Summary
Razzano et al. (2015) document the prevalence and treatment of 17 co-occurring physical health conditions in 457 adults attending publicly funded community mental health programs in four U.S. states. Using face-to-face National Health Interview Survey (NHIS)/National Health and Nutrition Examination Survey (NHANES) structured interviews, they find that 87% reported at least one of the 17 assessed conditions, 14 of 17 conditions were significantly more prevalent than in the general population, and treatment prevalence was below 70% for roughly half of ongoing conditions. Ordinary least squares (OLS) regression identifies racial/ethnic minority status and female sex as independent predictors of hypertension and diabetes, respectively, within this already-high-risk population.
Key Claims
- Pervasive multimorbidity: 87% reported ≥1 medical condition; 82% ≥2; 69% ≥3; mean 2.8 conditions per person (standard deviation [SD]=2.2). Only 13% had no physical comorbidity.
- Dramatically elevated prevalence vs. general population: 14 of 17 conditions significantly more prevalent. Largest excesses: non-viral liver disease 6.8% vs. 1.4% (6×); diabetes 21.3% vs. 8.3% (3×); chronic bronchitis 15% vs. 4.3% (3×); viral hepatitis 5.5% vs. 1.9% (2×); hypertension 44% vs. 26% (1.7×). Only arthritis, tuberculosis (TB), and HIV/AIDS did not differ significantly.
- Top 5 conditions: hyperlipidemia (45%), hypertension (44%), asthma (28%), arthritis (22%), diabetes (21%).
- Treatment gaps: Treatment prevalence below 70% for 8 of 17 conditions. Very low (≤50%) for: non-viral liver disease (30%), hepatitis A/B/C (33%), stroke (42%), cancer (50%), arthritis (53%). High treatment: diabetes (94%), hypertension (89%), asthma (81%).
- Race/ethnicity doubles hypertension and diabetes rates: Controlling for age, site, Medicaid, diagnosis, and health self-efficacy, racial/ethnic minorities were ~2× as likely as Caucasians to be diagnosed with hypertension and diabetes.
- Women and diabetes: Women were ~2× as likely as men to be diagnosed with diabetes, controlling for all other factors.
- Schizophrenia diagnosis predicts lower hypertension and arthritis: People with schizophrenia ~half as likely as those with other diagnoses to report hypertension and arthritis; authors attribute this to likely under-diagnosis from cognitive impairment rather than true biological protection.
- Age: Positively associated with 4 of the top 5 conditions (hypertension, diabetes, hyperlipidemia, arthritis); younger for asthma.
- Medicaid beneficiary status: Associated with higher asthma likelihood; lower hyperlipidemia likelihood.
- Health self-efficacy: Lower self-efficacy associated with higher asthma likelihood.
Concepts Introduced or Extended
- Multimorbidity — extends multimorbidity evidence base to serious mental illness (SMI) population; shows dramatically elevated rates vs. general population and documents within-SMI racial/ethnic and sex disparities
Entities Mentioned
- Lisa A. Razzano
- Judith A. Cook
- Supplemental Security Income — study population (SMI adults in publicly funded community mental health programs) substantially overlaps with the SSI-covered population; treatment gaps for comorbidities directly affect work capacity and SSI eligibility maintenance
Quotes
"Compared to the U.S. population, prevalence was significantly higher for 14 out of 17 medical conditions assessed."
"87% [of participants] reported at least one of the 17 conditions, with 18% reporting only one, 36% reporting 2–3, and 33% reporting 4 or more."
"Racial and ethnic minority group members were almost twice as likely as non-minority participants to be diagnosed with hypertension and diabetes, controlling for study site and other individual and system-level factors."
My Take
The paper's primary contribution is documenting that physical multimorbidity in the SMI population is not only widespread but also dramatically concentrated in specific conditions amenable to secondary prevention (hypertension, diabetes, hyperlipidemia). The treatment gap finding — 30–50% treatment prevalence for several ongoing conditions — is clinically significant but methodologically limited by self-report without chart verification. The racial/ethnic disparity finding within a population already characterized by extreme medical vulnerability is an important disparity-within-a-disparity result, though the cross-sectional design cannot distinguish whether minority SMI adults are more severely ill or face greater barriers to diagnosis and treatment. The sample is not nationally representative (convenience sample from 4 states; screening participants may be more health-conscious than the broader community mental health population), limiting generalizability. The schizophrenia "protection" finding almost certainly reflects under-diagnosis via cognitive impairment effects on self-report accuracy, not a true biological effect.