No maps existed of the interior of the buildings and the need to utilise all the spae in the buildings meant that there was no formal struture to or knowledge of the number of households on eah floor. To overome this, a soial worker employed by PARD skethed out the households on every floor of all four buildings, identifying 274. A 35% random sample of households was then generated, stratified by building and by floor. This resulted in a final sample of 97 households, with data olleted on 437 individuals living in those households. The sample size was seleted pragmatially, based on estimates of the time it would take health eduators to administer the survey and the number of households that they ould ollet data from eah day, within the timeframe of the study. Questionnaire development and administration A pilot questionnaire was developed after a review of the literature, questionnaires used by the World Health Organisation (WHO) Large Analysis and Review of European Housing and Health Status (LARES) projet ( who.int/housing/ativities/ _1) and previous work onduted in refugee amps, inluding the LIPRIL study. The pilot questionnaire was revised in Lebanon in onsultation with representatives of PARD, World Vision Lebanon (who sponsor projets onduted by PARD in the area) and PARD s health eduators. The final questionnaire overed three main areas: 1. Physial onditions in the household inluding floor area and size; aess to external ventilation, water, kithen and toilets; the presene of mould and damp; garbage olletion and the soure of energy for heating and ooking. 2. General onditions of the household inluding the number of residents; years of resideny in the buildings; soure of external finanial support; and availability of other residential plae outside Gaza. 3. Demographi and soioeonomi status and health of the residents inluding age, gender, physial disability, hroni illness, eduational attainment, employment and type of travel doument. When the questionnaire ontent was agreed, it was translated into Arabi. JZ then returned to Beirut to train three loal health eduators to administer it. These health eduators visited eah of the seleted households, ompleting the questionnaire with a member of the household, who answered on behalf of all the others in the household. The health eduator measured the size of the rooms and noted the presene of windows, mould and dampness. Analysis Questionnaire data were entered into Exel, then imported into SPSS version 10.0 for analysis (SPSS In., Chiago, Illinois, USA). Desriptive analyses were arried out using frequeny tables. Continuous variables were not normally distributed. Therefore median values were reported and omparisons analysed by Kruskall Wallis test. When appropriate, the hisquare test was used as a measure of assoiations between ategorial variable; the Mann Whitney or Kruskall Wallis tests were used for ordered ategorial variables, depending on whether the seond variable had two or more ategories. Two-by-two ontingeny tables were onstruted to explore the assoiation between hildren under 15 years of age and hroni disease with household and personal harateristis. Statistial signifiane was established using Fisher s exat test; 21 odds ratios and 95% onfidene intervals were alulated using an on-line alulator ( Signifiant independent variables (p,0.10) were entered into logisti regression models using bakwards elimination. Ethial approval This study was onduted as part of a Masters degree at the University of Glasgow. At the time the study was onduted, formal University ethial approval was not required. Ethial approval, as understood in the United Kingdom, was not required in the Lebanon. Formal approval to aess the 92 J Epidemiol Community Health 2008;62: doi: /jeh

