Domestic violence perpetration and victimisation and self-poisoning

in Sri Lanka: a protocol for a hospital-based case-control study

Vidusha H K D1,2, Bruna R3, Karen M3, Helen C3, Sharon C3,Ekanayake E H M R K2, Afra A L F2, Athapaththu W A M
A S2, Mushfira M S2, Wickramasuriya I P2, Knipe D*3,Rajapaksha T N*1,2
1
Department of Psychiatry, Faculty of Medicine, University of Peradeniya, Sri Lanka
2
South Asian Clinical research, Faculty of Medicine, University of Peradeniya, Sri Lanka
3
University of Bristol Medical School, United Kingdom

Abstract
Introduction
Domestic violence is a key risk factor for suicidal behaviour. There is a notable paucity of
evidence pertaining to the perpetration of domestic violence and its association with suicidal
behaviour. The aim of this study is to investigate the association between domestic violence
(victimisation and perpetration) and self-poisoning in Sri Lanka.
Methods and analysis
This is a hospital-based case-control study. Cases (n=130) will be individuals admitted to the
toxicology ward of the Teaching Hospital Peradeniya, Sri Lanka, for medical management of
self-poisoning. We will recruit controls (n=260) from other patients with unrelated
conditions or accompanying visitors presenting to the outpatient department of Teaching
Hospital Peradeniya. We will use unconditional logistic regression models to investigate the
association between domestic violence and self-poisoning.
Ethics and dissemination
We obtained ethics approval from the Ethical Review Committee of the Faculty of Medicine,
University of Peradeniya, Sri Lanka. The research assistants will be trained in administering
the questionnaire and ensuring participant safety. Results will be disseminated in peer-
reviewed articles, local media, and at national and international conferences
Strengths and limitations of the study
 This is one of the few studies to investigate the relationship between domestic violence perpetration
and suicidal behaviour.
 Hospital-control outpatients may have a different exposure distribution from the general population,
such as a greater prevalence of mood disorders and suicidal ideation, introducing the possibility of
selection bias. We will mitigate this by purposively recruiting individuals who accompany patients and
are less likely to be experiencing an acute health issue at the time of recruitment. Our previous
 investigations have shown that hospital controls recruited in this way are similar to community
controls.
 The reported rate of domestic violence (victimization and perpetration) may be underestimated due to
cultural stigma and recall bias. To mitigate against this, we will highlight that domestic violence is
common, and a broad statement will be made before particular concerns about domestic violence are
asked.
 This study will focus on hospitalisations following self-poisoning; other forms of suicidal behaviour will
not be investigated. However, previous investigations have shown that the vast majority of hospital-
presenting self-harm in Sri Lanka is due to self-poisoning.

Introduction

Suicide is an important public health issue globally. According to the World Health
Organization, more than 700 000 deaths occur due to suicide every year (1). Among global
suicide deaths, 77% occur in low and middle-income countries (1).

Sri Lanka had one of the highest suicide rates in the world in 1995, with a rate of 47 per
100,000 population (2). Since then the overall suicide rate has fallen and in 2019 the suicide
mortality rate in Sri Lanka was 14 per 100, 000 population (2), which is 1.5 times the global
average. However, rates of self-poisoning are still high, as in the most areas of the world (3).

Deaths due to suicide in Sri Lanka are more common in males than in females with a male to
female suicide ratio of 4:1 (3).
However, self-harm is more common among females, most likely due to several gender-
related risk factors,- including differences in the lethality of methods used, socially
constructed differences such as gender roles, responsibilities social status (4). Furthermore
pregnancy, miscarriages and postpartum periods are unique events to females which can
affect suicidality (5).

Women and girls are also more likely to be victims of domestic violence, which is known to
be associated with increased risk of self-harm, both internationally and in Sri Lanka (6) (7).
Domestic violence refers to physical, sexual or emotional abuse and controlling behaviours
(i.e. physical, sexual, economic, emotional, or psychological actions or threats of actions that
can influence another person) usually by a current or former partner, but also by a family
member or carer. Women are significantly more likely to be victims of domestic violence
than men and are more likely to be injured due to it (6). Being a victim of domestic violence,
contributes to feelings of helplessness, vulnerability, and increase the risk of self-harm and
suicide (8). Whilst domestic violence victimisation is more common in women, being a victim
of domestic violence increases the risk of self-harm in both sexes (9).

