Psychotic disorders, both affective (e.g., bipolar disorder) and non-affective (e.g., schizophrenia and related disorders), are associated with high levels of disability and poor outcomes [1–4]. While its prevalence is relatively low, non-affective psychosis falls within the top 20 leading causes of disability globally [1], responsible for 13.4 (95% UI 9.9-16.7) million years lived with disability [2]. In terms of outcomes, approximately one in seven patients (median 13.5%) experience long-term (2-year) recovery [3]. Suicide rates are also markedly higher among patients with psychotic disorders when compared to the general population [4]. Psychotic disorders place a substantial burden on the healthcare system. In Ontario, the direct costs of psychosis care were approximately $2.1 billion CAD in 2012 alone [5]. Indirect costs of psychotic disorders are also extremely high [6].

Early psychosis intervention (EPI) plays a key role in improving functional and clinical outcomes among those with first episode psychosis [7–10] and is widely recommended upon diagnosis [11]. In Ontario, health quality standards recommend that EPI should be available for all newly diagnosed patients, ideally within the first two weeks [12]. As of present, there are approximately 60 EPI programs across Ontario. However, there are many more EPI cases than are treated in these programs. In one Ontario catchment area, Anderson and colleagues found a rate of non-affective psychosis incidence of 33.3 per 100,000 per year, compared to the 18.8 per 100,000 per year who received EPI treatment in the same region, a nearly 2-fold difference in incidence vs. access to treatment [13]. While a proportion of the untreated cases may not have met the criteria for EPI program entry, these data suggest that not all individuals diagnosed with a psychotic disorder are receiving EPI.

Sociodemographic factors such as age, sex, socioeconomic status, and immigration status play a role in both the incidence of psychotic disorder [14–17] and access to EPI treatment [18]. Males, recent immigrants, those living in lower income areas, and those with a previous mental disorder are at greater risk of incident psychosis [14–17]. Accurate information on the regional incidence of first-episode psychosis is essential to close the gap between population-based incidence and treated incidence, which will, in turn, improve population-based outcomes in this population.

Ontario, the most populous Canadian province (~16 million), is made up of six health regions which vary in their sociodemographic makeup and healthcare needs [19]. The Toronto health region (population ~3.3 million) serves the most densely populated area of the province and contains a high concentration of specialised services. In contrast, the North East (population ~600,000) and North West (population ~240,000) health regions serve largely rural areas with higher proportions of Indigenous and Francophone residents. These regions experience healthcare inequities, such as those related to mortality and chronic disease, when compared to Ontario as a whole [20]. There are also lower rates of primary care attachment and higher specialist wait times in the Northern regions. Provincial psychiatrist supply varies from 62.7 per 100,000 in the Toronto area to 7.2 per 100,000 in lower-supply (largely Northern) areas [21]. These care inequities are perhaps driven by factors such as low population density, poor infrastructure, lack of trained providers, language barriers, and challenging weather conditions [20]. Previous research has shown regional variation in healthcare provision and clinical outcomes within the province [22–25], yet there is little evidence on the regional incidence of first-episode psychosis, both alone and by relevant sociodemographic factors, and variation in access to care [26].

Due to the importance of proper allocation of EPI services and Ontario’s regional variation in health outcomes, we conducted a population-based retrospective cohort study aiming: (1) to understand the variation in demographic and regional characteristics of individuals who may be suitable for EPI services in the province of Ontario, and (2) to determine healthcare utilisation after first diagnosis of a psychotic disorder.

We identified 425,724 individuals with first episode psychosis, 347,653 of which were outside of the study accrual window. Another 33,883 individuals were excluded because they could not be linked across databases (n = 11), did not meet the age criteria (n = 33,146), were non-Ontario residents (n = 519), were not eligible for OHIP (n = 182), or had problematic or missing data (n = 25), resulting in a final sample of 44,188 Ontarians aged 12-50 who were diagnosed with first episode psychosis between January 1^st 2017 and December 31^st 2021 (Supplementary Material Figure 1). Of these individuals, 41,257 had non-affective psychosis and 3,058 had affective psychosis (127 were diagnosed with both). Table 1 shows the baseline characteristics for both the non-affective and affective psychosis cohorts. Overall, affective psychosis disorders were most common in females (54.7%) and those aged 14-20 (24.5%), while non-affective psychosis disorders were more common in males (60.3%) and those aged 21-25 (20.4%). Higher numbers of incident cases were observed in the lowest income quintile, regardless of psychosis type (Table 1). One year prior to their incident psychosis case, approximately 30% of individuals had 2+ MHA-related visits to a family physician, regardless of psychosis type (Table 0).