3 Table 1 Charateristis of the households surveyed (number of households (%)) Gaza 1 Gaza 2 Gaza 3 Gaza 4 Total Total no. of households in the building No. of floors per building Mean no. of households per floor No. of households reruited into study No. of rooms within eah household Kruskall Wallis statisti , df 3, p = Room 21 (50.0) 1 (11.1) 23 (63.8) 4 (40.0) 49 (50.5) 2 Rooms 18 (42.9) 2 (22.2) 11 (30.6) 5 (50) 36 (37.1) 3 or 4 Rooms 3 (7.1) 6 (66.7) 2 (5.6) 1 (10) 12 (12.4) Aess to heating 21 (50.0) 7 (77.8) 14 (38.9) 8 (80.0) 50 (51.5) Presene of mould and dampness 25 (59.5) 6 (66.7) 13 (36.1) 8 (80.0) 52 (53.6) External ventilation 34 (81.0) 8 (88.9) 34 (94.4) 10 (8.9) 86 (88.7) Total no. of residents living in surveyed households No. of residents per household Kruskall Wallis statisti 5.219, df 3, p = or 2 Residents 14 (33.3) 0 (0) 10 (27.8) 3 (30.0) 27 (27.8) 3 5 Residents 16 (38.1) 4 (44.4) 17 (47.2) 5 (50.0) 42 (43.3) 6 or more Residents 12 (28.6) 5 (55.6) 9 (25.0) 2 (20.0) 28 (28.9) No. of people per room Kruskall Wallis statisti 2.666, df 3, p = Less than 3 23 (54.8) 7 (77.8) 19 (52.8) 5 (50.0) 54 (55.7) (14.3) 1 (11.1) 4 (11.1) 0 (0) 11 (11.3) 4 or more 13 (31.0) 1 (11.1) 13 (36.1) 5 (50.0) 32 (33.0) Floor area (m 2 ) per person x , df 3, p = (71.4) 5 (55.6) 23 (63.9) 7 (70.0) 65 (67.0) (28.6) 4 (44.4) 13 (36.1) 3 (30.0) 32 (33.0) Aess to water (no. (%)) Pipes as part of the building 0 (0) 5 (55.6) 0 (0) 0 (0) 5 (5.2) Individually established pipes 23 (54.8) 0 (0) 31 (86.1) 0 (0) 54 (55.7) Shared kithens and toilets 2 (4.7) 0 (0) 0 (0) 1 (10.0) 3 (3.1) No pipes 17 (40.5) 4 (44.4) 5 (13.9) 9 (90.0) 35 (36.1) Aess to kithen and toilet failities (no. (%)) Separate kithen within household 6 (14.3) 7 (77.8) 22 (61.1) 7 (70.0) 42 (43.3) Separate toilet within household 8 (19.0) 9 (100.0) 35 (97.2) 10 (100.0) 62 (63.9) Shower room within toilet in household 7 (16.7) 9 (100.0) 35 (97.2) 10 (100.0) 61 (62.9) Separate shower room in household 2 (4.8) 0 (0) 1 (2.8) 0 (0) 3 (3.1) Shared showers and toilets 25 (59.5) 0 (0) 0 (0) 0 (0) 25 (25.8) Showering within the room spae of the household 8 (19.0) 0 (0) 0 (0) 0 (0) 8 (8.2) Table 2 Charateristis of the residents (number of residents (%)) Gaza 1 Gaza 2 Gaza 3 Gaza 4 Total Gender (n = 181) (n = 53) (n = 159) (n = 44) (n = 437) Male 84 (46.4) 26 (49.1) 84 (52.8) 23 (52.3) 217 (49.7) Female 97 (53.6) 27 (50.9) 75 (47.2) 21 (47.7) 220 (50.3) Age (years) Kruskall Wallis statisti 3.899, df 3, p = (27.8) 15 (28.3) 53 (33.5) 17 (38.6) 135 (31.0) (53.3) 31 (58.5) 84 (53.2) 22 (50.0) 233 (53.6) (18.9) 7 (13.2) 21 (13.3) 5 (11.4) 67 (15.4) Citizenship Palestinian refugee living in Lebanon 126 (69.6) 51 (96.2) 111 (69.8) 35 (79.5) 323 (73.9) Palestinian refugee from Syria, Jordan or Egypt 12 (6.6) 1 (2.7) 9 (5.7) 0 (0) 22 (5.0) Lebanese itizen 21 (11.6) 0 (0) 12 (7.5) 4 (9.1) 37 (8.5) Syrian itizen 22 (12.2) 0 (0) 20 (12.6) 5 (11.4) 47 (10.8) Other 0 (0) 1 (1.9) 7 (4.4) 0 (0) 8 (1.9) Chroni onditions (n = 58) (n = 17) (n = 44) (n = 11) (n = 130) Cardiovasular disease 16 (27.6) 4 (23.5) 7 (15.9) 2 (18.2) 29 (22.3) Respiratory 11 (19.0) 4 (23.5) 9 (20.5) 1 (9.1) 25 (19.2) Musuloskeletal 10 (17.2) 1 (5.9) 7 (15.9) 3 (27.3) 21 (16.2) Diabetes 4 (6.9) 3 (17.7) 3 (6.8) 0 (0) 10 (7.7) Gastrointestinal 4 (6.9) 1 (5.9) 5 (11.4) 0 (0) 10 (7.7) Nervous system inluding epilepsy 4 (6.9) 0 (0) 3 (6.8) 0 (0) 7 (5.4) Renal 2 (3.4) 1 (5.9) 3 (6.8) 1 (9.1) 7 (5.4) Others 7 (12.1) 3 (17.7) 7 (15.9) 4 (36.4) 21 (16.1) J Epidemiol Community Health 2008;62: doi: /jeh

4 population was, however, obtained through PARD, the loal NGO. RESULTS There were 274 households in total. Residents in 67 of the 97 randomly sampled households agreed to partiipate (initial response rate 69%). The remaining 30 household units were either vaant (eight) or the residents were long-term absentees (22). In these ases, the preeding household was approahed. In this way, 97 households were reruited into the study. The number of households in eah building, and thus reruited into the study, varied (table 1). In most ases, the household respondents were adult women. Household harateristis There were marked differenes in the living onditions of the four buildings (table 