Deprivation and mental health

While there is consensus on the negative impacts of deprivation on mental health [11–14], the mechanisms by which deprivation impacts mental health as well as in the conceptualization of deprivation itself are heavily debated. Much of the research evidence conceptualises deprivation in economic terms, focusing on income-deprivation or other forms of material deprivation (eg: asset-deprivation) [9, 11]. Such measures of deprivation largely focus on individuals or households and do not take into account broader conditions and living standards, which may capture a fuller range of the effects of deprivation on health.

Sen [9, 10] shows the limitations of focusing on poverty or deprivation solely in terms of income-deprivation, which ignores the fact that individual characteristics (e.g., disability) and contextual characteristics (e.g., conflict) may interact to influence health. Alternatively, he proposes conceptualizing poverty as capability deprivation. Sen’s conceptualization [10] provides a useful starting point and links to a growing body of public health evidence in which relative measures of deprivation are taken into account in examining the relationship between deprivation and health [11]; here deprivation is viewed more broadly in terms of its effect on increasing vulnerabilities or reducing adaptive capacities [15, 16], which in turn can reduce opportunity and adversely affect health and wellbeing. These literatures highlight the importance of rethinking relative deprivation in terms of either social comparisons or the social distance that is created or expounded by inequalities [11, 14], which then negatively impact health through various mechanisms, including limiting access to health and social services and reducing psychosocial wellbeing [11, 12, 17]. The evidence illustrates that relative deprivation can predict poor mental health [18], depressive symptoms [12], psychological stress and poor self-rated health [19]. Similarly, previous studies noted the importance of the subjective assessment of relative deprivation, emphasizing how these subjective ratings are more strongly associated with poor physical and mental health [11], and depressive symptoms [12] than objective ratings.

Deprivation and mental health in conflict settings

Expounding on the relationship between conflict and mental health, an increasing body of literature has pointed to the ways in which conflict works to affect what Miller and Rasmussen [3, 4] term ‘daily stressors’, that in turn affect health. These daily stressors are “the stressful social and material conditions of everyday life that are common within settings of organised violence” (4, p.33). These conditions include poverty, social marginalization, inadequate living conditions, and changes in family structure and social functioning [3], which in turn adversely affect health and wellbeing. In a review of studies that drew on their initial framework to examine the effects of daily stressors on mental health in war and conflict contexts, Miller and Rasmussen [4] note that “much of the distress observed among war affected populations may, in fact, not be due to exposure to political violence per se, but to its ongoing impact on multiple domains of people’s lives” (p.34). Subsequent studies have incorporated ‘daily stressors’–which include material conditions, political conditions, and lack of freedom into their frameworks or approaches, and show that political violence stemming from conflict affect people’s daily life conditions and adversely impact mental health [20].

These daily stressors, while not necessarily framed in terms of deprivation, contain considerable overlap with conceptualizations of deprivation, especially in terms of poverty and living conditions. The literature discussing the relation between deprivation and mental health is sparse in low-income settings and particularly in conflict-affected areas [21, 22]. Two studies discuss the relation between deprivation and health outcomes among Palestinians living in refugee camps in Lebanon. One study highlights that material deprivation based on objective measures can predict poor self-rated health among Palestinian refugees [22], while the other study emphasizes that relative deprivation, and not the absolute measures of socioeconomic status, is a strong predictor of both self-rated health and self-reported chronic diseases among Palestinian refugee women [21]. Likewise, evidence from a study conducted with Syrian refugees living in Jordan shows that experiencing employment deprivation can negatively impact their health [23]. Researchers in Colombia [2] also demonstrate that structural violence is linked with economic deprivation, inequalities, and insufficient job opportunities. Furthermore, they point to inequalities in exposure to violence and their impacts during conflict, can have far-reaching and intergenerational impacts on living conditions and mental health.

This growing literature provides important insights into how various forms of stressors or deprivations can be exacerbated in conflict settings and how multiple forms of deprivation can overlap in people’s lives and interact with other disadvantages, with important implications for their health and wellbeing. These studies point to a widening scope in how deprivation is conceptualized, which has been echoed in more recent trends in multidimensional poverty and deprivation measurement [24–26]. Furthermore, they point to the importance of not only including a broader range of dimensions of deprivation spanning from the political to the economic, but also using measures that are meaningful to people and the contexts in which they live [2, 4, 11, 17, 20, 27–29]. As Sen reminds us, “human lives are battered and diminished in all kinds of different ways, and the first task in this perspective, is to acknowledge that deprivations of very different kinds have to be accommodated within a general overarching framework.” And that this framework “…must not try to overlook the pluralities that are crucially involved…(9 p.19)”.