Figures 1a and 1b display heat maps of incidence by health region for affective and non-affective cases, respectively. The overall rate of affective psychosis was 6.84 per 100,000 (Supplementary Material Table 3). The overall rate of non-affective psychosis was 92.24 per 100,000. Non-affective psychosis rates were highest in the North West and North East regions at 167.44 per 100,000 and 134.67 per 100,000, respectively, while affective psychosis rates were highest in the North East region at 14.23 per 100,000 (Figures 1a, 1b). In the North West region, the rate of affective psychosis was 3.51 per 100,000. Across health regions, incidence of cases diagnosed in an outpatient setting was approximately 2–3 times higher than that of those diagnosed in an inpatient setting. The highest rates of outpatient diagnoses took place in the North West and North East regions.

Table 2 displays the 6-month MHA-related outcomes after incident case diagnosis for the affective and non-affective cohorts, by health region. In the non-affective cohort, most patients had at least two outpatient psychiatrist visits and at least one outpatient primary care visit within the 6 months post-diagnosis. Across the province, 21.5% of the cohort had at least one ED visit and 26.5% had at least one hospitalisation during this time period. Compared to the provincial average, non-affective psychosis patients in the North East and North West health regions had a smaller proportion of outpatient psychiatrist and primary care visits and a larger proportion of ED visits and hospitalisations. The Toronto region had the highest proportion of outpatient psychiatrist visits (74.4% with one or more) and the Central region had the highest proportion of outpatient primary care visits (62.7% with one or more). Across the province, 0.8% of the non-affective psychosis cohort died within the first 6 months of diagnosis. Overall, similar results were seen for the affective cohort. However, the most common length of stay for hospitalisation was 8-14 days for the affective cohort and 1–7 days for the non-affective cohort.

Figure 2 displays the adjusted odds ratios (aOR) of 6-month MHA ED visit and hospitalisation, and adjusted rate ratios (aRR) of 6-month outpatient visits for incident non-affective psychosis cases, by health region (adjusted for sociodemographic characteristics; Supplementary Material Tables 4, 5). Compared to the referent health region of Toronto, both the North East and North West regions had higher odds of hospitalisation (aOR [95% CI] 1.25 [1.1,1.41], 1.16 [1.01,1.35], respectively) and ED visits (aOR 1.01 [0.90,1.12], 1.07 [0.82,1.24], respectively). The odds of individuals in the Central health region having an MHA-related ED visit within 6-months of their incident case diagnosis was 20% less (aOR 0.80 [0.74,0.87]) than those residing in Toronto. For outpatient visits, highest rates were seen across the health regions for primary care, followed by paediatricians (Figure 2). The lowest rates were seen for psychiatrist visits, especially in the North West (aRR 0.29 [0.27,0.30]) and North East (aRR 0.40 [0.38,0.41]) regions. No regional differences were observed for 6-month MHA-related health care use among patients with affective psychosis, except for primary care visits. Individuals had lower rates of primary care visits in most regions when compared to Toronto, with North East having the lowest rates (aRR 0.85 [0.74,0.96]) and Central having the highest (aRR 1.096 [1.007-1.193]) (Supplementary Material Table 5). Adjusting for previous health service utilisation in addition to sociodemographic characteristics did not produce meaningfully different results (Supplementary Tables 6, 7).

Our cohort contained 44,188 Ontarians aged 12-50 diagnosed with first-episode psychosis between 2017 and 2021. Within this population, we observed significant regional variation in incidence rates. Rates were higher in the Northern and Western health regions compared to the provincial average, ranging as high as 167 per 100,000 (non-affective psychosis; North West region). In addition to regional variation in incidence, we also identified differences in healthcare utilisation during the first six months following psychotic illness onset. Namely, the odds of hospitalisation and ED visits were higher among patients in the Northern health regions, while the Central and Toronto regions had higher rates of outpatient visits. Conversely, regions with high hospitalisation rates had the lowest per capita psychiatrist visit rates, whereas the regions with the lowest hospitalisation rates had the highest per capita psychiatrist visit rates.