1). Conditions were most rowded in Gaza 1 and inluded eight homes onstruted in the bakyard (fig 1). Half of the households (51%) surveyed had only one room (table 1). The median floor area of households differed signifiantly between the four buildings (Gaza 1, 18.1 m 2, interquartile range (IQR) 14.0 to 28.5; Gaza 2, 43.3 m 2, IQR 37.1 to 50.5; Gaza 3, 25.0 m 2, IQR 18.0 to 33.6; Gaza 4, 18.8 m 2, IQR 12.8 to 37.7; Kruskall Wallis statisti , df 3, p,0.0001). As the largest family groups were housed in Gaza 2, however, the median floor area per person was not signifiantly different between the four buildings (Gaza 1, 6.1 m 2, IQR 3.7 to 8.8; Gaza 2, 7.7 m 2, IQR 5.6 to 8.8; Gaza 3, 6.0 m 2, IQR 4.2 to 12.4; Gaza 4, 5.0 m 2, IQR 2.3 to 16.7; Kruskall Wallis statisti 2.301, df 3, p = 0.524). It was apparent, however, that the amount of spae was limited, with 43 households (44%) overall meeting the LIPRIL definition of overrowding and 27 (28%) meeting the WHO (Europe) definition 10 (table 1). There was a lak of aess to heating failities with half of all households (47; 49%) relying instead on blankets and lothing. Of the 50 households reporting a soure of heating, the most ommon was haroal or wood (18; 36%); eletriity (15; 30%); diesel/kerosene (nine; 18%); and gas/methane (eight; 16%). This again varied between buildings (table 1). Fifty-two households Figure 1 metal. Bakyard with households onstruted out of orrugated (54%) had mould and dampness; 11 (11%) reported having no window or door giving aess to outside air and light. The presene of heating was assoiated with mould and dampness (presene of mould and dampness in heated houses 33/50 (66%) versus presene of mould and dampness in unheated houses 19/47 (40%), p = 0.015, Fisher s exat test). Houses with external ventilation were less likely to have mould and dampness ompared with those with no external ventilation, although this was not statistially signifiant (presene of mould and dampness in houses with external ventilation 43/86 (50%) versus presene of mould and dampness in houses with no ventilation nine/11 (82%), p = 0.058, Fisher s exat test). Aess to water for domesti use varied, with over half (56%) only able to aess water through piping networks that they had onstruted (table 1). A further third had no aess to piped water of any sort. All respondents purhased drinking water, either ommerially or from the Lebanese muniipality (3%). Almost all households had their own toilets, with the exeption of Gaza 1 where only 19% did. Aess to personal kithen and shower failities was again poorest in Gaza 1 (table 1). Residents harateristis Of the 97 households, 40 (41%) reported household members living in the Gaza buildings for more than 15 years. Only 19 (20%) had lived there for less than five years. Most households (86; 89%) had no other plae of residene and 77 (79%) had no external finanial aid from sponsors or harity. Data were olleted on 437 people living in the 97 households (table 2). The population was young, with only 67 residents (15%) over 45 years of age. Eduational attainment was generally low, with 99 men (63%) and 103 women (57%) residents having left shool before seondary level (ie before the age of 16 years); 31 men (20%) and 43 women (24%) had never attended shool. Only 20 boys (13%) and 27 girls (15%) were reported to be urrently in fulltime eduation (ie between the ages of six and 18 years). There were no signifiant differenes in population demographis between the four buildings. Nineteen residents (4%) were reported to have a physial disability; 130 (30%) were reported to have a hroni ondition (table 2); 119 (27%) were taking mediation for a hroni ondition. Univariate analysis demonstrated that hildren under the age of 15 years were more likely to live in overrowded households, whether defined aording to the LIPRIL (odds ratio (OR) 3.43) or the WHO Europe (OR 3.24) definitions (table 3). Heating (OR 0.69), external ventilation (OR 0.82) and mould and dampness (OR 