Study limitations

The W-SOHO study is the largest prospective observational study on the outcome of schizophrenia in an out-patient setting. However, there are several limitations that must be considered when discussing the results. First, although the 37 countries participating in the W-SOHO study belong to 6 regions of the world with different economic and cultural characteristics, the countries are not necessarily representative of these regions and some regions, such as East Asia, had a relatively small number of participants. In addition, the centres or investigators participating in the study in each country may not be representative of the whole country. Second, although sociodemographic and clinical characteristics were assessed in individuals participating and were taken into account in the analyses, we cannot rule out that different types of individuals with schizophrenia were enrolled in different countries, that there were other confounding variables not recorded in the study and that service contexts and residual confounding may be influencing the results. Third, we did not collect detailed information on the cultural environment of the participants, which could have influenced outcomes, and limits the exploration of the reasons for the regional differences. Fourth,

Table 4

Baseline factors associated with achieving clinical and functional remission during the 3-year follow-up for the W-SOHO sample (n = 11 078)^a

	Clinical remission			Functional remission
	Odds ratio	95% CI	P	Odds ratio	95% CI	P
North Africa and Middle East (v. South Europe)	2.82	2.19–3.64	< 0.0001	0.89	0.66–1.20	0.4293
Central and Eastern Europe (v. South Europe)	0.91	0.78–1.05	0.1893	0.71	0.59–0.86	0.0004
East Asia (v. South Europe)	1.87	1.37–2.55	< 0.0001	1.02	0.75–1.39	0.8843
Latin America (v. South Europe)	2.50	2.11–2.96	< 0.0001	2.14	1.77–2.59	< 0.0001
North Europe (v. South Europe)	0.79	0.69–0.89	0.0002	1.34	1.15–1.56	0.0002
Female (v. male)	1.28	1.15–1.42	< 0.0001	1.60	1.42–1.81	< 0.0001
Age at first treatment	0.99	0.98–0.99	< 0.0001	0.97	0.96–0.97	< 0.0001
Duration of illness	0.98	0.97–0.98	< 0.0001	0.96	0.96–0.97	< 0.0001
Never treated (yes v. no)	2.01	1.62–2.50	< 0.0001	1.50	1.21–1.86	0.0002
Alcohol misuse (yes v. no)	0.98	0.83–1.16	0.8319	0.86	0.68–1.07	0.1664
Substance misuse (yes v. no)	0.78	0.65–0.94	0.0083	1.06	0.84–1.34	0.5973
CGI-SCH score
Overall severity	0.75	0.69–0.82	< 0.0001	0.99	0.90–1.09	0.8388
Positive	0.91	0.87–0.95	< 0.0001	0.99	0.94–1.05	0.7742
Negative	0.79	0.74–0.83	< 0.0001	0.95	0.89–1.01	0.0991
Depressive	1.08	1.03–1.12	0.0008	1.07	1.02–1.13	0.0087
Cognitive	0.85	0.81–0.90	< 0.0001	0.96	0.91–1.02	0.1916
Hostile behaviours (yes v. no)	1.19	1.06–1.33	0.0028	0.91	0.79–1.05	0.1829
Spouse or partner (yes v. no)	1.35	1.20–1.51	< 0.0001	2.16	1.90–2.45	< 0.0001
Independent housing (v. dependent housing)	1.08	0.97–1.21	0.1786	6.00	5.22–6.89	< 0.0001
Paid employment (v. unemployed/unpaid)	1.47	1.27–1.69	< 0.0001	5.66	4.94–6.48	< 0.0001
Socially active (v. no social activities)	1.22	1.10–1.35	0.0002	1.50	1.32–1.71	< 0.0001

CGI-SCH, Clinical Global Impression – Schizophrenia scale.

Values are odds ratios (OR) obtained from the logistic regression model. An OR <1 indicates a lower likelihood of achieving remission.

