Is SBIRT a Failure?! Not according to SAHMSA’s response to recent JAMA articles

You may have read about some recent JAMA articles that could be taken as critical of SBIRT. That would be an error, according to recently commentary by the Substance Abuse and Mental Health Services Administration:

Most likely many of you are familiar with the recently published studies on SBIRTpublished in the Journal of the American Medical Association (JAMA). Although these studies contain some very solid research data, their focus does not do justice to the benefits SBIRT brings to behavioral health. In case you have not seen the articles, the two studies (and the related editorial) are attached. SAMHSA believes these two papers are of substantial interest. However, it is important to look at the papers from the proper perspective.

Both papers focused on the Brief Intervention (BI) part of SBIRT, not the Screening part and not the referral to treatment (RT) part. These two studies were well done and elaborate. However, the value of SBIRT could not be challenged by either study. Furthermore, the value of SBIRT to behavioral health and primary care providers is the ability of SBIRT to identify when a patient is in need. When patients have chronic medical problems, there are multiple opportunities to address the issue of substance use. Roy-Byrne noted the majority of his participants had a single brief intervention contact, with only 47% receiving a follow-up booster call. Saitz reported a single session approach for his two test conditions.

During the 11 years since SAMHSA’s SBIRT program has been in existence, over 2 million people have been screened. Of those, only a small percentage screened positive for any “at risk” behaviors, with about 11 percent of those screened receiving a brief intervention. Without screening many of these people might have remained invisible. SBIRT gives providers and primary care physicians an opportunity to identify potential alcohol and substance misuse or abuse and, through brief intervention, an opportunity to use that “teachable moment” to educate patients and, potentially, change the behavior of “at risk” individuals for the better.

Roy-Byrne’s title “Brief Intervention for Problem Drug Use in Safety-Net Primary Care Settings” really is applicable to both papers. Ninety-one percent of Roy-Byrne’s participants were unemployed, while 81% of Saitz participants were on Medicaid or Medicare. Fifty-six percent of Roy-Byrne’s participants had greater than one ICD-9 Mental Illness code, while 46% of Saitz’s participants had a co-morbid mood disorder. Saitz required his participants to have an ASSIST score of greater than 4 in order to participate. That is understandable, since the World Health Organization (WHO) recommends that an ASSIST score of 4 to 26 should result in brief intervention and a score of 27+ should result in more intensive treatment.

It is important to remember brief intervention does not work for everyone. For many, learning the consequences of their “at risk” behavior or abuse can provide the wake-up call they need to either stop using or seek appropriate treatment. For individuals with more severe and complex substance use disorders, brief intervention will most likely not be sufficient to change their behaviors. For this group it is important that a treatment referral be made.

When dealing with complex patients with complex problems, is it reasonable to expect BI to “cure” the substance use disorder? No. The question for SBIRT is whether it is feasible to screen for drug use disorders in primary care, just as it is feasible to screen for alcohol use disorders. Both papers implicitly say “Yes.” Thus, if it is feasible, the next question is whether it should be done. We believe that if we are to promote integrated treatment, primary care providers (PCPs) must have the basic skills necessary to identify SUDs in primary care settings. SAMHSA’s SBIRT program accomplishes this. SBIRT is not a panacea, it is an important process that can help primary care providers identify alcohol and drug use problems. We have to wait for research on more representative populations to determine whether BI works.

In response to these articles, Dr. Clark authored a blog on SBIRT. You can access it via this link:

We invite any thoughts you have on this issue. Also, if you have seen any feedback/reactions in your setting, please share that information.

Thank you,


Leave a Reply