1.31) were not signifiantly assoiated with hildren (table 3). Children were less likely to suffer from a hroni disease themselves (OR 0.54), but were no more or less likely to live in a household where someone else had a hroni disease (OR 1.06). Children were signifiantly less likely to live in a household where someone had been resident in the Gaza buildings for over 15 years (OR 0.44). Signifiant univariate variables were entered into a logisti regression model. The LIPRIL definition of overrowding was used, as it was the most appropriate for the population under study. The final model (table 4) inluded overrowding (OR 3.26) and whether a member of the household had lived in the Gaza buildings for more than 15 years (OR 0.48). Individuals with hroni disease were less likely to live in overrowded housing. This differene was statistially signifiant when overrowding was defined using the LIPRIL definition (OR 0.61), but was not signifiant when defined by 94 J Epidemiol Community Health 2008;62: doi: /jeh

5 Table 3 Assoiation of household and personal harateristis with hildren under 15 years of age and with hroni disease Age Under 15 years (n = 135) N (%) Over 15 years (n = 300) N (%) Odds ratio (95% CI) p Value* Overrowding: three or more per room{ 107 (79.3) 158 (52.7) 3.43 (2.14 to 5.52), Overrowding:,8 m 2 per person{ 123 (91.1) 228 (76.0) 3.24 (1.69 to 6.2), Heating in the household 60 (44.4) 161 (53.7) 0.69 (0.46 to 1.04) External ventilation in the household 121 (89.6) 274 (91.3) 0.82 (0.41 to 1.63) Mould and dampness in the household 81 (60.0) 160 (53.3) 1.31 (0.87 to 1.98) Travel douments: Palestinian refugee 107 (79.3) 237 (79.0) 1.02 (0.62 to 1.68) Member of household in full-time employment 71 (52.6) 165 (55.0) 0.91 (0.60 to 1.36) Member of household eduated to seondary shool level or above 40 (29.6) 94 (31.3) 0.93 (0.59 to 1.44) Suffering from hroni disease themselves 29 (21.5) 101 (33.8) 0.54 (0.33 to 0.86) Member of household with hroni disease 110 (81.5) 242 (80.7) 1.06 (0.63 to 1.77) Member of household living in Gaza buildings for.15 years 36 (26.7) 136 (45.3) 0.44 (0.28 to 0.68), Chroni disease Yes (n = 130) N (%) No (n = 306) N (%) Odds ratio (95% CI) p Value* Overrowding: three or more per room{ 68 (52.3) 197 (64.4) 0.61 (0.40 to 0.92) Overrowding:,8 m 2 per person{ 99 (76.2) 253 (82.7) 0.67 (0.41 to 1.10) Heating in the household 71 (54.6) 150 (49.0) 1.25 (0.83 to 1.89) External ventilation in the household 120 (92.3) 276 (90.2) 1.30 (0.62 to 2.75) Mould and dampness in the household 71 (54.6) 170 (55.6) 0.96 (0.64 to 1.45) Travel douments: Palestinian refugee 107 (82.3) 237 (77.5) 1.35 (0.80 to 2.29) Member of household in full-time employment 58 (44.6) 179 (58.5) 0.57 (0.38 to 0.87) Aged over 45 years 42 (32.3) 24 (7.8) 5.61 (3.22 to 9.78), Member of household eduated to seondary shool level or above 31 (23.8) 104 (34.0) 0.61 (0.38 to 0.97) Member of household living in Gaza buildings for.15 years 67 (51.5) 105 (34.3) 2.04 (1.34 to 3.09) CI, Confidene interval. *Fisher s exat test. {LIPRIL definition. {WHO (Europe) definition. the WHO Europe guidelines (OR 0.67; table 3). Other household harateristis were not assoiated with the presene or absene of hroni disease. Chroni disease was assoiated with age (OR 5.61; table 3). Individuals living in a house where at least one person had been eduated to seondary shool level or above (OR 0.61) or where someone was in full-time employment (OR 0.57) were less likely to suffer from hroni disease. Again, signifiant univariate variables were entered into a logisti regression model, as was an interation term for age and the number of years living in the Gaza buildings. The final model (table 4) inluded age over 45 years (OR 5.32), whether a member of the household was in full-time employment (OR 0.58) and whether a member of the household had lived in the Gaza buildings for more than 15 years (OR 1.71). DISCUSSION Conduting health needs assessments among refugee and displaed ommunities is diffiult. 