Clinical Pearls · 2026

The Reported Strength of Association Between Symptoms and Quality of Life Depends on How Quality of Life Is Defined and Measured

Psychiatric symptoms and quality of life in schizophrenia

Meta-analysis of 56 studies, 190 effect sizes

Studies pooled

190

Effect sizes extracted

Symptom classes

Quality of life indicators

1966–2005

Literature searched

The question:

How strongly do psychiatric symptoms relate to quality of life in schizophrenia, and why does the reported strength vary across studies? Reported outcome: Pooled correlations with 95% confidence intervals (CIs).

The analysis drew on 56 studies, 52 published and 4 unpublished, found in a literature search running from 1966 to 2005, from which 190 effect sizes (ie, Pearson’s r) were extracted or computed. Symptoms were grouped into three classes and quality of life into four indicators, which the authors drew from a review of the measurement literature. Later work has reported consistent findings (Tolman and Kurtz, 2012; Chan et al, 2026).

Quality of life indicators

Subjective

Satisfaction with one’s own social and material wellbeing.

Objective

Observable social and material wellbeing, such as income or number of close friends.

General wellbeing

Satisfaction with life as a whole.

Health-related

Health and functional status.

Pooled Association With Composite Quality of Life

Analysis 1 of 3Overall pooled estimates across studies

How composite quality of life was derived. Composite quality of life was derived by averaging the quality of life indicators reported within each study, giving one composite effect size (Pearson’s r) per study. It is not a fifth indicator.

Positive symptoms43 studies · n = 3,998

−0.2095% CI −0.17 to −0.23

Negative symptoms44 studies · n = 4,114

−0.2595% CI −0.22 to −0.28

General psychopathology50 studies · n = 5,106

−0.3495% CI −0.31 to −0.36

0−0.10−0.20−0.30−0.40−0.50

Greater symptom severity was associated with poorer quality of life across all three symptom classes, and the pooled associations were modest.

General psychopathology accounted for a limited share of the variation

≤12%of variance in composite quality of life

<8%in longitudinal studies

Most of the variation was not accounted for by the symptom measures analysed.

The strongest of the three pooled estimates

−0.34

General psychopathology, the largest of the three pooled estimates.

It is also the broadest domain, covering a heterogeneous group of symptoms rather than one. The analysis did not establish which of them accounted for the association.

Pooled does not mean constant

A pooled correlation summarizes findings across studies. It does not represent one fixed relationship in every study or patient.

The next two analyses show where the reported strength differed: by quality of life indicator, and across study characteristics.

Clinical Pearls · 2026

Strength of Association With Symptoms Varied Across Quality of Life Indicators

Findings pooled across studies for specific quality of life indicators. These are distinct from the composite quality of life estimates in Analysis 1.

Pooled Associations by Symptom Class and Quality of Life Indicator

Analysis 2 of 3

Symptom classPearson’s r, 95% CI, and evidence base.	Subjective	Objective	General wellbeing	Health-related
Positive symptoms	−0.15 −0.11 to −0.19 19 studies · n = 2,256	−0.18 −0.13 to −0.24 15 studies · n = 1,150	−0.08 −0.03 to −0.14 12 studies · n = 1,198	−0.26 −0.21 to −0.31 15 studies · n = 1,256
Negative symptoms	−0.12 −0.08 to −0.16 20 studies · n = 2,359	−0.47 −0.42 to −0.51 16 studies · n = 1,207	−0.14 −0.08 to −0.20 11 studies · n = 1,154	−0.29 −0.24 to −0.34 15 studies · n = 1,256
General psychopathology	−0.29 −0.26 to −0.33 25 studies · n = 2,997	−0.26 −0.20 to −0.32 13 studies · n = 1,019	−0.27 −0.22 to −0.31 13 studies · n = 1,434	−0.42 −0.37 to −0.46 15 studies · n = 1,389

In the analysis, variation across quality of life indicators was significant for negative symptoms, Q_B(3) = 54.08, P < 0.0001, and general psychopathology, Q_B(3) = 18.88, P < 0.001. In the analysis restricted to studies reporting a single indicator, variation between indicators was not significant for positive symptoms, Q_B(3) = 1.15, ns. Repeated with all studies, that comparison was significant, Q_B(3) = 21.80, P < 0.001. The authors note the restricted analysis may have been biased. The four positive-symptom estimates themselves all exclude zero.

