What Does “53.5% Reported Benefit” Mean in a Medical-Cannabis Chart Review?

Short answer: it means that just over half of the patients in one retrospective Australian chart review were recorded as perceiving benefit from prescribed medicinal cannabis. It does not mean that the study proved cannabis caused improvement in 53.5% of all patients.

What the researchers reviewed

Morris and colleagues examined 157 medical records of adults with non-cancer diagnoses who had been prescribed medicinal cannabis. The patients were mostly women, with a mean age of 63. Neurological conditions accounted for 58% of the records and musculoskeletal conditions for 24.8%.

Balanced oral oils containing both THC and CBD were commonly prescribed. The average post-titration daily doses reported for those products were 16.9 mg THC and 34.8 mg CBD.

What changed in the records

The authors reported statistically significant changes over time in symptom ratings for pain, bowel problems, fatigue, sleep difficulty, mood, quality of life, breathing problems, and appetite. Perceived benefit differed by diagnosis and indication. Sleep had the highest reported indication-specific perceived benefit at 80%, followed by pain at 51.5% and muscle spasm at 50%.

Somnolence was the most frequently documented side effect, reported for 21% of patients.

Why “perceived benefit” is not the same as proven efficacy

A medical record can capture what a patient and clinician observed during care. That is valuable real-world information, but it is not the same as a randomized comparison. There was no untreated or placebo control group. Patients and clinicians knew treatment was being used. Doses and formulations were not uniform, and chart documentation is created for clinical care rather than a tightly standardized experiment.

Symptoms can also change because of natural fluctuation, other treatments, expectation, regression toward the mean, or changes in daily life. A before-and-after difference cannot automatically assign causation to cannabis.

What statistical significance does not answer

A small p-value addresses compatibility with a statistical model; it does not tell the reader whether the change was large enough to matter clinically. It also does not eliminate selection bias, incomplete follow-up, inconsistent recording, or confounding. The practical questions are how large the changes were, how many patients supplied follow-up data at each time, and whether improvement exceeded a meaningful clinical threshold.

Why the study still matters

Retrospective evidence can identify patterns worth testing prospectively. Here, the records suggest which symptoms, formulations, doses, and adverse effects deserve closer study. The finding that somnolence was common is also relevant to counseling, driving, work, and safety-sensitive activity.

The study provides a map for better questions; it is not a universal treatment verdict.

How to read a chart-review claim carefully

• Identify who was included and who was lost to follow-up.
• Separate patient-perceived benefit from validated outcome measures.
• Look for the size of the change, not only the p-value.
• Check whether formulation, dose, and co-treatments varied.
• Ask whether the design can distinguish treatment effects from confounding.
• Review adverse effects alongside claimed benefits.

The bottom line

The 53.5% figure describes documentation in one treated cohort. It supports the conclusion that many patients perceived benefit during care, while leaving the causal question unresolved. For an individual patient, the most useful approach is to define the target symptom, track function and adverse effects over time, and avoid treating a retrospective percentage as a prediction.

Source: Morris M, et al. “A Retrospective Medical Record Review of Adults with Non-Cancer Diagnoses Prescribed Medicinal Cannabis.” PubMed PMID 36836018; doi:10.3390/jcm12041483.

This article is educational and does not provide medical advice.