Abstract
Background
Gambling-related harm is increasingly recognised as a complex, multi-dimensional public health problem which variably impacts health, relationships, resources, and broader social functioning. Existing measures – often focused on diagnostic “problem gambler” thresholds – typically fail to capture the breadth of such harms. Moreover, there are very few instruments designed to capture the perspective of those harmed by someone else’s gambling – ‘affected others’ (AOs).
Aims
This study reports the content development and validation of the Gambling Harms Severity Index (GHSI-10) and GHSI for affected others (GHSI-AO-10); co-designed, person-centred, non-stigmatising instruments designed to assess impacts, severity and changes in gambling-related harm across diverse contexts and populations. Design, setting and participants We employed a multi-phase, mixed-methods design informed by best practice guidelines for patient-reported outcome development. Conceptualisation was constructed on a ‘holistic framework of gambling harm and recovery,’ derived from structured literature reviews and qualitative research. Item generation and refinement were co-produced with individuals with lived experience, AOs, practitioners, and academic experts. Psychometric validation was conducted via a convenience paid-for online sample from the UK (n= 3,315 for people that gamble; n= 3,017 for AOs).
Measurements
Statistical validation of GHSI-10 and GSHI-AO-10 involved a combination of classical test statistics (CTT) and Rasch Measurement Theory (RMT), alongside convergent validity versus existing measures of harm (e.g. PGSI) and divergence from measures of wellbeing (e.g. ICECAP-A).
Findings
Items were co-developed to assess harms across multiple components: wellbeing (mental, physical), relationships (personal, social/community), and resources (financial, occupational, ethical). This process refined the instrument to be non-stigmatising, person-centred and reflecting the language of lived experience, to reduce biases related to denial and social desirability, with a 3-month recall period, and a focus on harms that are dynamic and responsive to recovery pathways, rather than static, unchangeable harms (i.e. divorce, bankruptcy). Quantitative validation supported unidimensional measurement, good reliability (GHSI-10 α = 0.94, GHSI-AO-10 α = 0.95), convergent validity with existing measures (rho > .7), and criterion-related validity with measures of wellbeing (GHSI-10 rho < -.3, GHSI-AO-10 rho < -.23). Both GHSI-10 and GHSI-AO-10 met all RMT fit criteria, did not display infit or outfit, had well-ordered and fitting response options, and rationally ordered item locations.
Conclusions
The GHSI-10 and GHSI-AO-10 are psychometrically valid tools, co-designed to support research, service commissioning, and outcome evaluation across the gambling treatment and support ecosystem. They are grounded in diverse lived experience and aligned with public health principles.
Gambling-related harm is increasingly recognised as a complex, multi-dimensional public health problem which variably impacts health, relationships, resources, and broader social functioning. Existing measures – often focused on diagnostic “problem gambler” thresholds – typically fail to capture the breadth of such harms. Moreover, there are very few instruments designed to capture the perspective of those harmed by someone else’s gambling – ‘affected others’ (AOs).
Aims
This study reports the content development and validation of the Gambling Harms Severity Index (GHSI-10) and GHSI for affected others (GHSI-AO-10); co-designed, person-centred, non-stigmatising instruments designed to assess impacts, severity and changes in gambling-related harm across diverse contexts and populations. Design, setting and participants We employed a multi-phase, mixed-methods design informed by best practice guidelines for patient-reported outcome development. Conceptualisation was constructed on a ‘holistic framework of gambling harm and recovery,’ derived from structured literature reviews and qualitative research. Item generation and refinement were co-produced with individuals with lived experience, AOs, practitioners, and academic experts. Psychometric validation was conducted via a convenience paid-for online sample from the UK (n= 3,315 for people that gamble; n= 3,017 for AOs).
Measurements
Statistical validation of GHSI-10 and GSHI-AO-10 involved a combination of classical test statistics (CTT) and Rasch Measurement Theory (RMT), alongside convergent validity versus existing measures of harm (e.g. PGSI) and divergence from measures of wellbeing (e.g. ICECAP-A).
Findings
Items were co-developed to assess harms across multiple components: wellbeing (mental, physical), relationships (personal, social/community), and resources (financial, occupational, ethical). This process refined the instrument to be non-stigmatising, person-centred and reflecting the language of lived experience, to reduce biases related to denial and social desirability, with a 3-month recall period, and a focus on harms that are dynamic and responsive to recovery pathways, rather than static, unchangeable harms (i.e. divorce, bankruptcy). Quantitative validation supported unidimensional measurement, good reliability (GHSI-10 α = 0.94, GHSI-AO-10 α = 0.95), convergent validity with existing measures (rho > .7), and criterion-related validity with measures of wellbeing (GHSI-10 rho < -.3, GHSI-AO-10 rho < -.23). Both GHSI-10 and GHSI-AO-10 met all RMT fit criteria, did not display infit or outfit, had well-ordered and fitting response options, and rationally ordered item locations.
Conclusions
The GHSI-10 and GHSI-AO-10 are psychometrically valid tools, co-designed to support research, service commissioning, and outcome evaluation across the gambling treatment and support ecosystem. They are grounded in diverse lived experience and aligned with public health principles.
| Original language | English |
|---|---|
| Type | Preprint |
| Publisher | PsyArXiv Preprints |
| DOIs | |
| Publication status | Published - 3 Oct 2025 |
Bibliographical note
This manuscript is a preprint and has not been peer-reviewed. It reports new research that has not yet been formally evaluated by the scientific community. Any participant data have been anonymised in accordance with ethical approvals. Please cite this work as a preprint, acknowledging that subsequent versions may differ following peer review.Research Groups and Themes
- MGMT Marketing and Consumption
- Gambling Harms
