Problematic drug use is often identified via screening in non-specialised health and care settings, e.g. in primary care.
Brief interventions, primarily focusing on motivation, are recommended by NICE for problematic drug users who are not in contact with more formal drug services. These interventions have shown some success at reducing problematic alcohol use, however little is known about their effectiveness at reducing use of other substances.
JAMA has recently published two randomised clinical trials (RCTs) that assessed whether brief interventions in substance using primary care patients can reduce substance use (both illicit drugs and misused prescription drugs).
Trial 1: Roy-Byrne et al. (2014; Washington State, USA)
This RCT compared a one-time brief intervention (motivational interviewing) plus attempted telephone booster session, to enhanced care as usual.
- Recruited by screening from waiting rooms of seven “safety-net” public primary care clinics in Washington State, USA
- 18 years or older
- Self-reported users of an illegal drug or non-prescribed medication at least once in the 90 days prior to screening
- Exclusion criteria: past month attendance at formal substance use treatment; high risk of suicide; life-threatening medical illness; severe cognitive impairment; active psychosis
After screening, participants were randomised to condition. Social workers from the clinics received training to provide the intervention.
- The intervention condition involved a 30-minute motivational interview brief intervention
- Core tasks included exploring pros and cons of drug use, increasing confidence in ability to change, and discussing options for change
- Participants also received a hand-out depicting their Drug Abuse Screening Test score (DAST-10), and a list of substance abuse treatment resources
- The control condition received their DAST-10 hand-out, and the list of substance abuse treatment resources, with only brief explanation.
Assessments were conducted at baseline and then repeated at 3, 6, 9, and 12 months, by research assistants blinded to the condition.
- Self-reported past month drug use (of drug participants used most frequently used at baseline);
- Addiction Severity Index (ASI) drug use subscale score (integrating all drug use frequency and associated problems).
- Medical, psychiatric, social and legal subscales of the ASI;
- Administrative record: drug treatment; police arrest; deaths; emergency, outpatient & inpatient hospital visits.
They randomised 868 participants. Drop-out was low, with 87% of assessments at 12-month follow-up completed.
Statistical analyses (generalized estimating equations) demonstrated no treatment differences at 12-months follow-up, on both past month days of drug use (odds ratio, 1.20 [95%CI, 0.96 to 1.50]) and ASI drug use composite (β = 0.005 [95% CI, −0.005 to 0.016]).
There were also no intervention effects on any secondary outcomes.
The authors therefore concluded:
A one-time brief intervention with attempted telephone booster call had no effect on drug use in patients seen in safety-net primary care settings.
Trial 2: Saitz et al. (2014; Boston, Massachusetts, USA)
The second trial was similar to Byrne et al., however notably inclusion criteria required at least weekly drug use, or less frequently with a negative consequence. Recruitment was from an urban hospital-based primary care clinic.
Participants were randomised to either:
- Brief negotiated intervention;
- Motivational interviewing with follow-up booster session; or
- No brief intervention.
This trial also found no treatment differences between any of the conditions on past month drug use, or any of their secondary outcomes, after 6-months follow-up.
Similarly to above, the authors concluded:
…brief intervention for unhealthy drug use in primary care patients identified by screening appears unlikely to be effective for decreasing drug use or consequences.
Strengths of both studies
- Randomisation to intervention;
- Low attrition rates;
- Urine drug screens to corroborate self-reported drug use;
- Samples had high health co-morbidity rates, high poverty and unemployment rates. These samples are therefore representative of populations who are more likely to experience major problems related to their substance use. The failure to find an effect of brief intervention may demonstrate the complexity of substance use in disadvantaged social environments.
- Results may not generalise to less disadvantaged drug-user populations;
- In Trial 1, past month drug use declined for both the control and intervention group. Both conditions included five assessment sessions for all participants, which may alone have influenced drug use;
- Inclusion criteria were broad and included occasional users, with no requirement that participants were experiencing any drug-related problems. It may have been inappropriate to call the samples problematic drug-users, since many illicit drug users experience minimal negative consequences.
- The most common drug used was cannabis, which has lower rates of dependence and drug-related problems than opiates and stimulants. Design of interventions may be better directed at individual drug types or specifically at problematic drug use.
Implications for practice
Brief interventions may be seen as quick and simple solutions that can be employed within a primary care setting at minimal cost. Unfortunately these findings suggest they may not be effective for reducing drug use.