The aim of this study is to contribute to this growing literature by examining the effects of various forms of deprivation on mental health within the context of the oPt, where protracted conflict has shaped lives, health, and livelihoods [1, 5, 30, 31]. We further develop the conceptualization of deprivation and operationalize the expanded conceptualization of deprivation. In order to cover as broad and contextually relevant a range as possible, from available data, we focus on four key dimensions of deprivation: subjective deprivation, material (economic) deprivation, food deprivation (insecurity), and political deprivation. Furthermore, we outline key areas of deprivation that are likely to be of relevance to other contexts, particularly conflict settings, and contribute to the further development of multidimensional measures of deprivation. The findings from this study will provide quantitative evidence to inform a broader conceptualization of the effects of deprivation on mental health.

Study context

The West Bank (WB) and Gaza Strip (GS) are part of the oPt. While both regions have been experiencing severe and protracted political violence by the Israeli occupation, this violence has intensified in the GS since 2007 [5, 31, 32]. Israel implemented a siege on the Strip that restricts the movement of Palestinian living there in and out of GS, and continues to prohibit the entrance of many goods [32]. The continuous blockade on GS has negatively affected the economy and worsened the living conditions the GS [31, 32]. The unemployment rate in the GS reached 45% compared to 17% in the WB, 76.9% of the households in GS received some kind of assistance compared to 16.7% in the WB, 54.2% of the households in GS perceived themselves as poor compared to 13.6% in the WB (PCBS, 2020). Moreover, the healthcare system in the GS continues to suffer stunting, including the insufficiency of equipment and medications as a result of the Israeli siege [33].

Data

In our analysis, we rely on secondary data from the 2014 Socio-economic and Food Security survey (SEFSec) conducted by the Palestinian Central Bureau of Statistics (PCBS) in cooperation with the Food Security Sector (FSS). The survey sample consisted of 8,177 households in the WB and the GS. Our final analytic sample consisted of 7,723 households with complete information on all the key variables of interest. The sample is representative at the governorate and type of locality levels (i.e. urban, rural, and camp localities), which are the second and third level administrative units in the oPt [32]. The survey instrument consisted of three questionnaires: the household roster, the household questionnaire, and the individual questionnaire. The household roster includes basic demographic information for each member of the household, including age, sex, educational status, and relationship to the labor force. The household questionnaire includes detailed questions on the conditions of the household, some questions pertaining to exposure to political violence for the GS sample, questions on shocks faced by the household within six-month period prior to the survey, and questions pertaining to food insecurity. The individual questionnaire was completed by one randomly selected adult 18 years or older from each of the households and includes questions pertaining to mental health, subjective health, human insecurity, and distress.

Ethical considerations

The data that we use in this study is secondary data that we obtained in anonymized form from the Palestinian Central Bureau of Statistics (PCBS). Due to the secondary and anonymized nature of the data, this was deemed to be exempt from ethical review by the ethical review committee at the Institute of Community and Public Health at Birzeit University.

Analytical approach

Deprivation measures

Shocks

In addition to the deprivation measures, we also account for more acute shocks in our analysis. These acute shocks represent shocks experienced by households in the six months preceding the survey, and include economic, political, and health shocks as outlined below.

Limitations

This study utilizes data and empirical strategies, which permit the examination and expansion of deprivation as a concept and a quantitative measure, but there are several limitations to this study. Due to data availability at the time of analysis, this study only includes a cross-sectional survey dataset with one time-period, which prohibits the evaluation of causality. Our models highlight the association between our variables of interest but do not provide causal inference. There is likely an endogenous relationship between subjective deprivation and poor mental health, where people with poorer mental health are more likely to indicate subjective deprivation and vice versa. Nevertheless, demonstrating the association between our variables of interest is the first step to understanding how these variables may affect each other. The dataset is based on self-reported outcomes; however, the questions and choice responses are derived from validated instruments that have been adopted globally. Lastly, our conceptualization and operationalization of deprivation as a concept should be viewed as only a starting point to developing comprehensive measures and capture the intricate relationship between deprivation and mental health, particularly in conflict and war settings.

Multivariate models

We estimated the single-level model followed by the random intercept model with the random intercept at locality level. The log likelihood test ratio showed that in all four models the model fit improved where the LR test was significant at the p<0.001 level across all models when the random intercept was included. The multilevel models (Table 2) show a significant variance at locality level across all models. The greatest variance is at locality level across all models. More specifically the level of variance explained by locality is 7.5% in the combined model, 11.4% in the WB model, 3.1% in the GS model and 2.1% in the GS political model. This result is not surprising given the relative homogeneity of the GS compared with WB which is characterized by greater geographic spread and a regime of fragmentation through checkpoints, illegal settlement blocks, and variable access to services.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Models.

