Introduction
The risk of developing schizophrenia is strongly linked to genetics, with heritability estimated at ~80% based on twin studies (Hilker et al., Reference Hilker, Helenius, Fagerlund, Skytthe, Christensen, Werge, Nordentoft and Glenthøj2018). Candidate genes associated with schizophrenia have been identified (Farrell et al., Reference Farrell, Werge, Sklar, Owen, Ophoff, O’Donovan, Corvin, Cichon and Sullivan2015); however, results have not been documented in meta-analyses (Johnson et al., Reference Johnson, Border, Melroy-Greif, de Leeuw, Ehringer and Keller2017). More recently, the genetics of schizophrenia have been studied using genome-wide association studies (GWAS) that analyze millions of common genetic variations or single nucleotide polymorphisms (SNPs) to establish their association with schizophrenia (Legge et al., Reference Legge, Santoro, Periyasamy, Okewole, Arsalan and Kowalec2021). Based on the GWAS studies, it is possible to calculate polygenic risk scores for schizophrenia (PRSSCZ). A significant breakthrough identified 83 new loci and 128 SNP differences linked to schizophrenia (Consortium, 2014). Further, the most extensive GWAS study on schizophrenia identified 287 loci, focusing on genes like GRIN2A, SP4, STAG1, and FAM120A, which are expressed in central nervous system neurons and act as excitatory or inhibitory factors, playing a pivotal role in neuronal functions (Trubetskoy et al., Reference Trubetskoy, Pardiñas, Qi, Panagiotaropoulou, Awasthi, Bigdeli, Bryois, Chen, Dennison, Hall, Lam, Watanabe, Frei, Ge, Harwood, Koopmans, Magnusson, Richards, Sidorenko and O’Donovan2022).
GWAS studies on schizophrenia have shown that PRSSCZ can explain between 7.7% and 33% of the variance in liability to schizophrenia (Lee et al., Reference Lee, DeCandia, Ripke, Yang, Sullivan, Goddard, Keller, Visscher and Wray2012; Legge et al., Reference Legge, Santoro, Periyasamy, Okewole, Arsalan and Kowalec2021; Purcell et al., Reference Purcell, Wray, Stone, Visscher, O’Donovan, Sullivan and Sklar2009). Among those who have developed the illness, a high PRSSCZ is associated with more severe symptoms (Fanous et al., Reference Fanous, Zhou, Aggen, Bergen, Amdur, Duan, Sanders, Shi, Mowry, Olincy, Amin, Cloninger, Silverman, Buccola, Byerley, Black, Freedman, Holmans and Levinson2012; Jonas et al., Reference Jonas, Lencz, Li, Malhotra, Perlman, Fochtmann, Bromet and Kotov2019) and a more severe disease course (Jonas et al., Reference Jonas, Lencz, Li, Malhotra, Perlman, Fochtmann, Bromet and Kotov2019; Meier et al., Reference Meier, Agerbo, Maier, Pedersen, Lang, Grove, Hollegaard, Demontis, Trabjerg, Hjorthøj, Ripke, Degenhardt, Nöthen, Rujescu, Maier, Werge, Mors, Hougaard, Børglum and Mattheisen2016). Research also indicates that high levels of PRSSCZ increase susceptibility to other mental health conditions, including depression, anxiety, panic symptoms, and bipolar disorder, either in childhood or adulthood (Crouse et al., Reference Crouse, Carpenter, Iorfino, Lin, Ho, Byrne, Henders, Wallace, Hermens, Scott, Wray and Hickie2021; Richards et al., Reference Richards, Horwood, Boden, Kennedy, Sellers, Riglin, Mistry, Jones, Smith, Zammit, Owen, O’Donovan and Harold2019; Zheutlin et al., Reference Zheutlin, Dennis, Karlsson Linnér, Moscati, Restrepo, Straub, Ruderfer, Castro, Chen, Ge, Huckins, Charney, Kirchner, Stahl, Chabris, Davis and Smoller2019). In terms of lifestyle adversities, high PRSSCZ has been associated with a higher risk of smoking (Wang et al., Reference Wang, Lai, Lee, Su, Chen, Hsiao, Yang, Liu, Tsai and Kuo2020), alcohol consumption (Zheutlin et al., Reference Zheutlin, Dennis, Karlsson Linnér, Moscati, Restrepo, Straub, Ruderfer, Castro, Chen, Ge, Huckins, Charney, Kirchner, Stahl, Chabris, Davis and Smoller2019), and sleep disturbances (Reed et al., Reference Reed, Jones, Hemani, Zammit and Davis2019).