Table 5

Baseline factors associated with achieving clinical remission during the 3-year follow-up for each of the six regionsa

	Odds ratio (95% CI)
	East Asia	North Africa and Middle East	Latin America	Central and Eastern Europe	North Europe	South Europe
Female (v. male)	1.26 (0.62–2.58)	1.78 (1.03–3.08)*	0.85 (0.61–1.17)	1.40 (1.09–1.78)*	1.09 (0.89–1.34)	1.46 (1.23–1.73)*
Age at first treatment	0.97 (0.93–1.01)	0.98 (0.95–1.02)	0.98 (0.96–1.00)	0.98 (0.97–1.00)*	0.99 (0.98–1.00)	0.99 (0.98–1.00)*
Duration of illness	1.00 (0.96–1.03)	0.98 (0.95–1.01)	0.96 (0.94–0.07)*	0.97 (0.96–0.99)*	0.98 (0.97–0.99)*	0.97 (0.97–0.98)*
Never treated (yes v. no)	2.67 (0.29–24.57)	1.51 (0.50–4.52)	2.12 (1.08–4.19)*	2.42 (1.23–4.75)*	2.77 (1.84–4.16)*	1.57 (1.12–2.12)*
Alcohol misuse (yes v. no)	0.14 (0.03–0.76)*	0.80 (0.33–1.91)	0.73 (0.42–1.26)	1.83 (1.11–3.00)*	0.97 (0.70–1.35)	0.91 (0.70–1.17)
Substance misuse (yes v. no)	2.28 (0.27–19.04)	0.96 (0.28–3.23)	1.13 (0.62–2.08)	0.61 (0.31–1.19)	0.64 (0.45–0.90)*	0.86 (0.66–1.12)
Suicide attempts in past (yes v. no)	0.71 (0.35–1.43)	0.72 (0.41–1.26)	0.92 (0.66–1.28)	1.16 (0.88–1.54)	1.05 (0.84–1.30)	0.85 (0.70–1.02)
CGI-SCH score
Overall severity	0.92 (0.49–1.73)	0.69 (0.46–1.03)	0.95 (0.74–1.22)	0.59 (0.47–0.74)*	0.67 (0.56–0,80)*	0.79 (0.70–0.90)*
Positive	0.74 (0.49–1.11)	1.07 (0.84–1.36)	0.90 (0.77–1.04)	0.92 (0.82–1.02)	0.87 (0.80–0.95)*	0.92 (0.86–0.99)*
Negative	0.71 (0.48–1.04)	0.76 (0.59–0.99)*	0.83 (0.71–0.97)*	0.84 (0.73–0.97)*	0.82 (0.74–0.91)*	0.75 (0.68–0.81)*
Depressive	1.10 (0.80–1.52)	0.95 (0.78–1.17)	1.02 (0.91–1.15)	1.25 (1.12–1.39)*	1.06 (0.97–1.15)	1.09 (1.01–1.17)*
Cognitive	0.76 (0.55–1.05)	0.85 (0.65–1.09)	0.84 (0.72–0.97)*	0.92 (0.80–1.05)	0.87 (0.78–0.97)*	0.85 (0.79–0.92)*
Hostile behaviours (yes v. no)	0.87 (0.42–1.79)	1.06 (0.63–1.77)	1.14 (0.83–1.57)	1.03 (0.76–1.39)	1.67 (1.32–2.13)*	1.09 (0.91–1.29)
Spouse or partner (yes v. no)	1.33 (0.64–2.75)	1.89 (0.93–3.82)	1.81 (1.21–2.70)*	1.32 (1.01–1.72)*	1.20 (0.97–1.50)	1.34 (1.09–1.65)*
Independent housing (v. dependent housing)	0.71 (0.34–1.50)	0.90 (0.44–1.81)	2.19 (1.39–3.45)*	1.06 (0.81–1.39)	1.10 (0.89–1.37)	0.98 (0.82–1.18)
Paid employment (v. unemployed/unpaid)	2.96 (0.78–11.21)	0.87 (0.43–1.75)	0.64 (0.41–1.01)	1.28 (0.91–1.80)	1.74 (1.34–2.25)*	1.77 (1.40–2.25)*
Socially active (v. no social activities)	1.32 (0.66–2.63)	1.39 (0.81–2.39)	1.37 (1.00–1.86)*	1.03 (0.80–1.33)	1.16 (0.93–1.44)	1.28 (1.08–1.51)*

CGI-SCH, Clinical Global Impression – Schizophrenia scale.

a Values are odds ratios (OR) obtained from the logistic regression model. An OR <1 indicates a lower likelihood of achieving remission.

