Nearly half of patients in this study had polysubstance-related disorders, in accordance with results of previous studies, that range from 37.8-68.5% [13, 44–46]. Patients with one SRD, like cannabis-related disorder, are at high risk of developing multiple SRD [9]. Most had MD as well, as did 47-100% of patients treated in services for SRD in a recent review [55], which confirms that co-occurring SRD-MD is the norm among patients with SRD. A third of patients were also affected by co-occurring SRD-chronic physical illnesses, as previously reported [44]. Other studies confirmed that co-occurring SRD-MD, SRD-chronic physical illnesses, and polysubstance-related disorders increase ED use and hospitalization [9, 56]. Moreover, the high rates of residency in more materially and socially deprived areas confirmed the association between SRD and both poverty [57] and loneliness [58] previously reported. Finally, during the 12-month follow-up period in the year before use of acute care was measured, most patients did not consult their GP and psychiatrist, nor did they receive care from either community healthcare or addiction treatment centers for primary or specialized outpatient care, which supports results of previous studies suggesting little use of health and social services in this population [59–61]. Yet, the overall contribution of types of SRD to the frequencies of ED use and hospitalization was modest: 5.7 and 5.6% respectively, as compared with control variables, particularly clinical (71.1%) and service use variables (18.6%). Regarding variable effect sizes, about one third of variables in the ED use model, mainly clinical and service use variables, had medium effect sizes (RR from 1.27 for MD to 1.51 for the 12-17 age group), while 40% of variables in the hospitalization model, mainly clinical variables, had medium to high effect sizes (RR from 1.30 for 2-3 consultations with GP to 1.87 for chronic physical illnesses).

The first hypothesis that patients with polysubstance-related disorders would experience more ED use and hospitalizations versus other SRD groups was confirmed for ED use but not entirely for hospitalization. Polysubstance-related disorders are associated with poorer treatment adherence and health outcomes [19]. Those patients are also reported with more deliberate self-harm [62], which may explain their more frequent use of acute care relative to other subgroups. In this study, that patients most likely to be hospitalized were those with alcohol-related disorders, and whose SRD effect size (RR 1.38) was also highest, confirmed results of a previous study evaluating early hospital readmissions [63]. Higher hospitalization rates by patients with alcohol-related disorders compared to other SRD groups may be explained by their elevated rates of co-occurring SRD-MD and/or chronic physical illnesses (e.g. cardiovascular diseases, liver diseases) [64], alcohol withdrawal [65], and their greater tendency to seek care episodes than patients with other SRD [66].

Hypothesis two was also confirmed, as clinical variables contributed more to both frequency of ED use and hospitalization among patients with SRD, for whom medium to large effect sizes also emerged on MD related to ED use (RR 1.27) and on chronic physical illnesses related to hospitalizations (RR 1.87). Those with co-occurring SRD and chronic physical illnesses are recognized as frequent ED users [25] with high hospitalization rates [67, 68]. Studies have demonstrated that most ED visits and hospitalizations among patients with SRD or MD were for physical health reasons [6], while co-occurring SRD-MD are often associated with frequent use of acute care services [69, 70]. Longer duration of SRD increased the risk of developing co-occurring chronic physical illnesses [71], another explanation for frequent ED use and hospitalization. Finally, while behavioral addiction may provoke stress-related physical health problems (e.g. hypertension, insomnia, migraine) [72, 73], frequent ED use may be more strongly associated with suicidal behaviors in this group [74], resulting from bankruptcy, unemployment, interpersonal conflicts or other psychosocial problems [72, 75].

The third hypothesis was partially confirmed, as higher continuity of physician care was the only variable contributing to decreases in both ED use and hospitalization, while higher frequency of treatment (4+ interventions) in addiction treatment centers decreased the risk of hospitalization. However, more consultations with the usual GP and interventions provided in community healthcare centers produced the largest effect sizes (RR 1.48 for 4+ interventions provided in community healthcare centers related to ED use and RR 1.57 for 4+ consultations with the usual GP related to hospitalizations). As SRD are often chronic [76, 77], continuity of care is strongly recommended for reducing the risk of relapse [76, 78]. Some studies have found that higher continuity of care played a protective role against frequent ED use among patients with MD, including SRD [79, 80], as well as hospitalization [81]. High continuity of care is a key indicator of recovery for patients with SRD, particularly those with more health problems [78]. Decreased risk of hospitalization among patients who received more intensive interventions in addiction treatment centers suggests that these specialized services offer sufficient intensity of treatment to adequately address important health issues. However, such interventions may not resolve the crisis situations that lead patients with SRD to ED. The fact that frequency of ED use and hospitalization increased despite higher numbers of consultations with usual GP and psychiatrists as well as numbers of interventions in community healthcare centers seems to indicate that these services were insufficient or inadequate to prevent ED use and hospitalization among patients with multiple and complex needs. Community healthcare centers and specialized mental health services respond to the multiple biopsychosocial needs of highly vulnerable patients [70], whereas this study found that few intensive and diversified services were available.

Regarding sociodemographic variables, patients in the 12-17 year age group showed the largest effect sizes for both ED use (RR 1.51) and hospitalizations (RR 0.77). Younger patients with SRD are known to engage in higher-risk behaviors involving alcohol or drugs [82] compared with their older counterparts, making them more vulnerable to adverse outcomes, including overdose [82, 83]. They are frequently identified as victims of stigmatization by health professionals [84], leading them to prefer self-care or informal treatment [84] and to avoid accessing addiction treatment [85]. This may explain why patients under 45 years old in this study were more likely to use ED than the 45+ group. As well, increased hospitalization among those 45+ compared with the 25-44 and 12-17 age groups may be explained by the association between chronic physical illnesses, older age and SRD [86]. The preponderant use of ED and hospitalization among women with SRD versus men has been reported previously [9, 26, 46, 87]. Women are also more vulnerable to adverse outcomes associated with SRD [46, 88, 89], which may also explain their higher risk of acute care use in this study. Patients living in rural territories, where GP and outpatient services are frequently lacking, were also more likely to use ED and to be hospitalized [24]. Surprisingly, living in semi-urban versus urban areas decreased the risk of ED use but increased hospitalization rates. It is possible that primary care services were sufficiently robust in semi-urban territories to provide a viable alternative to ED, although the ED failed to prevent hospitalization among patients with more complex health problems. Finally, living in more materially deprived or unassigned areas increased the risks of ED use and hospitalization, possibly due to their strong association with unemployment, low income, food insecurity and homelessness among patients with SRD [57].