While PRSSCZ is known to predict multiple adverse outcomes, research on its association with subjective well-being has remained scarce. There are, however, two studies examining the association of PRSSCZ with positively toned traits such as creativity and social support. The studies have not only reported a correlation between high PRSSCZ and greater creativity (Power et al., Reference Power, Steinberg, Bjornsdottir, Rietveld, Abdellaoui, Nivard, Johannesson, Galesloot, Hottenga, Willemsen, Cesarini, Benjamin, Magnusson, Ullén, Tiemeier, Hofman, van Rooij, Walters, Sigurdsson and Sefansson2015) but also an association between high PRSSCZ and lower perceived social support in middle adulthood (Saarinen et al., Reference Saarinen, Hietala, Lyytikäinen, Hamal Mishra, Sormunen, Lavonius, Kähönen, Raitakari, Lehtimäki and Keltikangas-Järvinen2023).
There are many reasons why it is essential to study the relationship between PRSSCZ and subjective well-being. First, an increasing proportion of the population has been purchasing genetic testing to be aware of their genetic risks for disorders. In 2022, a total of 197,779 genetic tests were made available globally (Halbisen & Lu, Reference Halbisen and Lu2023). Furthermore, stigmatization is known to be strong in the context of psychotic disorders (Gronholm, Thornicroft, Laurens, & Evans-Lacko, Reference Gronholm, Thornicroft, Laurens and Evans-Lacko2017), even among those at risk for psychosis but without the disorder (Colizzi, Ruggeri, & Lasalvia, Reference Colizzi, Ruggeri and Lasalvia2020). If genetic testing indicated an increased genetic risk for schizophrenia, some individuals might disclose this information to their employer, even in the absence of clinical symptoms (Lawrence et al., Reference Lawrence, Friesen, Brucato, Girgis and Dixon2016). To prevent further stigmatization and negative stereotypical expectations among those aware of their genetic risk for psychoses, it is important to examine also potential positively toned outcomes of polygenic risk scores. Second, a large portion of research on PRSSCZ focuses on register-based data related to socioeconomic outcomes (e.g. unemployment status) or healthcare visits – that is, register-based measures in which individuals with high PRSSCZ are not directly heard. In particular, it is not known how they subjectively experience their life satisfaction. This is especially relevant, as people often seek medical care only when they perceive a decline in their quality of life rather than when the first symptoms appear (Gartland, Long, & Skevington, Reference Gartland, Long and Skevington2019). Thus, subjectively experienced quality of life plays a crucial role in help-seeking behavior.
In general, well-being measures can be roughly divided into two dimensions: affective well-being (the frequency of positive vs. negative effects) and cognitive well-being (cognitive evaluations of one’s life) (Diener, Suh, Lucas, & Smith, Reference Diener, Suh, Lucas and Smith1999). In our study, optimism referred to affective well-being and life satisfaction and self-acceptance to cognitive well-being. Life satisfaction is one of the strongest single indicators of well-being (Linley et al., Reference Linley, Maltby, Wood, Osborne and Hurling2009), reflecting the relationship between one’s life expectations and the life that has been realized (Pavot Reference Pavot and Diener1993). Optimism refers to a disposition toward positive future expectations and is associated with better recovery from stressful life situations (Carbone & Echols, Reference Carbone and Echols2017) and greater resilience in adverse circumstances (Gallagher, Long, & Phillips, Reference Gallagher, Long and Phillips2020). Finally, self-acceptance refers to a tendency to acknowledge one’s own qualities without the need for change or self-blame and predicts one’s ability to let go of things that have been lost (Prigerson & Maciejewski, Reference Prigerson and Maciejewski2008).
We investigated whether PRSSCZ predicts subjective well-being in terms of life satisfaction, optimism, and self-acceptance. We used the prospective, population-based sample of the Young Finns Study (YFS). Moreover, because the onset of a psychotic illness is known to cause a decrease in well-being (Fervaha et al., Reference Fervaha, Agid, Takeuchi, Foussias and Remington2016), we also examined whether PRSSCZ is related to subjective well-being among those not diagnosed with non-affective psychotic disorders. Since research on PRSSCZ and well-being is scarce (only two studies resulted in contradictory findings), we did not formulate a specific research hypothesis.
Methods
Participants
The Young Finns Study (YFS) is an ongoing prospective study that started in 1980 (baseline assessment). Follow-ups have been conducted in 1983, 1986, 1989, 1992, 1997, 2001, 2007, 2011/2012, and 2018–2020. Originally, a total of 4320 participants were invited (born in 1962, 1965, 1968, 1971, 1974, or 1977), and 3596 of them participated in the baseline study. The sample was designed to include a population-based sample of non-institutionalized Finnish children, representative of the most crucial sociodemographic factors. In practice, the sampling was conducted in collaboration of five Finnish universities with medical schools (i.e. Universities of Helsinki, Turku, Tampere, Oulu, and Kuopio). The design and methods of the YFS study are described in more detail elsewhere (Raitakari et al., Reference Raitakari, Juonala, Rönnemaa, Keltikangas-Järvinen, Räsänen, Pietikäinen, Hutri-Kähönen, Taittonen, Jokinen, Marniemi, Jula, Telama, Kähönen, Lehtimäki, Akerblom and Viikari2008).