* P < 0.05.

data were collected at 6-month intervals and limited information was gathered between assessment visits. Fifth, data were only collected over 3 years and, therefore, are unlikely to represent the full course of schizophrenia: some individuals may have experienced remission at a later time. Sixth, given the limitations of the ascertainment tools, our methods do not allow us to separate the effects of regional clinical practices (and therefore their ratings on the CGI-SCH) of the participating psychiatrists from the predictors of outcome analysed. Seventh, interrater reliability was not assessed given the large number of participating investigators. However, measures were chosen based on clarity and ease of use. Eighth, participants included in the analysis are those

Table 6

Baseline factors associated with achieving functional remission during the 3-year follow-up for each of the six regions

	Odds ratio (95% CI)
	East Asia	North Africa and Middle East	Latin America	Central and Eastern Europe	North Europe	South Europe
Female (v. male)	0.59 (0.33–1.05)	1.50 (0.77–2.90)	1.68 (1.22–2.31)*	1.18 (0.83–1.66)	1.41 (1.12–1.78)*	2.28 (1.81–2.86)*
Age at first treatment	0.99 (0.95–1.03)	1.01 (0.98–1.06)	0.97 (0.95–0.99)*	0.95 (0.93–0.97)*	0.95 (0.94–0.96)*	0.98 (0.97–0.99)*
Duration of illness	1.01 (0.98–1.05)	0.99 (0.95–1.03)	0.97 (0.95–0.98)*	0.95 (0.92–0.97)*	0.95 (0.93–0.96)*	0.96 (0.95–0.98)*
Never treated (yes v. no)	0.94 (0.16–5.41)	2.02 (0.59–6.92)	1.85 (1.15–2.97)*	2.80 (1.41–5.56)*	1.37 (0.93–2.03)	1.23 (0.82–1.84)
Alcohol misuse (yes v. no)	0.42 (0.07–2.58)	1.75 (0.56–5.45)	0.88 (0.49–1.58)	0.66 (0.30–1.45)	1.00 (0.68–1.48)	0.70 (0.47–1.04)
Substance misuse (yes v. no)	0.81 (0.10–6.49)	0.67 (0.14–3.27)	1.46 (0.81–2.64)	0.49 (0.15–1.62)	1.18 (0.79–1.76)	1.08 (0.73–1.60)
Suicide attempts in past (yes v. no)	1.05 (0.57–1.94)	0.84 (0.39–1.79)	0.94 (0.66–1.34)	1.14 (0.78–1.67)	1.05 (0.82–1.35)	0.92 (0.71–1.20)
CGI-SCH score
Overall severity	1.13 (0.69–1 87)	1.31 (0.81–2.11)	0.78 (0.61–0.99)*	1.00 (0.74–1.36)	0.87 (0.71–1.06)	1.07 (0.89–1.27)
Positive	0.88 (0.62–1.14)	0.99 (0.72–1.35)	1.09 (0.94–1.26)	0.88 (0.76–1.02)	0.98 (0.89–1.09)	1.02 (0.93–1.12)
Negative	0.82 (0.60–1.13)	0.84 (0.60–1.15)	1.13 (0.97–1.32)	0.81 (0.67–0.97)*	0.94 (0.83–1.06)	0.97 (0.86–1.09)
Depressive	0.97 (0.73–1.29)	1.18 (0.90–1.54)	1.06 (0.94–1.19)	1.10 (0.94–1.28)	1.05 (0.95–1.16)	1.14 (1.03–1.27)*
Cognitive	0.99 (0.73–1.35)	0.93 (0.66–1.32)	1.00 (0.87–1.16)	0.96 (0.80–1.16)	1.01 (0.89–1.14)	0.95 (0.86–1.05)
Hostile behaviours (yes v. no)	0.91 (0.49–1.71)	0.94 (0.48–1.81)	0.80 (0.58–1.10)	0.91 (0.60–1.37)	0.96 (0.73–1.27)	0.92 (0.72–1.18)
Spouse or partner (yes v. no)	0.88 (0.49–1.56)	3.61 (1.82–7.16)*	2.34 (1.67–3.28)*	2.29 (1.61–3.27)*	1.69 (1.33–2.15)*	2.69 (2.14–3.38)*
Independent housing (v. dependent housing)	2.70 (1.47–4.98)*	7.98 (4.07–15.66)*	5.52 (3.92–7.79)*	2.45 (1.70–3.54)*	6.47 (4.85–8.63)*	8.96 (6.96–11.52)*
Paid employment (v. unemployed/unpaid)	1.94 (0.99–3.78)	7.57 (3.73–15.36)*	3.54 (2.44–5.12)*	11.20 (7.86–15.96)*	6.29 (4.85–8.17)*	6.26 (4.85–8.07)*
Socially active (v. no social activities)	2.26 (1.17–4.38)*	1.06 (0.56–1.98)	1.41 (1.04–1.93)*	1.49 (1.03–2.15)*	1.96 (1.50–2.57)*	1.37 (1.08–1.73)*