Whilst the association between domestic violence victimisation and self-harm and suicide
has received some, albeit limited, attention, the association between domestic violence
perpetration and self-harm and suicide has received even less. Previous research has shown
an elevated risk of suicide in perpetrators of domestic violence (10)., but this research has
focused on high income countries and evidence from low middle income countries, like Sri
Lanka, are lacking. The aim of this study is to explore the relationship between domestic
violence (victimisation and perpetration) and self-poisoning in both sexes in Sri Lanka.

METHODS AND ANALYSIS

Study Setting
The study will be carried out at Teaching Hospital Peradeniya (THP), located in the Kandy
District, Central Province of Sri Lanka. The Kandy District has a total population of 1375382
and is 115km east from Colombo, the capital of Sri Lanka. Teaching Hospital Peradeniya is a
tertiary referral hospital with a catchment area that includes the North Central, North
Western and Sabaragamuwa Provinces. According to the 2019 Well Women Survey, the
prevalence of domestic violence in the Kandy district is 15% and the it is 17.4% in Sri Lanka
(11).

Patient and Public Involvement

The questionnaire used in this study was piloted with research assistants, who are familiar
with local patients, and patients in the toxicology unit and outpatient department of THP.
Piloting highlighted that the original questionnaire was too long, therefore it was shortened
and simplified.
A workshop will be conducted involving community members aged 18 years and over in
order to identify factors associated with domestic violence and to discuss associations of
domestic violence with self-harm such as alcohol and other drug use, financial burden and
depression.

Study Design

This study is a hospital-based case-control study. Cases will consist of both male and female
patients admitted for medical management of self-poisoning to the toxicology unit in THP.
All patients who present to THP due to self-poisoning are admitted to the toxicology unit for
observation and treatment, regardless of the severity of poisoning. We aim to recruit a
consecutive series of cases. This case definition excludes self-harm due to other methods, for
example, hanging and cutting; however, self-poisoning represents the most common method
of self-harm cases presenting to hospitals in Sri Lanka (12).

Controls will consist of males and females, recruited from either other patients attending the
outpatient department of THP for unrelated conditions, or people accompanying other
patients presenting at the outpatient department and clinic of the same hospital during the
study period. These outpatients frequently present with conditions such as respiratory
infection, diabetes, hypertension, pregnancy, and conditions unrelated to the outcome of
interest this study. Controls will not be age or sex matched to cases, but we aim to recruit
two controls for each case.

Inclusion and Exclusion criteria

All patients aged 18 years and above who have been admitted to the toxicology unit for
medical management of self-poisoning will be eligible for inclusion as cases for the study. We
will not include patients admitted for management of accidental poisoning. Accidental
poisoning will be identified initially from the patient's admission record and then verbally
confirmed by the patient via self-report to the data collectors.

Controls will be either other patients aged 18 years and above attending the outpatient
department of THP for unrelated conditions, or accompanying visitors aged 18 years or
above presenting at the outpatient department and clinic of the same hospital during the
study period. Cases and controls who are physically unable or too ill to participate, as well as
those who have been diagnosed with an intellectual disability or dementia, will be excluded
from the study.

Sample Size

Based on a study previously conducted in this hospital, we estimated roughly 50 cases will be
admitted for self-poisoning each month, and 50% of those will be male (13). Therefore, it
was estimated that there will be 150 cases over a 6-month period. Based on a recruitment
rate of 87%, reported in a previous study, we plan to collect data from 130 cases and 260
controls over 6 months (14). Assuming 20% of controls (odds=0.25) report having
experienced a domestic violence, we would be able to detect a two-fold difference in risk
with 82% power (alpha=0.05) (14).

Data Collection

All patients who are eligible for inclusion in the study will be approached during their
admission at the hospital, after receiving treatment, and informed about the study. The
reason and nature of the study will be clearly explained to all participants, and they will also
be given an information leaflet containing detailed information on the study. Those who give
informed written consent will be included in the study. Interviews will be conducted with the
participant in a confidential space as this will help to ensure that the interview can be
conducted in private and in the absence of accompanying friends or family members.

All cases and controls will be interviewed by a trained data collector. A clinically qualified
psychiatrist (DV) will train and supervise four data collectors on the importance of adhering
to the interview schedule, how to promote a safe environment for disclosure, building
rapport with study participants, and the actions to be taken when at-risk participants are
identified. This training will be delivered in eight sessions that will take place over two
weeks. They will be trained on the aforementioned areas, with a particular emphasis on the
consistency of interviewing approach among all data collectors by ensuring that the same
questions will be asked in the same manner to all the participants. Lectures, discussions,
and role-play will be used to accomplish this. We will use a pre-piloted questionnaire,
designed for the study, to conduct interviews in the participant’s preferred language- (i.e.
Sinhala, Tamil or English), in a confidential manner.

