Sjors Maarten Maria Lange

Course trajectories of subjective quality of life during follow-up

Figure 1 shows the pattern of transitions between high and low SQOL status between baseline and 5-year follow-up. 47% of participants reported a high SQOL status at baseline and 56% at 5-year follow-up. Stability of SQOL status (i.e., either high or low SQOL status at both baseline and at follow-up) was seen in 72%, while 28% changed from low to high SQOL status or vice versa. Figure 2 shows the course trajectories of total SQOL-change score and SQOL subdomain change scores during follow-up. Total SQOL-change scores showed considerable movement in individual SQOL trajectories over time, with 36% of patients reporting a positive change score and 20% reporting a negative change score. The remaining 44% did not show a significant change score.

Regarding the change scores of the SQOL subdomains, different course trajectories were found, with the highest proportion of participants reporting an improvement in the subdomains ‘daily life’ and ‘personal circumstances’ (for both subdomains 33.3%). The largest proportion of participants reporting a deterioration over time was seen in the sub-domain ‘satisfaction with physical and mental health’ (28%).

Fig.1 Movement between high (MANSA score ≥5) and low (MANSA score < 5) SQOL-status between baseline and follow-up Fig. 2 Histograms of total SQOL-change score and change score of satisfaction with daily life, physical and mental health, and personal circumstances Predictors of change scores between baseline and follow-up In univariate regression analyses, higher age at onset, lower depressive symptomatology, lower baseline SQOL score, and lower baseline SQOL subdomain scores were associated with SQOL-change scores at follow-up (p < 0.10) (Table 1). In multivariable, linear regression analyses (data not shown), after stepwise introduction of all predictor variables, higher age of onset remained significantly associated with SQOL-change score (Beta 0.23; p = 0.03). None of the other putative predictors was associated with SQOL-change score during 5-year follow-up. All analyses were adjusted for baseline SQOL: higher SQOL-score at baseline was negatively associated with SQOL-change score (Beta − 0.47; p < 0.001). Next, when conducting separate regression analyses for each SQOL subdomain (data not shown), with change scores of these subdomains serving as outcome variable, higher age at onset was associated with change score of satisfaction with physical and mental health (Beta 0.22; p = 0.04) and with change score of satisfaction with personal circumstances (Beta 0.25; p = 0.02) during 5-year follow-up. In all three SQOL subdomains, baseline scores predicted higher change score of that particular SQOL subdomain over time (satisfaction with daily life: Beta − 0.55 (p < 0.001); satisfaction with physical and mental health: Beta − 0.48 (p < 0.001); satisfaction with personal circumstances: Beta − 0.44 (p < 0.001)). Discussion Chapter 3 In this study, examining 5-year course trajectories of subjective quality of life in older patients with SSD, approximately half of the participants reported a high SQOL status. Concerning SQOL-change score, 56% of participants had a clinically relevant SQOL-change during 5-year follow-up. Next, we identified three SQOL subdomains, with the sub-domains of ‘personal circumstances’ and ‘daily life’ showing the highest proportion of improvement. Of all predictor variables, only two variables showed significant associations with SQOL-change score during follow-up. Higher age of onset was significantly associated with total SQOL-change score and the change scores of the subdomains ‘physical and mental health’ and ‘personal circumstances.’ Higher baseline SQOL scores were negatively associated with total SQOL-change score and change scores of all SQOL subdomains during 5-year follow-up. Since approximately half of the persons reported high SQOL-status at baseline and/or at follow-up, with 37% reporting high SQOL-status at both assessments, and 19% switching from a low to a high SQOL-status during follow-up, findings indicate that during 5-year follow-up, a substantial proportion of older persons with SSD report satisfaction with their quality of life. Previous cross-sectional studies in older patients with SSD also found relatively high levels of SQOL (Folsom et al., 2009; Bankole et al., 2007; Hoertel et al., 2019). During 5-year follow-up, we found positive change scores (indicating a significant improvement of total SQOL) in 36%, negative change scores (indicating a significant deterioration) in 20%, and no change of SQOL in 44% of older persons with SSD. The findings were remarkably similar in comparing these course trajectories with the New York City longitudinal study (NYC study) mentioned above. In both studies, around one-third (36% in the current study versus 33% in the NYC study) of patients reported a positive change of SQOL over time, no change was seen in, respectively, 20% and 24%, while almost half (respectively, 44% vs. 43%) of patients reported no clinically meaningful change in their SQOL evaluation (Cohen et al., 2017). The findings of both longitudinal studies indicate that although a large proportion of older SSD patients is stable in their SQOL and deterioration is possible, improvement of SQOL over time is an achievable goal in a sizable number of older SSD patients. Next, we identified three subdomains of SQOL and demonstrated differential course trajectories for total SQOL and these three subdomains. These findings suggest that whereas overall SQOL may be primarily stable, underlying subdomains may either deteriorate or improve. Differential course trajectories of SQOL subdomains were also found by others. For example, Jeste et al. (Jeste et al., 2011) demonstrated that patients with SSD show different physical, cognitive, and psychosocial aging trajectories, with improved psychosocial trajectories despite physical and cognitive deteriorating trajectories. Cohen et al. (Cohen et al., 2017) reported no significant group differences in SQOL subscale scores between baseline and follow-up. However, percentages of clinically relevant effect size changes for SQOL subdomains ranged from 51 to 60%. Hence, especially in an aging population, differentiation of SQOL into subdomains may be of clinical relevance (Boyer