Translating ethical and quality principles for the effective, safe and fair development, deployment and use of artificial intelligence technologies in healthcare.

OBJECTIVE: The complexity and rapid pace of development of algorithmic technologies pose challenges for their regulation and oversight in healthcare settings. We sought to improve our institution's approach to evaluation and governance of algorithmic technologies used in clinical care and operations by creating an Implementation Guide that standardizes evaluation criteria so that local oversight is performed in an objective fashion. MATERIALS AND METHODS: Building on a framework that applies key ethical and quality principles (clinical value and safety, fairness and equity, usability and adoption, transparency and accountability, and regulatory compliance), we created concrete guidelines for evaluating algorithmic technologies at our institution. RESULTS: An Implementation Guide articulates evaluation criteria used during review of algorithmic technologies and details what evidence supports the implementation of ethical and quality principles for trustworthy health AI. Application of the processes described in the Implementation Guide can lead to algorithms that are safer as well as more effective, fair, and equitable upon implementation, as illustrated through 4 examples of technologies at different phases of the algorithmic lifecycle that underwent evaluation at our academic medical center. DISCUSSION: By providing clear descriptions/definitions of evaluation criteria and embedding them within standardized processes, we streamlined oversight processes and educated communities using and developing algorithmic technologies within our institution. CONCLUSIONS: We developed a scalable, adaptable framework for translating principles into evaluation criteria and specific requirements that support trustworthy implementation of algorithmic technologies in patient care and healthcare operations.

Predicting urinary incontinence after surgery for pelvic organ prolapse.

PURPOSE OF REVIEW: Many women choosing to have surgery for pelvic organ prolapse also choose to undergo continence surgery. This review focuses on available evidence that clinicians may use to counsel patients when choosing whether to perform continence surgery and how predictive analytic tools improve this decision-making process. RECENT FINDINGS: Midurethral sling, Burch cystourethropexy and bladder neck sling are highly effective for the surgical treatment of stress urinary incontinence. Trials demonstrate that continence surgery may be routinely performed to reduce the risk of postoperative incontinence in women undergoing surgery for pelvic organ prolapse with or without preoperative stress urinary incontinence. Although these procedures are effective and well tolerated on average, media concerns, regulatory warnings and litigation reinforce the need for a balanced discussion regarding efficacy and potential adverse events directed at the individual patient during the preoperative visit. Advances in predictive analytics allow surgeons to quantitate individual risk using algorithms that tailor estimates for the individual patient and facilitate shared understanding of risks and benefits. These models are less prone to cognitive biases and frequently outperform experienced clinicians. SUMMARY: This review discusses how predictive analytic tools can be used to improve decisions about continence surgery in the woman planning to undergo prolapse surgery.

Perioperative adverse events after minimally invasive abdominal sacrocolpopexy.

OBJECTIVE: Our first objective was to compare peri- and postoperative adverse events between robotic-assisted laparoscopic sacrocolpopexy (RSC) and conventional laparoscopic sacrocolpopexy (LSC) in a cohort of women who underwent these procedures at a tertiary care center. Our second objective was to explore whether hysterectomy and rectopexy at the time of sacrocolpopexy were associated with these adverse events. STUDY DESIGN: This was a retrospective cohort study of women who underwent either RSC or LSC with or without concomitant hysterectomy and/or rectopexy from 2006-2012. Once patients were identified as either having undergone RSC or LSC, the electronic medical record was queried for demographic, peri-, and postoperative data. RESULTS: Four hundred six women met study inclusion criteria. Mean age and body mass index of all the women were 58 ± 10 years and 27.9 ± 4.9 kg/m(2). The women who underwent RSC were older (60 ± 9 vs 57 ± 10 years, respectively; P = .009) and more likely to be postmenopausal (90.9% vs 79.1%, respectively; P = .05). RSC cases were associated with a higher intraoperative bladder injury rate (3.3% vs 0.4%, respectively; P = .04), a higher rate of estimated blood loss of ≥500 mL (2.5% vs 0, respectively; P = .01), and reoperation rate for pelvic organ prolapse (4.9% vs 1.1%, respectively; P = .02) compared with LSC. Concomitant rectopexy was associated with a higher risk of transfusion (2.8% vs 0.3%, respectively; P = .04), pelvic/abdominal abscess formation (11.1% vs 0.8%, respectively; P < .001), and osteomyelitis (5.6% vs 0, respectively; P < .001). The mesh erosion rate for all the women was 2.7% and was not statistically different between LSC and RSC and for patients who underwent concomitant hysterectomy and those who did not. CONCLUSION: Peri- and postoperative outcomes after RSC and LSC are favorable, with few adverse outcomes. RSC is associated with a higher rate of bladder injury, estimated blood loss ≥500 mL, and reoperation for recurrent pelvic organ prolapse; otherwise, the rate of adverse events is similar between the 2 modalities. Concomitant rectopexy is associated with a higher rate of postoperative abscess and osteomyelitis complications.