We will collect data on suicidal behaviour, economic-stressors, gambling, drug use,

childhood adversity, psychiatry morbidity, alcohol use and domestic violence.

We will use a "domestic violence impact tool kit" to form the questionnaire to collect
domestic violence data (15). Individuals’ financial status data will be captured by a culturally
modified “Debit and finance survey" questionnaires (16). After piloting, both of the above
original questionnaires were simplified and shortened because participants found it difficult
to concentrate and focus on a lengthy questionnaire.

Data on psychiatric morbidity will be collected using the 9-item Patient Health Questionnaire
(PHQ-9), which is brief, one page self-administered questionnaire that is internationally
validated for the identification of depression, and has been validated for use in Sri Lanka
(17).
The Alcohol Use Disorders Identification Test (AUDIT), will be used to assess the alcohol
intake behaviour of the participants. It has been validated for the local population (18).
Data will be collected during non-visiting hours for patients in the toxicology unit to ensure
responses are not influenced by another person and for patient safety. Identification details
such as the name or the bed head ticket number (i.e. the number of the file that contains
patient’s clinical updates) and date of birth will not be recorded. Data entry will be done by
the data collectors. The lead researcher will conduct regular supervision of the entered data,
check data entry for errors and shadow researchers to ensure consistency in approach.
During the data collection, if a participant becomes distressed, is found to be depressed or
reports suicidal thoughts in the preceding two weeks, a referral to the doctor on call for the
psychiatry unit, or the psychiatric clinic will be done, as required. If a participant discloses
domestic violence, they will be informed about the support that is available to them.

The research database will be kept safely in a separate electronic device where only the
researcher has access to them. All data will be stored securely and confidentially i.e.-
hardcopy data in a locked cupboard in a secure room, and softcopy data with password
encryption.

Participant anonymity will be ensured in any future publications relevant to the study data.

Analysis Plan
The primary analyses will be based on complete cases only, excluding participants with
missing data. We will also conduct a sensitivity analysis with all cases to explore whether this
exclusion influenced our findings. All associations and descriptions will be stratified by sex.
This is due to the impact of domestic violence victimisation and perpetration is different in
men and women and because suicidal behavior is also different in men and women.

All analysis will be conducted using the STATA statistical software (version 18). We will report
mean (standard deviation) or median (inter-quartile range) for continuous variables and
numbers and proportions for categorical variables. The associations between exposures and
outcomes will be assessed using logistic regression models and odds ratios with 95%
confidence intervals will be reported. The main outcome will be self-poisoning, and the
exposures of interest will be domestic violence perpetration and victimization. Three
categories of domestic violence (i.e. physical, emotional and sexual violence) will be
considered. In the primary analysis we will have two exposures of interest and these will be
dichotomised into a binary (yes/no) variable to indicate presence of domestic violence
perpetration and victimisation. For our secondary analysis we will use a frequency score and
 we will calculate the frequency each type of domestic violence. We will also investigate the
potential effects of mental illness, alcohol harmful use, psychoactive substance use, income,
and ethnicity in a series of models.

Ethics

Ethical approval for this research was obtained from the Ethical Review Committee of the
Faculty of Medicine, University of Peradeniya, Sri Lanka. Further approval was obtained from
the director of Teaching Hospital Paradeniya and the Consultant Physician of the Teaching
Hospital Paradeniya's toxicology unit to collect data from the subjects.

Dissemination

The outcomes of this study will be disseminated through publication of findings in peer-
reviewed articles. The results will also be shared at both national and international
conferences, health symposia, and local policy forums, such as those attended by local
government and non-government organizations. All interested participants will have the
opportunity to obtain a copy of the research report and the study findings from the
investigators. We will also prepare a short lay report with key study findings, and will make it
available to participants upon request in local languages.