2 3 This is ompounded, in this situation, by the almost non-existent offiial status of this population who, for historial reasons, do not ome under the auspies of UNRWA, are not reognised by the Lebanese government and rely on NGO for support. The study demonstrated that the onditions under whih this long-term refugee population were living were omparable with those found in the slum housing of the United Kingdom in the last entury. 22 Children under 15 years of age were more likely to be living in overrowded onditions and, although not statistially signifiant, in housing that was unheated. Multivariate analysis retained the assoiation with Table 4 Assoiation of household and personal harateristis with hildren under 15 years of age and with hroni disease: logisti regression results Odds ratio (95% CI) p Value Dependent variable: hildren under 15 years Overrowding: three or more per room* 3.26 (2.01 to 5.27), Suffering from hroni disease 0.64 (0.39 to 1.05) Member of household living in Gaza buildings for.15 years 0.48 (0.30 to 0.76) Dependent variable: suffering from hroni disease Aged over 45 years 5.32 (2.99 to 9.47), Member of household in full-time employment 0.58 (0.37 to 0.90) Member of household eduated to seondary shool level or above 0.64 (0.38 to 1.05) Member of household living in Gaza buildings for.15 years 1.71 (1.09 to 2.67) CI, Confidene interval. *LIPRIL definition. J Epidemiol Community Health 2008;62: doi: /jeh

6 overrowding. Children were also part of a more reent wave of influx into the Gaza buildings, and may indiate that suh families end up living in poorer aommodation. As ould be expeted, individuals suffering from hroni diseases were older and were in households that had been part of the Gaza buildings for longer. Chroni disease also appeared to be assoiated with improved soial onditions, as those in full-time employment or eduated to seondary shool level or above were less likely to suffer from hroni disease. There were limitations to this study. The first was the lak of onsistent definitions, partiularly for overrowding. 16 By either definition used, however, a signifiant number of the households surveyed were overrowded. The seond was our inability to distinguish between the effets of poor housing on health and the possible onfounding effets of lower inome and lak of employment opportunities. Again, this is a reognised limitation of researh in this area The final limitation was the lak of ounterfatual, or ontrol, data. 23 Colletion of suh data was not possible within the parameters of this study and suh data are not routinely available. Comparison of the findings reported here with those of the LIPRIL study offers useful insights. For example, the LIPRIL report defined overrowded housing as that whih had three or more people to a room. Using this riterion, 44% of households in the Gaza buildings were overrowded ompared with 27% in offiial amps and gatherings in the Lebanon. The LIPRIL report also found that dwellings in the offiial amps had, on average, three rooms, exluding the kithen, toilet and hallways Here, half of the dwellings had just one room used for ooking, eating, sleeping and soialising. The median floor area per household of 24.8 m 2 was signifiantly less than the WHO minimum reommendation of 70 m 2 for a family of three to five. 10 There were also differenes between the four buildings. The LIPRIL survey indiated that over 90% of households had separate kithen and toilet failities Although the Gaza 2, 3 and 4 buildings were similar, Gaza 1 had muh poorer aess to suh failities. Many of the households also had mould and dampness, poor external ventilation and a lak of heating. All of these fators may ontribute to poor health, partiularly respiratory disease, with evidene pointing to old housing as being partiularly assoiated with poor health. Up to one third of the population did report some type of hroni ondition, inluding respiratory problems, although these were