A 2020 analysis estimated severe mental illness in 32.1% of a survey-defined group reporting methamphetamine and opioid use. That figure was an adjusted population estimate generated using a predictive model. It was not the percentage of people clinically diagnosed with methamphetamine-induced psychosis, and it did not establish which condition or exposure came first.
Who was included?
Shearer and colleagues pooled 2015–2018 National Survey on Drug Use and Health data for adults aged eighteen to sixty-four. The unweighted analytical sample included 156,870 respondents. Within it, 1,511 met the study’s opioid-use-only definition, 1,081 met its methamphetamine-use-only definition and 397 met both definitions.
Those subgroup counts are actual survey observations. They differ from weighted population totals and from the denominator used to interpret an adjusted percentage. The 32.1% estimate should not be converted into an observed case count simply by multiplying it by 397.
This was a cross-sectional observational analysis of historical U.S. civilian, noninstitutionalized survey data. It did not administer either substance or follow a newly exposed group to determine incident mental illness.
What did “opioid use” and “methamphetamine use” mean?
The definitions were intentionally asymmetric. Methamphetamine use meant any reported use in the preceding year. Opioid use meant any past-year heroin use or prescription opioid misuse on four or more occasions in the preceding month. It did not mean every instance of medically directed opioid treatment or every isolated episode of prescription misuse.
The group reporting both met these operational criteria. Its members were not necessarily exposed to both substances simultaneously. “Methamphetamine alone” meant without meeting the specified opioid-use definition; it did not mean no alcohol, cannabis or other drug exposure.
These distinctions affect the comparison population. An analysis using any past-year prescription opioid misuse, current weekly use or toxicology-confirmed simultaneous exposure would pose different questions.
How was severe mental illness determined?
The researchers used a predictive model validated by the Substance Abuse and Mental Health Services Administration. Inputs included the K6 Psychological Distress Scale, the World Health Organization Disability Assessment Scale, past-year suicidal ideation, past-year major depressive disorder and age.
This approach estimates a broad mental-health classification. It does not identify schizophrenia, a particular psychotic episode, methamphetamine-induced psychosis or intoxication at a specified time. The analysis did not provide the individual diagnostic history needed to attribute one person’s symptoms to one drug.
What were the estimates and their uncertainty?
Table 3 reported adjusted severe-mental-illness estimates of 20.7% for the opioid-only group, 22.8% for the methamphetamine-only group and 32.1% for the group reporting both. Their 95% confidence intervals were 17.3%–24.0%, 17.8%–27.8% and 25.0%–39.1%, respectively.
The comparison between the group reporting both and the opioid-only group met the table’s p≤0.01 criterion. The authors described the 32.1% estimate as approximately 55% higher than 20.7%. That is a relative comparison, corresponding to an absolute difference of about 11.4 percentage points. It is not a 55-percentage-point increase or a measured causal effect.
The table adjusted estimates for age, sex, income, education, rural versus urban residence, race/ethnicity and housing instability. Adjustment helps compare groups on measured characteristics. It cannot establish temporal order or remove unmeasured differences.
Did every health measure show a difference?
No. Estimated disability was high across all three groups: 37.7%, 40.0% and 42.1%, respectively. The disability row did not identify statistically significant between-group comparisons. Its broad definition included reported difficulty with seeing, hearing, concentrating, mobility, self-care or doing errands alone.
The authors also reported no statistically significant group differences in diabetes, kidney disease or cancer. These null findings should remain visible rather than being absorbed into a claim that every health outcome was worse in the combined-use group. Nonsignificance does not prove equivalence or absence of disease.
Why can’t the association establish which condition caused which?
Drug-use reports and health characteristics were assessed cross-sectionally. The authors explicitly could not determine the timing of drug-use initiation relative to health-condition diagnoses. Mental-health difficulties could precede, follow or share underlying causes with substance-use patterns.
Recall, disclosure and unrecognized exposure can affect self-report. The survey excluded incarcerated people and people who were homeless and unsheltered, populations relevant to estimating the broader burden. The findings also cannot be assumed to describe today’s drug supply or treatment environment.
For a client-ready interpretation, retain the study’s drug definitions, modeled classification, adjusted percentages and historical sampling frame. The research identifies a population with potentially substantial service needs. It supplies neither a diagnosis for an individual nor evidence that methamphetamine caused that individual’s mental illness, disability or impairment.
Scientific source
Shearer RD, Howell BA, Bart G, Winkelman TNA. Substance use patterns and health profiles among US adults who use opioids, methamphetamine, or both, 2015–2018. Drug and Alcohol Dependence. 2020;214:108162. doi:10.1016/j.drugalcdep.2020.108162. Historical cross-sectional human survey analysis; full author manuscript read from the supplied research collection.
Leave a Reply