Prognostic Disclosure in Acute Myeloid Leukemia (AML): A Qualitative Study of Patient Preferences and Physician Practices

Background: Many patients with high-risk AML have a poor understanding of their prognosis. However, little is known about patients’ preferences for receiving this information, nor about how oncologists (MDs) routinely approach conveying it.

Methods: AML inpatients receiving chemotherapy completed semi-structured qualitative interviews about their experience with receiving an AML diagnosis, and any associated prognostic disclosure. MDs were also interviewed about their practices of conveying diagnostic and prognostic information. We used an inductive content analysis approach, a standard method in qualitative research.

Results:We enrolled 19 patients and their 5 primary MDs. Mean patient age was 65.6, and all had high-risk disease due to either age >59, complex cytogenetics, or secondary AML/relapsed disease. Measured conservatively, at least 5 patients (26%) did not have a functional understanding of the term "prognosis." Several key themes emerged, including: (1) patients’ thirst for more information, (2) physicians use of a standard approach to prognostic disclosure, and (3) variable patient information preferences regarding type and delivery method. Most patients wanted diagnostic and prognostic information to be delivered in a straightforward and honest manner, and most MDs reported presenting information in this way. However, some patients thought their MDs delivered this information very bluntly and without adequate preparation or context; others wanted more information, while a minority wanted less. Ten patients (53%) reported feeling so shocked about their diagnosis that their ability to comprehend information was impaired. While all MDs reported always discussing prognosis, 9 patients (47%) said they did not receive prognostic information. Some patients thought this reflected difficulty in calculating an accurate prognosis, or that MDs preferred not to give patients this information. Patients reported varying comfort levels with receiving prognostic information as numbers or statistics. MDs also had varying comfort levels with giving information in this way. Five patients (26%) reported receiving numerical prognostic information; 4 of these (21%) had asked specifically for numbers. MDs did not elicit patients’ preferences regarding type or nature of information delivery. Rather, several described using a standard "scripted response" in most cases. This "script" was only sometimes modified or supplemented, based on informal assessments of patients’ understanding, or because of questions patients asked.

Conclusions: Patients have varying preferences for prognostic information delivery but MDs do not routinely assess patients’ preferences, instead delivering a standard amount and type of information with little variation or adaptation. While MDs report always discussing prognosis, many patients do not report receiving this information, and they describe emotional considerations as limiting their ability to process information. A more formal assessment of AML patients’ informational needs and preferences is recommended at diagnosis, along with follow-up assessments of their understanding and information retention thereafter.