Seven years after laparoscopic radical nephrectomy: oncologic and renal functional outcomes.

OBJECTIVES: To compare the long-term oncologic and renal function outcomes in patients undergoing laparoscopic (LRN) versus open radical nephrectomy (ORN). METHODS: The medical records of 116 patients undergoing radical nephrectomy for pathologically confirmed renal cell carcinoma before January 2000 were reviewed. Of these 116 patients, 63 underwent LRN and 53 ORN. The oncologic and renal functional data were obtained from the patient charts, radiographic reports, and direct telephone calls to the patients or their families. RESULTS: The median follow-up was 65 months (range 19 to 92) in the LRN group and 76 months (range 8 to 105) in the ORN group. LRN was successfully completed in all patients without open conversion. The mean tumor size was 5.4 cm in the LRN group and 6.4 cm in the ORN group (P = 0.007). The 5-year overall survival (78% versus 84%, respectively; P = 0.24), cancer-specific survival (91% versus 93%, respectively; P = 0.75), and recurrence-free survival (91% versus 93%, respectively; P = 0.75) rates were similar between the LRN and ORN groups. At 7 years, the overall survival (72% versus 84%; P = 0.24), cancer-specific survival (91% versus 93%; P = 0.75), and recurrence-free survival (91% versus 93%; P = 0.75) rates were also comparable. No port site recurrence was noted in the laparoscopic group. The long-term renal function outcomes were similar in the LRN and ORN groups, with serum creatinine increasing by 33% and 25%, and the estimated creatinine clearance decreasing by 31% and 23% from baseline, respectively. Chronic renal insufficiency developed in 4% of patients in each group. CONCLUSIONS: The results of our study have shown that LRN and ORN have comparable long-term oncologic and renal functional outcomes.

Anatomic relationships of infracoccygeal sacropexy (posterior intravaginal slingplasty) trocar insertion.

OBJECTIVE: The purpose of this study was to describe the distances of the major bony, vascular, neurologic, and visceral structures to the path of the infracoccygeal sacropexy trocar and to determine the point of trocar entry into the vagina. STUDY DESIGN: Infracoccygeal sacropexy trocars were inserted bilaterally into 6 fresh frozen cadavers. Dissection was performed and the maximal length of the vagina, ischiorectal fossa, and pararectal spaces were measured bilaterally. Mean distances with 95% CIs to important anatomic structures were made from fixed points along the trocar's path. RESULTS: The path of the trocar began dorsal and lateral to the anus, passed through the ischiorectal fossa, iliococcygeus muscle, into the pararectal space, and into the posteriolateral vagina. Along this course, the mean distance (95% CI) to the pudendal vessels at the exit of Alcock's canal was 2.8 cm (2.1 to 3.4 cm) and rectum was 0.5 cm (0.2 to 0.9 cm). The closest inferior rectal vessel was 0.1 cm (0 to 0.3 cm). In the pararectal space, the mean distance to the ischial spine was 2.6 cm (1.7 to 3.5 cm). In 12 of 12 trocar passages, the inferior rectal branches of the pudendal artery and the rectum were within 1 cm or less of the trocar. The mean distance of trocar entry into the vagina was only 4.8 cm (4.3 to 5.4 cm) proximal to the hymenal ring compared with a mean total vaginal length of 8.7 cm (8.0 to 9.3 cm). CONCLUSION: This anatomic study suggests that the rectum and the inferior rectal branches of the pudendal neurovascular bundle may be at risk of injury during infracoccygeal sacropexy trocar placement. Additionally, this procedure appears to provide support to the mid-posterior vaginal wall, not the vaginal apex.

Functional bowel and anorectal disorders in patients with pelvic organ prolapse and incontinence.