However both trials were conducted in the USA. The relationship and engagement patients have with healthcare may differ to the UK, potentially affecting generalisability of findings to the NHS.
Roy-Byrne P, Bumgardner K, Krupski A, Dunn C, Ries R, Donovan D, West II, Maynard C, Atkins DC, Graves MC, Joesch JM, Zarkin GA. Brief intervention for problem drug use in safety-net primary care settings: a randomized clinical trial. JAMA. 2014 Aug 6;312(5):492-501. doi: 10.1001/jama.2014.7860. [PubMed abstract]
Saitz R, Palfai TP, Cheng DM, Alford DP, Bernstein JA, Lloyd-Travaglini CA, Meli SM, Chaisson CE, Samet JH. Screening and brief intervention for drug use in primary care: the ASPIRE randomized clinical trial. JAMA. 2014 Aug 6;312(5):502-13. doi: 10.1001/jama.2014.7862. [PubMed abstract]
RT @Mental_Elf: Brief interventions for substance misuse in primary care http://t.co/SnZrEnYCYF
Brief interventions for substance misuse in primary care: Claire Mokrysz reports on two RCTs in JAMA that find… http://t.co/2xRzXRP7tG
Sally Adams liked this on Facebook.
Brief interventions for substance misuse in primary care http://t.co/Zp1EVK531o via @sharethis
Olivia Cialdi liked this on Facebook.
Today @claimokr reports on 2 @JAMA_current RCTs of brief interventions for #SubstanceMisuse in primary care http://t.co/4qfbb2Jnpu
The Mental Elf liked this on Facebook.
Hi @UnhealthyAlcDrg We’ve blogged about your ASPIRE RCT http://t.co/4qfbb2Jnpu Please read & comment when you can. Cheers, André
@Mental_Elf done–thanks for notice and thoughtful coverage
New RCTs cast doubt on use of motivational interviewing for unhealthy drug use in primary care patients http://t.co/4qfbb2Jnpu
.@Mental_Elf once again we see that most bright ideas for interventions don’t work
.@david_colquhoun @Mental_Elf and isn’t this a problem for encouraging local innovation without appropriate evaluation?
@david_colquhoun @Mental_Elf RCTs are of BRIEF motivational interviewing. Changing drug use in 60 minutes is (evidently) unrealistic.
@david_colquhoun @mental_elf Motivational interviewing is even more than a bright idea it also has the political correctness sparkling light
@david_colquhoun @mental_elf MI is like Peer Support in psych: Few dare to think it mightn’t be helping in any form. http://t.co/4RM6uCg7Zs
Hi @mike_ashton_ have you seen this re: MI & treatment http://t.co/CN9M3Jc425
Mental Elf: Brief interventions for substance misuse in primary care http://t.co/4hY91Y3RzR
Two recent @JAMA_current RCTs show brief interventions no better than control for substance misuse in primary care http://t.co/4qfbb2Jnpu
Don’t miss: Brief interventions for substance misuse in primary care http://t.co/4qfbb2Jnpu
Brief interventions for substance misuse in primary care shown to be ineffective
http://t.co/Mn7aaCjVWH via @sharethis
Kehksha Azam liked this on Facebook.
Thx for posting thoughtful review. As the author of one study report, happy to respond.
