Introduction

The lasting internal migration in China probably can be called the most magnificent population reshuffling in modern history. Thanks to the loosening of household registration regulation ever since the implementation of “opening-up policy” in the late 1970s and early 1980s [1], along with the modernization of agriculture which resulted in a huge surplus of labor in rural regions [2], lots of inlanders keep swarming into mega cities, especially those reside in the southern and eastern coastal areas of China during the past three decades. According to the 6^th National Population Census data, by the end of the year 2010, there were more than 221 million internal migrants in China, over 80 percent increase in amount when compared with the inception of 21^st millennium [3], yet a continuous growth is expected in the near future.

In China, internal migration usually occurs without a change of hukou (household registration) status. To a great extent, hukou status decides one's full access to employment, medical insurance, housing stipend, social welfare and education within the registration area [4]. In this instance, internal migrants are often labeled as “vulnerable individuals” in host cities [5]. Ever since the late 1990^th, studies which focused on health issues of internal migrants started to burgeon. But nearly all of them were focusing on physical health, like infectious diseases (such as AIDS and tuberculosis) and reproductive health of women [6], whereas studies on psychological well-being were scarce. However, some existing studies in this field had already disclosed a disturbing trend: the overall mental health status of internal migrants is alarming, especially the prevalence of depression. For example, one study which carried out in the capital city of an inland province (Sichuan), Qiu et al found that 23.7% of 1,180 internal migrants reported depressive symptoms [7], and another study which implemented in an economically prosperous coastal city of China, among 4,088 internal migrants, a comparable 21.4% reported depressive symptoms [8].

Depression is closely related to deliberate self-harm (DSH). According to western literatures, about two-thirds DSH patients are suffering from depression, simultaneously [9], [10]. Take the large size of population and high prevalence of depressive symptoms into consideration, DSH behaviors among Chinese internal migrants might be more common than that in general population. In Chinese culture, mental illnesses are often attached to stigma [11], thus usually DSH individuals won’t seek help from health facilities unless under urgent situations. Therefore, the predominant part of light DSH cases remains undetectable. However, considering of the fact that severe DSH usually results in much more devastating outcomes such as disabilities or even death, and is more intimately related to mental health distresses rather than impulsion [12], it is of a much greater intervention importance than light self-inflicted injuries. The major aim of our study is to investigate self-harm behaviors among Chinese internal migrants, by looking into the risk factors that associated with severe DSH.

Method

Results

Characteristics of DSH migrants

From the year 2005 to 2010, altogether 547 self-harm internal migrants were identified. After went through quality review, 121 of them were excluded because of incompleteness of vital information, finally we got 426 cases to analyze. Among those cases, females took the majority (66.2%), younger individuals who aged between 15 to 29 years accounted for nearly two-thirds of all subjects. Temporary and salaried manual worker were the predominant professions, with a combined proportion of 70 percent. About 80 percent subjects were resided in cities. DSH cases were evenly distributed in four seasons. Over 70 percent cases chose either stabbing/cutting or poisoning to hurt themselves. Severe DSH cases accounted for one-third of all subjects (Table 1).

Download:

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

Table 1. Characteristics of identified DSH migrants, 2005–2010, Zhjiang, China.

https://doi.org/10.1371/journal.pone.0080667.t001

Distributive features of self-harm severity

Gender difference in self-harm severity was significant, as the proportion of severe self-harm was much higher among male cases. There was a palpable trend that along with age increase, the proportion of severe self-harm was increasing dramatically, for example, among cases aged 15 to 29 years old, the proportion of severe self-harm was 26.8%, whereas among cases aged 45 to 65 years old, such proportion was 63.6%, instead. Residence place was also significantly associated with self-harm severity, as the proportion of severe self-harm was much higher among DSH migrants who resided in county areas. Preceding alcohol intake was an associated factor, too. A greater proportion of severe self-harm was discovered among subjects who did not take alcohol before the implementation of self-harm behaviors (Table 2).

Download:

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

Table 2. Distributive features of self-harm severity among identified DSH migrants, 2005-2010, Zhejiang, China.

https://doi.org/10.1371/journal.pone.0080667.t002

Model fitting results

During model fitting process, we initially put all significant variables based on Chi-square tests (gender, age, residence place, self-harm method, underlying cause, and preceding alcohol intake) as independents into the logistic regression model, but fitting result was not ideal (goodness-of-fit χ² = 124.05, df = 90, p = 0.01). Then we examined possible interactions between independents, at the setting significance level (p<0.10), we detected out interactions between gender and two other variables (residence place and self-harm method). After putting those interactions into the original model, underlying cause was not significantly associated with self-harm severity anymore, so we dropped it to construct the final regression model. After dropping underlying cause, partial-regression coefficients of other independents showed minute changes, fitting result suggested that this final model was satisfactory (goodness-of-fit χ² = 48.02, df = 54, p = 0.28).

The estimates of the final logistic regression model were enumerated in Table 3: compared with DSH migrants who aged between 15 to 29 years old, the odds of severe self-harm in those aged 45 to 65 years old was 3.44 (95% CI: 1.30, 9.15) times; preceding alcohol drinking was inversely related to severe self-harm, with an odds ratio of 0.40 (95% CI: 0.22, 0.74); although the main effect of gender indicated that compared with females, the odds of severe DSH was 8.55 (95% CI: 3.54, 20.67) times among males, after taking those two identified interactions into consideration, in every possible combination of residence place and self-harm method, there was no significant difference in the odds of severe DSH between males and females, and the male-to-female odds ratio ranged from 0.28 (95% CI: 0.01, 6.49) to 2.18 (95% CI: 0.30, 15.80).

