Patients with addiction complain of cravings so frequently that sometimes it’s difficult to make sense of them. They are a common symptom and one that may serve as a predictor of whether a patient will relapse.

What are cravings? At the most basic level, cravings are thoughts and urges to use a substance. Studies have shown, however, that cravings have much greater texture and are the product of multiple factors (Jimenez M et al, Eur Addict Res 2009;15(3):135–142).

These include positive reinforcement (the rewarding aspects of substance use), negative reinforcement (the distressing emotions that substance use takes away), and impaired control (impulsive or compulsive substance use).

Classically, cravings have been described as either positive or negative, depending of what the patient is feeling and expecting (Tiffany ST, Alcohol Res Health 1999;23(3):215–224). Positive cravings are associated with the patient’s anticipation of pleasure (the “buzz” or “high”), whereas negative cravings, also known as “relief cravings,” address negative affect, such as problems with mood, anxiety, stress, or withdrawal symptoms.

Research has shown that addiction involves maladaptive memory, which has led some to conceptualize cravings “as the memory of the pleasant rewarding effects of drugs of abuse superimposed on a negative emotional state” (Koob GF and Le Moal M, Annu Rev Psychol 2008;59:29–53). Cravings are further described as either Type 1 if they are triggered by drug-related cues and stimuli, or Type 2 if a negative emotional state is also involved.

All of this is fairly consistent with the wisdom of Alcoholics Anonymous. People in recovery have long known that “people, places, and things” (Type 1 cravings) can lead to relapse. The “Big Book” also famously describes alcoholics as “restless, irritable, and discontented” (reflecting negative affect or Type 2 cravings).

Assessing a Patient’s Cravings

In addition to asking about a patient’s cravings, you can use validated ratings scales to measure them, as more addiction professionals are starting to do.

Scales and measures are assuming an ever-greater role in mental health and addiction treatment. Part of this is a result of the Patient Protection and Affordable Care Act of 2010, which has ushered in the era of “accountable care” predicated on measurable treatment outcomes.

Another driver is sound clinical practice. Clinicians generally strive to create treatments plans with specific goals—such as symptom reduction—that can be measured. It’s pretty hard to track progress toward those goals without some sort of scale.

There are many instruments for measuring alcohol cravings. Various other scales and measures are available for cannabis, cocaine, nicotine, and opioids (Heishman SJ et al, Drug Alcohol Depend 2009;102(1–3):35–40; Sussner BD et al, Drug Alcohol Depend 2006;83(3):233–237).

Penn Alcohol Craving Scale (PACS). I like the PACS, a self-report instrument with good validity (Flannery BA et al, Alcohol Clin Exp Res 1999;23(8):1289–1295). PACS consists of five questions about the frequency, duration, and intensity of alcohol cravings over the past week. Total scores can range from zero to 30, where higher scores represent more intense cravings. Here’s a sample item:

How often have you thought about drinking or about how good a drink would make you feel? (0 = Never; 6 = Nearly all the time)

The Obsessive Compulsive Drinking Scale (OCDS). The OCDS is another self-report instrument, often used in clinical trials (Anton RF et al, Alcohol Clin Exp Res 1995;19(1):92–99; Anton RF et al, Arch Gen Psychiatry 1996;53(3):225–231). It contains 14 questions, which makes it impractical for routine clinical use. However, six questions specifically deal with cravings, and you could pull one or two of those items to help assess cravings in your patient.

Do Cravings Lead to Relapse?

A number of studies have examined the relationship between alcohol cravings and addiction relapse. The most recent investigation was performed at the Mayo Clinic using a prospective cohort of 314 adults with alcohol dependence that received residential substance abuse treatment (Schneekloth TD et al, Am J Addict 2012;21:Suppl 1:S20–S26).

Researchers asked about such factors as drinking histories, the presence of psychiatric comorbidity, the severity of depressive symptoms, the number of prior treatments, and whether anti-relapse medications were prescribed at program discharge. Researchers also measured participants’ alcohol cravings using PACS. Data were obtained from patients at program admission, weekly during treatment, at program discharge, and quarterly following release to the community.

PACS scores and the number of prior treatments were the only factors that predicted relapse when all the above variables were combined. The risk of relapse increased by 6% for each one-point increase in discharge PACS scores. In other words, the more intense cravings were at discharge from treatment, the more likely the person was to relapse.

The strongest predictor of success was a PACS score of less than seven on admission: 72% of these patients remained abstinent during the first six months following treatment. Conversely, about 60% of patients with an admission PACS score ≥7 or a discharge PACS score ≥4 relapsed within the six-month timeframe.

This study replicated prior work using PACS and OCDS that found that craving scores were a better predictor of drinking behavior than prior drinking patterns (Flannery BA et al, J Stud Alcohol 2003;64(1):120–126). Other studies have found that rate of change in cravings scores can discriminate abstainers from those who relapse (Anton 1996, op.cit; Kranzler HR et al, Alcohol Clin Exp Res 1999;23(1):108–114). This separation occurred within the first two weeks of treatment.

Psychologist Stephen Tiffany, PhD, a major figure in the field of craving research, has noted that the issue is messier than suggested by some studies (Tiffany ST and Wray JM, Ann NY Acad Sci 2012;1248:1–17). He suggests that cravings per se do not appear to predict prognosis and need to be considered in context. From his standpoint, the question isn’t whether cravings predict drug use, but rather under what conditions, in which patients, do cravings contribute to relapse? This returns us to the whole business of Type 1 and Type 2 cravings, which are much more challenging to measure than simple thoughts and urges.

Four Clinical Recommendations

To my mind, cravings are too common and distressing to simply ignore. I recommend following these four steps to address cravings in your patients:

1. Ask patients with addiction about their cravings at every clinical encounter.


2. Strongly consider using validated ratings scales to measure those cravings. I suggest using a question or two from validated scales like PACS.


3. Pay close attention to patients whose scores are high and show little change over time. They probably merit more careful clinical supervision and possibly prescription of anti-craving medications.


4. Further assess patients who report significant cravings for situationally-bound drivers (those “people, places, and things” that can lead to relapse) and negative affect. Many of these triggers can be modified or addressed through psychosocial interventions and various medications.