Psychotic symptoms predict health outcomes even after adjusting for substance use, smoking and co-occurring psychiatric disorders: findings from the NCS-R and NLAAS

Psychotic disorders are associated with medical pathologies, and a recent study has shown that these associations extend into the sub-threshold regions of psychosis. Moreno et al (1) found that psychotic symptoms conferred risk for several lifetime physical health problems, including angina pectoris/cardiovascular problems, asthma/pulmonary problems, arthritis, tuberculosis, vision/hearing problems, and mouth/teeth problems. Also, psychotic symptoms were related to risky lifestyle behaviors, such as alcohol consumption and smoking, although these outcomes may have actually functioned as potential mediators. Further, psychotic symptoms are often accompanied by other psychiatric disorders (e.g., 2), which may likewise mediate relationships with medical conditions. Building on the work of Moreno et al, we used data from the U.S. to test whether substance use, smoking and co-occurring psychiatric disorders mediated the relationship between psychotic symptoms and physical health problems.

We analyzed two surveys conducted in the U.S.: the National Comorbidity Survey – Replication (NCS-R), using a nationally representative sample; and the National Latino and Asian American Study (NLAAS), using a national area probability sample with supplements for adults of Latino and Asian national origin. Both surveys adopted multi-stage clustered sampling designs. Details on the sampling strategy and interview procedures have been provided elsewhere (3). Respondents were included if they were assessed by the non-affective psychosis screen of the Composite International Diagnostic Interview (CIDI), which was administered to a random sub-sample of the respondents of the NCS-R (n+2322), and all respondents of the NLAAS (n+4644). Participants were excluded if they were missing data for any of the variables of interest. The final sample for this study consisted of 6,917 respondents.

Respondents were asked to report the lifetime presence of six specific psychotic experiences, which included two types of hallucinatory experiences (visual and auditory) and four types of delusion-like experiences (thought insertion, thought control, telepathy, and feelings of persecution). Responses were excluded if the experience took place in the context of falling asleep, dreaming or substance use. Respondents were asked (yes/no) if they had ever had the following five conditions in their lifetimes: arthritis/rheumatism, chronic back/neck problems, other chronic pain, stroke, and heart disease.

Demographic covariates included age, sex, ethnicity, income-to-poverty ratio, education, insurance coverage, and region of the country. Substance use included diagnoses of alcohol abuse and dependence, as well as drug abuse and dependence, ascertained through the CIDI. Co-occurring psychiatric disorders included CIDI diagnoses of anxiety disorders (agoraphobia with and without panic disorder, generalized anxiety disorder, panic attacks, panic disorder, post-traumatic stress disorder, and social phobia) and mood disorders (major depressive episode, major depressive disorder, and dysthymia). Smoking was measured by an item that prompted respondents to identify with four mutually exclusive categories: “current smoker”, “ex-smoker”, “only a few times”, or “never”. “Current smoker” and “ex-smoker” were combined, as well as “only a few times” and “never”, to form a dichotomous variable.

All analyses were conducted using the complex sample features of STATA SE 13. Analyses were two-tailed, alpha+0.05. Design-based analyses were used to estimate standard errors that accounted for the complex multistage clustered design of the NCS-R and NLAAS samples. All statistical estimates were weighted to account for individual-level sampling factors, including non-response and unequal probabilities of selection. Odds ratios were calculated using blocked hierarchical logistic regression. First, bivariate logistic regression analyses were used to determine the impact of psychotic experiences on health outcomes. In the second block, the logistic regression analyses were repeated with adjustments for potential demographic confounders. In the third block, analyses were repeated with adjustments for demographic confounders and substance use. In the fourth block, clinical variables were added to the demographic variables to determine whether the effects of psychotic experiences were independent of other mental health conditions. In the fifth block, smoking was added to demographic variables. The final block consisted of demographic variables and all potential mediators. Analyses were repeated with the exclusion of individuals who self-reported a history of a schizophrenia diagnosis, and results did not vary significantly.

After adjusting for demographics, psychotic experiences were associated with the increased likelihood of reporting arthritis/rheumatism (OR: 1.80, p+0.000), back/neck problems (OR: 1.98, p+0.000), headache (OR: 2.08, p+0.000), heart disease (OR: 2.36, p+0.024), other chronic pain (OR: 1.94, p+0.001), and stroke (OR: 1.72, p+0.143). Based on separate models adjusted for individual confounds – co-occurring psychiatric disorders (affective and anxiety), smoking, and substance use – we found that each of the confounds partially mediated the associations between psychotic experiences and physical health problems, especially co-occurring depressive and anxiety factors (data available upon request). The full model (adjusted for demographic variables and all potential mediators) showed that psychotic experiences still predicted some health outcomes: arthritis/rheumatism (OR: 1.50, p+0.001), back/neck problems (OR: 1.56, p+0.006), headache (OR: 1.64, p+0.000), and heart disease (OR: 2.14, p+0.023). The effect was no longer statistically significant in the full model for other chronic pain (OR: 1.44, p+0.060) and stroke (OR: 1.40, p+0.395).

In conclusion, psychotic experiences were associated with negative health outcomes independent of a psychotic disorder diagnosis, and the effect persisted for arthritis/rheumatism, back/neck problems, headache, and most strongly for heart disease, even with the inclusion of other mediating variables, among which co-occurring psychiatric disorders proved to be the strongest. These findings support screening for health conditions (particularly heart disease) among individuals with psychotic symptoms, regardless of psychiatric diagnosis.