Macarthur Competence Assessment Tool For Treatment

How the Macarthur Competence Assessment Tool For Treatment Actually Works in Clinical Practice

The MacArthur Competence Assessment Tool for Treatment is a semi-structured interview instrument designed to evaluate a patient's decision-making capacity concerning medical treatment. It measures four subscales: understanding, appreciation, reasoning, and expressing a choice. Each domain is scored on a 0 to 4 scale, giving you a quantifiable breakdown rather than a single pass/fail judgment. It was originally developed by Grisso and Appelbaum and has been used across hospitals, forensic settings, and research sites for decades. The structure is intentionally flexible, but getting it right requires practice. Most people rush through it and end up with scores that don't hold up under scrutiny.

Macarthur Competence Assessment Tool For Treatment

Here is how I actually administer it in practice. You start by establishing rapport, then move through each subscale with open-ended questions that require the patient to explain their situation in their own words. You are not asking yes or no questions. You are asking them to demonstrate comprehension. Understand the condition, explain the proposed treatment, describe the alternatives, and articulate the risks and benefits. That is the core sequence. The scoring comes after you have the full interview. You listen first, score second. A lot of people flip that order and accidentally influence the patient's responses mid-interview, which invalidates the results.

The scoring system

Each of the four subscales contains multiple items. Understanding typically has nine items, appreciation has six, reasoning has four, and expressing a choice has two. Each item is scored from zero to four based on the depth and accuracy of the patient's response. A zero means no relevant information or incorrect information. A four means comprehensive and accurate understanding. The total score ranges from zero to fifty-six. There is no universally agreed cutoff score, which is one of the more frustrating aspects of this tool. Different institutions and jurisdictions use different thresholds. Some use a total score of thirty-four or above as the threshold for competence. Others rely on subscale-specific cutoffs. You need to know what your facility or legal jurisdiction requires before you administer it.

Practical problems I have encountered

One edge case that comes up repeatedly involves patients with aphasia or expressive language deficits due to stroke or traumatic brain injury. The MacCAT-T relies heavily on verbal expression, so a patient who comprehends the treatment fully but cannot articulate it clearly will score artificially low on the understanding and reasoning subscales. This is a known limitation that the developers acknowledged but never fully addressed in the manual. My workaround for that situation is straightforward. I use the nonverbal comprehension items where available, I allow written responses, and I cross-reference with a validated aphasia screening tool like the Western Aphasia Battery before finalizing the score. If the patient can match treatment components to conditions using picture cards or simple sorting tasks, I note that in the clinical report as supporting evidence of preserved understanding despite low verbal scores. It is not perfect, but it prevents systematic underestimation of capacity in this population. Another problem I deal with often is cultural and linguistic mismatch. The standard interview questions assume a certain level of health literacy and familiarity with medical framing that simply does not exist across all patient populations. A patient might answer "I don't know" to an appreciation question not because they lack capacity, but because they were raised in a culture where deferring to the doctor is the normative response. Scored literally, that looks like impaired appreciation. It is not.

The fix here is documenting the context explicitly in your assessment report and adjusting your questioning strategy. I spend extra time clarifying what each question is asking and rephrase it in plain language without changing its meaning. I also note any interpreter involvement and whether the patient was more coherent in their native language. These details matter legally.

Common pitfalls that make your assessment worthless

The biggest mistake I see is treating the MacCAT-T as a checkbox exercise. It takes approximately twenty to forty-five minutes depending on the patient's cognitive status and cooperation level. If you are finishing it in ten minutes, you are not doing it correctly. The tool requires sustained engagement and follow-up probing when answers are incomplete or vague. Another pitfall is conflating competence with consent. Competence is a clinical determination about decision-making ability. Consent is a legal process that follows from that determination. A patient can be competent and still refuse treatment, or they can be found incompetent and still have some supported decision-making options available depending on your jurisdiction. Mixing these up creates real legal exposure. A third issue is the appreciation subscale. This is the most subjective domain and the most vulnerable to rater bias. Two trained clinicians can score the same patient's appreciation differently if one is more conservative and the other is more lenient. To minimize this, use structured rating anchors, document your rationale for each score, and consider dual-rater scoring when the stakes are high.

What the tool does not handle well

The MacCAT-T was not designed for acute intoxication, delirium, or severe psychiatric psychosis. Patients in those states may perform below their baseline capacity, and the tool does not account for temporary fluctuations. You should always assess whether the patient's current state reflects their typical cognitive functioning before administering the instrument. If they are intoxicated or delirious, reschedule. A score obtained during an acute confusional state is clinically meaningless and legally inadvisable to rely on. There is also limited validation data for certain populations, particularly adolescents and individuals with intellectual disability. The original norms are based on adult psychiatric and medical patients. Applying it to younger or cognitively different populations without considering the psychometric limitations is a reasonable path toward an incorrect conclusion.

Accessing the instrument

The MacCAT-T is copyrighted and requires purchase for use. It is available through University Press at the University of Massachusetts Amherst, and through several clinical assessment distributors. The full manual and test materials typically cost between two hundred and four hundred dollars depending on whether you need the examiner manual alone or the complete kit with forms. Training is recommended but not formally required, and the developer website offers some self-guided materials. If you are using this in a research context, there may be different licensing terms. Check with the copyright holder before assuming educational use is free. Several universities have institutional licenses that cover faculty and graduate students, so check your institution's psychology or medical assessment resource office first.

Final notes on administration

The most reliable scorers are those who complete supervised training and practice with at least ten to fifteen cases before using the tool independently. Inter-rater reliability improves significantly after that threshold. Before that point, you are more likely to produce inconsistent results than useful ones. Keep your scoring sheets secure. They contain protected health information and serve as legal documents in many contexts. Write legibly, date each administration, and record the clinical context including medications, time of day, and any environmental factors that might have affected performance. A well-documented assessment survives scrutiny far better than a perfectly scored one with no supporting narrative.