Ten Years of Gambling Treatment Records to Be Mined in Swedish Comorbidity Study Protocol

A research team in southern Sweden has laid out a detailed plan to comb through a decade of clinical records from the country’s first specialized gambling disorder treatment unit, aiming to map the physical and psychiatric illnesses, medications, referral routes and gambling patterns of everyone who sought help there over ten years. The study protocol, published in Harm Reduction Journal on 16 October 2026, describes a retrospective clinical and register study covering all treatment-seeking patients at the unit from its opening in December 2015 through November 2025, an estimated total of around 1,200 adults.

The unit under study sits in Malmö, the largest city of Skåne county, and serves an uptake area of 1.4 million inhabitants. It was the first Swedish health care treatment unit for gambling problems, launched in 2015 within the psychiatric services of Region Skåne, and it treats adults aged 18 years or older. The protocol’s authors, led by Anders Håkansson, with Yassir Abdul Rahim, Molly Miles and Carolina Widinghoff, plan to extract data from patient records by manual reading and to link each treatment episode to regional health care registers that capture all other primary and specialized care contacts in the region, including outpatient and inpatient episodes and diagnoses for both physical and mental health disorders.

The motivation, the authors write, is that gambling disorder was the first non-substance-related addictive disorder to be recognized and can severely affect health and psychosocial well-being, yet treatment seeking is often low and the documentation of clinical needs and medical comorbidities is sparser than for substance use disorders. A majority of individuals with a manifest gambling disorder never seek treatment, and among people with emerging or less severe problems, treatment seeking is rare. Barriers described in the literature include stigma, shame, patient misrepresentation and structural obstacles, and detection of gambling problems may be more challenging in primary care than for other health conditions.

Existing evidence already points to a heavy burden of comorbidity among clinical patients. In an early follow-up of the first 16 months of the same treatment unit, 58% of patients were reported to suffer from at least one additional mental health disorder. Research has linked gambling disorder to depressed mood, anxiety and an increased risk of suicidal behavior, including death from completed suicide, and clinical data suggest suicidal behavior may be particularly common when a psychiatric disorder is comorbid with the gambling disorder. Over-indebtedness also appears strongly linked to gambling disorder, affecting housing and socioeconomic conditions of affected individuals and their families.

Physical health has received less attention than mental health in gambling research, but the protocol notes growing evidence that individuals with gambling disorder face more physical health issues than the general population, including higher rates of neurological, respiratory, endocrine and digestive diseases. A national Swedish register study suggested links between gambling disorder and both cardiovascular and endocrine disease, as well as an increased risk of problem gambling after bariatric surgery for obesity. Cardiovascular effects from the gambling situation itself have been proposed, and more hazardous lifestyle habits have been described in people with gambling problems.

The planned study asks a set of specific research questions. Which physical and psychiatric comorbidities are seen in patients seeking treatment for gambling disorder, and how do they differ between routes of referral? How do demographic and socioeconomic factors, such as over-indebtedness, housing and employment status, differ between patients with and without physical comorbidities? How common are tobacco use and alcohol-related problems, and do they vary with gambling patterns and referral routes? The team also plans to examine how common voluntary self-exclusion from gambling is, how often such self-exclusion is breached, and how both vary across gambling types, comorbidities, age, gender and socioeconomic status.

Self-exclusion receives particular attention because it is a common harm reduction measure through which individuals ban themselves from gambling for a period of time in order to control their behavior. In clinical contexts, self-exclusion has been seen to be both commonly applied and commonly breached, through overseas or otherwise irregular gambling patterns. In relatively small study samples, patients who self-exclude, and especially those who breach their ban, have been more likely to suffer from mental health disorders, which the authors say makes it particularly relevant to study this tool in relation to mental health comorbidity.

Methodologically, the study will draw on manual reading of health care documentation following a structured protocol, with variables covering legal gender, age, housing and working conditions, socioeconomic characteristics, comorbid physical and psychiatric diagnoses coded according to ICD-10, prescribed medications of all types categorized by the Anatomical Therapeutic Chemical system, routes of referral, emergency psychiatric contacts, tobacco use, identified alcohol or drug problems requiring treatment, and gambling-specific data including history of self-exclusion and all reported gambling types. This includes, alongside traditional gambling types, the minority of patients reporting gambling for money or monetary values within a videogame context, a pattern the authors describe as a novel but relevant challenge. Three to four assessors with clinical experience, licensed physicians, nurses or social workers, are planned to carry out the manual extraction, with overlapping reading assignments and 10% of each assessor’s records reviewed independently by another assessor to allow inter-rater reliability to be described using kappa statistics; disagreements will be resolved by joint assessment until consensus.

