Cross-sectional study finds elevated gambling disorder in Danish forensic psychiatric patients

A Danish multi-site cross-sectional study found that 46% of forensic psychiatric patients had gambled during the previous year, while 12% met the study’s threshold for severe gambling problems indicative of gambling disorder. The proportion with severe problems was more than six times the 1.8% recorded in an age- and gender-matched national sample, indicating a marked imbalance between the severity of gambling problems in this clinical population and the rate observed in the comparison group. The study, published in Frontiers in Psychiatry, surveyed adults receiving forensic psychiatric services between June 2024 and January 2025 across 19 units at six hospitals in Denmark’s Central and Capital Regions.

Researchers recruited 214 participants and compared their responses on the nine-item Problem Gambling Severity Index, or PGSI, with data from 2,090 people matched by age and gender to a recent Danish national gambling survey. The PGSI ranges from 0 to 27 and classifies responses as non-problem gambling, low risk, moderate risk, or problem gambling at scores of 8 or above, a category considered indicative of gambling disorder. Almost half of the patients reported gambling in the past year, but harmful gambling was concentrated among a smaller group.

Overall, 30% of forensic psychiatric patients had a PGSI score above zero, including 8% classified as low risk, 10% as moderate risk and 12% as problem gamblers. Among the 98 patients who had gambled, 66% met the study’s definition of gambling problems at some level, and 27% reached the problem-gambling threshold. The general-population comparison showed a different pattern: past-year gambling was slightly more common in the matched sample, at 54% versus 46%, while the overall distribution of PGSI risk levels differed significantly between the groups.

Clinical records suggested that gambling disorder was seldom identified despite the survey results. A meta-analysis cited by the researchers estimated problem-gambling prevalence of 8.7% among people with psychotic disorders, compared with 12% in this forensic psychiatric sample, although the study said broader evidence on overlapping conditions could help contextualize the elevated rate. Among patients who had gambled, the researchers divided the group according to PGSI score, comparing 50 lower-risk participants with scores of 2 or less against 47 higher-risk participants with scores of 3 or more; one of the 98 gamblers lacked PGSI data.

Higher-risk gambling was associated with being male, reporting weaker financial conditions, and reporting an urge to commit crime because of gambling. All 10 patients who said they had felt compelled to engage in criminal activity during the past year were in the higher-risk group, equivalent to 21% of that group, while no lower-risk patient reported such an urge. The cross-sectional design cannot establish whether gambling contributed to offending or whether other factors influenced both gambling severity and reported criminal intent.

Spending and financial indicators also separated the two risk groups. No lower-risk patient reported spending more than 1,000 Danish kroner, approximately 184 euros, on gambling in the previous month, whereas 42% of higher-risk patients did. Higher-risk patients were more likely to have consumer loans, with 42% reporting current loans compared with 21% of lower-risk gamblers, and less likely to set aside money each month, at 41% versus 76%.

Monthly income did not differ significantly between the groups. The authors said these patterns were concerning because financial stability is important to recovery and rehabilitation in forensic psychiatric populations, but the survey did not establish a causal relationship between gambling severity and the reported financial conditions. Some higher-risk patients were already taking steps to limit their gambling.

National gambling self-exclusion services were being used by 47% of higher-risk gamblers, compared with 12% of lower-risk gamblers, and 53% of the higher-risk group reported gambling problems before entering forensic psychiatric care. Yet fewer than three patients had ever sought help from forensic psychiatry staff for gambling, and only nine of the 214 participants reported ever seeking gambling help of any kind. In another indication of possible receptivity, 67% of higher-risk gamblers requested personalized normative feedback after completing the PGSI, compared with 36% of lower-risk gamblers.

The study did not test an intervention, so it cannot establish whether such feedback changes gambling behavior or clinical outcomes. Online gambling was prominent in the findings. It was the only gambling modality used by 61% of patients who gambled, while another 16% used both online and offline channels; slot machines were the most frequently reported online activity.

The study’s broader context was Denmark’s liberalized gambling market, in which licensed online operators and growing access to offshore gambling have expanded opportunities to wager. The authors also noted that forensic psychiatric patients have internet access even in medium-secure units. These features make online and institutional contexts relevant to further investigation, but the survey did not establish that online availability, institutional conditions or other environmental factors caused the elevated prevalence.

Several limitations constrain how widely the results can be generalized. Although 486 eligible people were invited, 214 participated, a participation rate of 44%, while 56% declined. Another 105 patients who were considered too unwell or otherwise unable to complete the survey were excluded, and no non-responder analysis was possible under data-protection constraints.

The findings also relied on self-report and therefore may be affected by social-desirability bias and the assistance some participants required. Gambling was reported more often by people who completed questionnaires independently, although their mean PGSI scores and distribution of risk categories did not differ significantly from those who received assistance. The cross-sectional design, multiple comparisons, wide confidence intervals around some estimates, and differences between countries in gambling regulation and forensic psychiatric practice mean that the observed associations require replication with longitudinal designs and diagnostic interviews.

Within those limits, the study identifies a clear mismatch between survey-based gambling severity and formal clinical recognition. One in eight participants met the PGSI threshold for gambling disorder, yet gambling disorder was seldom recorded in the diagnostic data available to the researchers. The findings support further evaluation of systematic assessment, staff training and tailored treatment in forensic psychiatric settings, while the cross-sectional evidence does not show that these measures improve care, recovery or recidivism outcomes.

Whether gambling is associated with subsequent offending, how institutional environments shape gambling behavior, and which interventions are effective remain open questions for prospective research.

Subject of Research: Prevalence and correlates of gambling problems among forensic psychiatric patients in Denmark

Article Title: Prevalence and correlates of gambling problems among forensic psychiatric patients: evidence of elevated rates from a multi-site study

Article References: Eriksen, J. W., Sørensen, L. U., Kristiansen, S., Terkildsen, M. D., Hansen, R. M., Marcussen, T., Brandt-Christensen, M., Storgaard, A., Wellnitz, K. B., & Frostholm, L. (2026). Prevalence and correlates of gambling problems among forensic psychiatric patients: evidence of elevated rates from a multi-site study. Frontiers in Psychiatry, 17, Article 1931287. https://doi.org/10.3389/fpsyt.2026.1931287

Image Credits: AI Generated

DOI: 10.3389/fpsyt.2026.1931287

Keywords: gambling disorder, forensic psychiatry, problem gambling, cross-sectional study, severe mental illness, PGSI, Denmark, online gambling, financial hardship, criminal intent, Prevalence, correlates