Download:

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

Table 3. Estimates of the final logistic regression model, 2005-2010, Zhejiang, China.

https://doi.org/10.1371/journal.pone.0080667.t003

Residence place and self-harm method showed gender-specific association with severe DSH. In female subjects, resided in counties was related to increased odds of severe DSH (adjusted odds ratio: 3.01, 95% CI: 1.55, 5.84), and compared with stabbing or cutting, poisoning was observed 4.68 (95% CI: 2.22, 9.87) times odds of severe DSH. However, in male subjects, both residence place and self-harm method were not significantly related to severe DSH (Figure 1).

Download:

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

Figure 1. Gender-specific adjusted odds ratio of severe DSH by residence place and self-harm method, 2005-2010, Zhejiang, China.

https://doi.org/10.1371/journal.pone.0080667.g001

Discussion

In this study, we intended to explore associated factors of self-harm severity among DSH internal migrants in China. To our best knowledge, there are no existing literatures which ever addressed this topic before. Based on model fitting result, a positive association between age and severe self-harm has been identified. Considering of the intimate relationship between DSH and mental disorder, especially depression, such result may suggest that older migrants are more prone to be tortured by serious psychological problems. Some previous studies had pointed out that since the chance of permanently residing in host city is slim to most internal migrants, the majority of them will eventually choose to head back home at the rear of the productive age (usually from 50 to 60 years old), maybe in order to guarantee the full access to medical care (which is also determined by hukou status), in case of frequent ailments in the coming later age [14]. In this instance, migrants who still linger in host cities in their 50s or 60s are probably confronted with more serious financial predicaments, such reality can bring about a huge psychological pressure, which may adversely affect mental health. Moreover, internal migrants tend to leave home at a younger age, thus those older ones have already stayed in host cities for a considerably long period. In order to save living expenses as much as possible, most married internal migrants choose to live alone, longer separation from family can definitely do harm to mental health, too.

Although many existing literatures indicated that alcohol consumption is a major risk factor for suicidal behavior and DSH [15], [16], the association between alcohol involvement and DSH severity has seldom been discussed. One previous study suggested that as both DSH and alcohol use are risk factors for completed suicide, drinking may further increase the risk of impulsive behavior during DSH process, and cause aggravated injuries in turn [17]. However, our finding was quite in the opposite, as an inverse association has been detected between preceding alcohol drinking and severe DSH among study subjects. We think coping strategy theory can make a plausible explanation to this phenomenon: even if often be labeled as “maladaptive coping” or “avoidance coping”, alcohol drinking still can be an effective way of anguish venting, which is expected to mitigate the feeling of desperation among psychologically pressured individuals, and reduce the risk of severe DSH eventually.

In this study, we found that residence place showed significant association with severe DSH only among female subjects, such result may suggest that the psychological well-being of female internal migrants is more susceptible to social environment. Compared with city-dwelled counterparts, we think two major reasons may jointly contribute to higher odds of severe DSH among female internal migrants who reside in county. One is income difference, for job opportunities in cities are usually much higher paid than those in counties, so female migrants who reside in counties are in face of greater economic pressure, which can cause worse mental health. In addition, the mindset of city dwellers in China is usually more open and compatible, whereas local residents in small places are comparatively conservative and exclusive, so female migrants who reside in big cities may have a much better chance to integrate themselves into social environment, and confront less discrimination. Previous studies have found that both dysfunctional social connection and discrimination were significantly related to higher level of depression among Chinese internal migrants [18], [19], therefore, it is not surprising to find a deteriorated mental health status among county-dwelled female migrants.

Two hypotheses can be generated based on our study results, and need to be verified by future studies. One is that, as poisoning was positively associated with severe DSH only in females when compared with stabbing/cutting, we hypothesize that gender-related difference in poisonous substance choosing when performing self-harm behaviors may partly contribute to this discrepancy. The second is that, as we did not identify a significant association between gender and severe DSH, this finding may suggest that although it has been reported that the general mental health well-being of male internal migrants was much more worrying than their female counterparts [20], the proportion of gravely stressed individuals was not significantly different between two genders.

Those major findings above indicate that among Chinese internal migrants, older ones, and females who reside in counties may be more likely to commit severe self-harm behaviors, thus merit intense focus. In view of the catastrophic impacts of severe DSH to either family or the entire society, population-based surveys are imperatively needed, to broadly explore the overall mental health well-being, especially depression and its associated risk factors among Chinese internal migrants, in order to construct targeted intervention measures and effectively implement them as soon as possible.

Several limitations of this study should be noticed. First of all, to some research variables, the estimated 95% confidence interval of adjusted odds ratio was large, which means the sample size of certain subgroups was small, and that will definitely affect the precision of estimation in this study. Secondly, when constructing logistic regression model, the study population we used was DSH internal migrants, so to a certain extent, the associations we detected between self-harm severity and related factors can be somewhat different from the estimations had we studied the general population of internal migrants instead. But considering of the fact that DSH is still a rare event among general migrants, if we intend to find possible risk factors at population level, we might need an extremely large sample size to guarantee enough events, thus our method is definitely a more realistic and efficient way to perform preliminary screening. Thirdly, because of data limitation, in this study, we can not determine the place of origin for study subjects, considering the reported discrepancy in psychological well-being between rural-to-urban and urban-to-urban internal migrants in China [21], the associated risk factors of severe DSH can also be different between those two subtypes, thus should be further investigated. Finally, all DSH migrants we studied were collected from a localized area, although Zhejiang province is one of the major recipients of internal migrants in China, the generalization of our study results should be cautiously made.

Acknowledgments

We thank all participants of Zhejiang provincial hospital-based injury surveillance system, especially 28 sentinel hospitals and relevant county CDCs, for their onerous work in data collecting, verifying and reporting.