Table 4.

Comparison of POPS against methods found in our literature review of preoperative prediction of postoperative pain.

Study	Retrospective cohort size	Prospective cohort size	Patient interaction	Distribution of surgical services	Outcome	Performance
POPS (ours)	243,274	365	No	Orthopedic: 24.3% General: 15.0% Urology: 10.5% Gynecology: 9.5% Thoracic: 7.5% Neurosurg.: 7.1% Surg. Oncology: 5.9% Other: 20.3%	Moderate and severe pain on postop days 0–4	0.72–0.79 AUC
Armstrong et al. 2023	17,079		Yes	General: 62.0% Urology: 15.6% Thoracic: 7.5% Orthopedic: 5.5% Neurosurg.: 3.1% Gynecology: 1.7% Vascular: 0.7% Other: 3.9%	Severe pain on first postop day	0.66 optimism-corrected c-statistic
Van Driel et al. 2022	344	150	Yes	Orthopedic: 45.3% General: 20.9% Vascular: 19.5% Other: 14.2%	Persistent post-surgical pain at 3 months	0.7 AUC
Rehberg et al. 2017		198	Yes	Surg. Oncology: 100%, only breast cancer surgeries	Max pain on first postop day	0.82 AUC
Hur et al. 2021	112,989		No	Unreported, 13 common procedures	30-day refill, new persistent use	0.66–0.68 AUC
Janseen et al. 2008	549	1035	Yes	Unreported, non-cardiac surgery	Severe pain 1 h after surgery	0.65 AUC
Kallman et al. 2003		1416	Yes	Unreported, non-cardiac surgery	Severe pain 1 h after surgery	0.73 AUC
Sommer et al. 2010		1490	No	Unreported, non-cardiac surgery	Mean pain (VAS) > 4	0.74–0.78 AUC

AUC Area under the receiver operating curve.