"First, factors related to death investigation might affect rate estimates involving specific drugs. At autopsy, the substances tested for, and circumstances under which tests are performed to determine which drugs are present, might vary by jurisdiction and over time. Second, the percentage of deaths with specific drugs identified on the death certificate varies by jurisdiction and over time. Nationally, 19% (in 2014) and 17% (in 2015) of drug overdose death certificates did not include the specific types of drugs involved. Additionally, the percentage of drug overdose deaths with specific drugs identified on the death certificate varies widely by state, ranging from 47.4% to 99%. Variations in reporting across states prevent comparison of rates between states. Third, improvements in testing and reporting of specific drugs might have contributed to some observed increases in opioid-involved death rates. Fourth, because heroin and morphine are metabolized similarly (9), some heroin deaths might have been misclassified as morphine deaths, resulting in underreporting of heroin deaths. Finally the state-specific analyses of opioid deaths are restricted to 28 states, limiting generalizability."

Source

Rudd RA, Seth P, David F, Scholl L. Increases in Drug and Opioid-Involved Overdose Deaths — United States, 2010–2015. MMWR Morb Mortal Wkly Rep 2016;65:1445–1452.