References
1. Suicide [Internet]. [cited 2023 Dec 28]. Available from: [Link]
room/fact-sheets/detail/suicide
2. Sri Lanka - Suicide Mortality Rate (per 100,000 Population) - 2021 Data 2022 Forecast
2000-2019 Historical [Internet]. [cited 2021 Oct 3]. Available from:
[Link]
[Link]
3. De Silva V, Hanwella R, Senanayake M. Age and sex specific suicide rates in Sri Lanka
from 1995-2011. Sri Lanka J Psychiatry. 2013;3(2):7.
4. Vijayakumar L. Suicide in women. Indian J Psychiatry [Internet]. 2015 Jul 1 [cited 2021
Nov 16];57(Suppl 2):S233. Available from: /pmc/articles/PMC4539867/
5. Reid HE, Pratt D, Edge D, Wittkowski A. Maternal Suicide Ideation and Behaviour During
Pregnancy and the First Postpartum Year: A Systematic Review of Psychological and
Psychosocial Risk Factors. Front psychiatry [Internet]. 2022 Mar 24 [cited 2023 Dec
 28];13. Available from: [Link]
6. de Brouwer AM, Kaitesi U. Sexual violence. Elgar Companion to Int Crim Trib Rwanda.
2016;171–201.
7. (PDF) Intimate partner violence in Sri Lanka. Bioscience Trends, 4, 90-95 [Internet]. [cited
2024 Feb 27]. Available from:
[Link]
_Lanka_Bioscience_Trends_4_90-95
8. Correia CM, Gomes NP, Couto TM, Rodrigues AD. REPRESENTATIONS ABOUT SUICIDE OF
WOMEN WITH HISTORY OF DOMESTIC VIOLENCE AND SUICIDE ATTEMPT 1.
2014;23(1):118–25.
9. McManus S, Walby S, Barbosa EC, Appleby L, Brugha T, Bebbington PE, et al. Intimate
partner violence, suicidality, and self-harm: a probability sample survey of the general
population in England. The Lancet Psychiatry [Internet]. 2022 Jul 1 [cited 2023 Dec
29];9(7):574–83. Available from:
[Link]
10. Knipe D, Vallis E, Kendall L, Snow M, Kirkpatrick K, Jarvis R, et al. Suicide Rates in High-
Risk High-Harm Perpetrators of Domestic Abuse in England and Wales. Crisis [Internet].
2023 [cited 2023 Dec 29]; Available from: [Link]
11. DCS. Women ’ s Wellbeing Survey - 2019. 2019.
12. Rajapakse TN. A review of the changing patterns of suicide and deliberate self-harm in Sri
Lanka. Sri Lanka J Psychiatry. 2017;8(1):3.
13. Knipe D, Silva T, Aroos A, Senarathna L, Hettiarachchi NM, Galappaththi SR, et al. Hospital
presentations for self-poisoning during COVID-19 in Sri Lanka: an interrupted time-series
analysis. The Lancet Psychiatry [Internet]. 2021;8(10):892–900. Available from:
[Link]
14. Bandara P, Page A, Senarathna L, Kidger J, Feder G, Gunnell D, et al. Domestic violence
and self-poisoning in Sri Lanka. Psychol Med [Internet]. 2020 [cited 2021 Oct 4];1–9.
Available from:
[Link]
violence-and-selfpoisoning-in-sri-lanka/621C6F4F2AFD0FB1D99ADE20BF464B14
15. En WWP, Outcome I, Toolkit M. T3 Client END of intervention self-evaluation form. :1–8.
16. Debt-finance-survey-qs-Jan2022-v2.
17. Hanwella R, Ekanayake S, De Silva VA. The validity and reliability of the sinhala translation
of the patient health questionnaire (PHQ-9) and PHQ-2 screener. Depress Res Treat.
2014;2014.
18. De Silva P, Jayawardana P, Pathmeswaran A. Concurrent validity of the alcohol use
disorders identification test (AUDIT). Alcohol Alcohol. 2008;43(1):49–50.
Author’s contributions
The authors confirm contribution to the paper as follows:

study conception and design: Vidusha H K D, Knipe D, Rajapakse T N, Bruna R Karen M, Helen C,
Sharon C; data collection will be done: Vidusha H K D, Afra A L F, Athapaththu W A M A S, Ekanayake
E H M R K, Wickramasuriya I P; data analysis will be carried out by Viduhsa H K D, Knipe D, Bruna R,
Rajapakse T N ;draft manuscript preparation: Vidusha H K D, Knipe D, Rajapakse T N, Bruna R. All
authors reviewed and approved the final version of the manuscript.

Funding statement
This work was supported by Elizabeth Blackwell Institute for Health Research- Wellcome
Trust

Competing Interest Statement