The Declaration of Helsinki has been followed throughout the study. The study design has been approved by the ethical committees of all the Finnish universities conducting the study. All the participants or their parents (participants aged <18 years) provided informed consent before participation.
From the full sample of 3596 participants, we included all the participants who had data available on variables under investigation: that is, participants who had been genotyped in 2011 for calculating their polygenic risk for schizophrenia and who had data available on subjective well-being measures in at least one measurement point: optimism in 2011, life satisfaction in 2007 and/or 2011, and self-acceptance in 1997, 2011, and/or 2012. Additionally, the analyses were adjusted in a stepwise manner for covariates: age, sex, quality of early family environment (1980 and/or 1983), adulthood socioeconomic factors (2011), and adulthood health behaviors (2001, 2007, and/or 2011). This resulted in a stepwise decrease in the sample size when adding covariates in each model. Thus, the final sample size ranged from 878 to 1866 participants in the main analyses.
Measures
Polygenic risk score for schizophrenia (PRSSCZ)
Polygenic risk for schizophrenia was calculated based on the most recent GWAS study on schizophrenia (Trubetskoy et al., Reference Trubetskoy, Pardiñas, Qi, Panagiotaropoulou, Awasthi, Bigdeli, Bryois, Chen, Dennison, Hall, Lam, Watanabe, Frei, Ge, Harwood, Koopmans, Magnusson, Richards, Sidorenko and O’Donovan2022). In the calculation, we used the PRS-CS method (Ge et al., Reference Ge, Chen, Ni, Feng and Smoller2019), which infers posterior SNP effect sizes under continuous shrinkage (CS) priors using GWAS summary statistics and an external LD reference panel. The results of the late available GWAS on schizophrenia (Trubetskoy et al., Reference Trubetskoy, Pardiñas, Qi, Panagiotaropoulou, Awasthi, Bigdeli, Bryois, Chen, Dennison, Hall, Lam, Watanabe, Frei, Ge, Harwood, Koopmans, Magnusson, Richards, Sidorenko and O’Donovan2022) were used as SNP summary statistics, and HapMap 3 EUR was used as an external LD reference (Altshuler et al., Reference Altshuler, Gibbs, Peltonen, Altshuler, Gibbs, Peltonen, Dermitzakis, Schaffner, Yu, Peltonen, Dermitzakis, Bonnen, Altshuler, Gibbs, de Bakker, Deloukas, Gabriel, Gwilliam, Hunt and McEwen2010).
Subjective well-being
Life satisfaction was measured on a scale adapted from the Operation Family Study Questionnaire (Makkonen et al., Reference Makkonen, Rönkä, Timonen, Valvanne and Österlund1981) in 2007 and 2011 (participants were 30–49 years old). The scale has been used also previously (Hintsa et al., Reference Hintsa, Kivimäki, Elovainio, Keskivaara, Hintsanen, Pulkki-Råback and Keltikangas-Järvinen2006; Keltikangas-Järvinen & Heinonen, Reference Keltikangas-Järvinen and Heinonen2003; Makkonen et al., Reference Makkonen, Rönkä, Timonen, Valvanne and Österlund1981). The self-assessment questionnaire contains three items measuring life satisfaction in three roles: as a spouse, employee, and parent (e.g. ‘Rate yourself as a parent’). Responses were given on a 5-point Likert scale (1 = very dissatisfied, 5 = very satisfied). In addition, the response scale included a response alternative of 0 (‘Does not apply to the respondent’) if, for example, the participant had no children or was not in working life. These values were coded as missing values. We calculated an average score (i.e. a life satisfaction index) over the three items for all the participants who had responded to at least one of the items in both follow-ups. Thus, higher values of the life satisfaction index indicated higher life satisfaction. Life satisfaction scores seemed to be moderately stable over time (r = 0.46 between the measurement years).
Optimism was measured in 2001 (when the participants were 24–39 years old) with the Revised Life Orientation Test-Revised (LOT) questionnaire (Scheier, Carver, & Bridges, Reference Scheier, Carver and Bridges1994). The questionnaire contains six self-rated items (e.g. ‘I always have an optimistic attitude to the future’), including three reversed items (e.g. ‘If something can go wrong with me, it certainly will’). The items were answered on a 5-point Likert scale (0 = strongly disagree, 4 = strongly agree). In this study, an average score was calculated over the items for all the participants who had responded to at least half of the items. The LOT questionnaire has shown adequate validity and reliability (Scheier et al., Reference Scheier, Carver and Bridges1994). In our dataset, its internal reliability was good (Cronbach’s α = 0.78) in accordance with previous studies (Heinonen, Räikkönen, & Keltikangas-Järvinen, Reference Heinonen, Räikkönen and Keltikangas-Järvinen2005; Serlachius et al., Reference Serlachius, Pulkki-Råback, Juonala, Sabin, Lehtimäki, Raitakari and Elovainio2017).