et al., 2013; Petkari et al., 2019). Identifying those SQOL subdomains that have shown to be prone to improvement or deterioration over time could provide valuable information for older SSD patients’ treatment approaches. Since differential course trajectories of SQOL subdomains were demonstrated, we further explored course trajectories of underlying individual SQOL items in post hoc analyses. For example, the subdomain ‘physical and mental health’ consisted of two items: satisfaction with mental health and satisfaction with physical health. Post-hoc exploration of these items' course trajectories showed that 36% of participants reported improved satisfaction with mental health, whereas 48% reported a deterioration of their physical health satisfaction. Hence, also within subdomains, underlying items may show different course trajectories. These findings are in line with previous reports, showing different course trajectories of mental and physical SQOL (Folsom et al., 2009; Brink et al., 2020). The paradoxical improvement of mental functioning despite a deterioration of physical health while getting older was previously demonstrated in both SSD patients and patients with other psychiatric diagnoses, as well as in the general population (Priebe et al., 2010; Solomon et al., 2010; Stone et al., 2010). These findings indicate that improvement of mental health remains an achievable goal in older SSD patients. On the other hand, close observation and active treatment of physical health should remain an important goal in the later stages of patients lives. It is well conceivable that subjective standards people use to answer questions about SQOL-change over time (Plagnol, 2010). Older SSD patients may become more adapted to the impact of illness, which could change the effects of specific SQOL subdomains on overall SQOL evaluation (Boyer et al., 2013; Plagnol, 2010). It could well be that younger SSD patients evaluate their SQOL in light of day-to-day positive symptoms and their ability to adapt to these symptoms, but that older SSD patients have developed coping strategies, thereby giving less weight to positive symptoms in their SQOL evaluation. Instead, for older SSD patients, factors covering personal circumstances like religiousness, personal safety, and finance may become more important in evaluating their SQOL. Along another line of thought, it has been suggested that improvement of SQOL in older SSD patients could also partially be explained through a survivor effect (Jeste et al., 2011), with the possibility that higher-functioning patients may be overrepresented in cohorts of older persons with SDD. Although multiple predictor variables of higher SQOL in older SSD patients were identified in cross-sectional investigations, the longitudinal NYC study found only modest correlations between a limited number of predictor variables and SQOL at follow-up (Cohen et al., 2017). In the current investigation, identifying associations between predictor variables and SQOL-change score during follow-up, we found a small but positive association between higher age at onset and SQOL-change score. This finding implies, for participants in our cohort, a higher age of onset predicts a positive change of SQOL over time in a substantial portion of participants. We also found a small and negative association between baseline SQOL score and SQOL-change score over time. These associations between age at onset and SQOL and baseline and SQOL were reported before (Cohen et al., 2017; Kao et al., 2010; Karow et al., 2014; Rotstein et al., 2018). However, since only two putative predictors reached significance, findings should be interpreted with caution and warrant replication. The findings of this study need to be reviewed in the context of the following strengths and limitations. The present report is the first longitudinal study of older SSD patients employing a catchment area-based design that included both community-living and institutionalised patients, regardless of age of onset of SSD. However, we do not have any information about our patients’ SQOL evaluation between both assessments, limiting insight into pleiomorphic course trajectories during the 5 years of follow-up. Although in the present study, drop-out over time was relatively low (n = 15 (20%) of participants, who were still alive at follow-up assessment, did not participate), our baseline cross-sectional study suffered from non-consent with 109 out of 172 eligible patients (62%) participating. Due to the relatively small number of patients included, we could not conduct a confirmatory factor analysis to validate our findings of the calibration sample (Petkari et al., 2019). Because the aim of the current study was the assessment of change of SQOL during follow-up, only associations between predictor variables and change of SQOL during follow-up were assessed, and not predictors of higher SQOL at follow-up. This approach could be a reason for the failure to uncover significant associations between previously identified predictor variables and change of SQOL. Finally, considering the relatively high number of persons with high SQOL-status, a possible ceiling effect of the MANSA may have obscured any significant longitudinal effect, next to the possibility of regression to the mean. For future research, a Tobit regression model could avoid this effect (J. Tobit, 1958). To conclude, we demonstrated heterogeneous course trajectories for overall SQOL and underlying SQOL subdomains. Insight into the course trajectories of these domains may improve in-depth understanding of SQOL in SSD patients. Despite the deterioration of physical health (partially inevitably related to aging), improvement of SQOL is achievable in older SSD patients. To optimally support the improvement of SQOL subdomains, close observation and treatment of, for example, physical health should be an essential focus in the care of older SSD patients. Getting older with SSD patients does not imply a stable-end-state of SQOL. Current findings emphasise the importance of an optimistic and active treatment approach for older SSD patients. With indications that overall SQOL can improve in late-life SSD, SQOL subdomains that are sensitive to improvement or deterioration should be further explored in future research fueling the focus of treatment and care in older SSD patients.