It is notable that both studies were done in disadvantaged populations. However, this has not been a limitation of alcohol screening and brief intervention (SBI)—in other words, is the idea that drug BI works different depending on social context but alcohol BI does not? This is plausible but then two other observations are important. 1) there are no convincing drug SBI studies showing efficacy in any population so the state of the science is—we have good evidence it does not work (at least in disadvantaged populations) and no evidence it works in other populations. Seems a poor evidence base upon which to disseminate a nationwide policy as is being done in the US (google SAMHSA SBIRT). Also, Bernstein et al in a different setting but a severe disadvantaged population at the same Boston site found efficacy for BI in heroin and cocaine users…so disadvantaged seems an unlikely explanation here… see here for summary of drug SBI trials http://journal.frontiersin.org/Journal/10.3389/fpsyt.2014.00121/full 2) if drug SBI doesn’t work in people who have social stresses and possibly mental health conditions or in minorities…then, seriously, what good is it as a universal preventive strategy?? We must do better. Also note a systematic review by Kaner E et al. that found no efficacy of alcohol or drug SBI when people used more than one substance or had a co-occurring mental health condition (Kaner EFS et al. Ment Health Subst Use. 2011;4(1):38–61). Again….not promising since that circumstance is common…
My discussion in the article addresses the question of whether maybe drug SBI would work in people with more problems. Of course when presented with a study that finds no efficacy, advocates of SBI have said ‘well maybe it will work in more severe’ or ‘maybe it will work in less severe.’ Requiring a certain level of severity or lack or presence of comorbidity removes the likelihood that this will ever be a very useful universal strategy. SBI is supposed to be a preventive strategy—so it should work in milder cases. But it doesn’t for drug. Maybe it is because someone who is using drugs is already doing something they know they are not supposed to be doing and without perceived harm. But then of what use is drug SBI? Maybe for more severe? But no one really believes that 10-15 minutes of BI will be the one intervention that changes someone’s more severe drug use—the idea then is to refer to some sort of treatment. But experience in primary care SBI programs suggests that people don’t complete such referrals. I should note that in Massachusetts where our study was done, insurance/payment is really not a barrier to treatment for many reasons I can’t go into here. And regardless of all of this…fewer than 1 in 5 identified by screening as using drugs in primary care settings have dependence as their level of severity. So even if it did work in the more severe, we would have an intervention that is useless (and potentially harmful as one puts the fact of illegal behavior into identifiable medical records) for 82% of patients who receive it…
Yes, most people in both studies were using cannabis as their main drug. The fact is that is who is identified by screening in primary care. remember, primary care is a place people have to make and keep appointments to attend, and they are seeking care to improve their health. That is a select group of drug users (who one might expect would respond to BI…). That said, our cannabis users used drugs on many more days per month (around 18 days) than cocaine and opioid users (who used 5-7 days per month). So the cannabis users may have actually been more severe—many did report problems from it. Referral success is more related to motivation than to the drug itself.
It should be pointed out that in both studies a second BI was offered. Of course it cannot be forced and shouldn’t be. But this means that the designed BI was 2 conversations. The fact that often one, not two, happened is notable. But it is unclear how to make a second one happen, and to make it effective…
Lastly, regarding the assertion that assessments might have lead to changes in drug use…This is often talked about. But…note that in our trial >90% were still using drugs 6 mo later. Also, if it is so difficult to demonstrate that skilled counseling can change drug use why would w e believe that assessment might do it? Third, in the alcohol SBI literature, two studies (Daeppen et al and D’Onofrio et al) have found NO effect of assessment in people randomized to assessment only. Assessment effects in studies designed to find them tend to be found in treatment seekers (see Clifford and Maisto et al) not in SBI studies. An unlikely explanation for the findings. And if anyone still believes in assessment effects, then why don’t we just abandon all of the skliled counseling and just assess people? The fact that no one does this leads me to believe that people like to lob the assessment effect as an explanation for a negative finding they believe should have been positive but they don’t really believe it…
So the state of affairs is that drug SBI as a universal approach has little evidentiary support. Thus we need something better (as pointed out by the blogger there were few methodological limitations (by the way in our study we did hair testing which covers 90 days so is quite useful) and the results were consistent—many people would read these and consider the questions posed to have been answered). Something better to address this problem in general healthcare should be the starting place for a discussion.
So what the outcomes are saying is the research was flawed from the proposal stage…?
Your comment regarding cannabis being used with minimal negative consequences is disingenuous and subjective, considering the sample consisted of people who were either using “weekly or or less frequently with a negative consequence”. If this is indeed the case you should have extrapolated the figures which related to cannabis specifically and which were significantly related to so-called minimum negative consequences. This comment is in direct contradiction of your findings and cannot be under-pinned by your findings, so you should not slip in an unscientific comment like that. It is misleading and not objective or clarifying to state cannabis causes “minimal negative consequences”. The term minimal is a relative and subjective term unless you can explain what has been categorized as minimal. Your inference in this case being that minimal is somehow acceptable which really is a matter of personal opinion. In South Africa, A 2017 SACENDU study shows that the number of people under the age of 20 who have sought treatment for substance abuse in 2017 have sought that treatment far and away the most for cannabis use by a factor of 10 (922 versus 97 for alcohol abuse, which was next in line). Please stick to the facts and do not slide your own personal opinions in to your reports.