Self-acceptance was measured with the Self-Acceptance versus Self-Striving Scale which is a subscale of the Temperament and Character Inventory (TCI) (Cloninger Reference Cloninger, Przybeck and Švrakić1994). The scale was conducted in the follow-ups of 1997, 2001, and 2012 (when participants were aged 20–50 years). The scale contains 11 statements that were responded to a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). The items included also reversed statements (e.g. ‘I wish I was smarter than anyone else’) which were reversely scored. Thus, higher scores on the scale referred to higher self-acceptance. In this study, the scale had high internal reliability (Cronbach’s α = 0.82–0.83 in different measurement years). The scores of self-acceptance were relatively stable over the 15-year follow-up (r = 0.59–0.67 between the measurement years).
Psychiatric diagnoses
In our supplementary analyses, we excluded participants who had been diagnosed with a non-affective psychotic disorder. Participants’ psychiatric diagnoses that had required hospital care were collected up to 2017 from the Care Register for Health Care (also known as the Finnish Hospital Discharge Register). In 2017, the participants were 40–55 years old and, thus, older than the typical onset age of schizophrenia (Ochoa et al., Reference Ochoa, Usall, Cobo, Labad and Kulkarni2012). In the register, diagnoses were given by the prevailing diagnostic classification (ICD-8, ICD-9, or ICD-10). ICD diagnoses were converted to DSM-IV diagnoses, and this conversion is described in more detail elsewhere (Sormunen et al., Reference Sormunen, Saarinen, Salokangas, Telama, Hutri-Kähönen, Tammelin, Viikari, Raitakari and Hietala2017). Diagnoses were grouped into the following categories: (1) non-affective psychotic disorders, (2) substance-related disorders, (3) affective disorders (mood and anxiety disorders), and (4) personality disorders. Participants with several psychiatric diagnoses were categorized into only one of the categories in the following priority order: non-affective psychoses (DSM-IV 295, 297, 298), personality disorders (DSM-IV 301), affective disorders (mood and anxiety disorders, DSM-IV 296, 300, 311), and substance-related disorders (DSM-IV 291, 303, 292, 304, 305). The register is shown to cover most psychiatric diagnoses (Sund, Reference Sund2012) and has been used previously for research purposes (Suvisaari, Haukka, Tanskanen, & Lönnqvist, Reference Suvisaari, Haukka, Tanskanen and Lönnqvist1999). In this study, we used a dichotomous variable of non-affective psychoses (0 = no, 1 = yes).
Covariates
The quality of the early family environment was assessed using three cumulative risk scores for (1) stressful life events, (2) adverse socioeconomic circumstances, and (3) unfavorable emotional atmosphere. Information on the early family environment was prospectively assessed in 1980/1983 with questionnaires presented to the parents. The cumulative risk score for stressful life events included changing residence, parental divorce, parental death, parental hospitalization in the past 12 months, and the child’s hospitalization due to illness or injury. The cumulative score for adverse socioeconomic circumstances included parents’ low occupational status, low educational, unstable employment situation, and overcrowded living conditions. The cumulative score for unfavorable emotional atmosphere included emotional distance between the child and parent, parental tolerance toward the child, strict disciplinary actions toward the child, parental life dissatisfaction, parents’ mental disorder, and parents’ frequent alcohol intoxication. When calculating the cumulative risk scores, each single variable (e.g. family income) was first standardized within each age cohort (mean = 0, SD = 1), and then an average score of these standardized variables was calculated. Thus, regardless of the original measurement scale, each factor was equally weighted in the cumulative risk scores (e.g. emotional distance between child and parent and parental life dissatisfaction were similarly weighted). The cumulative risk scores have been used also previously and reported there with further details (Saarinen et al., Reference Saarinen, Lyytikäinen, Hietala, Dobewall, Lavonius, Raitakari, Kähönen, Sormunen, Lehtimäki and Keltikangas-Järvinen2022).
Adulthood socioeconomic factors (2011) included annual income and educational level. Annual income was assessed with a 13-point scale (1 = < 5000 €, 13 = >60 000 €). Educational level included three classes (1 = comprehensive school including the first nine school years; 2 = occupational school or high school; 3 = academic level such as university or college). Each socioeconomic variable was added as a separate time-invariant covariate to the analyses.
