From Diagnosis to Surgical Error: Addressing Inconsistencies in Medical Documentation

9/11/2024

Wrong-site surgery, including operations on the wrong body part (site) or side (e.g., left vs. right), is the most common type of surgical error and can lead to severe, permanent harm or death. The Joint Commission reported 112 surgical errors out of 1,411 incidents in 2023, a 26% increase from 2022. Wrong-site surgery accounted for 62% of these cases.

Wrong-site surgeries occur most frequently in orthopedics, neurosurgery, and urology. According to the study, the most common types of procedures that involved wrong-site surgery were spine surgery, including spinal fusion and excision of intervertebral discs, arthroscopy, and surgeries on muscles and tendons. Patient injuries resulting from these errors include the need for additional surgery, pain, worsened injury, total loss, and death, etc., Only 60% of the cases were settled.

The top contributing factors to wrong-site surgery were:

Failure to follow policy/protocol
Failure to read medical records
Selection/management of surgical treatment
Inconsistent documentation
Known complications or technical issues
Communication among providers

However, inconsistencies in medical documentation do not always indicate errors. A diagnosis that does not specify a body part or side might still be accurate if it justifies a particular surgery. An example could be a diagnosis of diabetes-related gangrene without specifying the body part or side, paired with a right lower leg amputation.

To further explore this issue, we conducted a preliminary retrospective analysis of 8.93 million inpatient records from the U.S. CMS in 2020. We categorized records based on whether the diagnoses and surgeries involved the same or different body parts, and whether they specified right side, left side, or both. We particularly focused on cases where site and side records were inconsistent, resulting in 1,064 records with right-side procedures and left-side diagnoses, and 1,106 records with right-side diagnoses and left-side procedures. Each medical record was reviewed by physicians to identify whether the inconsistencies were clinically justifiable or actual errors, based on the following criteria:

All diagnoses did not justify the necessity of the surgery
Diagnosed and surgical body parts were unrelated, not near each other, or not connected
Diagnosed and surgical body parts were similar but on opposite sides

The findings revealed that in cases of right-side procedures with left-side diagnoses, 49% of inconsistencies were clinically justifiable, while 51% were actual errors. For right-side diagnoses with left-side procedures, 45% of inconsistencies were justifiable, with 55% were actual errors.

These results raise concerns about whether the errors are just mistakes in the documentation or if they actually happened to patients. With approximately 51 million inpatient surgeries performed annually in the U.S.—about 1.62 surgeries every second—this high frequency highlights the need for us to implement more effective approaches to prevent surgical errors.

According to the 2024 Global Patient Safety Report from the World Health Organization, only 38% of countries worldwide have implemented reporting systems for preventable and highly destructive medical errors, known as 'Never Events'—medical errors that should never occur. Many of these errors remain underestimated and overlooked.

Beyond retrospective reviews, we should also consider using modern technology to prevent surgical errors in real-time. This includes proactive prediction and analysis of diagnoses and clinical evidence to guide accurate surgical decisions and documentation.

Overcoming the limitations of traditional rule-based systems and evaluating clinical justification—especially in cases where diagnoses are vague, incomplete, or lack clear specifications of body parts or sides—will mark significant progress in the development of medical AI.

Reference:

From Diagnosis to Surgical Error: Addressing Inconsistencies in Medical Documentation

AESOP Talks