OBJECTIVE: The purpose of this study was 1) to determine the prevalence of functional bowel and anorectal disorders as defined by the Rome II criteria in patients with advanced pelvic organ prolapse (POP) and urinary incontinence (UI), and (2) to determine if the extent of prolapse on gynecologic examination is related to the subtypes of constipation or any functional anorectal pain disorder. STUDY DESIGN: Three hundred and two consecutive female subjects presenting to a tertiary urogynecology clinic were enrolled. Demographic, general medical, and physical examination information, including POPQ measurements and a standardized sacral neurologic evaluation, were collected. The prevalence of functional disorders of the bowel, rectum, and anus as defined by the Rome II criteria were collected using the Rome II Modular questionnaire. Relationships of functional disorders to various components of the vaginal examination were reviewed. RESULTS: Thirty-six percent (108/302) met the criteria for constipation, including the following subtypes: 19% outlet constipation, 5% functional constipation, 5% constipation predominant irritable bowel syndrome (IBS), and 7% IBS-outlet. Nineteen percent (56/302) of subjects had IBS or 1 of its subtypes. Functional diarrhea was seen in 6% (17/302), fecal incontinence in 19% (58/302), and anorectal pain disorders in 25% (77/302). After controlling for age, parity, diabetes, constipating medications, and previous pelvic surgery, there were no differences in the prevalence of constipation or any of its subtypes between patients with UI and those with stage 3 or 4 POP. Fecal incontinence was independently associated with UI (adjusted odds ratio [OR] 6.3; 95% CI 2.6-19.1), but not advanced POP. Neither overall stage of POP nor stage of posterior vaginal prolapse was significantly associated with any of the functional bowel disorders, including constipation and its subtypes. Perineal body measurement was significantly longer in patients with outlet type constipation (mean 3.5 +/- 0.6 cm vs 3.1 +/- 0.9 cm, P < .01) and in those with proctalgia fugax (mean 3.4 +/- 1.0 vs 3.1 +/- 0.8, P < .05). CONCLUSION: There is a high prevalence of constipation and anorectal pain disorders in women with urinary incontinence and pelvic organ prolapse. However, patients with stage 3 or 4 pelvic organ prolapse have similar rates of constipation compared with those with urinary incontinence. Constipation and its subtypes are not related to the stage of pelvic organ prolapse. It appears that either constipation is not a significant contributor to prolapse, or constipation contributes equally to the development of both urinary incontinence and pelvic organ prolapse.

Incidence and predictors of prolonged urinary retention after TVT with and without concurrent prolapse surgery.

OBJECTIVES: The purpose of this study was to describe the time to adequate voiding, incidence of urinary retention, and predictors of voiding efficiency and urinary retention after tension-free vaginal tape (TVT) with and without concurrent prolapse surgery. STUDY DESIGN: Medical records of patients who underwent TVT between August 1999 and July 2003 were reviewed. Urinary retention was defined as the need for urethrolysis, urethral dilation, or postoperative catheterization for >6 weeks. Linear and logistic regression models were used to determine predictors of time to adequate voiding and urinary retention. RESULTS: Two hundred sixty-seven patients were available for analysis; 66% had concurrent prolapse repair, 4% had concurrent laparoscopically assisted vaginal hysterectomy (LAVH), and 30% had an isolated TVT. TVT with and without concurrent prolapse repair or LAVH were statistically similar with respect to median days to voiding (8 vs 5) and the rate of urinary retention (11.2% vs 11.3%). Overall, 4.9% underwent urethrolysis, 1.9% received urethral dilation, and 4.1% required prolonged catheterization. Increasing age, decreasing BMI, and postoperative urinary tract infection were independent predictors of time to adequate voiding. Previous history of incontinence surgery was the only independent predictor of urinary retention (Adjusted odds ratio [AOR] 2.96, 95%CI [1.17-7.06]). CONCLUSION: Concurrent prolapse surgery does not appear to significantly alter postoperative voiding efficiency or increase the risk of prolonged urinary retention compared with TVT alone.

Endometriosis of the liver containing mullerian adenosarcoma: case report.

We present a case of a liver endometrioma in a postmenopausal woman. After failed management with leuprolide acetate, the mass was resected and contained focal areas of mullerian adenosarcoma. This is a rare case of mullerian adenosarcoma that appeared to arise within an endometrioma of the liver.