Adulthood health behaviors included alcohol consumption, physical activity, and smoking that were assessed in 2001, 2007, and 2011. Alcohol consumption was assessed with self-report questions on consumption of 1/3 cans or bottles of beer, glasses (12 cl) of wine, and 4 cl shots of liquor or strong alcohol during the past week. The total alcohol consumption index has been used also previously and described in detail elsewhere (Juonala et al., Reference Juonala, Viikari, Kähönen, Laitinen, Taittonen, Loo, Jula, Marniemi, Räsänen, Rönnemaa and Raitakari2009). Physical activity was assessed by inquiring about participants’ frequency and intensity of leisure-time physical activity and participation in organized sports exercises. The index of physical activity has been described elsewhere (Rovio et al., Reference Rovio, Yang, Kankaanpää, Aalto, Hirvensalo, Telama, Pahkala, Hutri-Kähönen, Viikari, Raitakari and Tammelin2018). Smoking was assessed with a 5-point Likert scale ranging from 1 = (daily smoking) to 5 (never smoked). We formed a dichotomous variable (1 = daily smoking, 0 = not daily smoking). For each domain of health behaviors (alcohol consumption, physical activity, and smoking), an average score over the follow-ups was calculated for all participants who had data available in at least one measurement year.
Statistical analyses
Statistical analyses were conducted using SPSS (version 29). First, pairwise correlations between the study variables were examined using Pearson/Spearman correlation coefficients. In addition, attrition over the prospective follow-up was analyzed using independent samples t-tests and χ2 tests. In attrition analyses, ‘included participants’ consisted of the participants who were included in at least one statistical analysis, while ‘non-participants’ referred to those who were excluded from all the analyses due to missing values in the study variables.
Next, linear regression analyses were used to investigate whether PRSSCZ predicts life satisfaction or optimism. A total of three models were made: Models 1 were adjusted for sex and age, Models 2 also for the cumulative risk scores of early family environment, and Models 3 also for adulthood socioeconomic factors (annual income, occupational level, educational level) and health behaviors (alcohol consumption, physical activity, and smoking status). These covariates were selected because it is known that early family environment (Flèche, Lekfuangfu, & Clark, Reference Flèche, Lekfuangfu and Clark2021), socioeconomic factors (Pinquart & Sörensen, Reference Pinquart and Sörensen2000), and health behaviors (Grant, Wardle, & Steptoe, Reference Grant, Wardle and Steptoe2009) are associated with well-being and may act as potential confounders. The assumptions of linear regression analyses were fulfilled: the residuals followed an approximately normal distribution and the variances of the residuals were approximately homogeneous. We also calculated values of R squared change for PRSSCZ, indicating the percentage of the variation in subjective well-being that was explained by PRSSCZ.
Finally, we predicted the trajectory of self-acceptance by PRSSCZ using growth curve models. The models contain both fixed and random effects. Fixed effects can be interpreted similarly to regression coefficients and, in this study, were estimated for PRSSCZ and all the covariates. In addition, age2 was included as a fixed effect since self-acceptance was assumed to possibly have curvilinear changes over age. Random effects included variance of repeated measures and variance of intercept (between-individual variance in the constant term). The structure of the random effects covariance matrix was defined to be scaled identity. Regarding control variables, we added covariates in a stepwise manner The results were illustrated with graphs drawn with Stata SE version 18.
Results
Sample statistics
The descriptive statistics of the study variables are shown in Table 1. Participants were on average 41.3 years old (in 2011), and 52.1% of them were biologically male. Most of the participants had a second-level education (42.0%). Approximately 1.2% of the participants had been diagnosed with a non-affective psychotic disorder. On average, the participants seemed to be rather satisfied with their lives (mean = 3.96). In addition, self-acceptance seemed to increase over age. The pairwise correlations between the study variables can be found in Supplementary Table 1.
Table 1. Descriptive statistics of the study variables
Attrition analyses showed that included participants and non-participants (i.e. those excluded from analyses due to missing values) did not differ in terms of subjective well-being such as life satisfaction (p = 0.63), optimism (p = 0.53), or self-acceptance (p = 0.57–0.86). Also, there was no attrition bias in age (p = 0.11), sex (p = 0.14), health behaviors (p = 0.10–0.98), or cumulative risk scores for stressful life events (p = 0.68) or unfavorable emotional family atmosphere (p = 0.76). Participants had, however, slightly lower scores of PRSSCZ (p < 0.05) and were less likely to have been diagnosed with a non-affective psychotic disorder (p < 0.001) than non-participants. In addition, participants had slightly higher socioeconomic conditions both in childhood (p < 0.001) and adulthood in terms of higher education (p < 0.001) and higher income (p < 0.05) when compared with non-participants.
Main analyses
The results of the linear regression analyses when predicting life satisfaction are presented in Table 2. To summarize, PRSSCZ did not predict life satisfaction in Model 1 (p = 0.15), Model 2 (p = 0.18), or Model 3 (p = 0.37). Table 3 shows the results of linear regression analyses when predicting optimism. Again, PRSSCZ did not statistically significantly predict optimism in any of the three models (p = 0.82 in Model 1, p = 0.71 in Model 2, and p = 0.54 in Model 3).
Table 2. Results of linear regression analyses when life satisfaction was predicted by the polygenic score for schizophrenia (PRSSCZ)
Note: B refers to an unstandardized coefficient.
a Female as the reference group.
b Cumulative risk scores of early family environment.