The regional health care registers, the protocol notes, are complete and do not typically suffer from missing data on primary and secondary diagnoses or prescribed pharmaceuticals, and all individual data are linked to an identity-specific Swedish social security number that includes dates of birth and legal gender. Chart review data can be missing, for example where admittance routines were not followed or assessments were discontinued at a patient’s request or due to severe mental distress, which clinical experience from the unit suggests is very rare. In a previous study of a limited part of the time frame, the type of gambling involved was missing in 13% of patient records, predominantly before the introduction of a quality register that is likely to have improved structured documentation.

Analysis plans call for descriptive statistics to characterize the population, and logistic regression adjusted for age and legal gender to examine associations between comorbidities and gambling-related variables, demographic characteristics, psychosocial living conditions, prescribed medications and referral routes. Conservatively, regression models will include only items showing significant associations in uncontrolled binary comparisons, with age and legal gender added to correct for demographic bias given the large gender differences in gambling types and the marked age difference between women and men in this clinical setting. A binary correlation matrix will be inspected before modeling to avoid including variables correlated above 0.7, Bonferroni corrections will be applied in multiple group comparisons such as between women and men, and missing data will be reported for each variable but not replaced, with sensitivity analyses where more than 20% of responses are missing. Changes over the study period, including those potentially associated with the 2018 Swedish legislation shift, will be assessed using year of treatment entry as a continuous or categorical variable. All analyses will use SPSS version 29, with significance set at p < 0.05.

The long time frame is central to the design. The authors argue that after ten years of running a dedicated unit, conclusions can be drawn about who seeks and is referred for treatment, and the study may highlight circumstances that should have prompted other providers to screen for or refer suspected gambling problems. The period spans changes in the Swedish gambling market, including the 2018 legislation making it mandatory for regional health care and municipal social services to offer support and treatment for gambling disorder, the 2019 legal re-regulation of the gambling market, and the COVID-19 pandemic, and the protocol states the study can detect differences over time in the distribution of gambling types across these shifts.

The register component is presented as a distinctive strength. Few settings have access to large-scale register data allowing clinical data to be run against a wide range of other treatment providers within a geographical area, and Sweden has a history of such registers, including for the study of gambling disorder patients. The contribution of this project, beyond combining registers, is the possibility of manually extracting detailed patient record data, so that register data can be compared with descriptions of gambling patterns and psychosocial characteristics that cannot readily be detected in automated register sources.

The authors are explicit about the limits of what the study can show. Because it involves only individuals who chose to seek treatment, it cannot lead to conclusions about people with a corresponding degree of gambling problem who refrain from seeking treatment, and these groups may differ, with treatment seekers likely having higher problem severity, greater awareness of their problem, or more motivational pressure from others. The project therefore cannot draw conclusions about the general population with gambling disorder. Context matters here: around 2% of the Swedish population are believed to suffer from moderate-risk or problem gambling according to the most recent nationwide survey data published, from 2014, and the total number of individuals diagnosed annually for gambling disorder in Sweden has recently been reported at approximately 1,000, with a similar and likely overlapping number reported from social services, representing a small minority of inhabitants likely to have a gambling problem at a diagnostic level.

The study was pre-registered at ClinicalTrials.gov with identifier NCT07488871 on March 9, 2026. The research group reports a close collaboration with the national patient and peer support organization, which will review the project through data analysis and interpretation of conclusions, and a communication plan is established for disseminating results to political and authority decision makers, media, and stakeholders in health care and social services. The authors frame the work as laying the ground for improved and earlier detection of gambling problems in diverse clinical settings, by describing the full range of sociodemographic conditions and comorbidities, in mental as well as physical health, among clinical patients with gambling disorder. As a protocol, it sets out planned methods and analyses; no findings from the ten-year data extraction are reported in the published document, and no datasets were generated or analysed during the current study.

Subject of Research: A study protocol for a ten-year retrospective clinical and register study of comorbidity and gambling patterns in treatment-seeking patients with gambling disorder in Sweden

Article Title: Study protocol: seeking treatment for gambling – a 10-year retrospective clinical and register study of gambling patterns and physical and psychiatric comorbidity in patients with gambling disorder

Article References: Håkansson, A., Rahim, Y. A., Miles, M., & Widinghoff, C. (2026). Study protocol: seeking treatment for gambling – a 10-year retrospective clinical and register study of gambling patterns and physical and psychiatric comorbidity in patients with gambling disorder. Harm Reduction Journal, 23(1), Article 165. https://doi.org/10.1186/s12954-026-01565-0

Image Credits: AI Generated

DOI: 10.1186/s12954-026-01565-0

Keywords: gambling disorder, study protocol, Sweden, comorbidity, mental health, physical health, self-exclusion, health registers, treatment seeking, addiction psychiatry, Skåne, retrospective study