CGI-SCH, Clinical Global Impression – Schizophrenia scale.

a. Values are odds ratios (OR) obtained from the logistic regression model. An OR < 1 indicates a lower likelihood of achieving remission.

* P < 0.05.

requiring a treatment change in routine clinical practice, which allowed us to study treatment outcomes but are obviously not representative of the overall patient population. Ninth, attrition was highest in the regions with the highest remission rates, which could explain some of the findings if attrition was higher in individuals with severe schizophrenia. Finally, our definition of clinical remission required a low level of symptoms for at least 6 months, consistent with the definition proposed by Andreasen. ^{Reference Andreasen, Carpenter, Kane, Lasser, Marder and Weinberger28} However, our definition of remission was based on the CGI-SCH, which is a valid but less specific measure of clinical severity than other scales such as the Positive and Negative Syndrome Scale (PANSS). ^{Reference Kay, Fiszbein and Opler29} Previous analyses have shown a good agreement between this and Andreasen’s definition. ^{Reference Haro, Ochoa, Gervin, Mavreas and Jones27}

Regional differences in outcomes

With these limitations in mind, the results of the W-SOHO study show that the clinical outcomes of schizophrenia seem to be worse in Europe compared with other regions. Remarkably, the regional differences were different for functional remission. The frequency of clinical remission was lower in the three European regions (60–65%) than in East Asia, Latin America, and North Africa and Middle East (79–84%). Participants living in the latter three regions had a much greater likelihood of achieving clinical remission than those living in South Europe. Moreover, the variations in clinical remission rates between regions were not accounted for by differences in baseline clinical and sociodemographic characteristics evaluated in the study. Economic development, cultural factors such as family support or other country characteristics may explain the differences. Differences in economic development are not necessarily translated to differences in the course of schizophrenia. For example, South Europe and Central and Eastern Europe show similar clinical remission rates but have different levels of economic development. Our findings support the earlier WHO studies reporting differences in outcomes between regions. ^{1–Reference Hopper and Wanderling6}

However, regional differences in functional remission followed a different pattern. Although it was more likely for individuals in Latin America to achieve functional remission compared with South Europe, there were no clear differences with East Asia or North Africa and Middle East. A new pattern emerged when compared with Central and Eastern Europe and North Europe; Central and Eastern Europe seemed to have a lower functional remission rate compared with South Europe, whereas North Europe tended to have a higher functional remission rate.

When comparing the descriptive and regression differences between the regions, we wanted to highlight that the logistic model showed that individuals with schizophrenia in Central and Eastern Europe were significantly less likely to achieve functional remission than those in South Europe. However, this was not detected in the descriptive analysis, probably due to the fact that some social functioning variables (independent housing and having a spouse/partner) are confounding variables.

Although this is somewhat speculative, differences in remission rates seem to be the result of economic, cultural and environmental factors more than differences in schizophrenic disorder. The same diagnostic criteria were applied in all regions and similarities in predictors of outcome were seen across the regions, which may indicate similar characteristics of the disorder. The reasons for the better clinical outcome in low- and middle-income countries are unknown but may be related to differences in the balance between treatment and vulnerability experienced by the individuals. ^{Reference Nuechterlein and Dawson30} Differences in functional remission between regions were mostly driven by differences in independent living and paid employment. Thus, these differences in functional remission rates may be influenced by differences in access to accommodation, the presence of rehabilitation services and social benefits, the development of specific policies for individuals with severe mental disorders and the level of societal stigma about mental illness.