Table 3. Results of linear regression analyses when optimism was predicted by the polygenic score for schizophrenia (PRSSCZ)
Note: B refers to an unstandardized coefficient.
a Female as the reference group.
b Cumulative risk scores of early family environment.
As an additional analysis, we investigated whether sex or age could modify the associations of PRSSCZ with life satisfaction or optimism. For this, the interaction term of PRSSCZ*sex or PRSSCZ*age was separately added to the model. None of the interactions were statistically significant (p = 0.12–0.80), indicating that the association of PRSSCZ with life satisfaction and optimism was non-significant in both sexes and over the age range of our sample (30–49 years).
The results of the growth curve models are presented in Table 4. Age (p < 0.001) and age squared (p < 0.001) were significant predictors of self-acceptance, indicating that self-acceptance increased in a curvilinear way over age (p < 0.001). PRSSCZ did not predict the trajectory of self-acceptance in any of the three models (p = 0.23 in Model 1, p = 0.40 in Model 2, and p = 0.60 in Model 3). As an additional analysis, we examined sex or age modified association of PRSSCZ with self-acceptance. For this, the interaction term of PRSSCZ*sex or PRSSCZ*age2 was added to the model. None of the interactions were statistically significant (p = 0.61–0.99), indicating that the association of PRSSCZ with a trajectory of self-acceptance was non-significant in both sexes and over the age range of our sample (20–50 years). The results of the growth curve models are illustrated in Figure 1.
Table 4. Results of growth curve models when predicting self-acceptance by the polygenic risk score for schizophrenia
Note: B refers to an unstandardized regression coefficient.
a Female as the reference group.
b Cumulative risk scores of early family environment.
Figure 1. Estimated means of self-acceptance over age separately for individuals with low (−1 SD) or high (+ 1 SD) scores of polygenic risk for schizophrenia (PRS).
Sensitivity analyses
As additional analyses, we repeated the main analyses in a subsample of participants who had not been diagnosed with non-affective psychotic disorder. This was done since the onset of a psychotic illness is known to commonly cause a significant drop in subjective well-being (Fervaha et al., Reference Fervaha, Agid, Takeuchi, Foussias and Remington2016). Thus, here we examined whether PRSSCZ is associated with subjective well-being (life satisfaction, optimism, and self-acceptance) in individuals without a history of non-affective psychosis.
The results can be found in Supplementary Tables 2–4. To summarize, the main results were mostly replicated: in any model, PRSSCZ did not predict optimism (p = 0.73 in Model 1, p = 0.61 in Model 2, and p = 0.49 in Model 3), life satisfaction (p = 0.15 in Model 1, p = 0.19 in Model 2, and p = 0.40 in Model 3), or self-acceptance (p = 0.28 in Model 1, p = 0.43 in Model 2, and p = 0.67 in Model 3).
Finally, we also reran the analyses so that all participants with psychiatric diagnoses required hospital care were excluded from the sample (i.e. non-affective psychoses, mood disorders, substance use disorders, personality disorders). The results were replicated: PRS was not associated with optimism (p = 0.628 in Model 1, p = 0.563 in Model 2, and p = 0.436 in Model 3), life satisfaction (p = 0.065 in Model 1, p = 0.090 in Model 2, and p = 0.252 in Model 3), or self-acceptance (p = 0.286 in Model 1, p = 0.464 in Model 2, and p = 0.672 in Model 3).
Power estimation
Given the null results, it was necessary to estimate statistical power in our analyses. Therefore, we conducted post hoc power analyses to determine the minimum effect sizes detectable in our sample. Based on the sample sizes (n = 1502 for life satisfaction, n = 1414 for optimism, and n = 1866 for self-acceptance), a recommended power estimate of 80%, and a statistical significance threshold of p = 0.05, the minimum detectable squared partial correlations of PRSSCZ were 0.0051, 0.0055, and 0.0042 for life satisfaction, optimism, and self-acceptance, respectively. These correspond to effect sizes of 0.0052, 0.0056, and 0.0042, respectively. Taken together, our sample size allowed us to detect associations between PRSSCZ and well-being indicators with effect sizes as small as 0.0042–0.0052.
Discussion
This study investigated, for the first time, the association between the polygenic risk for schizophrenia (PRSSCZ) and subjectively experienced well-being. Contrary to our hypotheses, PRSSCZ did not predict aspects of subjective well-being such as life satisfaction, optimism, and self-acceptance. We found no associations between PRSSCZ and well-being in any model with stepwise inclusion of covariates (age, sex, early family environment, adulthood socioeconomic factors, adulthood health behaviors), nor when including or excluding individuals who had developed non-affective psychotic disorders. Additionally, PRSSCZ was not associated with subjective well-being in either sex or in any age between 20 and 50 years. In sum, our study showed that individuals with a higher PRSSCZ reported being just as satisfied, optimistic, and self-accepting as those with a lower PRSSCZ.