Pattern of the point estimates

The strongest associations were generally seen with health-related quality of life, except for negative symptoms. Individual cells within a row were not compared with one another.

The largest value reflects the instrument used

The large −0.47 between negative symptoms and objective quality of life was strongly influenced by studies using the Quality of Life Scale, which returned −0.52 against −0.22 for studies using other objective measures. The authors attributed its size to that instrument rather than to objective quality of life as a construct. The same instrument effect appeared in first-episode psychosis: −0.52 with the Quality of Life Scale versus −0.23 with other scales (Watson et al, 2018).

Health-related measures often contain symptom items

Some health-related quality of life measures include symptom content, which the authors note may contribute to stronger observed associations with symptoms. They recommend caution when interpreting these relationships.

Clinical Pearls · 2026

Association Strengths Differed Across Study Design, Setting, and Stage of Illness

Moderator analyses.

Each study was coded for its design, to explore the attenuating effect of examining the relationship over time; for its treatment setting, because different factors may influence quality of life for inpatients and for outpatients; and for stage of illness, because symptoms may affect first-episode and chronic patients differently. Every moderator was then tested on the composite index rather than on any single indicator.

Moderator Analysis of Quality of Life Associations

Analysis 3 of 3

Between-study moderator analysesComposite quality of life

Study designLongitudinalCross-sectional

General psychopathology

−0.28−0.35

Q_B(1) = 4.57, P < 0.05

Negative symptoms

−0.20−0.27

Q_B(1) = 3.90, P = 0.05, marginal

The two estimates did not differ for positive symptoms. Q_B(1) = 0.15, ns

Treatment settingInpatientOutpatient or communityBoth comparisons Q_B(1) > 7.87, P < 0.01

Positive symptoms

−0.12−0.28

Negative symptoms

−0.22−0.32

The two estimates did not differ for general psychopathology. Q_B(1) = 0.62, ns

Stage of illnessFirst episodeChronicCompletely collinear with treatment setting in this research

Positive symptoms

−0.08−0.21

Between-study difference:
Q_B(1) = 4.51, P < 0.05

The two estimates did not differ for negative symptoms or general psychopathology. Q_B(1) < 0.76, ns

Interpretation Considerations

Comparisons between groups of studies, not effects within individual patients

Study design

The authors indicated that cross-sectional studies tended to overestimate the predictive utility of symptomatology compared with longitudinal studies.

Treatment setting

The outpatient and inpatient analysis was restricted to a reduced set of studies, because 10 studies commingled the two groups.

Stage of illness

First-episode and inpatient studies were completely collinear with treatment setting in this research, so whether the weaker first-episode association reflects treatment setting or recent onset could not be determined.

The moderator analyses used composite quality of life and should not be applied directly to any single quality of life indicator. Subgroup correlations are not reported in the source for the comparisons that were not significant.

Clinical pearl

A symptom and quality of life correlation is not a fixed quantity. Its reported magnitude varied with the quality of life indicator, and in separate between-study analyses with study design, treatment setting, and stage of illness.

The field has not converged on a single definition or instrument, so the measure named in a study’s methods section is part of how its result should be read.

Sources and Disclosures

Meta-analysis caveats. Effect sizes are correlations and describe association rather than causation.

Sources: Eack SM, Newhill CE. Psychiatric symptoms and quality of life in schizophrenia: a meta-analysis. Schizophr Bull. 2007;33(5):1225–1237. doi:10.1093/schbul/sbl071. Chan A, Lu A, Menon T, et al. Association between negative symptoms and health-related quality of life and functional outcomes in persons with schizophrenia: a systematic review. Schizophr Res. 2026;288:95–103. doi:10.1016/j.schres.2025.12.019. Tolman AW, Kurtz MM. Neurocognitive predictors of objective and subjective quality of life in individuals with schizophrenia: a meta-analytic investigation. Schizophr Bull. 2012;38(2):304–315. doi:10.1093/schbul/sbq077. Watson P, Zhang J-P, Rizvi A, et al. A meta-analysis of factors associated with quality of life in first episode psychosis. Schizophr Res. 2018;202:26–36. doi:10.1016/j.schres.2018.07.013. Review the original sources for full detail.