Our results indicate that there are several other baseline predictors of outcome. We found that women were more likely to achieve remission compared with men. This is consistent with many reports that women with schizophrenia experience better outcomes than men. ^{Reference Grossman, Harrow, Rosen and Faull31,Reference Usall, Ochoa, Araya and Marquez32} Younger age, shorter duration of illness and no previous treatment for schizophrenia were also associated with a better chance of achieving remission, whereas substance misuse was associated with a lower chance of clinical remission, especially in North Europe. Our findings are consistent with systematic reviews and meta-analyses, which found that a shorter duration of untreated psychosis is associated with better symptomatic and functional outcomes in high-income and low- and middle-income countries. ^{Reference Marshall, Lewis, Lockwood, Drake, Jones and Croudace33,Reference Farooq, Large, Nielssen and Waheed34} Although it has been reported that comorbid substance misuse is highly prevalent in schizophrenia and associated with poorer clinical outcomes, ^{Reference Volkow35} most of the evidence is based on studies in Western countries; the prevalence and impact of substance use disorders among people with schizophrenia in low- and middle-income countries has not been well studied. The frequency of alcohol or substance misuse was low in the overall W-SOHO population at baseline compared with other samples, ^{Reference Buhler, Hambrecht, Loffler, an der Heiden and Hafner36} and its role as an independent predictor of remission varied across regions. Further work on substance misuse as a factor influencing outcome of schizophrenia across a wide range of countries is needed.

Higher symptom severity at baseline in terms of positive, negative, cognitive and overall symptoms was associated with less likelihood of clinical remission, but the severity of these symptoms was not significantly associated with functional remission. However, a higher level of depressive symptoms at baseline was associated with a higher likelihood of achieving both clinical and functional remission in the logistic regression analyses of the overall W-SOHO population. This supports previous findings that high levels of depressive symptoms at baseline predict favourable short-term outcomes in individuals with schizophrenia. ^{Reference Oosthuizen, Emsley, Roberts, Turner, Keyter and Keyter37} However, other researchers have found that people with depressive symptoms have poorer long-term functional outcomes. ^{Reference Conley, Ascher-Svanum, Zhu, Faries and Kinon38}

Social functioning variables were important prognostic factors for remission in all regions. Participants with a spouse/partner, in paid employment and who were socially active at baseline were more likely to achieve clinical and functional remission, supporting previous findings that better baseline social functioning is associated with recovery (when defined as achieving symptomatic plus functional remission). ^{Reference Wunderink, Sytema, Nienhuis and Wiersma17} The direction of the causality, however, may not be clear. For example, although working appears to help people recover from schizophrenia, ^{Reference Warner39} the converse may also be true, i.e. individuals who maintain work are those who have a good prognosis. In the W-SOHO population at baseline, the frequency of paid employment was low (19%), ranging from 16% in East Asia to 23% in North Europe. This is similar to the employment rates reported for people with schizophrenia in Western countries, ^{Reference Marwaha and Johnson40} which vary both between and within countries. However, fully dissecting the role of social functioning on outcomes in schizophrenia is complicated because clinical changes can have an impact on social functioning. ^{Reference Warner39} There are also high rates of stigma and discrimination against people with schizophrenia across countries, ^{Reference Thornicroft, Brohan, Rose, Sartorius and Leese41} which can have an impact on their social functioning. ^{Reference Yanos, Roe, Markus and Lysaker42}

The W-SOHO study has shown that there are cross-national differences in outcomes among out-patients with schizophrenia. Outcomes in terms of remission seem to be better for people living in low- and middle-income regions, especially Latin America. In general, the regional variation in outcome persisted even after adjusting for clinical and sociodemographic variables at baseline. Several predictive factors were identified suggesting that outcome differences are related to cultural and environmental factors rather than to differences in the disorder itself.