While previous evidence has shown associations between PRSSCZ and other mental disorders, including depression, anxiety, panic symptoms, bipolar disorder (Crouse et al., Reference Crouse, Carpenter, Iorfino, Lin, Ho, Byrne, Henders, Wallace, Hermens, Scott, Wray and Hickie2021; Richards et al., Reference Richards, Horwood, Boden, Kennedy, Sellers, Riglin, Mistry, Jones, Smith, Zammit, Owen, O’Donovan and Harold2019; Zheutlin et al., Reference Zheutlin, Dennis, Karlsson Linnér, Moscati, Restrepo, Straub, Ruderfer, Castro, Chen, Ge, Huckins, Charney, Kirchner, Stahl, Chabris, Davis and Smoller2019), as well as adverse health behaviors such as alcohol consumption (Zheutlin et al., Reference Zheutlin, Dennis, Karlsson Linnér, Moscati, Restrepo, Straub, Ruderfer, Castro, Chen, Ge, Huckins, Charney, Kirchner, Stahl, Chabris, Davis and Smoller2019) and sleep disturbances (Reed et al., Reference Reed, Jones, Hemani, Zammit and Davis2019), we found no association between PRSSCZ and subjective well-being. Since it is well-known that experienced life satisfaction largely depends on the standards individuals set for their lives (Pavot Reference Pavot and Diener1993), individuals with a higher PRSSCZ may possibly set more modest life expectations and find satisfaction from smaller and more attainable goals. Our results also align with previous reports that psychological ill-being and well-being constitute distinct dimensions (Ryff et al., Reference Ryff, Dienberg Love, Urry, Muller, Rosenkranz, Friedman, Davidson and Singer2006), indicating that the presence of mental symptoms may not necessarily reduce one’s sense of well-being.
Previous studies have reported a decrease in subjective well-being in individuals at familial risk for psychosis (Ellersgaard et al., Reference Ellersgaard, Gregersen, Ranning, Haspang, Christiani, Hemager, Burton, Spang, Søndergaard, Greve, Gantriis, Jepsen, Mors, Plessen, Nordentoft and Thorup2020; Sin et al., Reference Sin, Murrells, Spain, Norman and Henderson2016) and also after the onset of a psychotic disorder such as schizophrenia (Fervaha et al., Reference Fervaha, Agid, Takeuchi, Foussias and Remington2016). We found that individuals at polygenic risk for schizophrenia do not experience lower levels of life satisfaction, optimism, or self-acceptance. This may have many explanations. First, a crucial factor contributing to the reduced well-being of psychotic patients seems to be stigmatization (Degnan, Berry, Humphrey, & Bucci, Reference Degnan, Berry, Humphrey and Bucci2021) that, in turn, requires awareness of the disorder or an elevated risk for it. Unlike individuals with either familial risk or symptoms of psychosis, individuals with a high PRSSCZ may not necessarily be aware of their risk for schizophrenia. Second, the offspring of psychotic patients are known to be exposed to illness-related adversities, such as parent relapses, difficulties in maintaining daily routines, or being emotionally present (da Silva et al., Reference da Silva, Peixoto, Martin, Galera, Vedana, de Freitas and Zanetti2022; Strand, Boström, & Grip, Reference Strand, Boström and Grip2020). Those with a high PRSSCZ, on the other hand, may not necessarily have close family members with psychotic disorders and may not encounter similar illness-related sources of stress.
A relevant question is whether our null results might result from limited statistical power. According to our power analyses, our sample size allowed us to detect associations between PRSSCZ and well-being indicators with effect sizes as small as 0.0042–0.0052, corresponding to minimum detectable squared partial correlations of 0.0042–0.0055, respectively. If there had been an association with an effect size smaller than that, it would be reasonable to question whether such an association holds any practical significance. Additionally, our sample has enabled the detection of statistically significant associations with psychosocial outcomes in our previous studies: for example, individual differences in PRSSCZ have explained variance in magical thinking (Saarinen et al., Reference Saarinen, Lyytikäinen, Hietala, Dobewall, Lavonius, Raitakari, Kähönen, Sormunen, Lehtimäki and Keltikangas-Järvinen2022), social support from close networks (Saarinen, Hietala, et al., Reference Saarinen, Hietala, Lyytikäinen, Hamal Mishra, Sormunen, Lavonius, Kähönen, Raitakari, Lehtimäki and Keltikangas-Järvinen2023), and accelerated epigenetic aging in the interaction of social dispositions (Saarinen, Hietala, et al., Reference Saarinen, Marttila, Mishra, Lyytikäinen, Raitoharju, Mononen, Sormunen, Kähönen, Raitakari, Hietala, Keltikangas-Järvinen and Lehtimäki2023). Taken together, while the risk of type II error should be considered, we argue that our null results are unlikely to be solely due to a lack of statistical power.