self-reported. This was higher than those reported for offiial amps, where the figure was 19% There was, however, a lak of statistially signifiant assoiations between the presene of a hroni ondition and these markers of poor housing. This may be aused by a lak of power in the sample size or the timing of the survey, onduted in early summer. The available evidene on the relationship between old and damp housing, however, omes from northern Europe and it may be that the impat of old housing is less in a warm limate suh as the Middle East. One weakness of the study was the omission of any questions regarding residents mental health. Poor housing impats on individuals mental health and an ontribute to depression In addition, given the overall situation in whih these people find themselves, it is likely that mental health issues would be a major omponent of their overall ill-health, as reently reported by Karam et al 25 for the Lebanese population as a whole. The study was unable to survey all of the households in the Gaza building, beause of time and resoure onstraints. It does demonstrate, however, that with the right support and knowledge, aess an be negotiated and that members of the What is already known on this subjet Gathering health and environmental data on refugee and displaed populations for publi health purposes is often diffiult One suh situation is that of the Palestinian refugees living in the Lebanon, of whih are housed in offiial UNRWA amps Although there is some information on the living onditions of those refugees, little is known about the living onditions of those in unoffiial amps What this study adds Palestinians living in an unoffiial displaement entre experiene housing onditions similar to those of the slum housing in the United Kingdom in the last entury In the absene of routine systems to ollet suh data, researh may be the only way to ollet robust data for future planning Poliy impliations Refugees and internally displaed individuals are frequently missed from the olletion of data about environmental and health onditions. This study shows that suh data olletion is possible and that the findings an be used to inform future planning and servie provision. ommunity are willing to partiipate in suh a survey. This was learly failitated by the role and knowledge of JZ, who is herself a Palestinian refugee and had previously worked for PARD as oordinator of their health linis, giving the study important insight, loal knowledge and redibility. The data olleted in this survey are of the type routinely available in western ountries, suh as ensus-type data on population denominators, health and environmental onditions. In the absene of suh soial infrastrutures, however, suh data may only be obtained through researh. Indeed, the findings from this study have already been used by the NGO to seek further funding for infrastruture improvements and to raise awareness of the onditions within suh amps. This study demonstrates, for the first time, the very poor onditions that Palestinian refugees in unoffiial amps are living under. Whereas those in the offiial UNRWA amps also experiene poor living onditions and ill-health, these unoffiial refugees have even poorer onditions, with severe overrowding, lak of external ventilation and lak of aess to basi needs suh as kithens and toilets the norm. This study also demonstrates that with the right support and knowledge it is possible to ollet evidene in a rigorous and robust manner, leading to opportunities to inform future poliy and planning within the Gaza buildings. Aknowledgements: In Beirut, Lebanon, the authors would like to thank the administrative and management boards of PARD for their support of this projet; Ms Rita Hamdan and Mr Ahmad Halimeh for providing the politial approvals and offie failities; Ms Inaam Khaled, Ms Intissa Ibrahim and Ms Maryam El Khatib for 96 J Epidemiol Community Health 2008;62: doi: /jeh

