Similar to prior studies, in this chart review, approximately 15% of patients had an AD on file8,9. Over a quarter of the patients were readmitted in the following year. Our data does not find an association between readmission and having an AD on file, which was surprising as readmitted patients have more contacts with the health system and are potentially a sicker portion of the cohort. This is equally surprising considering one-third of our readmitted patients had a diagnosis of potentially lethal cancer.
Our mortality rate was < 5%. Other studies of patients undergoing elective surgeries had mortality rates ranging from < 1% to approximately 3%, however these studies examined in-hospital and/or six month mortality 1,2,5. The overall low mortality rate may be more of a reflection of the safety of modern surgery rather than improved health of our patients. Recent data suggests that patient frailty, which takes medical co-morbidities as well as cognitive and functional measures into account, may be a more powerful risk factor for post-operative outcomes13,14. Frailty may be a better marker to identify which patients may benefit most from AD completion prior to surgery.
Only 17% of patients who died in the cohort had an AD on file prior to surgery and 63% of patients who died had an AD on file prior to death. While there was a statistically significant relationship between having an AD on file and death, some of those were likely in response to the patient’s declining condition. Further review revealed that patients who died were high risk based on having cancer and other advanced diseases at the time of surgery. These findings suggest a missed opportunity for identifying patients with high one-year mortality risk so that ACP conversations can occur and ADs can be completed in advance.
Almost 40% of patients who died were resuscitated prior to their death. The perioperative evaluation is a missed opportunity to explore whether life sustaining treatments were consistent with the patient’s goals. Studies have demonstrated that the more elements of goals of care that are discussed, the better agreement between patient’s preferences and goals of care15. Moreover, ACP is correlated with decreased aggressive measures at the end of life and increased concordance with patients’ end of life wishes16.
There are several limitations to this study. It was conducted within one health system, including one tertiary referral center, which may attract a higher risk surgical population. Data was collected over only one year and was observational. Finally, the high percentage of surgeries performed for cancer affected the likelihood of patients proceeding to surgery with minimal discussion as most preoperative clinics are more likely to recommend proceeding to cancer surgery without delay than for elective procedures.
In conclusion, the finding that a minority of patients who die within a year of major surgery have an advance directive highlights a missed opportunity to conduct advance care planning in a perioperative clinic. Future studies should include the patient perspective on how patients undergoing elective procedures would have benefited from the opportunity to discuss advance care planning along with the implementation of advance care planning in preoperative clinics and subsequent measure of effect on patient expectations and experience.