A study of spatiotemporal epidemiology of self-harm in Kwai Tsing

Abstract

Self-harm is an increasing concern in Hong Kong with an average annual incidence of 116 per 100,000 populations from 2002 to 2011. There is also a major concern about the repetition of self-harm and the risk of subsequent suicide. We studied Kwai Tsing district because it has a high incidence of self-harm with more than 2,000 episodes during 2004-2008.

There is a lack of studies on spatial and temporal factors involved in self-harm, especially within small urban areas. We studied the spatiotemporal epidemiology of self-harm in Kwai Tsing, focusing on service accessibility and attempters’ perception of those services to enhance our understanding of area-level patterns and features that would help direct prevention strategies and service enhancements for self-harm. The thesis analyzed five topics: 1) ascertainment of self-harm in Kwai Tsing, 2) spatial and temporal variations of self-harm with area-level deprivation, fragmentation, and financial hardship, 3) pure spatial and space-time clusters of self-harm in Kwai Tsing, 4) spatial accessibility of community services and self-harm, and 5) self-harm attempters’ perception of community social services and the implication on service provision.

The capture rate of self-harm cases in emergency departments (23%), inpatient records (66%) and integration of both records (80%) highlighted the need for enhancing the capture of self-harm in a hospital setting. The results identified spatial heterogeneity of self-harm (mean SMR = 0.99; SD 0.66) and social deprivation index (mean SDI = 20.32; SD 9.11) across geographical study units (LSB, n=58) in Kwai Tsing. We also found significant associations between SMR and social deprivation (r = 0.611, P<0.001), fragmentation (r = 0.448, P<0.001), and financial hardship (r = 0.651, P<0.001). Five areas with high SMR were located, which were not limited to those areas that were socially deprived. Besides, most likely pure spatial (19 LSB, RR =1.69; P <0.001) and space-time clusters (2004-2006, 16 LSB, RR= 1.74, P<0.001) were identified. For community services, the overall mean adjusted accessibility index for elderly (age ≥65 years) and youth (age 15-24 years) services were estimated as 0.0003 (95% CI 0.00033 – 0.00042) and 0.0002 (95% CI 0.00021 – 0.00023), respectively. Shortcomings and uneven distribution of provider (center) -to-population ratios of community social services for elderly (age > 65) and youth (age 15-24) were observed, which supported the need to enhance the locality of community services. A semi-structured interview was carried out on 11 self-harm attempters to explore their perception on service availability, accessibility, affordability, and acceptability. The results supported the need to revise the scope of services, to enhance the skills and trust between caregiver and clients, and to extend service hours to cater to the needs of clients after work hours.

In conclusion, this study identified spatial and temporal characteristics of self-harm in a small area of Hong Kong. The incorporation of the evidence from service accessibility and feedback from self-harm attempters could enhance the support of self-harm patients along the continuum of care.