7 onduting the questionnaire data olletion; Mr Walid Taha for produing the maps of the Gaza buildings; Mr Bernard Hillenkamp for assisting with the development of the questionnaire; and PARD staff for their help and enouragement. The authors would also like to thank Mr Harper Gilmour, Publi Health and Health Poliy, University of Glasgow, for statistial support and advie, partiularly with the sampling strategy. Finally, the authors would like to thank the peer reviewers for their extensive and thoughtful omments that led to a substantial revision of this paper. Funding: This work was onduted as part of a Masters in Publi Health degree at the University of Glasgow. This was funded by the Karim Rida Said Foundation (London). Additional support was provided by World Vision Lebanon. Competing interests: JZ was previously employed by PARD and worked as a health o-ordinator in the Gaza displaement entre. Three are no other ompeting interests. REFERENCES 1. MKee M, Janson S. Fored migration. The need for a publi health response. Eur J Pub Health 2001;11: Salama P, Spiegel P, Brennan R. No less vulnerable: the internally displaed in humanitarian emergenies. Lanet 2001;357: Banatvala N, Zwi AB. Publi health and humanitarian interventions: developing the evidene base. BMJ 2000;321: UNRWA. Establishment of UNRWA (last aessed 3 Marh 2007). 5. UNRWA. Lebanon refugee amp profiles lebanon.html (last aessed 3 Marh 2007). 6. Shiblak A. Resideny status and ivil rights of Palestinian refugees in Arab ountries. J Palestine Stud 1996;25: Al-Natour S. The legal status of Palestinians in Lebanon. J Refugee Stud 1997;10: Haddad S. Palestinians in Lebanon: towards integration or onflit? haddad.html#the%20palestinian%20community%20in%20lebanon (last aessed 5 February 2007.) 9. Lowry S. Housing. BMJ 1991;303: World Health Organization. Guidelines for healthy housing. Environmental Health Series. Copenhagen: Regional Offie for Europe, Evans J, Hyndman S, Stewart-Brown S, et al. An epidemiologial study of the relative importane of damp housing in relation to adult health. J Epidemiol Community Health 2000;54: Gemmell I. Indoor heating, house onditions, and health. J Epidemiol Community Health 2001;55: Makenbah JP, Howden-Chapman P. Houses, neighbourhoods and health. Eur J Pub Health 2002;12: Thomson H, Pettirew M, Morrison D. Health effets of housing improvement: systemati review of intervention studies. BMJ 2001;323: Dedman DJ, Gunnell D, Davey Smith G, et al. Childhood housing onditions and later mortality in the Boyd Orr ohort. J Epidemiol Community Health 2001;55: Offie of the Deputy Prime Minister. The impat of overrowding on health and eduation: a review of evidene and literature. London: Offie of the Deputy Prime Minister, Thomson H, Pettirew M, Morrison D. Housing improvement and health gain: a summary and systemati review. Oasional paper no. 5. Glasgow: MRC Soial and Publi Health Sienes Unit, Al-Madi Y, Bashour N, Jaobsen LB, et al. Diffiult past, unertain future. Living onditions among Palestinian refugees in amps and gatherings in Lebanon. In: Ugland OF, ed. Oslo: FAFO Institute for Applied Soial Siene, 2003 p Tiltnes ÅA. Falling behind. A brief on the living onditions of Palestinian refugges in Lebanon. Oslo: FAFO Institute for Applied Soial Siene, 2005 p Velupillai Y. Health needs assessment of Palestinians living in the refugee amps in Lebanon. Glasgow: University of Glasgow, Altman DG. Pratial statistis for medial researh. London: Chapman and Hall, O Donnell CA, Watt GCM, Zabaneh JE. Children, housing and health: from Glasgow slums to displaed persons. J Epidemiol Community Health 2004;58: Purdon S, Lessof C, Woodfield K, et al. Researh methods for poliy evaluation. No 2. London: National Centre for Soial Researh Department for Work and Pensions Researh Working Paper, Olsen NDL. Presribing warmer, healthier homes. BMJ 2001;322: Karam EG, Mneimneh ZN, Karam AN, et al. Prevalene and treatment of mental disorders in Lebanon: a national epidemiologial survey. Lanet 2006;367: J Epidemiol Community Health 2008;62: doi: /jeh