Regarding limitations, there was participant drop-out over the follow-up from the baseline measurement in 1980, resulting in a sample size of 1866 participants in our study. Our attrition analyses, however, did not identify any systematic drop-out with respect to most of the study variables. That is, participants and non-participants (i.e. those excluded from the analyses due to missing values) did not differ in indicators of subjective well-being, age, sex, health behaviors, or most qualities of early family environment. However, included participants were less likely to have been diagnosed with a non-affective psychotic disorder, had slightly lower PRSSCZ scores, and were living in slightly more advantaged socioeconomic conditions when compared with those who dropped out. Drop-out has been a well-documented issue in other prospective studies as well. When examining participant drop-out, previous reports on prospective datasets have generally found that it does not cause significant bias in results (de Graaf et al., Reference de Graaf, Bijl, Smit, Ravelli and Vollebergh2000; Tambs et al., Reference Tambs, Rønning, Prescott, Kendler, Reichborn-Kjennerud, Torgersen and Harris2009). Researchers have concluded that ‘differential loss to follow-up rarely affects estimates of association’ (Saiepour et al., Reference Saiepour, Najman, Ware, Baker, Clavarino and Williams2019).
The concept of subjective well-being encompasses many different definitions, making it challenging to establish a unified definition. First, some studies have included the absence of negative affect, such as sadness or envy, in their well-being measurements. However, we did not include assessments of negative affect, as they may confound with affective disorders that are known to correlate with PRSSCZ (Richards et al., Reference Richards, Horwood, Boden, Kennedy, Sellers, Riglin, Mistry, Jones, Smith, Zammit, Owen, O’Donovan and Harold2019). Second, while there have been attempts to assess subjective well-being through neural correlates, such as ‘brain fingerprints’ (Jung et al., Reference Jung, Pae, An, Bang, Choi, Cho and Lee2022; Saarimäki et al., Reference Saarimäki, Gotsopoulos, Jääskeläinen, Lampinen, Vuilleumier, Hari, Sams and Nummenmaa2016), this method appears to require further development before it can produce reliable estimates of emotional well-being. Third, while some studies have evaluated well-being on the basis of external measures, such as wealth, recreational activities, or access to healthcare. A key strength of our study is that we were able to assess well-being based on individuals’ own subjective experiences over many follow-ups.
In a broader context, previous research has focused on the associations of PRSSCZ with a range of negative developmental pathways, including mental disorders, lifestyle challenges, domains of reduced functioning, and an accumulation of other adversities. This study provides a novel positive perspective by demonstrating that individuals with high PRSSCZ are, on average, as satisfied with their well-being as those with a lower PRSSCZ.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291725000911.
Data availability statement
The Cardiovascular Risk in Young Finns (YFS) dataset comprises health-related participant data, and their use is therefore restricted under the regulations on professional secrecy (Act on the Openness of Government Activities, 612/1999) and on sensitive personal data (Personal Data Act, 523/1999, implementing the EU data protection directive 95/46/EC). Due to these legal restrictions, the data from this study cannot be stored in public repositories or otherwise made publicly available. However, data access may be permitted on a case-by-case basis upon request. Data sharing outside the group is done in collaboration with the YFS group and requires a data-sharing agreement. Investigators can submit an expression of interest to the chairman of the publication committee (Prof. Mika Kähönen, Tampere University, Finland, [email protected]).
Funding statement
This study has been financially supported by the Emil Aaltonen Foundation (grant 220255). The Young Finns Study has been financially supported by the Academy of Finland: grants 356405, 322098, 286284, 134309 (Eye), 126925, 121584, 124282, 129378 (Salve), 117797 (Gendi), and 141071 (Skidi); the Social Insurance Institution of Finland; Competitive State Research Financing of the Expert Responsibility area of Kuopio, Tampere and Turku University Hospitals (grant X51001); Juho Vainio Foundation; Paavo Nurmi Foundation; Finnish Foundation for Cardiovascular Research; Finnish Cultural Foundation; The Sigrid Juselius Foundation; Tampere Tuberculosis Foundation; Emil Aaltonen Foundation; Yrjö Jahnsson Foundation; Signe and Ane Gyllenberg Foundation; Diabetes Research Foundation of Finnish Diabetes Association; EU Horizon 2020 (grant 755320 for TAXINOMISIS and grant 848146 for To Aition); European Research Council (grant 742927 for MULTIEPIGEN project); Tampere University Hospital Supporting Foundation; Finnish Society of Clinical Chemistry; the Cancer Foundation Finland; pBETTER4U_EU (Preventing obesity through Biologically and bEhaviorally Tailored inTERventions for you; project number: 101080117); CVDLink (EU grant no. 101137278) and the Jane and Aatos Erkko Foundation.
Conflict of interest
The authors declare none.
Ethical statement
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Open access
