I am seeing a 52-year-old male in my jail medical clinic who was booked yesterday on a felony DUI charge. He says he drinks "a lot of beer" but denies having a drinking problem. He is cranky and not really cooperative. He does not want to be here. However, he does tell me that he did not sleep much last night and did not feel like eating breakfast. I note that he has a mild hand tremor and a heart rate of 108. According to the (the most common tool used in the United States to assess the severity of alcohol withdrawal since 1989) my patient needs no treatment. But this is wrong! In actuality, my patient is experiencing moderate withdrawal and should be treated immediately.
Using CIWA is like using a wrench to hammer in a nail. It can be done, but it is not really efficient or accurate. A different tool (a hammer) could drive the nail much more quickly and effectively. CIWA is simply not the right tool to assess alcohol withdrawal. We should be using something better.
consists of 10 items each scored on a scale of 1-7 (one is scored 1-4). You add up the total and use the resulting number to determine the severity of withdrawal and what treatment you will administer. But the problems with CIWA are numerous. Let's begin!
1. CIWA requires cooperation from the patient. Seven of the 10 CIWA scores are obtained by asking the patient about a symptom and having them rate the severity of the symptom. For example, one CIWA question has you ask about headache and the patient must score their headache from 0 (not present) to 7 (extremely severe). The assumption is that my patients can communicate in English and are cooperative. Such was the case in the original CIWA study. But in my jail? Not so much. Many of my alcohol withdrawal patients do not speak English. Some have dementia, are developmentally delayed, or are otherwise impaired. Some give deliberately misleading answers, perhaps hoping to "score" some Valium (which we use to treat alcohol withdrawal). And some, like the cranky patient above, simply do not cooperate. What are we to do if we cannot communicate well with our patient or do not trust their answers?
2. Nine of the 10 CIWA scores are totally subjective. Either the patient or the evaluator has to estimate the degree of some symptom. For example, the degree of tremor is estimated by having the patient extend their arms and judging the severity of the tremor on a seven-point scale. If a patient has a mild tremor, is it a 1, 2, or 3? Different observers score these differently. Adding up differences for nine scales can (and does) result in big discrepancies. The original CIWA study found excellent "inter-rater reliability." But these were medical professionals who were specially trained and knew that they were part of a study. I have not found excellent inter-rater reliability among jail staff. Some jails use correctional officers to calculate CIWA scores. How is their inter-rater reliability?
3. Some CIWA scores are duplicates of each other. In the setting of alcohol withdrawal, what's the meaningful difference between "anxiety" and "agitation?" I'm not sure myself. My patient is cranky but he's not really anxious or agitated. How do I score that? Guess-timate?
4. Trivial symptoms are scored as being as important as serious symptoms. An example of a serious symptom is a patient who yesterday was "oriented and could do serial additions" but today is "disoriented for date." The score for this significant change is 1 point. Compare this to a patient who yesterday was "mildly anxious" but today is "guarded so anxiety is inferred." This relatively trivial change is worth three points.
5. In actual practice, CIWA is not used as written. For example, when scoring sensorium with CIWA, one is supposed to ask the patient to do serial additions. I have seen many CIWA assessments done and have not often seen that done. Similarly, one is supposed to ask specific questions for many of the other scores, such as this question for tactile disturbances: "Have you had any itching, pins and needles sensations, and burning, any numbness or do you feel bugs crawling on or under your skin?" When is the last time you asked that strange, convoluted question? Instead, we (appropriately) use simple, easy-to-understand language like "Do you feel anything weird on your skin?"
6. CIWA does not ask questions that I think are important. As I have written before, I think it is critically important that my cranky patient did not sleep last night and did not eat breakfast. Neither merits a score on CIWA.
7. CIWA ignores abnormal vital signs. In my experience, the heart rate correlates strongly with the severity of withdrawal. My patient is tachycardic. CIWA would have me ignore his abnormal vital signs.
8. CIWA does not treat some patients with withdrawal. CIWA recommends no treatment for patients scoring less than 10 -- like my patient. But this is wrong! We have a highly effective treatment for alcohol withdrawal. Most alcohol withdrawal patients will get worse with time before they get better. Treating early is more effective than treating later. Why in the world would we want to withhold treatment? This is like telling a wheezing asthma patient, "I could treat you now and make you feel better but instead I'm going to wait until you get worse and then I'll treat you."
So, despite CIWA, I'm going to treat my cranky patient for alcohol withdrawal. He'll feel better and probably be less cranky when he is reassessed this afternoon. In my next article, I will talk about alternatives to CIWA.
Jeffrey E. Keller, MD, FACEP, is a board-certified emergency physician with 25 years of experience before moving full time into his "true calling" of correctional medicine. He now works exclusively in jails and prisons, and blogs about correctional medicine at .