To successfully plan for future EPI programs in Ontario, it is essential to understand regional needs based on the actual number of cases requiring intervention using a systematic, population-based approach. Our population-based study found significant and meaningful regional variation in the incidence of first episode psychosis. In 2022, Rotenberg and colleagues published an investigation of incident psychosis cases in Ontario, also identifying notable regional variation. In general, the authors found elevated rates of psychotic disorder in the metropolitan areas of Toronto, Southeastern Ontario (e.g., Kingston), Southwestern Ontario (e.g., Guelph, Windsor), and Northern Ontario (Sudbury) compared to non-metropolitan rural areas [26]. Overall, improving our understanding of these regional incidence variations will help address the gap between psychosis incidence (demand) and EPI service availability (supply) across the province. Regional variation was also present in health care utilisation within six months post-diagnosis. Patients in the Northern health regions had lower rates of mental health-related outpatient psychiatrist and primary care specialist visits and higher odds of mental health-related ED visits and hospitalisations. Other regions, such as the Central and Toronto regions, had higher rates of outpatient mental health visits. These patterns are important to consider from a healthcare system planning perspective, where resources could be saved by diverting care to outpatient services. Our finding of relatively low use of outpatient care is consistent with findings from the Southwest region of Ontario, where prior studies have shown less than half of individuals who appear to be eligible for EPI services receive them [18].

As discussed above, we observed differences in both incidence of disease and post-diagnosis care for Northern populations compared to the rest of the province. There is a need for more psychiatrists and family physicians in these regions to help increase outpatient follow-up and decrease burden on acute care centers. Rotenberg et al. explored physician follow-up after hospitalisation among patients with schizophrenia, finding that the greatest proportion of individuals who received timely follow-up from both psychiatrists and physicians (psychiatrists/general practitioners) often lived in health regions with the largest metropolitan centers (Toronto, Ottawa, and Hamilton) [28]. There is also drastic variation in the availability of psychiatrists, with the vast majority practicing in Toronto and very few in the Northern regions [21]. Various strategies have been proposed to improve access to general and specialized outpatient care in the Northern regions, such as establishing health centres with interprofessional teams, implementing telemedicine or mobile clinics in remote areas, and providing education and training to retain Northern-born professionals [20].

Incidence of psychosis was high in the Northern health regions. The proportion of individuals who identify as Indigenous in these regions is also high [19]. Other jurisdictions have reported higher rates of psychosis among Indigenous populations compared to non-Indigenous populations. In a New Zealand cohort study, rates of psychosis were higher among the Indigenous population compared to the non-Indigenous population after adjusting for age, sex, deprivation, and urbanicity [29]. Reasons for the increased incidence of psychotic illnesses in regions with large Indigenous populations should be a focus of further study. In terms of access to mental health care, Indigenous individuals face barriers including lack of awareness of available services, lack of access to services (including long wait times, lack of Indigenous providers, and language barriers), mistrust of formal services, and stigma surrounding services [30]. In Canada, a growing body of literature has demonstrated inequities in mental health and substance use rates among Indigenous peoples compared to the general population [31–33]. We also identified differences in rates of psychosis between the two Northern health regions, which could be related to the proportion of residents identifying as Indigenous in the North East (13.6%) compared to the North West (26.4%) [19].

A strength of our study was its use of population-based health administrative data, which allowed us to gain a better understanding of regional incidence in Ontario. These data may be useful for informing future EPI program resource allocation. This study also had several limitations. Administrative health data may have missing or inaccurate values; however, these issues are likely non-differential. In addition, while we used a pre-existing algorithm to identify psychosis cases, this technique may not be entirely sensitive or specific [27]. Lastly, factors that we could not measure may also play a role in first-episode psychosis. For example, we were unable to look at the role of race/ethnicity and data on income were only available at the area level.

We observed substantial regional variation in incidence of first-episode psychosis in Ontario. Specifically, rates were highest in the Northern and Western health regions. We also identified variation by other sociodemographic characteristics, as well as regional variation in post-diagnosis health service utilisation. Regions with high hospitalisation rates had the lowest per capita psychiatrist visit rates, whereas the regions with the lowest hospitalisation rates had the highest per capita psychiatrist visit rates. Additional research is needed to further understand regional variation and to isolate the effects of specific factors. This type of work can be applied to other jurisdictions, promoting the use of evidence-based decisions for EPI policy. These data can also be applied to the future development of EPI services in Ontario, helping to better align EPI resources based on measured regional need.