As Dr Wertheim has suggested, it would be ideal if we could collect “all relevant information.” However, at this point, we collect as much information as is practically feasible. The lack of data regarding physician counseling is a limitation, but the finding that 38% of hysterectomy cases had no documentation referencing other treatments before hysterectomy is concerning.

It is not possible for the clinical reviewers in the Michigan Surgical Quality Collaborative (MSQC) to track down every paper clinic note. However, they go to extensive efforts to identify whether other treatments were considered, attempted, or declined. As physician practices increasingly become part of hospital systems, there is fortunately greater access to clinic notes through electronic health records.

In addition, the MSQC clinical nurse reviewers, who routinely identify 30 day postoperative complications by communicating with the surgeon’s office staff, routinely ask whether there is any evidence about whether other treatments are mentioned in the clinic record. A reference, even in general, to options for other treatment, consideration of any hormonal therapy, or a dilatation & curettage in the surgical history is sufficient evidence for the MSQC reviewer to indicate that other treatments were considered before hysterectomy.

This analysis is limited by a lack of information regarding patient counseling and by a lack of complete access to every clinic note, but we also contend that there is an important signal regarding the underutilization of other treatments.

As for surgical appropriateness of hysterectomy, it is not surprising that questions are raised when no pathology is present in 18% of specimens. We recognize that negative pathology is not conclusive evidence of inappropriate surgery and that adjudicating appropriateness is not possible at this point. For this reason, we do not conclude how many were actually inappropriate.

As we work to determine an acceptable proportion of negative pathology hysterectomies, it is worthwhile noting that a protocol encouraging the use of other treatments for benign gynecological symptoms decreases the number of hysterectomies and the number with negative pathology.

In summary, we readily recognize the limitations of our analysis and will continue to explore whether documentation of other treatments and negative pathology are important markers for quality improvement in hysterectomy care.