The prevalence, incidence, and gender and age-speciﬁc
incidence of problem gambling: results of the Swedish
longitudinal gambling study (Swelogs)
Auckland University of Technology, Auckland, New Zealand,
Public Health Agency of Sweden, Östersund, Sweden
and School of Public Health and Health Sciences,
University of Massachusetts, Amherst, MA, USA
To estimate the prevalence, incidence and gender and age-speciﬁc incidence of problem gambling in the Swedish
Longitudinal cohort study with linkage to register data.
Stratiﬁed random sample aged 16–84 years at baseline (n = 8165) re-assessed a year later (n =6021).
Problem gambling (life-time and past 12 months) was measured by the South Oaks Gambling
Screen–Revised (SOGS-R). Past 12-month (current) problem gambling was also measured by the Problem Gambling Se-
verity Index (PGSI).
The SOGS-R combined current pathological and problem gambling prevalence rate (PR)
was 2.1 [95% conﬁdence interval (CI) = 1.8–2.4] at baseline and 1.7 (1.4–2.0) at follow-up, approximately half the corre-
sponding life-time estimates.[Correction added on 22 Dec 2017, after ﬁrst online publication: In the preceding sentence, the
SOGS-R combined current pathological and problem gambling prevalence rate (PR) was incorrectly reported as being dou-
ble the corresponding life-time rate. It has been corrected in this version.] PGSI combined current problem and moderate-
risk gambling PRs were 2.2 (1.9–2.5) at baseline and 1.9 (1.6–2.2) at follow-up. Combined incidence rates (IRs) were 1.0
(0.8–1.3) (SOGS-R) and 1.4 (1.1–1.7) (PGSI), with more than three-quarters being new cases. While ﬁrst-time IRs did not
vary by gender, males had a higher relapse IR and proportionately more females were new cases. The young adult IR was
more than double the older adult IR; similar proportions were new cases.
The actual incidence of problem
gambling relapse in Sweden is likely to be higher than estimated. The proﬁle of problem gambling in Sweden is likely to
change over time, with increased proportions of women and older adults.
Keywords Age, epidemiology, gender, incidence, prevalence, problem gambling.
Correspondence to: Max Abbott, Auckland University of Technology, AG141, North Campus, 90 Akoranga Drive, Northcote, Auckland 0627, New Zealand.
Submitted 19 June 2017; initial review completed 2 August 2017; ﬁnal version accepted 30 October 2017
General population prevalence rates (%) of problem gam-
bling are typically within the 0.5–3.0range,withthree
to four times as many people experiencing subclinical levels
of pathology and harm [1,2]. There are strong associations
between gambling participation and problem gambling
[3,4]. Males, young adults, low-income and non-married
people are almost universally at elevated risk . Problem
gamblers’ socio-demographic proﬁle has changed
somewhat over time . In jurisdictions where electronic
gaming machines (EGMs) have been distributed widely,
gender differences have often diminished [7,8].
Since the late 1980s, Swedish state-owned gambling
organisations increased the range and number of gambling
products . EGMs were reintroduced, and casinos and in-
ternet gambling became available from domestic and un-
regulated foreign operators . Gambling expenditure
rose markedly then slowed, increasing by 13% in real value
between 1998 and 2008, remaining since at 3% of house-
hold disposable income .
Pathological gambling was conceptualised initially as a
chronic or chronically relapsing disorder. Early measures
such as the South Oaks Gambling Screen (SOGS) 
reﬂected this . The SOGS was adapted in 1990 to in-
clude current and life-time measures [14,15]. Subsequent
© 2017 Society for the Study of Addiction Addiction, 113,699–707