https://doi.org/10.1371/journal.pgph.0001239.t002

Overall, in all the multivariate models and analyses the variable that has the most impact on GHQ12 score (poor mental health) is subjective deprivation. People who reported some deprivation have scores that are about 1.69 points higher compared with those who have not reported any deprivation; among people who are very deprived, the score increases by 4.23 points. This pattern is statistically (p<0.001) and substantively significant in terms of magnitude. In terms of the other forms of deprivation, material deprivation is the second most important predictor in the overall model and in the WB model (this would include subjective economic status, wealth quartile, and economic shocks), then the political, and finally food deprivation (insecurity and consumption). In terms of the political, experiences of political shocks and greater human insecurity are associated with worse mental health (GHQ12 increases by 0.52 per political shock and 0.043 per point on human insecurity scale, both significant at p<0.001). Women have higher scores than men, older people have higher scores than younger people, and people with post-secondary education have lower scores than people with secondary education. People who reported the health stressor (one item consisting of having an illness that limits function) also have higher scores on the GHQ12 across models.

Combined model

The first set of analysis included combined the samples for WB and GS (GS) as shown in Table 2. Subjective deprivation has the largest effect on GHQ12 score. The results indicate that people who have some deprivation have scores that are about 1.69 points higher compared with those who don’t have any deprivation (p<0.01). The score among people who are very deprived is 4.23 points higher compared with not deprived (p<0.01). Material deprivation consisting of subjective economic status classified as rich, middle, poor or very poor; wealth quartile variable (poorest quartile, second quartile, third quartile and the richest quartile); and economic shocks are the second set of determinants of poor mental health. People who perceived themselves as very poor have about 2.29 points higher GHQ12 score compared to people who identified as middle class (p<0.001). People who perceived themselves as poor have an increase in the score by 0.65 points (p<0.001) also in comparison with middle wealth people, and those who reported themselves as rich have 0.66 (p<0.01) lower score of GHQ12. The scores of GHQ12 for poorest quartile, second quartile, and third quartile increase by 0.76, 0.48 and 0.49 (p<0.01) respectively compared to the richest quartile. People suffering from economic shocks have an increase in GHQ12 score by 0.23 points (p<0.001) per shock. Food insecurity has a significant positive association with GHQ12 score (β = 0.187, p<0.001), while more food consumption is associated negatively with GHQ12 score (β = -013, p<0.001). In other words, food insecurity is associated with poorer mental health, while greater food consumption is associated with better mental health.

People who have experienced political shocks have worse mental health, where each experience is associated with about an 0.52 point increase in GHQ12-score (β = 0.52, p<0.001). Similarly, greater human insecurity slightly increases GHQ12-score by 0.043 points (p<0.001). The health stressor (one item which is having an illness that limits function), is another predicting factor for mental health with those who reported having the health stressor have 1.36 points higher GHQ12 score compared to those who did not (p<0.001). Age is positively associated GHQ12 score, where each year of age is associated with about a 0.1 increase in score (β = 0.089, p<0.001), i.e. older age is associated with poorer mental health. Women have higher scores than men (β = 0.25, p<0.001), and people who have post-secondary education have significantly lower scores than people with secondary education (β = -0.57, p<0.001). People who reported having one household member or more employed have lower GHQ12 scores (β = -0.38 p<0.001) compared to those who do not have employed members.

Separate models of the WB and GS

As in the combined model, subjective deprivation is the strongest predictor of GHQ12 score in the models for the WB and GS, with similar coefficient sizes. People who feel very deprived have GHQ12 scores that are almost 4 points higher than people who do not feel deprived (β = 3.94 in the WB (model-2) and 3.63 in the GS (model-4), p<0.001). Upon closer examination of the various dimensions of deprivation, we find that, in general, experiences of deprivation have an adverse effect on mental health, with some variation between the WB and GS in terms of the size of effect on GHQ12 score.

Health stressor and sociodemographic determinants

People who reported that they have an illness that limits function have GHQ12 scores that are 1.09 and 1.38 points higher than those who did not report an illness that limits function in the WB and GS respectively (p<0.001). In terms of other sociodemographic characteristics, age is positively associated with poor mental health in both settings, although with a stronger effect in the GS (β = 0.043 in the WB and β = 0.17 in the GS, p<0.001). In both settings, post-secondary education was a protective factor against poorer mental health (β = -0.56 in WB, and β = -0.75 in the GS, p<0.001). In the WB, women had scores 0.479 points higher than men (p<0.001). Household size was positively associated with GHQ12-scores in the GS (β = 0.058 in the GS, p<0.05).

The key differences in the WB and GS appear to be in the magnitude of the effect of economic predictors, where they have a greater impact on GHQ12 score in the WB. Furthermore, in the GS, extreme poverty appears to be what counts in negatively affecting mental health. Food insecurity is significant in both settings, but greater in magnitude in the WB. Political determinants are statistically significant in both settings with similar coefficients; however, the range of scores are wider and overall scores higher in the GS. In all models we controlled for household size being both a measure of network and of burden on resources as well as a proxy for number of children often identified as a source of support. The variable was only mildly significant in the GS model.